THELEVERAGEFACTOR: HowtheInvestorCanProfitfromChangesinCorporate Risk ByJ.D.Ardell íhşmƌmorvşoƍşmvşmolƏƌr,şJƌƈhşŴhƋrƌwsşŴrƋƌll,şwƏoşmƈƋƌşilşloşlƏƌşripƌolƋşƈƎƌşoƍş 9ůşƈhƋşlƈuƎƏlşmƌşƌvƌrvlƏihƎşíşkhowşƈƉoulşpƈliƌhƊƌ,şƏohorşƈhƋşƋiƎhilvşş ş ş ş ş ş ş ş ş ş ş ş ŽrƌƍƈƊƌ`şş şŴşƏƈhƋƍulşoƍşƉooksşwillşmƈkƌşƈşƋiƍƍƌrƌhƊƌşihşvourşliƍƌ.şşíƍşlƏƌşrƌƈƋƌrşisşƋiliƎƌhlşihş uhƋƌrslƈhƋihƎşlƏƌşmƈlƌriƈlşƊohlƈihƌƋşƏƌrƌih,şlƏisşmƈvşƉƌşohƌşoƍşlƏƌm.şşƄƏilƌşlƏƌş ƉookşisşƊollƌƎƌlƌvƌlşƈhƋşrƌlƈlivƌlvşƊomplƌx,şlƏƌrƌşƈrƌşslƌpşƉvslƌpşihslruƊliohsş lƏƈlşƌhƈƉlƌşlƏƌşsluƋƌhlşorşihvƌslorşloşƌƈsilvşlƌƈrhşƉolƏşƈppliƌƋşslƈlisliƊsşƈhƋş ƊorporƈlƌşƍihƈhƊƌş·şihşƈhşƌhvirohmƌhlşlƏƈlşisşƊohƋuƊivƌşloşmƈkihƎşsouhƋş ihvƌslmƌhls.şşíhşolƏƌrşworƋs,şilşƌhƊourƈƎƌsş¨ŸƈrhihƎşwƏilƌşvouşƈrƌşLƌƈrhihƎ".şş Ƃllimƈlƌlv,şvouşmƈvşƍihƋşvoursƌlƍşmƈkihƎşƉƌllƌrşƋƌƊisiohsşlƏƈhşsomƌşƌxƌƊulivƌsşwƏoş ƌƈrhşohƌşƏuhƋrƌƋşlimƌsşvourşsƈlƈrv.şşƄƏilƌşliƍƌşisşholşƈlwƈvsşƍƈir,şilşisşsomƌlimƌsş mƌrƊiƍul.şşŻowƌvƌr,şliƍƌşisşƈlsoşƈhvlƏihƎşƉulşƊƌrlƈihşƈhƋşƈhvşƋƌƊisiohşvouşmƈkƌş hƌƌƋsşloşƉƌşƊorroƉorƈlƌƋşƉvşƈsşmuƊƏşihƍormƈliohşƈsşpossiƉlƌş·şihƊluƋihƎşopposihƎş viƌwpoihls.şşŴlwƈvsşƊƏƈllƌhƎƌşƈulƏorilvş·şƌspƌƊiƈllvşvourşowhŠş TABLEOFCONTENTS i 1.-Introduction:ATaleofTwoCompanies,orthree,orfour...1 SECTION1:THETHEORYOFCAPITALSTRUCTURE11 2.-Leverage11 BusinessRisk,12 FinancialRisk,13 Leverage:ADefinition,13 BasicRisksandProportions,14 BalancingLeverage,16 TheAnatomyofFinancialLeverage,17 TheAnatomyofOperatingleverage,20 TotalRisk,22 LeverageMeasurement,23 Theoryvs.Reality:FinancialLeverage,24 Theoryvs.Reality:OperatingLeverage,24 TheoryandReality:TotalLeverage,25 LeverageManagement,27 SourcesofVariation,28 ReturnonEquity,30, TheForcesBehindLeverage,31 Appendix:Streamlining?,33 3.-CapitalStructure35 TheCostofCapital:MarketOrientationandPracticalApplication,36 TheAdvantagesofDebt,38 RiskandDebt,40 MathematicalOptimization:Theoryvs.Reality,43 TheModigliani-MillerPropositions,45 TheOptimizationProblem,46 TheProposedIdealanditsInherentProblems,48 CapitalStructureandtheCostofCapital,50 TheConceptofaWeightedAverageCostofCapital,53 EarningsandCapitalStructure,55 CapitalStructureLogic,56 ChangesinCapitalStructureandStockPrices,57 Four“Postulates”,59 ShareLimitations,60 AdaptedMeasurements,61 Explicitvs.ImplicitCosts,62 TABLEOFCONTENTS ii ImplicitCostsofDebt,64 TheImplicitCostofEquity,64 TheMomentofTruth,66 Appendix:TheNetOperatingIncomeApproachtoStockValuation,68 4.-TheCostofDebt71 TheProblemofShort-termCredit,71 InterestExpenseInequalities,73 Risk,Return,andtheSignificanceofShort-termCredit,74 TheCorporateCostofDebt,76 TheNominalCostofDebtandtheCostofBankruptcy,78 TheCostofBankruptcy,80 TheProbabilityofDefault,80 CommercialRatingsSystems,82 TypesofBankruptcy,84 TheAmountofLoss,85 TheRoleofDebtinCapitalStructureOptimization,90 TheOptimalAmountofDebt,92 Long-termDebtandtheAmountofLoss,93 GraphicDepiction,93 Appendix:MakingSureThatWhatYouSeeisWhatYouGet,95 5.-TheCostofEquity99 Modifications,100 TheReinvestmentRateandOpportunityCosts,103 EvaluatingtheCostofEquity:Methodology,104 TheCAPMasaBuildingBlockforCapitalStructureAnalysis,110 AnEstimateofRiskandnotPrediction,117 AnomaliesPertainingtotheUseoftheCostofEquityinCapitalStructure,119 AdaptiveExpectationsversusRationalExpectations,120 TheDecompositionofBeta,122 BalancingLeverage,RiskandtheCAPM,127 BetaandLeverage,128 ASimpleEqualizer,129 ConcludingComments,131 Appendix:TheMathematicalRelationshipBetweenLeveredandUnleveredCompaniesin TermsoftheCAPM,133 6.-TheCapitalDynamicandOtherTools138 ThePercentageTrap,138 TheWeightedAverageCostofCapital,139 TheTheoreticalShapeoftheWeightedAverageCostofCapitalCurve,141 TABLEOFCONTENTS iii AccelerationRates,142 TheWACCandRisk,143 TheMechanicsoftheWACC:RiskAdjustment,143 CorporateMethod,144 RiskAdjustedMethod,145 TheMarginalCostofCapital,145 DecisionMakingandtheMarginalCostofCapital,147 EconomicProfit,EVA,andtheCapitalDynamic:UtilizingtheOpportunityCost,149 ElementsinanEVACalculation,150 TheCapitalDynamic,151 TheRelationshipbetweenEVAandtheCapitalDynamic,152 EconomicProfitandCorrelation,152 ManagementofEconomicProfit,154 ComponentMovementsoftheCapitalDynamic,157 TheComparativeCapitalDynamic,159 Appendix:TheEfficiencyofEVAversusROE,160 Appendix:TheCostofCapitalandWhatTheInvestorNeedstoKnow,161 7-FundamentalsandCapitalStructure168 DuPontAnalysis,170 ComparingROEComponents,175 ModifyingandEnhancingDuPontAnalysis,181 TheReturnonCapitalRatio,186 8.-CapitalStructureandtheBusinessCycle189 CommonElementsofBusinessCycles,189 TheYieldCurveandInterestRateBehavior,190 GraphsThatUnitetheTwoTheories,192 StrategicConsiderations,198 TheBusinessCycleandtheCostofEquity,200 TheCapitalAssetPricingModelandSensitivityAnalysis,202 CircumventingtheOptimalCapitalStructure,208 TheGameofCapitalStructure“Gothcha”,209 IdealizedTrends,210 SectorRotation,212 SectorLogic,213 IndustryResponsetotheBusinessCycle213 EconomicSignals,217 9.-OperatingRisk219 FixedCostsandEconomics,220 TheCaseofCompaqComputer,223 TABLEOFCONTENTS iv TheNatureofCostsandMargins,224 FixedCostsandtheBreakevenpointforSales,227 CompaqComputer:TheRestoftheStory,229 OperatingLeverageandPrediction,231 CharacteristicsofOperatingLeverage,232 OperatingTrendsandReversals,235 TheQualityofanOperatingMargin,235 ARiskyProposition:ConfidenceIntervals,237 OperatingBeta,240 TheUnleveredBetaEquation,241 TheArdcoBarbellCompany:AnExampleofUnleveredBeta,241 “MomandPopStore”Betas:CompaniesWhoareNotontheMarket,244 OperationsResearchfortheInvestor,245 TwoMasters:FisherandBuffett,247 ABriefOperatingAnalysisofFed-ExandStaplesfortheYear2000,249 StaplesandFed-ExOperatingHistories:1994-1999 AnalysisandStatistics TheConfidenceIntervalTool OperatingMargin 10.-OperatingMomentum258 ReasonsforStudy,259 OperatingMomentumSensitivity,264 Regression,269 TheGeneralElectricSolution,271 ClassicalMicroeconomicsandOperatingMomentum,275 11.StrategicCapitalRequirements279 TheRealitiesofFunding,279 TheProperAmountofCapital,280 TheDebt/EquityTradeoffandEVA,282 EVA/CapitalDynamicBasedImprovement,285 IncrementalEquityImprovement,287 TheIrrationalityofRationingCapital,287 IncrementalDebt,288 DividendsandRetainedEarnings,288 CapitalFundingFromEVA:TwoMethods,288 Method1:SolvingfortheOptimalEquity Method2:SolvingfortheOptimalNetIncome DeterminingCapitalProportionsandRequirements,290 ProjectedAnalysis,291 FinancialEngineering:SettingCapitalRequirementsFromEVA,291 ConocoPhillips2005-2006:aRealWorldExample,291 TABLEOFCONTENTS v Method1:CapitalProportionsfromOptimalEquity Method2:CapitalProportionsfromOptimalNetIncome FinancialEngineeringMethod TheAdditionalFundsNeededEquation,296 TheModifiedAdditionalFundsNeededEquation,299 DegreeofAFNLogic,303 TheNeedforQualitativeAssessment,306 TheProblemwithOptimality,306 MergerMania,308 MergerGrowthIllusionandEVA,311 SECTION2:BUILDINGCAPITALSTRUCTUREMODELS315 12.-TheEconomicProfitLaboratory:ComputerApplications315 SetUp,315 SectionOne,315 SectionTwo,316 SampleData,318 SensitivityversusOptimization,319 ProvingtheCapitalDynamic/EVAHypothesis,320 SettingtheConstraints,320 TriggerPoints,321 SettingtheTriggerPointModule,322 TheEarningsSolution,323 OptimizationandCorrelation,324 RaisingCapitalEffectively,324 TheConnectionBetweenCapital,StockPrice,andEVA,326 SensitivityAnalysis:TheEffectofChangesinOperatingIncomeandCapital,327 EstablishingGuidelines,331 Appendix:SpreadsheetExamples,333 13.-TheMarginalBenefitsEquation:AnExperimentalModel335 TheModeledConcept,336 TheMarginalBenefitsEquation,336 DefaultProbabilityandBankruptcy,338 TheInterestBenefitsMechanism,339 CheckingResultsAgainstaViableStandard,340 DefaultMechanics,340 StrategicImplications:FinancialLeverage,341 StrategicImplications:OperatingRisk,343 SpreadsheetConstants,344 TABLEOFCONTENTS vi SpreadsheetLogic,346 DynamicVariables,347 ModelSetUp,348 TheProcess:EntryVariables,349 TheProcess:OptimizingwithSolver,350 TheResults:ThreeExamples,350 EVADiscrepancies,354 Appendix:ListofFormulasandSpreadsheetConstruction,355 14.AnIntroductiontoResidualEconomicProfitTheory:Usinga ConstantDividendDiscountModel359 AnIntroductiontoResidualEconomicProfitTheory,359 OpportunityCost,360 ValuationModels,361 DividendTheory,362 ResidualEconomicProfit,364 TheDividendTrap,365 ModelOptimization,366 ModelBackground,366 ModelSetUp,368 ModelAdaptations,369 TheCaseof:CanYouTopThis?,369 ComparingSpreadsheets,370 Appendix:ThreeSpreadsheets,373 SECTION3:REALWORLDCASES376 15.-AnalyticalTools:PracticalApplication376 TheTolerationofImprecision,376 ErringontheSideofConservatism,378 BriefMethodologiesforDeterminingtheCostofEquity,379 TheHurdleRate,382 TheEVA/CapitalDynamic,383 TheWeightedAverageCostofCapital,383 ComparingRisk:JustificationforTwoCostsofEquity,386 ChangesintheCAPM,386 TheComparativeCapitalDynamic,388 TheMarginalBenefitsEquation,388 LeverageStateAnalysis,391 The“LookAhead”Bias,395 MicroAnalysis:QuarterlyObservation,396 NaiveExtrapolation,397 EarningsPressure,399 TABLEOFCONTENTS vii Appendix:DividendDiscountModels,402 16.-Kimberly-Clark-”TooMuchofaGoodThing”:Economic ProfitandMarginalBenefitsAnalysis404 Underpinning1:PositionintheBusinessCycle,404 Underpinnings2,3and4:OpportunitiesforAnalysis,405 TheLeverageState,406 ChangesinEconomicProfit,411 TheExtremeConsensusMethod,411 EconomicProfit,416 TooMuchofaGoodThing,417 MarginalBenefitsAnalysis,418 BasicMethodology,419 TaxBenefitsforKimberly-Clark,420 AmountofLossforKimberly-Clark,420 TheProbabilityofDefault,421 TheCostofBankruptcy,422 MarginalBenefits,423 Confirmation,424 Altman’sZScore:BookValueVersion,424 Kimberly-Clark’sZScore,426 InvestmentConclusion,427 Appendix:ExtrapolatedRisk:When“Normal”isTooRisky,428 17.-“FullSteamAhead”:AnAnalysisofConocoPhillips,2002- 2006431 TheContext,431 TheDecision,434 PricePerformance,435 AnticipatingPerformance:LeverageStates,436 QuarterlyLeverageResults,438 InterpretingRegression,439 EstablishingaComparativeCostofEquity,441 ContrastingtheRequiredReturnwiththeExpectedReturn,445 TheEVA/CapitalDynamic,446 NaiveExtrapolation,448 TheMarginalBenefitsFunction,451 TheComparativeCapitalDynamic,434 EarningsPressure,456 Appendix:SelectedFinancialDataforConocoPhillips,2001–2006,459 TABLEOFCONTENTS viii 18.-MicrosoftVersusConocoPhillips:ComparingCompaniesin DifferentIndustries460 ApplesandOranges:MicrosoftVersusConocoPhillips,461 CommonGround,463 AMarketDisconnectandEventualReconciliation,467 IndustryCompetition:ChevronandtheComparativeCapitalDynamic,469 PercentageofNewRetainedEarnings,473 SECTION4:CORRELATIONANDPROBABILITYSTUDIES- 475 19.-OperatingIncomeCorrelationStudies475 Name,Premise,DataPointsandStructure,476 Categories,477 CompaniesintheSample,479 FundamentalVariables,479 AllVariables,480 StatisticalResults,481 SpearmanRankCorrelations:NextYear’sMidrangeStockPrice,484 Interpretation,485 MethodologicalCriticism,488 20.-ChangesinCapitalStructureandTheirEffectonStockPrices -491 TheValidityofLeverageFactors,491 ConnectingtheDots:EarningsandDividendGrowthandtheCostofEquity,492 EarningsAcceleration,493 StatisticalValidity,495 ABriefStudy,496 ThreeAssumptions,498 Expectations,499 InterpretationandResults,500 HypotheticalCausation,502 TheHazardsofPlayingDetective,503 TheArgumentforCapitalRationing,505 SpearmanRankCorrelationandIndividualInterpretation,510 21.-ProbabilityandCapitalStructure518 TABLEOFCONTENTS ix TheEfficientMarketsHypothesis,519 Screens,519 TheReturnonCapital,520 LeverageStates,521 IndustryAverages:Lemmingsvs.Leaders,522 TheLeverageStateRatios,523 Combinations,524 TheMechanismofLeverageStates,525 MatchingtheLeverageStatetotheBusinessCycle,528 ProbabilityandDiversification,530 SalesandBeta,531 ProbabilityandAnticipation,532 PrincipleComponentsAnalysis,534 Staticvs.Forward-LookingRatios,536 TheQuickPayoff,540 BarrRosenbergandResponseCoefficients,541 22.-TechnicalAnalysisandCapitalStructure544 MajorForces,545 TheBaneofVolatility,547 Self-fulfillingProphecy,548 Ex-PostPerformance,549 StochasticConformance,552 CapitalStructuralism:Quasi-TechnicalAnalysis?,334 FightingWords:“TheEfficientMarketsHypothesis”,555 TheArtandScienceofForecasting,557 MovingAveragestoUseWisely,558 TheExponentialMovingAverage,559 BriefInterpretation,561 PrimaryTrendsandSecondaryTrends,561 23.-StatisticsPrimer563 TheMean,ModeandMedian,564 TheVarianceandAssortedAdaptations,566 TheStandardDeviation,567 TheCovariance,567 DownsideRisk,568 AnnualizedVolatility,569 EstimatedVolatility,570 SampleStandardDeviation,571 TheMean–Variance,571 TheCoefficientofVariation,572 WorstCaseScenarios,572 AccountingforAdditionalRisk:CombiningStandardDeviations,574 TABLEOFCONTENTS x Prediction,576 ConfidenceIntervals,576 TheNormalDeviate,578 UpdatingtheMeanandStandardDeviation:MovingAverages,580 AccountingforSmallSampleSize:TheStudent“t”Distribution,581 Regression,582 RelationshipsinLinearRegression,584 TheCoefficientofDetermination,585 MakingPredictions,585 GrowthRates,586 GeometricApproximation,589 AccurateGrowthRatesfromLogarithms,588 SpearmanRankCorrelation:NonParametricStatistics,589 SampleSize,592 SELECTEDREFERENCES 1 1 “ATALEOFTWOCOMPANIES”-orthree,orfouror... ThecomparisonbetweenWallStreetanda“jungle”wasnevervalid.Iflifeina stateofnaturecanbedescribedas“nasty,brutishandshort”,nowhereistheDarwinian conceptof“survivalofthefittest”championedmore.Whileconsumptionofsmallfishbya largeroneseemsanappropriatemetaphor,itistriteandmyopicenoughtodivertour attentionfromthetruth:theconceptsofmutualbenefitandsharedgainaremore prominentthananyvestigesofunilateralconquest. Infact,ifanyallegoryisespeciallyapplicabletocharacterizing“theStreet”,itisthe BiblicalstoryofJonahandthewhale.Beliefinmiraclesnotwithstanding,“theBigFish” swallowingasmallercreatureisonlypartofthestory.Jonahturnsthecrisisintoan opportunity;heis“spat”outontoabeach,andbeginspreachingtothekingdomof Nineveh,changingtheirmisbegottenways. Likethewhale,thefinancialcommunitymustcontinuetogrowtoremain competitive.However,likeJonah,thereislifeafterconsumption.Despiteitsreputation fora“bottomline”mentality,WallStreetcreatesmutualbenefitsbysharingrisk.When individualswithsimilartolerancesforriskpooltheirresources,thepotentialreturnrises, orthechanceforanunacceptableoutcomedecreases,orboth;theembodimentofthis conceptistheinherentlimitedliabilityofthecorporationitself.This“symbiosis”isthe productnotofsomehierarchicalstructurethateliminatescompetition,butamathematical processthatcombinesriskandreturninthemostefficientmannerpossible. Thefollowingstoryrelatesascenariothatisquitecharacteristicofcompetition betweenmoderncorporations.Onecompany’srisk-seekingmarketingstrategynarrows itsfocustoapointwhereitscash-flowiscompromised.Anothercompanywhoserevenue streamismorediversified,seeksriskinitsmanneroffunding,andendsupborrowing 2 moneytotakeoverthefirstcompany.Thistensionbetweentwotypesofrisk–onethat affectsrevenuesandonethataffectsfunding–formsthecruxofallcapitalstructure decisions.Leverageismerelythemeasuredmanipulationoftheserisks,whileanoptimal capitalstructurepointstotheirsuccessfuluse.Ineffect,corporatesharepriceswill maximizewhenthesetworisksareperfectlyreconciled. Liketheproverbial“whale”,largercorporationscanswallow-upsmallerones becausetheyhaveawiderarrayofoptionswhencombiningrisk.Lessriskintheirrevenue streamallowsthemtotakemoreriskinotherareas–especiallyintheirsourcesoffunding. Ontheotherhand,anymiscalculationineitherriskbyasmaller“Jonah–type”company willhaveafar-reachingimpactbecauseoftherelativesizeofthefirm.Ultimately,larger, “predator”companiescanexploitanimbalancebetweenrisksinsmallercompanies- usuallyamatteroftiming. CurtinMathesonScientificwasaqualitydistributorofscientificproductsforover twenty-fiveyears.Afterthefamedstatistician,DrW.EdwardsDeming,reportedonthe precisionqualitycontrolofJapanesefirms,CurtinMatheson’sexecutivesbecamedisciples ofPhilipCrosby,oneoftheearlyandpremieradherentsoftheconceptofquality leadership. Aboomingmarketinhealthcareproductsinthemiddle1980sproducedhigh profitsandsomepricingpower,andCurtinMathesonshifteditsfocusawayfrom industriallaboratoryequipmentandtowardtheburgeoningdiagnostictestingfield.The marketinginitiativewassteadfast:thefirmwouldattempttocarveoutanicheforitself basedonahighlevelofserviceandfastidiousproductknowledge.Sellingtotherapidly consolidatingHMOs(healthmaintenanceorganizations)wouldensurecosteffectiveness andhighprofitability. Bythetimearecessionhitintheearly1990’s,however,competitionhadalteredthe healthcarelandscape.Ashifttohighervolumeandlowerpricesnecessitatedtheclosingof distributioncentersandtheconsolidationofcustomerservice.Cut-backsbecameeven 3 morerampantwhenfearsofnationalizedhealthcaregrippedtheindustry.Pricecontrol wouldreverttoagovernmententityandsalesrepswouldbecompetingforcontractsthat wouldyieldalmostnothing.CurtinMathesonScientificbecamevulnerableintheonearea onwhichtheyconcentrated-healthcare Healthcareproductdistributionhastwocharacteristicsthatmakeitespecially attractiveforacquisitionbylargerfirms.Althoughprofitmarginshadbeenshrinking, significantcash-flowwaschanneledthroughveryhighrevenues;thepriceoftheinventory wasbuffetedbytechnologicalscarcity-bloodanalyzersthatsoldfor$200,000,forexample. Moreover,thedemandforhealthcareproductsissteadyenoughtocushionotherrisk takingventures;eveninadownturn,revenuesarestable. FisherScientificwasanoldnemesisofCurtinMatheson.WhenCurtinMatheson beganconcentratingonthehealthcaremarket,Fishermovedintheoppositedirection, focusingonspecialtychemicalsandindustrialproducts-amovethatlimitedtheir exposuretoanindustrywithdecliningmargins.Bythemiddle1990s,Fisherspottedan opportunity;theflounderingCurtinMathesonwasripeforatakeover.Fishergarnereda loanfromaCanadianbankandpaidCurtinMatheson’sEnglishholdingcompany, Fisons’,approximately350milliondollars.Managementbecamejitteryaboutlosingthe companytheyhadsoadeptlybuilt,evenasatelltalesignquashedanyrumorsabout maintainingtheCurtinMathesonScientific(“CMS”)brandintegrity:theFisherlogo beganappearingoneveryproduct,fromsharpscontainerstobeakersandtesttubes. FisherScientificInternationalwaslistedontheNewYorkStockExchange,butthey werenota“big”playeronWallStreet.Theirstocksoldforaboutelevendollarsashare, whichwasconsideredpaltryinthehyper-inflatedmarketofthelate90s;therewasnothing “romantic”aboutspecialtychemicalsandanalystsmaintainedalowkeycoverageonthe company.However,Fisherhadasolidreputationinthescientificcommunity.Infact,the firmhadbeenaroundsincethelateindustrialrevolutionofthe1800’s,buildingprofitable vendorrelationshipsthathadproducedalongtrackrecordofconsistentsales. 4 TheintegrationofCurtinMathesonintoFisherwentsmoothly.Asaninterested participant,Icouldnothelpbutnoticesomeoddities.Althoughwelaughedathowrapidly Fisherputtheirbrandnameon“our”products,thereseemedtobesomedisconnect betweensalesandoperations.WhilethetwoentitieswerecloselyintegratedinCurtin Matheson,emphasizinganemotional,“Japanese-like”commitmentandunitybetween teammates,theFisherapproachwasveryclinical,likebrokeringacommodity,whichof course,healthcareproductshadbecome.ThemainoperationscenterwasinPittsburgh, butallexecutivedecisionswereconveyedfromasmalltownonthecoastofNew Hampshire,calledHampton.Thedichotomiesposedmorequestionsthantheydid answers. Infact,FisherScientificwasrapidlybecomingastrategically-runfinancial powerhousethatexpertlynegotiatedrisk.Bytheonsetofthenewmillennium,thefirm wasextremelywell-diversified,carryingover250,000items.Fisherhaddivisionsinsafety, healthcare,chemicals,electronicsandevenhadasupplycenterforradioactivematerialat theLosAlamosnuclearfacilityinNewMexico.CurtinMathesonwasjustoneacquisition thatfueledthisdiversification,albeitthelargestatthetime.Byhavingatleastonedivision thatwouldreactfavorablytoachangingeconomyatanyonetime,theriskofFisher’s cash-flowsweredecreased,anditsrevenuebasewasmaintained.Whilemosthigh-tech companieswerestrugglingwithrevenuesofapproximately$250,000peremployeebythe year2000,Fisherhadastreamofabout$340,000inafieldthatwasnotparticularlycapital intensive-distribution. Althoughrevenueswerehigh,Fishercouldnotgeneratethetypeofinternalfunding thatsupportedbothexistingoperationsandaprogramofdiversifiedgrowth simultaneously;marginswerejusttoolow.Thefundingforacquisitionscamefromdebt- alotofit.Byearly2001,Fishercarriednegativeequity.Stockwasneverissuedfor purchases,andretainedearningswereinsubstantial.Ontheotherhand,thevarious 5 integrationsofacquisitionswereexpensiveevenasFisher’slong-termdebttocapitalrate approachedeighty-fivepercent. Cash-flow,however,remainedveryhigh,evenduringtherecessionthatbeganin 2001.CreditorstookonelookatthesizeandvariationofFisher’srevenuestreamandgave themthe“greenlight”.Thefirmrespondedbyrenewingloansatlowerinterestrates, courtesyoftheFederalReserve.Equitywaskepttoaminimum. Ultimately,whenacquisitionsbegantopayoff,thestocksoared,butitdidnotmove onthebasisofsalesorprofits.Thestockbarelymovedatallinfact,exceptforasingle situation:whenanynewsorrumorofanacquisitionoccurred,thestockwouldjumpoutof itsusualstabledormancyandtakeofflikearocket.Sincesomesmallacquisitionoccurred atleasttwiceayear,thestockwasagoodadditiontoanyportfolio;itsonlyvolatilitywas selfgenerating.Inthemeantime,Fisherbegantoparedownitsdebtandissueequity, causingthestocktosoarevenhigher.TheyboughtbiotechsuppliersinSwedenandtest equipmentcompaniesintheUnitedStates.By2006,theyhadmergedwithThermo- Electron,acompanythathadnolong-termdebtwhatsoever.Theyendedupcalling themselves,“Thermo-FisherScientific”(TMO).Andinevitably,theyalsohadthelast laugh-theyseemedpoisedtostartthewhole“process”overagain. Inanutshell,thestoryofFisherScientificprovidesavaluablelessoninmanaging capitalstructure.Fisherhadtwotypesofriskthatwereinpotentialconflict:businessrisk sometimescalledeconomicriskor“operatingrisk”,andfinancialrisk.Businessriskisthe variationinrevenue,costsandoperatingincomethatstemsfromthetypeofindustry; someindustriesreacttoinflation,recessions,foreigncompetition,andothereconomic factorsdifferentlythanothers.Ontheotherhand,financialriskisalmostentirelyself- generated,andstemsfromthevariationinnetincomefromthedecisiontousedebt.In essence,financialriskisexpressedasthepotentialfordefaultingoninterestpaymentsand principal.Itworkstogetherwithbusinessriskthroughthevariabilityofoperatingincome; 6 anadequateandsteadyoperatingincomecankeepfinancialriskverylowbecausethereis lessprobabilityofdefault. FisherScientifictreateditsoperatingcash-flowslikeaportfolio,addingand droppingproductlinesthatwouldmakeitlessrisky.Evenasmarginsdeclined,itsreturn onequity(ROE)increasedbecauseitneverfundedwithitsownmoney.Byfinancingwith debt,butsimultaneouslydecreasingtheriskofitsoperatingincome,Fisherconfiguredthe risk-returntradeoffinitsfavor.Alternatively,thedecisionbyCurtinMathesontofocuson healthcaretotheexclusionofotherdivisionsmadethecompanyatakeoverprospect. Withfewbarriersofentry,theindustryinvitedintensecompetition;marginsdeclined,and thecompanywasleftwithariskieranddepletedcash-flow. CompanieslikeFisherScientificarequiteordinary.Theyneverhavethetypeof sensationalresultsthatmakesthemthedarlingsofspeculators.Theyrarelymakethe eveningnews.Andyet-herewasacompanywhosestockwassellingat$11asharein1997 onlytorisetoapeakof$77eightyearslater.Inthatperiod,itwasonlyabouthalfas volatileastherestofthemarket. Onemisconceptionthatstudentsandinvestorssharealikeisthatabusinessis supposeto“maximize”profits:the“bottomline”mentalityisalmostanendemic archetypeandyetrarelyoccursineconomicbehavior.ImagineacashflowforCompanyA of60,70,65,90,and110.NowcompareittothecashflowofCompanyB:60,60,65,75, and75.Whichwouldyouprefer?Mostpeoplewouldpickthefirstbecausethechanceof gettingahighnumberisgreater.However,theflowfromCompanyBismuchsteadierand byseveralmathematicalgaugesofriskhasabetterrisk-returncharacteristicthan CompanyA’s.Infact,thedifferenceissmall,butmaybecompellingenoughforinvestors tochooseCompanyBasaninvestment.WhiletheaverageinCompanyAismuchgreater thanin“B”(79vs.67),theriskisfargreateralso. Capitalstructuralismisnotaboutdirectlymaximizingprofitsthroughprograms likeanewmarketingcampaignor“zerobasebudgeting”ortheimplementationofnew 7 technology;ittakesafarmoresubtleapproach.Itchoosesacourseofactionfromseveral alternativesthatbalancestheriskofdifferenttypesoffundingwithreturnsthatexceed theircost.Thus,thegoalofminimizingthecostofcapitalisimplementedthroughthe capitalbudgetingprocess;thecostofthemixtureofdebttoequitywilldeterminethe plausibilityofeachprojectbecauseofthenecessityofexceedingcapitaloutlayswith returns.Theloweristhecostofcapital,thegreaterthenumberofprojectsthatwillbe potentiallyprofitable. Riskandreturnaresointertwinedthatitispropertorefertothemasastatistical “distribution”withtwoparameters,ratherthanasseparatecategories.Asanexample, consideranequityissue,themarketingofmoresharesofstocktoraiseadditionalfunds. Thecharacteristicsofriskandreturnforsuchanissuearemuchdifferentatthebeginning ofarecoverythanattheendofabullmarket-forboththeissuingcompanyandthe investor.Althoughtheinvestorisnotencouragedto“time”investmentsovertheshort term,someawarenessofthecorrelationbetweensectorperformanceandthebusinesscycle isessential.Capitalstructureisdependentontherelationshipbetweeninterestratesand theequitymarket,whicharedependentonthestateoftheeconomy.Therefore,timeisa unifyingfactorbetweenriskandreturnandencouragestheirinterdependence.The investorisleftinaprecariousposition.Ontheonehand,heorsheisencouragednotto timethemarketbecauseitisnotsuccessfullydoneoveralongperiod.Ontheotherhand, timeistheessentialcomponentinallrisk-returndistributions-frominvestmenthorizons tothechoiceofwhichinvestmentstomake.Byidentifyingandinvestinginfirmswho repeatedlymovetowardtheiroptimalcapitalstructures,capitalstructuralismresolves someofthisconflict. Whataboutvariation?Randomfluctuationisthebaneofanyanalyst.Nomatter howpreciselyonemeasuresthedeviationsinpastperformance,currentandfuture behaviorofaninvestmentseemstodefyformulation.WhileWallStreetprizescertainty, long-termviabilityisnevercertain.Themarketkeepschangingandtheresponsetoworld 8 eventsisembeddedincorporategainsandlosses.Capitalstructuralismisflexibleenough toencompasschangebecauseitneverdefinesoptimalityasarigidsetofconditions.Each industryhasaparticularresponsetoeconomicfactorsthatproducesadifferentoptimum levelofproportionaldebtandequity.Someindustrieshavebetterrisk-return characteristicswithoutanydebtatall.Otherscancompetewithfirmsthathavetwoor threetimesitsprofitmarginssimplybecausetheyknowhowtousedebtjudiciously.Since capitalstructureisdependentonthebusinesscycle,itrespondstosocietaltrends, demographicchangesandpoliticalrisksbetterthanthevarious“systems”thathavemade theirwayintotheinvestmentliterature.Ineffect,capitalstructurereflectsthereasons whyacertainentityisinbusinessinthefirstplace:togrowandmakeaprofit. Ultimately,ouranalysisattemptstoputadollarpriceonrisk.Whilethemarket respondstoinformationinstantaneously,weattempttomeasureitscontentbeforeit becomesmeaningful.Wecandefinecostinthreedifferentways,allofwhichareusedto evaluaterisk: • 1)TheNominalCost-Thisisthe“upfront”,accountingcostofanactionwhichwillbe reportedinfinancialstatements • 2)The“Real”Cost-Thisisthecostofanactionwitheconomicconditionsfactoredin. Ifmynetincomeis$100andtheinflationrateisfivepercent,thenmy“real”net incomeisprobablyonly$95.IfIhavetax“lookbacks”of$20figuredintothat$100, thenmyeffectivetaxratewasreducedandIwillhavetomakeamuchgreaternet incomeinthenextyear. • 3)TheRiskAdjustedCost-IfIkeepallofmymoneyinacheckingaccountwhenthe marketisrisingbyfifteenpercentayear,Iwillbepenalizedfornotputtingmore moneyintothemarket.Theriskadjustedcostisthecomparativecostoftakingone actionoveranother,creatingeitheragain(opportunitygain)oraloss(opportunity loss).Itismostrelatedtowhatcanbetermed,“thegoingmarketprice”.Incapital structure,thisriskadjusted,“opportunitycost”ismoreimportantthananyother 9 becauseitlooksatanarrayofalternativesandchoosesacourseofactionthatattempts tocreatethelargestpossibleopportunitygain. Therefore,manyofthecostsweincurincapitalstructurearenotrepresentativeofa physicalassetandpassedonfromapreviousowner,butaretheresultofachoiceofactions withwhichwehavecomparativeinformation. Theintegrativeapproachofthistextistopositiontheanalyst,theinvestorandthe financialmanagerfromthesameviewpoint:heorshegaugestheriskofoperatingcash flowsandbalancesthatobservationwiththechoiceofalternativesourcesofcapital, repeatedlymakingcomparisonsbetweentheindustry,thesector,andthegreatereconomy. Underthepremisethatcapitalstructureistheinterfacebetweencomparativeaccounting andthemacroeconomy,thestudentreceivesanoverviewofcorporatefinancethroughthe attemptedreconciliationofriskandreturn.Ineffect,thedifferencebetweenstudent, investor,analystandmanageriscloudedbecauseeachperspectiveisdirectedbytheneed toseekanddiscoveroptimality. Thetextrequiressomefamiliaritywithstatisticsandcomputerspreadsheetsbutnot anextensivebackgroundineither.Thereisachapterdedicatedtostatistics,andmost spreadsheetshavestepbystepinstructions.Theflowofthetextisasfollows: • 1.TheoreticalBackground:Capitalstructuretheoryisexaminedthroughprevious researchwithanemphasisonintegratingtheevaluationofriskandreturn. • 2.ModelBuilding:Amathematicalconceptionofcapitalstructureisbuiltthrough computermodelsandtheadaptationofexistingformulas.Eachofthespreadsheet modelshasbeenusedtoevaluatecorporatebehavior. • 3.CorrelationStudies:Examinationoftherelationshipsbetweenstockpriceand capitalstructurevariablesgivesinsightintothebehaviorofsomemajorcorporations. Whilenodefinitiveconclusionsaredrawn,tendenciesthatsupportcapitalstructure theoryareexaminedusingtheSpearmanrankcorrelation. 10 • 4.CaseStudies:ApplicationofcapitalstructureanalysistoKimberly-Clarkand ConocoPhillips,aswellasMicrosoftandChevrondisplaytheeffectivenessofthe techniques. Oneofthegreatphilosophersoftheearlytwentiethcentury,WilliamJames,might haveappreciatedthepersonalcomputerrevolution.Hewhochampioned“thecashvalue ofideas”andthephilosophyofpragmatismmighthavefoundsolaceinamachinethat testedtheviabilityoftheory.Whileweoftenlackthepoliticalframeworktoimplement ideas,atleast“theinformationage”hasmadethemavailable,whichiscertainly“halfthe battle”. Forstudents,theauthorhopesthatthisbookwillunifyfinancialthoughtintoa comprehensible“whole”andencouragetheactualizationof“justtheory”.Forinvestors, theauthorhopesthatthisbookwillhelpthemseebeyondthesuperficialityofconventional wisdomwiththeknowledgethatthecashvalueofanyideaisalmostalwaysfoundinits underlyingstructure.Finally,fortheexecutive,theimperativeisplacedoninnovative thinking:atime-testedsolutionistheoutgrowthofanewperspective. (BacktoTableofContents) 11 SECTIONI:THETHEORYOFCAPITALSTRUCTURE 2 LEVERAGE Inafundamentalsbased“bottomup”analysissystem,theincreaseandacceleration ofsalesisparamount.Theaxiom,“nothinginbusinesshappenswithoutasale”appears selfevident.However,capitalstructuralismoftenseemstodenytheneedforgreater returnsbyfocusingonrisk,eventotheexclusionoflarge,unevenstreamsofincomethat mightupset“corporateequilibrium”.Thisfrictionbetweenmarketingstrategyand absoluteriskisreconciledbyastrongadherencetotheprinciplesofleverage. Capitalstructureanalysisadjustsforboththeamountandvariabilityofsalesby firstevaluatingoperatingincomeasafunctionofsales,andsecondly,bychoosingthe amountoffundingfromseveralalternativesourcesbasedontheriskofthisevaluation. Sinceitmakescomparativechoicesfromamacroeconomicperspective,capital structuralismisa“topdownapproach”;businessriskandcreditavailabilityput restrictionsonallavailablechoices.Forexample,achoicetoaddanewproductlinemay notcometoimmediatefruitionforacompanywhopayshighinterestratesandhas excessivedebtonitsbalancesheet.Whiletheideasthatgeneratehighreturnsoftencome fromadetailedmarketingplanthatformsafoundationforthebusiness,therisksincurred bytheplanareoftenimposedfromabove:governmentregulations,competitor’sactions, andthefluctuationsinherentinatypicalbusinesscycle.Thegistofcapitalstructure analysisistoresolvethisconflictbetweenwhatcanbeproducedandwhatwillbe produced,byreconcilingriskwithreturn.Accordingly,thetoolstomanagethisresolution areencompassedbytwodistinctmeasurements:operatingleverageandfinancialleverage. 12 BUSINESSRISK Operatingleverageisonemeasureofbusinessriskalsoknownaseconomicrisk.As anexample,considertheattributesoffarming.Tostayinbusiness,thefarmermustbe concernedaboutthecostofseed,irrigation,storage,andtransportation.Demandforhis orhercropisdependentonweather,foreigncompetitionandtheavailabilityofsubstitutes. Thevariabilityofinputs(costs)andoutputs(demandandquantityproduced)form economicrisk.Highfluctuationsindemandoftencauselargeswingsinthepricesthata farmercancharge.Whensuppliers’pricesalsovary,thedoubleedgedswordcreatesan environmentofhighbusinessriskWithoutanyideaofhowmuchtopayvendorsorhow muchtochargecustomers,planningmustbetotallycontingentontheunexpected,an immediatebarter-likenegotiationwhereuncertaintyprevails.Littlegrowthwilloccurin suchanenvironmentbecausenoinvestorwantstocommitcapitalwithoutconfidenceina minimumreturn. Manyeconomistsbelievethatbusinessriskisareflectionoftheleveloftechnology inanindustry.Becausefixedcostsmustbepaidregardlessofthelevelofdemand,higher fixedcostsimplythatmorebusinessriskisincurred.Whencompetitivepressuredemands thatspecificqualitystandardsaremet,thosestandardsareanoutgrowthofthelevelof technologyrequiredbytheindustry.Fixedassetsthathavelongdepreciablelivesarevery costly,butnecessarytomeetthesecompetitivepressures.Considerforamoment,the shrinkwrappingonaCD.WouldacustomerbuyahandwrappedCDwhentheindustry standardistowrapit“astightasadrum”?Moreover,addingfixedcoststoanyoperation raisesthebreakevenpointforsales,evenwhenthetotalcostisthesame.Oncethe percentageoffixedcostsisincreased,moresalesmustbegeneratedtocoverthem. However,whenanoperationhasahigherproportionoffixedcosts,andsalesareadequate, moreunitsofproductionwillbespreadamongthesameamountofcosts;theresultisa higheroperatingprofit.Thissinglekernelofcorporaterisk,affectsallotherelementsin 13 thechain:demandschedules,variabilityofincome,theprobabilityofdefaultandthe methodsandsourcesoffundingprojects. FINANCIALRISK Financialleverageisonemeasureoffinancialrisk,whichistheriskincurredbya firmforitsdecisiontousedebtfinancing.Companiesfaceachoiceoffundingprojects withequity(retainedearnings,commonstockandpreferredstock)ordebt(bonds,bank loans,commercialpaper).Whendecidingtoincreasetheamountofdebt,thefirm increasestherisktoexistingshareholdersbecauseearningsbecomepartiallychanneled towardcreditorsintheformofinterestpayments,andawayfromthepotentialforhigher dividends;thevariabilityofincomeisincreased.Inreturn,shareholdersreceivethe possiblerewardofhigherearningsonapersharebasisbecausefewershareswillbe outstandingwhendebtisusedinplaceofequity.Consequently,thefirmincreasesits chanceofbankruptcywhenitincursmoredebt;itcandefaultoninterestpaymentsif earningsarenothighenoughtocoverthem.Thisriskcanbedecomposedintotwobasic elements:1.Theamountofpotentialloss-theclaimsthatcreditorshaveonafirm.2. Theprobabilityofloss-acomplexinteractionbetweensales,earnings,andliabilitiesthat determinessolvency. LEVERAGE:ADEFINITION Ifthechoicetotakeondebtsoundsdire,thestudent/investorwillturnthisdecision intoaprofit-makingventurebydeterminingthecrucialdifferencebetweenstrategyand obligation.Firmsthatareobligedtoincreasefinancialleverageinordertocushionpoor demandhaveradicallydifferentcharacteristicsfromthosewhooptimizecapitalstructure. Infact,manywell-runfirmslowertheiroverallriskbecauseofthechoicetousemoredebt; theriskentailedbythecostofhigherinterestpaymentsismuchlessthantheprobabilityof newcash-flow.Indeed,the“primerate”issetlowenoughtoattractthebestcustomers withoutburdeningthemwithworriesofinsolvency. 14 Ifwethinkofleverageasaproportionoftwodifferentcomponentsofthesamerisk, eachseekingtobalancetheother,wecanformageneraldefinition.Inphysics,asmall forceappliedatonepointcanbalanceorcontrolamuchlargerforceatanotherpoint.A child’ssee-sawistheclassicexampleofthisprinciple:whenafiftypoundchildbalancesa twohundredpoundadultandthenjumpsoff,theadultdropswithathud.Ifweviewthe smallerforce(thechild’sweight)asthedenominatorofaratio,andthelargerforceasthe numerator,itissimpletoobservehowachangeinonecomponentaffectsthechangeinthe other,dependingontheirrelativeamountofassociation.Inafinancialcontext,wespeak of“leverage”whenasmalleramountofonevariablehasalargereffectontheother.In mathematicalterms,weputthe“derivedcomponent”inthenumerator,andthe“source component”inthedenominator,anddeterminethechangeinboth.Inourexample,the child’sweightwasthesourcecomponent,whichhadanexaggeratedeffect(derived)onthe adult’sweight. Ineconomics,weusuallyviewleverageintermsofinputandoutput,butinfinance, weaddtheelementofconnotativerisk:welookforotherassociationsthatthelevelof leveragemayaffect.Forexample,ifwediscovera“new”laborsavingmethodinwhich twopeoplecanaccomplishthesameamountofworkastwenty-two,themethod undoubtedlyhasalotof“leverage.Ofcourse,leveragealmostalwaysexactsaninherent “cost”andinourexample,thetwolaborerswouldatleasthaverisingexpectationsabout wages,ifnotactualdemands.Secondly,sinceeachremainingpersonismoreresponsible fortotalproduction,moreriskisinvolved;losingonepersonmaycutproductioninhalf. Therefore,leveragealwaysimpliessomerisk-returntradeoff,whichneedstobeidentified. Leveragecanonlybeincreasediftheriskshavebeenfullyvetted. Whilethereareothermeasuresofriskbesidesleverage,fewdisplaytheintegration ofriskandreturnbetterthanthebalancebetweenfinancialandoperatingleverages.In fact,capitalstructuretheoryisfoundeduponthisintegration:behindeverystrategic 15 decision,thatchangesthepriceofastock,layssomethreadofleverage.Eachtimethat capitalisallocatedforanygivenproject,ultimateprofitabilitydependsonleverage. BASICRISKSANDPROPORTIONS WallStreetdoesnotlikeuncertainty.Ifthereisonequalitytocultivateintheworld offinanceitisconsistency;whenamarketiscoherent,thefinancialcommunitycanmake plansaroundexpectationsandpredictions.Leverage,however,impliesvolatilityanditis whentwodifferenttypesofvolatilityaremixedthatalevelofreturnisderived.Cash,for examplehasalmostzerovolatility,andverylittlereturnwhenkeptinthatform.Atthe otherextremearecertaincommoditiesthatcanskyrocketovernight,onlytoleaveafutures ownerpoorerafewweeksout.Operatingleverageisconceptuallymeasuredas%∆ OperatingIncome/%∆Salesandimpliestheinherentvolatilityofachangeinsales creatingachangeinEBIT(earningsbeforeinterestandtaxes).Whenoperatingleverageis large,moreriskisincurred;thepossibilityofhighprofitsisgreaterinanupturnwhen salesarelarge,butinadownturn,withlowerdemand,profitsarejeopardized.Lower operatingleveragecreateslowervolatility,butalsolessprofitwhenapositiveshiftin demandoccurs. Thesamethemeofvariabilityappliestothefinancialleverageratio,whichis theoreticallydefinedas,%∆NetIncome/%∆OperatingIncome(EBIT).The student/investorwillnotethattheterm“debt”isnowheretobefoundintheratio,butthat interestpayments,aswellasanytaxadvantagessuchpaymentsentail,areimplicitwhen operatingincomebecomesnetincome.Again,thederivedcomponent(netincome)is affectedbyachangeinthesourcecomponent(operatingincome).Bothtaxesandinterest paymentsmustbedeductedfromEBITbeforeitistermed“earnings”,anditisthe magnitudeofthesedeductionsthatwillcausevariability. Areferencethatconfusesbothinvestorsandstudentsalikeisthatsomeacademic literaturedesignatesseveraldifferentratiosbytheterm,“financialleverage”.Among thoseproportionsare:Assets/Equity,Long-termdebttocapitalandEBT/EBITwhichis 16 earningsbeforetaxesdividedbyearningsbeforeinterestandtaxes,andisanintegralpart ofanycapitalstructureanalysissystem.Thecommonpointineachoftheseratiosisthat theyrepresentsomeformofdebtfinancing,althoughsomespecificallymeasureinterest expenseandothersmeasureaparticularliabilityorcategorythereof.Weusetheratio,% ∆NetIncome/%∆OperatingIncome(EBIT)becauseitbestexpressestheconceptual integrationofriskandreturnwhencombinedandbalancedwithoperatingleverage.In fact,wecanformameasureoftotalriskifwemultiplythetwoleveragestogether,(%∆ OperatingIncome/%∆Sales)x(%∆NetIncome/%∆OperatingIncome),which convenientlybecomes(%∆NetIncome)/(%∆Sales)Obviously,thepremiumforWall Streetistohaveashighalevelof(%∆NetIncome)/(%∆Sales)-totalleverage-as possiblewithouttheinherentvolatilitysuchalevelimplies. Leverageisalwaysameasureofpotentialvolatility,whichcanbedampenedby maintainingasteadyleveragefigure.Forexample,ifoperatingleverageisnormally3/1. whichisconsideredhigh,operatingincomereactsheavilytochangesinsales.However,if thesamechangeinsalesisenactedyearafteryear,investorswouldbeoverjoyedbythe stability;theywouldknowexactlywhattoexpectandwhen.Theproblemwithhigh leverageisitsinabilitytoadjustitselftoachangingeconomy.Demandcyclesinvariably change,andwhentheydo,thehigherleveragefirmssufferthemost-morefixedcosts entailthenecessityofhighersalestocoverthem.Whentheeconomylagsandsalesabate, fixedcostsmuststillbepaiddespiteaslowerproductioncycle. BALANCINGLEVERAGE Sinceincurringdebtimpliestheriskofdefault(failuretopayinterestexpenseina timelymanner),asteadiersourceofincomeallowsafirmtoeitherengagemoredebt,or maintainitsexistingdebtwithlessrisk.Thesourceofincomeisevaluatedforitsamount andconsistency,whiledebtisevaluatedforthesizeofinterestpaymentsandtheamountof principal,especiallyinrelationtoassetsandnetworth.Thebalancebetweenthe comparativeamountsofincomeanddebtarethengaugedintermsofriskandreturn. 17 Aproperbalanceofleveragewillyieldtaxbenefitsbecauseinterestisatax deductibleexpense.Secondly,thereisthepotentialforariseinbothearningspershare andsharepricebecausefundingisachievedwithlesssharesofstockoutstandingthanif donethroughanequityissue;moreincomeisspreadoverfewershares.Thirdly,debtcan shiftthebalanceofcontrolawayfromshareholdersandtowardcreditorsinalegalfashion, i.e.,moredebtcanmakeacompanyprohibitivelyexpensiveforatakeoverandevenshift controltoasupportiveinvestoror“whiteknight”incaseofahostileattempt. Themainriskofleverageistheriskofdefault,whichdetractsfromafirm’s viabilityinseveralways. • 1)Theprobabilityofdefaultaffectsthecostoffuturefinancingbyincreasingboth interestratesandthecostoffloatinganequityissue. • 2)Defaultmayendangerdividends,andrestrictivecovenants(bondcontracts)may limittheirgrowth. • 3)Creditorsmayhaveaclaimonassetsandrestrictincomepotential. • 4)Theprobabilityofdefaultaffectsmarketvolatilityforthestock Addedtotheriskofcreditdefaultistheriskofincomevariability,whichisafunctionof payingoutinterestexpense.Thetworisks,businessriskandfinancialrisk,are inseparable:boththeamountofdebtandthechanceoflossarepredicatedonasteady operatingincome.Anydownturninsaleswouldbemagnifiedwithahigheroperating leverage,creatingvolatilityinearningsandmakingthefirmmorelikelytodefaulton interestpayments. THEANATOMYOFFINANCIALLEVERAGE Thevolatilityofnetincomeisanaturaloutgrowthofitsdependenceonboth operatingincomeandtheamountofinterestexpense.Keepingoperatingincomeconstant andchanginginterestwillincreasethevolatilityofchangesinnetincome.Whatisnot apparentfromthefinancialleverageratio,%∆NetIncome/%∆OperatingIncome,is 18 thatEPS(earningspershare)canchangebecauseofexogenousfactors-themanipulation ofthenumberofsharesoutstandingbyfinancialmanagement. Ifthemantraoffinancialleverageistodo“morewithless”,itispartiallyachieved byincreasingthepotentialvalueofeachsharebyissuinglessofthem.Wheninterestand taxesareheldconstant,thevariabilityofnetincomeremainsconstantaswell.Theability tolimitsharesaddsanewlevelofvolatilitytofinancialleverage.Ineffect,netincomethat isnormallyderivedfromthedeductionofinterestandtaxesfromoperatingincome,is spreadoverfewershares,increasingboththereturnandthevolatilityofthatreturnona persharebasis.Netincomemaynotvarymorethanitdoeswithoutdebt,butlimitingthe amountofshareswillmakethepersharefiguremorevolatileandmoreprofitable. Toillustratethisconcept,considerthefollowingflowsofoperatingincome,thefirst groupwithnodebt,andthesecondgroupwith$20ofinterestexpense.Taxesare30%in bothgroups. Table2-1 NODEBT EBIT 90 100 110 120 130 TAX 27 30 33 36 39 NET INCOME 63 70 77 84 91 Table2-2 WITH DEBT EBIT 90 100 110 120 130 INTEREST 20 20 20 20 20 TAX 21 24 27 30 33 NET INCOME 44 56 63 70 77 19 Wenowmeasurethemeanofeachflowandusethesamplestandarddeviationasa measureofrisk: Table2-3 ACTUALFLOWOF FLOW1(NODEBT) FLOW2(DEBT) NETINCOME 63 44 70 56 77 63 84 70 91 77 MEAN 77 63 SAMPLESTD. DEVIATION 11.07 11.07 Althoughthedebtladenflowhasasmallermean(63)thanthenodebtflow(77),its essentialriskisthesame-11.07-asanabsolutevaluefornetincome.Itmaybeless preferablebasedonthecoefficientofvariationwhichmeasuresthestandarddeviation dividedbythemean(11.07/63)vs.(11.07/77)-thelowerthefigure,thebetter.However,if weproceedtoobservethefourseparatechangesinnetincomeforeachflow,wefindthat thedebtladenflowismuchmorevolatile,butgrowsatafasterrate: Table2-4 YEARTOYEAR CHANGES FLOW1(PERCENT) FLOW2(PERCENT) FLOW1-NODEBT 11.1 14.29 FLOW2-DEBT 10 12.5 9.09 11.1 8.33 10 MEAN 9.63 11.97 SAMPLESTD DEVIATION 1.194 1.853 Nowassumingthattensharesareoutstandingforthefirstflow,whilefiveareoutstanding forthesecond,wehaveEPScalculationsforeach: 20 Table2-5 EARNINGSPERSHARE FLOW1(NODEBT) FLOW2(DEBT) 10SHARES-NODEBT 6.3 9.8 5SHARES-DEBT 7 11.2 7.7 12.6 8.4 14 9.1 15.4 MEAN 7.7 12.6 SAMPLESTD DEVIATION 1.106 2.24 Notethatbothmeanandstandarddeviationincreaseonlessincomeinflow2thaninflow 1.SinceEPSbehaviorisoftenaproxyforsharepricebehavior,thepotentialtoincrease priceisbalancedbytheriskofvolatilemovementinbothdirections.Anypoorlymanaged capitalstructurewillhavetoomanysharesoutstandinginadditiontodebt,causinga higherriskofdefaultanddilutionofEPSatthesametime.Suchacombinationmoves companiesawayfromtheoptimalproportionofdebttoequity.Sincecreatingequity reducessomevolatility,manyfirmsadheretothecreedthat“moreisbetter”,especiallyif theequityisbeingusedasabargainingchipforexecutivecompensationorforacquisitions. However,theproportionofdebttoequitymustbeidealinorderforthestocktorise,and manyofthesefirmsissueequitywithoutsuchforesight. THEANATOMYOFOPERATINGLEVERAGE Likethefinancialleverageratiowhereavariabilityfactor(interest)causesvolatility inthenumerator(netincome),operatingleveragealsocarriesanimplicitandundisclosed variabilityfactor-fixedcosts.Inthedegreeofoperatingleverage,%∆EBIT/%∆Sales, itisthechangeinfixedcoststhatcontributetoearningsvolatility.Inessence,bymaking costslessvariablewithsales,afinancialparadoxoccurs:steadier,“fixed”costscontribute moretovariabilitygiventhesamelevelofsalesandoperatingincome.Higherfixedcosts requireahigherlevelofsalestobreakeven,butvariablecostsareabsorbedbythesaleof 21 eachadditionalunit.Moreover,anybusinessfluctuationswillhaveanexaggeratedeffect onthosefirmswithhigherfixedcosts;profitsmaysurgeduringanupswingbutfixedcosts muststillbepaidduringadownturn,makingitmoresevere.Perhapsthebestexampleof oversensitivitytothebusinesscycleoccurswithbiotechcompaniesthatspendalotfor researchanddevelopment,havelittledebt,andseehighprofitsinanexpansionbutnear insolvencywhenadownturnoccurs.Thevariabilityisasmuchafunctionofthetypeof costasitisoffluctuatingdemand. Thefundamentalkeytounderstandingoperatingleverageistorecognizetheeffects ofincreasingfixedcostsasapercentageofsales.Companiesoftenwanttoautomate remedialtaskstoreducelaborcostsortoaddaspecialcompetitivequalitytotheproduct. Besidestheinherentcostofmachinery,someoperatingcostsmustbepaidevenwhen nothingisproduced(insurance,storage,maintenance).Thus,acompanymustincrease salesbyaspecificamounttocoverthesecostsevenifthevariablecostperunitof productionremainsunchanged.Thetrueriskofincreasingoperatingleveragestemsfrom thepossibilityofnotincreasingthebreakevenpointforsales.Anexampleofhowthis worksin“breakevennotation”isasfollows: Table2-6 VARIABLE SYMBOL VALUE SALES S 1000 PROFIT PFT 300 FIXEDCOSTS FX 100 VARIABLECOST VC 600 BREAKEVEN EQUATION PFT=S-VC-FX 300 Inlaterchapters,wewilldecomposesalesintotheproduct,(PricexQuantity).To illustrateanincreaseinfixedcosts,whilemaintaininga60%variablecostproportion(that is:VCremains0.6S),wecanpluganincreasedirectlyintothebreakevenequation.We 22 willincreasefixedcoststo200,aonehundredpercentincrease.PFT=S-VC-FXor1000- 600-200=PFT=200.Thesmallerprofitcanberaiseduptoitsoldlevelbyalgebraically solvingforsales: X-(.6X)-200=300.Salesisthesymbol“X”and0.6(X)isvariablecosts.SolvingforX, weobtain1250asthenewlevelofsalesneededtomaintaintheoldprofitlevelof300.A onehundredpercentincreaseinfixedcostsrequiredatwenty-fivepercentincreaseinsales toabsorbit.Noticealsothat0.6(1250)=750.A$100increaseinfixedcostsinstigatedan increaseinvariablecoststo$750,andalsoadecreaseintheratioVC/FXfrom“6” (600/100)to“3.75”(750/200). Onamicrocosmicproductionlevel,eachrevenue-generatingprocedureinaplant hasitsownunit-dependentoperatingleverage.Inthiscase,variablecostsarespreadover eachunitproduced,andsalesareafunctionofthepriceoftheunitmultipliedbythe quantityoftheunit.Fixedcosts,however,arenotunitdependent;theyareindependentof thelevelofproduction.Atthislowestlevel,operatingleverageisrelativelystablebecause mostprocessesarestandardized.Itisintheraretimeswhenfixedcostschangeasa percentageofsales(higherrent,moresalariedemployees,etc.)orasnewtechnologyis added,thatoperatingleveragechanges. Onamacrocosmiccorporatelevel,operatingleverageismorevolatile.Newproduct linesandprocessesareconstantlychangingoperatingleverage.Anacquisitionofaservice typebusiness,forexample,willprobablyloweramanufacturer’srisk.Ontheotherhand, theswitchfromdistributinganitemtomanufacturingitrequiresanentireshiftfrom formermethodsoforderingandstorage,aswellasthepurchaseofcapitalequipment.The higherprofitsthatarecitedwillbeaccompaniedbythehigherriskofmorefixedcosts. Ultimately,anytimethatasegmentofthebusinessisdiscontinuedorchanged,thereisa risk-relatedeffectinvolved–somechangeinoperatingleverage. TOTALRISK 23 Thehigherleveloffixedcostsassociatedwithmoreoperatingleveragerequires extensivecapitalfunding.Anyprocessthatneedsmoremachineryalsoneedsareliable sourceofsteadyfinancingtoreplaceandrepairequipmentandfundnecessaryshiftsin production.However,creditorsdonotwanttoextendloanstocompaniesthatevince incomevariability.Theydesireacustomerwhohassteadycash-flowandisnotsensitiveto businesscyclefluctuations.Thus,thereexistssomeleveloftotalrisk,theproductof operatingandfinancialleveragesthatdeterminesthesourceoffunding.Iftotalriskistoo high,oroperatingleveragebyitselfishigh,creditorswillonlyextendloansathighinterest ratesornotatall.Thefirmswiththehighestoperatingleveragesendupfinancingwith retainedearningsorequityissues,whichputsthemmoreatriskduringadownturn. Theleveragedbuyoutisperhapsthebestexampleofhowthetwoleveragesinteract. Theadditionofcompanieswithlessoperatingriskisamethodofdiversificationthatworks intandemwithfinancialleverage.Iftheriskofdefaultcanbeminimizedbyspreadinga parentcompany’sfixedcostsovermoreunitsofproduction,moredebtcanbeincurred (duetolowertotalrisk),andlessequitysharesissued.Theresultwillbeahighershare priceforthe“merged”company.Inshort,thebrokersofaleveragedbuyoutuselessof theirowncapital(equity)andallowthecorporationtoassumethe“limitedliability”of greaterdebt. LEVERAGEMEASUREMENTS Usingconventionalmethods,itisdifficulttoobtainarealisticoperatingleverage figurefromthefinancialstatementsofacompany.Notonlywillproductionchanges obscurethe“true”number,butaccountantsoftenhavedifficultyattributingfixedand variablecoststospecificunits:manycostshaveboth“fixed”and“variable” characteristics.Additionally,thefinancialleverageratioisinherentlyunstablebecause likeoperatingleverage,wearemeasuringtheratiooftwochanges:meaningful measurementisdifficultwhenlargevariabilityisincurredthroughtheuseofpercentage changes.Toanalyzecapitalstructureinthiscontext,weattempttofindaconcreteproxy, 24 afigurethatwillmirrortheanalyticalvalueoffinancialleveragewhileremainingstable enoughtobufferthevolatilityofchangesinoperatingleverage. THEORYVS.REALITY:FINANCIALLEVERAGE Thefinancialleverageratio,%∆NetIncome/%∆EBIT,hasaconcrete counterpartthatismoreamenabletodirectmeasurementwhichis,EBIT/(EBIT-Interest Expense).Thestabilityofthisratiomakesitidealforcomparison,anditcaneasilybe convertedtoatimesinterestearned(TIE)ratiowhichisusedtocalculatedefaultratings forbonds. Theuniquecharacteristicofthefinancialleverageratioisitsdualcapacity:itisnot onlyacomparativeratingtoolfordefault,butitpredictsthepressureonearningsandthe relationshipbetweennetandoperatingincomes.Wheninterestexpenseremains unchanged,theconcreteratiopredictstheexactearningsfiguresoneyearhence.Toview howthisworks,examinethefollowing: Inyear1,operatingincome(EBIT)is100,interestexpenseis20andthetaxrateis30%. Thus,thefinancialleverageratiois100/(100-20)=1.25.Toconfigurenetincomeinyear 1,wesubtract20from100=80,whichwemultiplyby(1-taxrate)toobtain56[80(0.7)= 56].Foryear2,wepickatrandomanoperatingincomeincreasefromyear1,say40%, makingyear2’sEBITequalto140.Thenetincomecalculationis140-20=120,and120x (1-0.3)=84.Theincreaseinnetincometooperatingincome,%∆NetIncome/%∆ EBITis((84/56)-1)/((140/100)-1)=50%/40%=1.25.Thus,bykeepinginterestexpense constant,netincomebecomesfullypredictable! THEORYVS.REALITY:OPERATINGLEVRAGE Thetheoreticalrelationshipofoperatingleverage,%∆EBIT/%∆Salesalsohasa concretecounterpartthatiscalculablewhentheanalysthasfullknowledgeofassigned costs.Theratiois:(Sales-VariableCosts)/(Sales-VariableCosts-FixedCosts).Since investorsrarelyhaveaccesstosuchspecificcostbreak-downsbetweenfixedandvariable, itisalmostunusableonthatlevel.However,onacorporate“needtoknow”basis,itcanbe 25 utilizedincombinationwiththefinancialleverageratiotoformameasureoftotalrisk,and toanticipatechangesinEPS.Thestudent/investorshouldnotethatthenumerator(sales- variablecosts)isreferredtoasthe“contribution”whilethedenominatorisactuallyEBIT, restatedwithitscomponentparts.Alsonotethattheonlydifferencebetweennumerator anddenominatorisin“fixedcosts”. Besidesmeasuringoperatingrisk,theconcreteversionofoperatingleveragealso hasapredictivecapacity:giventhatvariablecostsareastablepercentageofsales,and fixedcostsremainunchanged,theratiowillpredict%∆EBIT/%∆Salesexactlyoneyear intothefuture.Thefollowingexamplewillexhibitthisrelationship: Inyear1,theHardseatBicycleCompanyhas1000insales,variablecoststhatare0.6times salesandfixedcostsof100.Determinenextyear’sEBITiffixedandvariablecostsremain stable. Year1OperatingIncome(EBIT)isSales-VariableCosts-FixedCosts=300.Operating leverageis:(1000-600)/(1000-600-100)=1.33 Year2willyieldachangeofEBIToverthechangeinsalesof1.33,nomatterthelevelof sales.Ifwepickasalesincreaseatrandom,say67.6%,thefollowingvaluesapply: Table2-7 YEAR2 SALES=1000(1.676)=1676 VARIABLECOSTS=0.6(1676)=1005.6 FIXEDCOSTS=100 EBIT=S-VC-FX=1676-1005.6-100=570.4 CHANGEINEBIT=(570.4/300)-1=90.1333% CHANGEINSALES=(1676/1000)-1=67.6% 90.1333/67.6=1.33INOPERATINGLEVERAGE Thus,byapplyingoperatingleveragetoaknownchangeinsales,operatingincome becomesfullypredictable! THEORYANDREALITY:TOTALLEVERAGE 26 Sincerealitydictatesaconstantlychanginginterestexpense,fixedcostpercentage andvariablecostrate,analystsarehardpressedtomakepredictionsfromleverageratios. However,eachratioaccuratelypredictsthepressureonearningsifthestatusquois maintained.Whencomparedtoindustryaverages,theratioscangaugerelativeriskand exhibitthepressuretoconformtothosestandards.Ifmultipliedtogether,weproduce% ∆NetIncome/%∆Sales,whichappearstobeadynamicversionoftheclassicratio,profit margin(NetIncome/Sales).Thefullyconvertedequationis: (EBIT/(EBIT-InterestExpense))x((S-VC)/(S-VC-FX))= %∆NetIncome/%∆EBIT)x(%∆EBIT/%∆Sales)= %∆NetIncome/%∆Sales Indeed,marginsandleveragearecloselyrelatedandmoreleveragewillcontributetoa largermargin,butthetwoshouldnotbeconfused.Leverageistheprecursortoamargin becauseitexhibitsthedynamicmovementnecessarytochangeit.Whenmanagement attemptstohaveacontrollabletotalleverageratio,predictingthenextearningscycle becomesaremedialequationaslongassalesareforecastcorrectly: (OldNetIncome)x(1+(TotalLeveragex%∆Sales))=(NewNetIncome). Conceptually,morenetincomecanarisefrommoreleverageorgreatersales,butthe leveragecomponentscannotviolateindustrystandardsordysfunctionwilloccurthat upsetsafirm’sequilibrium-toomuchdebtforthelevelofincome,costoverruns,lackof capacityutilizationetc.Infact,manystockrun-upswilloccurpreciselybecausea companyissuccessfullydefyingtheoddsandnotsuccumbingtothenegativeassociations thatoccurwithtoomuchleverage.Whilemeetingindustrystandardswilldetermine“ball park”figuresfortheleverageratios,eachmanagementteamhasauniqueflexibilityin changingthemasaresponsetocompetitivepressure.Forexample,ifindustrystandards foroperatingleveragearetraditionallyhigh,diversifyingthefirmwithacquisitionsthat havealoweroperatingleveragewouldbufferthefirmfromaneconomicdownturn.A firmwhoincreaseditsfinancialleverageatthebeginningofarecoveryandcouldaffordto 27 doso,wouldbetakingadvantageoflowerinterestratesandultimatelyincreasenet income.Thepremiumistofindacombinationofleveragethatwillimmunizethefirm frombusinessfluctuations,meetthethresholdsoftheindustry,andyetstrategically contributetoalargeincreaseinsales. LEVERAGEMANAGEMENT Forafirmthatissolvent,fivecomponentsarethekeytocontrollingleverage:fixed costs,variablecosts,sales,interestexpenseandtaxes.Ofthefive,taxpolicyandinterest expensearethemostcontrollable;theotherfactorsaregreatlyaffectedbytheindustry, vendorsandthegeneraleconomy.Whiletaxpolicyandinterestratestrend,theywillnot surgesuddenlyupordown,causingdysfunction.Ontheotherhand,salesandcostscan fluctuatewildlydependingonthestateoftheindustryandeconomy.Thus,financial leverageismuchmoreamenabletomanagementthanoperatingleverage.However,the risksincurredinmanagingfinancialleveragelieoutsidethecomponentsofthe measurement;therelationshipbetweeninterestexpenseandoperatingincomeare determinedbythesometimes“uncontrollable”elementsofoperatingleverage-vendor prices,afavorableeconomy,andtheleveloftechnologyintheindustry.Inessence,one cannotmanagefinancialleveragewithoutdeterminingthestabilityofcash-flow,whichis derivedfromoperatingleverage.Onlywhensalesaresteadycanmanagementuse financialleverageasastrategictooltoincreasenetincome.Itisthissymbioticrelationship betweenstablesales,costs,earningsanddebtthatenablesacorporationtolimitthe amountofsharesissuedandraisethepriceofthestock-anddosowithaminimumofrisk. Fromamacrostandpoint,financialleveragetrendsmorethanoperatingleverage becauseitreflectsmanagementstrategyandmustconformtothenecessityofraising capitalinlargeincrements.Managementwillmatchtheneedforcapitalwithanticipated cash-flows.Foraprojectthatisexpectedtopayoffthroughanumberofyears,itissimply morecosteffectivetoraiselargeamountsofdebtwhentheconditionsarerighttodoso- whenthefirmdoesnotalreadyhaveexcessivedebtandwheninterestratesarerelatively 28 lowcomparedtoothersourcesoffunding.Moreover,thistrendingcharacteristicisaboon toinvestors.Bywatchingtheflowoffundsintoafirm,theinvestorseesabuild-upofrisk whichmustbeultimatelyfollowedbyoneoftwoscenarios:eitherthefirmincreases return,oritwallowsindebt-takingonevenmoredebt,divestingassetsetc.Ineffect,the objectiveofcapitalstructureanalysisistodiscriminatebetweenthesetwooutcomesand choosetheformerbeforeitoccurs. SOURCESOFVARIATION Variationinsalesisthecommonfactortovariationinbothtypesofleverage.If salesareespeciallysteady,operatingleveragewillberelativelylowandmorefinancial leveragecanbeafforded.However,thereisarisk-returntradeoffinanyindustrywith consistentsales;largerreturnsoftenaccompanymoreriskyoperatingleverageandthose companieswilltradethebenefitsoffinancialleveragetofundmostlywithequity-retained earningsandcommonstock.Again,therootofthefunctionliesinfixedcosts:firmswith higherfixedcostsmustconstantly“uptheante”andincreasesalestocoveradditional investmentsintechnology.But-thesamehighneedforcapitaltofundfixedassetscreates pricingpowerinthoseindustriesbecausethelargeinvestmentandaddedexpertiseactas barriersofentrytotheindustry.Theresultisahigherprofitmarginandalowerasset turnoverthaninotherindustries.Firmswhomakefifteenandtwentypercentprofit marginsarenotfundedinthesamemannerasagrocerystorechainmakingtwopercent- but-thegrocerychainmaybeasuperiorinvestmentifitbalancesriskandreturnmore adeptly.Moreonthissubjectiscontainedinthechapteronoperatingrisk. Withinnarrowparameters,therearemyriadmethodsofcombiningoperating leveragestoreduceriskorincreasereturn:outsourcinganddiversifyingintorelated productsoftenreducesriskaslongascorecompetenciesremainstrong;manufacturing insteadofbuyingapartalwaysincreasesbothriskandpotentialreturn;consolidating processesinonelocationincreasesriskandreturn;diversifyingoneproductlineinto differentstandardsofqualityappealstoagreatercustomerbase,butoftenreducesrisk. 29 Thisbrieflistoftechniquesisbynomeansexhaustive,andrepresentsjustasmallexample ofthepossibilitiesavailabletoreducevariationandincreasereturn. Inthecaseoffinancialleverage,besidesthederivedvariationinoperatingincome fromsalesandfixedcosts,liesthevariationininterestexpense.Whileinterestexpenseis partiallycontrollablebytheamountofdebtafirmincursandthetypeandsourceofloans, itisalsoafunctionofthestateoftheeconomy.WhentheFederalReservecutsoradjusts boththediscountrateandthefederalfundsrate,thosefirmswhofundwithdebtaremost affected.Theriskofchangesininterestrateswillaffecttheneedtorefinanceatalower rate,oralternatively,willaffecttheleveloffundingforfutureprojectsiftherateisgoing up.Thissensitivitytoboththeinternalriskdynamicsoftheindividualcompanyandthe stateoftheoveralleconomymakesthefinancialleverageratioaprimebarometerfor changesinbothcapitalstructureandstockprices. Thefinancialleverageratiowillnotvarybyarelativelygreatamount.Ifitgoesup ordownbytentofifteenpercentitisconsideredalargechangeThisrelativestability makeschangessignificantandallowstheinvestortogaugeriskbyusingitincombination withmorevolatileratioslikethetheoreticalconstruct,%∆EBIT/%∆Sales.Thislatter ratioisfullyavailabletoallinvestorsbutsuffersfromlargejumpsinmeasurement,which makesitlessamenabletointerpretation.Ratherthanseekoutathresholdnumberfor risk,theinvestorusesthecombinationtolookatchangesinbothratios.Anincreasein% ∆EBIT/%∆Sales,forexample,mayindicateupwardsearningpressureandisusually accompaniedbyanincreaseinoperatingmargin.Ontheotherhand,anincreaseinthe financialleverageratiousuallyindicatesthatmoredebthasbeenincurred.Together,the combinationofchangeswillindicateageneraldirectionforriskandreturninthefirm. Wheneachratioisbrokenintoitscomponentparts,%∆Salesforexample,theinvestor startstoseekoutmoredefinitivereasonsforthebehavioroftheratioandthedirectionof risk. 30 Thevariationininterestandoperatingincometellonlyapartialstory.To understandfinancialleverage,onemustlookfortheimplicitdynamicsaswellasthe obviousinteractions.Partofthereasonforfundingwithdebtisderivedfromadesireto keepthenumberofsharesoutstandingtoaminimum.However,nowhereineither leverageratioisthisriskdefined.TheriskofdilutingEPSandmarketpriceisespecially highinthosecompanieswhouseminimaldebtfinancingandneedtoissuestockto maintainaleveloffixedassets.Thus,anabsoluteleverageratioof3.5,forexample,gives uslittleinformation.Wenotonlyneedtounderstandthechangesinleverage,wemust combinethatinformationwithseveralotherindicatorstoformatruepictureofcapital structure. Thepremiumincapitalstructureanalysisistodetectmovementtowardanoptimal proportionofdebttoequity.Ifacompanyfinanceswithnodebtatall,thenthe managementofincomeandequityformsasimilarrisk/returnimperative.Bycomparing theindustryaveragesofthelargestcompetitorsoverthespanofabusinesscycle,some paradigmofanoptimalproportionisformed:usuallyastockwillpeakatleastonce duringthecycle,yieldingsome“guesstimate”ofoptimalitywhenthatsectorisfavored. However,relyingonaveragesisrisky:characteristicsofcorporationsandindustries change,andtheinvestorisencouragedtocoordinateinformationfromseveralsources; focusingononeconcretenumbercanbemyopicatbestanddisastrousatworst.Thus,the investorneedsto“diversify”hisorheranalysisandassumea“balancedapproach”by evaluatingtheriskofearnings.Observingthecomponentpartsofthequotient,Net Income/Stockholders’Equity,alsoknownasthereturnonequityorROE,isonemethod ofthistypeofextensiveanalysis. RETURNONEQUITY Maximizingthereturnonequityisveryclosetooptimizingcapitalstructure.The riskcharacteristicsofmaximizingeitherEPSorROE,however,willnotbethesamelevels 31 neededtomaximizethesharepriceofthestock.Tocreateasustainablegaininastock,the right“infrastructure”needstobeinplace,whichisformedbystrategicmovementtoward anoptimalproportionofdebttoequity.Thisstrategicmovementisabalancingact betweenleverageriskandthecostofcapital,butalsoincludesintegrationbetween scheduleddemandandthebusinesscycle.Ineffect,theriskofmaximizingROEistogrow attoorapidarate,whicheventuallypropelsthestockdownward,andsocapital structuralismoptstoachievethehighestlevelofROEwiththeleastamountofrisk. Thereturnonequityhasasitsfoundation,thesamecomponentsthatformleverage measurements.InsteadofmeasuringthechangeinsalesandEBIT,theROEequationuses theabsolutevalueoftheratio,EBIT/Sales,alsoknownas“operatingmargin”.Insteadof measuringEBIT/EBIT-Interest,theROEequationturnsitupsidedownandmeasures theinverse,EBT/EBIT.Fixedcostsareoftenmeasuredbytheratio,Assets/Sales,whichis often,calledthe“capitalintensity”ratio.TheROEequationturnsthisintoanother inverse,knownastheassetturnoverratio,Sales/Assets.Theseratioswillbeexplainedin detailinsubsequentchapters.Fornow,thestudent/investorshouldrealizethatriskand returnarefirmlyinterconnectedonbothconceptualandmathematicallevels.Thevery sameelementsthatreduceriskwillsometimesincreasereturn;theobjectiveistoitemize thecharacteristicsofeachcomponentandseeksomecombinationoffactorsthatproduces anoptimum. THEFORCESBEHINDLEVERAGE Inthecontextofcapitalstructure,absolutelevelsofriskandreturnareless meaningfulthantheintegratedrelationshipbetweenthem.Wegaugethatrelationshipby observingtheirmovement.Atwentypercentincreaseinnetincomeaccompaniedbya thirtypercentincreaseinequity,maybemoredeleteriousthanatenpercentincreasein earningsthatisaccompaniedbylowinterestdebtfunding.Eachchangeincapital structureneedstobeexaminedinthecontextofchangesintheeconomy,productlines, industryandinternaldynamicsofthecompany.Sinceleverageisthebackboneof 32 corporaterisk,settingupasimplequadrantanalysiswillexhibitthefundamentalforces. Wewillsubstitutetheterminology“incomerisk”foroperatingleverageandtheterm“debt risk”forfinancialleverage.Wewillthendelineatetheincreaseordecreasebya“+”ora “-”respectively. Table2-8 INCOMERISK ++ -- DEBTRISK (Increasedebtandincome risks) (Reducedebtandincome risks) +- -+ (Increasedebt,reduce income) (Reducedebt,increase income) Obviously,the“+-“quadrant-increasingdebtriskswhilereducingincomerisksappears tobethemostdangerous.Gaugingtheotherthreequadrantsisamatteroffactoring severalvariables,andobservingthedegreeofrespectiveincrease.Infact,the“+-” quadrantshouldhaveboththemostriskandmostreturn,becausefinancialleverageisa functionofoperatingincomewhoseriskisderivedfromtheconsistencyofsales.Bythat chainoflogic,theleastriskyquadrantwouldbethe“-+”quadrantinthelowerrighthand corner.Inthisquadrant,defaultrisk(notpayinginterestinatimelymanner)wouldbethe leastbecauseincomewouldcoverdebtbetter.Thebestchoicefortheinvestormaybethe firstquadrant(upperleftcorner)becausedefaultriskseemsneutral,buttheeffecton incomemaybegreater.Noneofthesecombinationsare“setinstone”,andeachmustbe examinedforindividualrisksandreturns. (BacktoTableofContents) 33 APPENDIX:STREAMLINING? Whenmostbusinessesthinkof“streamlining”anoperation,theythinkofreducing costsandincreasingprofits.Acommonmethodistoconsolidaterepetitivefunctionsand avoidduplicationbyaddingtechnology-usuallyintheformofamachine.While "naysayers"decrythedehumanizationofwork,thebusinesspersonmustaskifthereturn isworththerisk.Thefollowingscenarioshoulddemonstratethateveninthebestof circumstances,moreefficiencywillleadtogreaterrisk. TheSkidmarkTireCompanyhas$1000(million)insales,avariablecostrateof0.6 salesandfixedcostsof$100.Theiroperatingleverageandprofitareasfollows: OPERATINGLEVERAGE)1000-600/1000-600-100=400/300=1.33 OPERATINGINCOME)1000-600-100=300 Anewtypeofrubberisbothlessexpensive,andstronger,butneedstobeappliedwith machinerycosting$100more.Costsavingswouldbringvariablecostsdownto0.4sales andimproveoperatingincomeby$100attheexistinglevelofsales.Thenewoperating leverageis: NEWOPERATINGLEVERAGE)1000-400/1000-400-200=1.5 NEWOPERATINGINCOME)1000-400-200=400 Whilethisscenariowouldrepresentadreamcometrueforsomeoperations managers,operatingleverageincreasedto1.5from1.33,indicatingmorerisk.Managers mightreply,“Butwhere’stherisk?”Infact,withoutathoroughexaminationofsales volatility,thisprojectshouldnotbeimplemented.Ifsalesdeteriorateonaseasonalbasis (astheydointhetireindustry),moreoperatingleverageandtemporarycostsavingswould leadtolessprofitandnotmore;fixedcostsmustbepaidregardlessofthelevelofsales. Technologyandriskgotogether.Somereadersofthisbookmaybeoldenoughto rememberwhenentireindustrieswerecomputerizedinthelate1970sandearly80s.While computersincreasedproductivity,systemswouldgodownfordaysatatimeanditwas importanttohaveacontingencybackup-a“papertrail”.Anytimeproductiveactivityis 34 centralizedinoneprocessorlocation,moreriskisincurred-eventhoughpotentialprofits increase. (BacktoTableofContents) 35 3 CAPITALSTRUCTURE Economicshasadualisticnature.Whetheritisahumanquestforcertaintyor simplyaneedtoachievebalance,eachconceptcanbedefinedbyitsopposite:supplyvs. demand,inflationvs.deflation,Keynesianvs.monetarist,debitvs.creditandultimately, riskvs.return.Rarelydoesonefinancialthreadweaveitselfthroughtherandomchaosof opposingideas,andholisticallyembracetheirreconciliation.Thatthreadiscapital structure. Studentsoftenleavecollegewithasetofidealsthatreadilyflourishinasanitized, isolatedlaboratory,onlytofalterwhentestedandstressedbyreal-worldrandomvariation. Ratherthanabandonthoseideals,moststudentswilladaptthemtothevicissitudesof modernfinance.Forexample,ifonegraduatedbefore1970,bothhighinflationandhigh unemploymentoccurringatthesametimewasinconceivable:thePhillipscurveprofessed atradeoffbetweentheseeconomicstatesanddirectcorrelationwasinfrequent.However, bythetimetheyear1980rolledaround,inflationandunemploymenthadbeensorampant thateconomistschangedtheirownconceptofcausation;expectationsofinflationcarriedas muchmathematical“weight”asanyotherhypothesizedcause. Whenstudentsbecome“investors”,thereiseventualdisenchantmentwith “fundamentals”becausestocksseemtohavea”mindoftheirown”andrarelyrespondto suchanalysis.Investorsbecomediscouragedbecausesome“magic”combinationofsales andearningsfailstobeatacompetitorwhoisbarelyfunctioning.Often,thefrustration with“chasingearnings”willturnintoapenchantfortechnicalanalysis,whichatfirst appearstoyieldlegitimateresults-untiltheinvestorrealizesthatheorsheismerely followingrandompatternsmadewithinthecontextofarisingmarket.Inthiscase,“the trendisthefriend”,butthechangesinpatternsandtransitionsareunpredictable.Indeed, ifthereturnwereashighastoutedbythesoftwaredealerswhosellit,technicalanalysis 36 wouldattractmajorcompanieslikeBoeingandGEwhowouldhappilyforegotheireight orninepercentprofitmarginsinpursuitof“safe”fortypercentreturns. Capitalstructureisfirmlyentrenchedinacademictradition,butisflexibleenough toapplytorealworldsituations;itsstudycanleadtobothrevelationandfinancial remunerationbecauseitisinteractivewithsomanyeconomicdisciplines.Thebasic concept,however,isnotsounusual:movementtowardafirm’soptimalproportionofdebt andequityfundingtendstopropelthestockupward.Whatismoredifficulttograspisthe balancebetweenriskandreturnthatallowsthistooccur.Theabilitytoforecastisnotas importantastheabilitytocoordinateinformationandidentifyfirmswhoseleverageis conducivetogreaterearnings.Ineffect,mostanalysisconcentratesonearningsbecauseit isthemostcorrelatedfundamentaltostockprice.Alternatively,capitalstructureanalysis concentratesonthecontextofearningsbecauseitisconcernedwithbothriskand sustainability;thevariationofthereturnhasasmuchimportasthereturnitself.Without thisdomainofrisk,profitappreciationcanbebothdeceptiveandtransient,andbeguiling enoughtolosemoneyover. THECOSTOFCAPITAL:MARKETORIENTATIONANDPRACTICAL APPLICATION Thetraditionaldefinitionofthecostofcapitalisconceptuallyvague.Definedasthe amountofreturnthatabusinesscouldmakeonalternativeinvestmentsofsimilarrisk,the costofcapitalencompassesseveralimplicitfactorsthatcomplicateitspracticaluse. Foremostamongtheseisthegaugingof“similar”risk,andtheneedtopriceallsourcesof capital–debt,equityandassociatedvariations–atthemarketrate.Secondly,thecostof capitaldoesnotalwayshavean“upfront”accountingcost.Itisconsideredan “opportunitycost”thatiscomparativeinnature,anditsonly“cost”maybethegreater 37 risktakentogeneratemoreincome.Comparativeactionsthatarenotpursuedmayhave asmuchsignificanceastheactualcourseofaction. Theoretically,allsourcesofcapitalarepricedatwhatthemarketcurrentlydictates andimplicitintheanalysisofcomparativeinvestmentsisthebreakdownofthesecapital componentsintorelativepricelevels.Byinterfacingindividualcorporateriskwiththe pricesconfiguredbythecurrentstateofthecapitalmarkets,aspecific“requiredrateof return”willbedetermined.Fordebt,thereturnisthemostcurrentinterestratewhichis multipliedbyareciprocalofthetaxrateandthenbythemarketvalueofafirm’sdebt. Sincethepriceofafirm’sdebtwillchangeexponentiallydependingontherelativeincrease ordecreaseinthenewlynegotiatedrate,thescopeofthecalculationisbeyondthepurview ofmostinvestors.Forequity,investorswillattempttodeterminetheexpected appreciationofstockswithequalriskandattachthisratetothemarketvalueofthefirm’s stock.Eachvaluecanbemultipliedbytheproportionofitsrespectivecomponentinthe capitalstructure.Ifafirmhasa30/70percentdebttoequity,then30%ismultipliedby themostcurrentinterestrateandthenbytheincurredtaxadvantageof(1–taxrate)to produceapercentagecostofdebtThisfigureisaddedtotheproductofthepercentage costofequityandtheproportionofequityinthecapitalstructure(70%inthisexample). Together,therespectivecostsofequityanddebtareproportionedbytheirrelativeweights inthecapitalstructuretoformanaggregatecostofcapital. Problemsarisewhentheinvestorattemptstocorroboratetheorywithreality. Stockpricescanbequitevolatileandanybodywhofollowsthemarketcantestifytothe futilityofgaugingacorporation’srequiredamountofequityfrommarketvaluesalone. Similarly,investorsarenotprivytonegotiationswithcreditorsoverinterestratesnorare 38 marketvaluesofdebtalwaysdeterminable.Moreover,themarketwillfrequentlymisprice riskovershortperiods.Thus,determiningcapitalproportionsfrommarketvaluescan evenbedangerouslymisleading. Whyusemarketratestopricethecostofcapital?Inordertogaugetheriskof alternativeinvestments,theinvestorneedsacommondenominator.Withmarketrates,a corporationcanobservethedirectgainsorlossesinfollowingaspecificcourseofaction.If, forexample,acorporationcanissuestockatahighpricebutchoosestoissuehighcoupon ratebondsinstead,itmayincurwhatistermedan“opportunityloss”–ameasurable outlayofinterestexpenseoverandabovethecostofequity.Aslongasadollaramountcan beattachedtoanystrategicaction,alternativescanbecomparedandthemostcost effectivepathcanberealized Whiletheuseofcapitalimplieslong-termplanningandobligation,thevolatilityof marketratesmakesthecostofcapitalarelativevalue:itcanbedesignatedas“improved”, or“lower”thancompetitor’srates,butitneedsotherfinancialinformationtocorroborate it.Todeterminecapitalproportions,forexample,thehistoryoftheindustry,thetypesof assets,andtheexpectedsizeandstabilityofearningsmustbecounterpoisedtothecurrent marketcostofcapital.Ineffect,theabsolutesizeofthecostofcapitalislessimportantthan itsrelationalvalueanditscontext.Therefore,adaptationsofthecostofcapitalare improvisedthroughoutthistexttoassisttheinvestorinbettergaugingrisk.Anappendix inchaptersixcoverssomeofthetheoreticalunderpinningsindetermininga“realcostof capital”,andwhyusingsomebookvalues-suchasinterestexpense–mayoffertheinvestor apracticalanalog. THEADVANTAGESOFDEBT 39 Implicitinthedefinitionofdebtisitsinherentadvantage-thereceptionof immediatefundswithpayment“postponed”untilalaterdate.Anyamountofcash-flowis morevaluableinthepresentthanitisinthefuturebecauseitcanbeinvestedandearn interest.Wheneveranaccountreturnsmoreonaloanthanitcosts,anetadvantageoccurs However,thetimingoftheinflowsisofevengreaterimportance.Notonlyarereturns greaterwhenthepayofffromaninvestmentoccursfaster,buttheriskoftheloanis diminishedbecausethefirmhasadequatecashtoserviceinterestpayments.Like“justin time”inventorysystems,mostbusinessesrecognizetheimportanceofreceivingcash-flows atthemomentabillisdue.Ineffect,profitsareenhancedandriskisdiminishedif paymentismadepromptlybutnottoosoon.Whenservicingdebt,profitsreceivedin JanuaryandFebruaryaremuchmorevaluablethanthesameprofitsreceivedin NovemberandDecembersimplybecausethe“lagtime”isnotproductive.Duringthislag, interestmaybedueontheloan,increasingitsrisk;nocashinflowsarebalancingthe outflows. Besidesthetimevalueofmoney,debtisevaluatedbycomparisontootherfirms.A loanthatreturnsmorethanitscostmaybeconsidered“ineffective”,ifafirm’speersare returningoneandahalftimesasmuch.Thus,industrystandardsareanimportant elementinevaluatingdebt.Performanceindicatorslikethereturnoncapitalare significant,butevenmoreindicativearetheaverageratiosofdebttoequity,andthe amountofinterestexpenserelativetolong-termdebt.Whenbothreturnsandleverage ratiosarehoveringaroundtheindustrystandard,riskandreturnwillbecommensurate withafirm’speers.However,whenthesemeasurementsarebelowthatstandard,the investorcanlookforoneoftwooutcomes:eitheracompensatingbounce(turnaround) thatlowersriskandpropelsthestockupward,ormaintenanceoftheposition,whichmay beaccompaniedbyaselloff.Thislatterpositioncanmakeorbreakinvestorsbecause maintenanceofthepositioncanbeasignofeitherfaithinaninvestmentpayoff,or 40 uncertaintyovertheinevitable“costoverruns”thatmayfurtherscareinvestors.A numberofmechanismsareinplacethatconfersotheradvantages: • 1.Interestisataxdeductibleexpense.Interestexpenseisroutinelydeductedfrom operatingincomebeforetaxableincomeiscalculated.Infact,thetaxsavingsarethe costofinterestmultipliedbythetaxrate. • 2.Along-termloan(overoneyear)grantstaxadvantagesuntiltheloanispaidoffand mayhaveastabilizingeffectonthefirmifcash-flowisespeciallytight. • 3.Sinceinterestistaxdeductible,thecostofdebtisalmostuniformlylessthanthecost ofissuingequity,whichcandilutemarketprice. • 4.Investmentwithdebtmaylimitthenumberofsharesoutstanding;performancecan beenhancedwithoutburdeningshareholdersifearningspershare(EPS)rise. • 5.Issuingdebtratherthanequityhelpsmaintainexistingcontrolofthecompany.Less newsharesrestrictsvotingpowertothelargestcurrentshareholders.Moreover,more debtonthebookswilllimittakeoverattemptsbecauseprospectivebuyersdonotwant tobeburdenedwiththeobligationsofleverage,i.e.,interestexpense,restrictive covenants,sinkingfunds,etc. Sincealldebtfundsaspecificlevelofassets,takingondebtwillsavethefirman amountequaltothetaxratemultipliedbytheamountofbonds.Ineffect,thegovernment conferstaxadvantagestoencourageinvestment,andthenplanstomakeupthedifference innewincometaxeswhentheinvestmentgeneratesprofits.Withoutthistaxbenefit,the optimalproportionofdebttoequitywouldbelesssignificant,dependingontheriskof insolvencyalone. RISKANDDEBT Whileitseemsadvantageoustopileupdebtandusethetaxsavingsforother investments,acorporationwhodoessoflirtswithdefault.Earningsareoftencyclical,but interestpaymentsmustbetimely,andoncedebtisincurred,theprobabilityofbankruptcy increasesexponentially.Infact,theamountoffixedcostsinaproductioncycleregulates 41 bothoperatingriskandtheamountoffinancialleveragethatcanbeincurred.Thosefirms withveryhighfixedcostsmayhaveearningsthatfluctuatemorethanfirmswithlower fixedcosts,andhaveahigherriskofdefaultduringeconomicdownturns.Theirriskof bankruptcyishighersimplybecausetheirproductioncyclesaremorereactivetoeconomic conditions.Thus,thesefirmswithhigh“economic”riskarepoorcandidatesforfinancial leverage. Onemajorproblemwithcapitalstructureanalysisisenumeratingthecostof bankruptcy.Theterm“bankruptcy”hasmanydefinitionsandcoversawidebreadthof legalstatesandfinancialconditions.Thistextapproachestheconceptfromtheperspective ofthecorporatecommonshareholderandalwayspresumesthelossofshareholdervalue. Itdesignatesarelationshipbetweenassets,liabilitiesandmarketvalue,whichbyitsvery natureisprobabilistic;marketvalueisaffectedbypsychologyasmuchasassetsare affectedbyinflation.However,themaintenetneedstobeexamined;thereissomecostof bankruptcythatiscomposedofatleasttwoassociatedelements:someamountoflossand someprobabilityofdefault.Althoughtheremaybemanyotherfactors,theseformthe baseofagenericmodel. Oncethe“genericpremise”isaccepted,thereisconsiderabledifficultyincreatinga relationshipbetweenfixedcostsandassetstructuretoobtainan“amountofloss”.There aresomeassetsthatcanbesoldintheeventofliquidationandsomeassetsthatdonot directlyaffecttheoperatingcapacityofthefirm.Moreover,themarketvalueofthestock abovethelevelofassetsisbasedontheexpectedabilitytogenerateincomeinthefuture. Thus,theproperconfluencebetweenseveralassetclassesandtheamountofintrinsic shareholdervalueformsanamountofloss. Defaultprobabilitiesarecreatedfromhistoricaldistributions.Whenthe relationshipbetweentypicalinputtedvariableslikeassetsize,orsalesstabilitychanges, theprobabilityisnolongervalid.Ineffect,onceadefaultprobabilityisreadyforthe market,itisnolongeronehundredpercentaccurate.Thefinancialcommunitydevelops 42 limitsoftolerancethatacceptsimprecisionasagivenvariable.Ultimately,capital structureanalysisrestsuponthistolerancefora“ballparkfigure”,butmusthavedata thatiscoordinatedtoalerttheanalysttothepresenceofmorerisk.However,thereliance ondefaultprobabilitiesisinitself“risky”(the2007“creditcrunch”)becausethereisrisk thatcannotbeenumerateduntiltheeventoccurs.Theeconomicandpoliticalrisksthatare outlierstoany“system”,areveryillustrativeofthisconcept;theymayaffectafirm’s performancemorethananyinternallygeneratedvariablelikesalesorassets.Thus,to forma“genericcostofbankruptcy”,wedependondefaultprobabilitiesandwemultiply thembyanestimatedamountofloss,buteachcomponentpartishypotheticaland tentative. Withanincreaseinbankruptcycostscomesaconcurrentriseinthecostofcapital. Investmentbanksdemandmoreinterestonbondsandhigherflotationcostsforstocksto mirrorthegreaterriskinacompany.Thecostofeachsourceoffundingisinterdependent onthemarketforalternativesources.Infact,proportionalincreasesindebthaveafixed tendencytoraisethecostofbothequityanddebtasriskbecomeshigher.Thisinteraction betweenriskandthecostofcapitalisneverstaticandformsthebasisofcapitalstructure analysis.Whenanoptimalproportionofcapitalsourcesisachieved,cost,risk,andthe interrelationshipbetweeneconomicoutlookandcorporateperformancewillbebalanced; thefirm’sstockpricewillbemaximized. Inevitably,seekinganoptimalcapitalstructureturnsintoagameofstrategicrisk. Sincethecostofcapitalcutsintoprofitability,firmswithtoomuchdebtareusuallycash- poor,low-earnerswithdiminishedmarketvalues.Theirinabilitytomaintaincash-flow perpetuatesachainofloansinwhichoneloanretiresanotherwithlittlepaymentof principal.However,thosefirmswhohavegreaterresourcescanaffordtousemoredebtin theircapitalstructureswhichcauseslessstrainonexistingshareholders,enhancing marketvalue.Thisstrategicuseofleverage,theutilizationofdebtwithrecourse,allows fewersharestobeissued,contributingtoperformanceonapersharebasis.Whenprojects 43 becomeprofitable,thereislessshareholderinvestmentbutgreaterreturn,andthemarket priceofthestockincreases. MATHEMATICALOPTIMIZATION:THEORYVERSUSREALITY Thetheoryofcapitalstructureispredicatedonthebalancebetweenthetaxbenefits ofdebtononesideofafunctionandbankruptcycostsontheother.Thisequality optimizestheproportionofdebttoequityatthepointwherethechangeintaxbenefits equalsthechangeinbankruptcycosts.Ineffect,wetrytomaximizethefunction(Tax benefitsofdebt)-(BankruptcyCosts).Ifwecalculatethefirstderivativeofthefunction andsetittozero,thefunctionisatanoptimumand∆ ∆∆ ∆Taxbenefits=∆ ∆∆ ∆Bankruptcycosts. Theproblem,however,isnotintheconceptofthefunction,butinthedefinitionof variablesthatinteract.Aspreviouslymentioned,theterm“bankruptcycosts”isuniqueto therealmoftheamountofloss;eachcorporateentitylosessomethingdifferent.Moreover, someofthevariablesdependonprobabilityandsomearedeterministic;theyare interdependentneverthelessandcreatebothvariationanduncertainty.Toafinancial executive,themathematicalsituationisakintoforecastingthepathofahurricaneand decidingtomakepreparations;theeffortsmaybewasted,butthepotentialoutcomeof inactioncanbedevastating. Ultimately,theneedformathematicalcertaintyisnotassignificantastheneedto realisticallygaugeriskandmovethefirmintherightdirection.Infact,thepotentialfor priceappreciationinastockismuchgreaterwhenafirmisalongdistanceawayfromits optimalcapitalstructure,butresolutelymovingtowardit.Sinceastockpricemaximizes whenafirm’scapitalstructureisoptimal,thereislittleroomforittomove-upordown. Inthiscase,investorsmightevendemandthatthefirmtakemoreriskbymovingaway fromtherelatively“safe”worldofastockpriceoptimumandengageinmergersand acquisitions.WallStreetrewardscompanieswho“defytheodds”byoutperformingthe marketinsomespecialway.Consequently,firmswhoseemladenwithdebtbutbeginto escalatesalesandearningsaregenerallyobservedtobealongdistancefromanoptimal 44 capitaltarget;theirhigherriskofbankruptcycoupledwithsubsequentimprovement substantiatesmoreinvestment. Thetrueoptimalcapitalstructureisinastateofflux.Infact,itrespondstoso manydifferentvariablesthatitperpetuallychanges-almostinstantaneously.Thereason behindthisvariationisthatitisdependentonexternalrelationshipsoutsidetheinternal controlofthefirm.Whileleveragefactorsmaydetermineanestimateoftheamountand sourcesoffunding,thefoundationofcorporateriskistheinterrelationshipbetweenlong andshort-terminterestrates,andtheequitymarket.Theserelationshipschangedaily. Thus,ifafirmwereshutdownforaweekend,itsoptimalcapitalstructurewouldchange eversoslightly,evendependingontheperformanceofforeignmarkets. Fortunately,theanalystdoesnotneedamathematicallypreciserenditionofan optimalcapitalstructuretomakeeffectiverecommendations.Severalusefulindicators existtoaidindeterminingwhetherthefirmismovinginafavorabledirection.Classic measurementslikethereturnonequity(ROE),andEVA ®1 (economicvalueadded)canbe modifiedtoreflectactivitytowardanoptimaltarget.Althoughthestudent/investoris encouragedtocreateaworkingmodelandobtaina“ballpark”estimate,suchexercises willbeladenwithatleastthreediscrepancies:1)Alackofadequatevariables2)Artificial constraintsand3)Toomanyconstants.However,sucheffortscanbeinformative,evenif imperfect.Theycanhelptheanalysttounderstandthecomponentsofcorporaterisk,and especiallyanychangesthereof. 1 EVA ® istheregisteredtrademarkofSternStewart,Inc. 45 THEMODIGLIANI-MILLERPROPOSITIONS Mostofourpresentknowledgeofcapitalstructureismerelyanextensionofthe researchdonebytheteamofMertonMillerandFrancoModiglianiinthelate1950s.In fact,thefunction,(TaxBenefits)-(BankruptcyCosts)issimplyarelaxationofaconstraint thattheyusedtodeterminetheincrementalvalueofaleveragedfirmoveranunlevered one.IntheirfamousequationV(l)=V(u)+TB,thevalueofaleveragedfirmwasgreater thanthevalueofonewithanallequitystructurebyafactorofTB,whichwastheamount ofthefirm’sbondsmultipliedbytheirtaxrate.Sinceinterestisataxdeductibleexpense, intheabsenceofbankruptcy,afirmcanincreaseitsvaluesimplybyincurringmoreand moredebt.Withoutsuchtaxadvantages,thevalueoftheunleveredandleveragedversions ofafirmwouldbeequal,whichformsthecruxofMiller/Modigliani’sPropositionI:In theabsenceoftaxesorbankruptcy,thevalueofafirmisindependentofitscapital structure. Thesecondproposition,propositionII,provesthatincreasingtheproportional amountofdebtincreasesthecostofequity.Sincetherealrateofreturnondebtmust coverinterestpaymentsbeforeitshowsprofitability,itishigherthantherateofreturnon equitybya“riskpremium”thatpushesthecostofequityupward.Whenthisproposition isexaminedinaworldoftaxes,wecanconcludethattheoptimalcapitalstructureis composedofonehundredpercentdebt.Aslongasinterestistaxdeductible,intheabsence ofbankruptcy,thecostofdebtwouldalwaysbelowerthanthecostofequity;freecash- flowwouldalwaysbegreaterwiththeuseofdebt.Infact,thisextreme“corner”solution formsthefoundationforallfurtherextrapolationsofcapitalstructuretheory.Inessence, werelaxconstraintsandaddvariablestoderivearealistichypothesis.Byadding bankruptcycosts,forexample,wemerelyregulatethetendencytouselowercostdebt. ThelogicbehindpropositionIIwasthatthecostofequityreactedtotheincreasein debtbyrising.This“riskpremium”wasreconciledwithequitybyahigherrequiredrate ofreturn;thenecessityofcoveringinterestpaymentsincreasedthecostofequity.Without 46 taxesandbankruptcy,thecostofdebtwouldbeahorizontalstraightlinethatfellwell belowtheangledlineofthecostofequity. Figure3-1 Eachincrementalunitofdebtpushesupthecostofequitybecauseitincreasestherequired rateofreturn.Weassumethattheleveredcompanyismakingatleastasmuchasthe valueofitsinterestpaymentmorethantheunleveredfirm,ortherewouldbenoreasonto incurdebtinthefirstplace. THEOPTIMIZATIONPROBLEM Whiletaxbenefitsarecomposedofasimplelinearfunction(long-termdebt multipliedbythetaxrate),thecostofbankruptcywilladoptashapethatisdefinedbyits variables.Insomemodels,withaconstantamountofloss,ittakesontheshapeofthe defaultprobability.Inothermodels,theshapemaybedefinedbyanexponentially increasinglossoralineardefaultprobability;eachmanifestationofbankruptcyis fundamentallyunique.Withoutastandardized“costofbankruptcy”,solvingforan optimalamountofdebtdependsonthecombinationofvariablesthatmakeupthe function. Debt/EquityRatio Costof Capital CostofDebt AverageCostofDebtandEquity CostofEquity 47 Attheveryleast,thecostofbankruptcymusthaveaslopethatisgreaterthanTB (TaxRatexBonds)ortaxbenefitswillexceedbankruptcycostsateverylevelandthe modelisnotoperable.Secondly,thecurveshouldemulatearisk-returntradeoffand increaseatanincreasingrateatsomepoint;amodified“S”shapeisidealbecauseitshows athresholdamountofbankruptcycosts,followedbyarapidincreaseandthenaleveling offwhenactualbankruptcyoccurs.Lastly,someoftheshapeofthecostofcapitalcurve shouldbeincorporatedintothecostofbankruptcy.Whilebankruptcycostsdonot perfectlymirrorthecostofcapital,theyshouldshowincreasingcostsforgreaterriskjust ascreditorschargehigherinterestratesformoreriskyloans. ThelinearityoftheassortedcostsofcapitalintheMiller/Modiglianipropositions wasafunctionoftheconstraints-especiallytheabsenceofbankruptcycosts.Inreality, thecostofcapitalisregulatedbyseveralrelationshipsthatcreateajaggedcurveandlimit theamountofdebtthatafirmcanincur.Thecostofcapitalcurvewillatfirstdecreaseas moredebtcreatesbothtaxbenefitsandalargerEPS,throughfewersharesissued.Itthen movesrapidlyupwardsasinterestexpensebecomesonerousandthefirmmovescloserto bankruptcy.Thus,thetheoreticalunderpinningsofthecostofcapitalwillcreateacurve thatinthelongrun,mimicsthecostofbankruptcycurveascreditorschargemoreinterest atanincreasingrate;pastaspecificpoint,eachadditionalunitofdebtwillincreaserates sorapidlythatfinancialleveragewillnolongerbecosteffective. Thelimitingconstraintinanyoptimizationmodelistheamountofcapital.Whena modelattemptstooptimizetheamountofdebt,itwillmatchtheamountofcapitalwiththe boundariesofthedefaultprobability.Thegreatestaccuracyisachievedwithinnarrow boundsandmaynotincludethegivenamountofcapital.Forexample,manydefault algorithmswillallowagreateramountofdebtthanthecapitallimitsimplybecausethey arenotconfiguredforextremevalues:defaultprobabilitiesarecreatedfromaveragesand maynotbeaccuratewithinthelevelofdebtthatisnormallyusedbythecompany. 48 Somemodelsshelvetheconceptofmarketvaluealtogetherandconcentrateonthe amountofassetlossthatoccursinatypicalbankruptcywithintheindustry.Fromtwenty tosixty-fivepercentofassetsisacommonfigure,andthesemodelswillbeconfiguredfor specificsectors.Defaultalgorithmsmaybe‘customized”tomeettheneedsoffirmswithin anindustryandwillbemoreaccuratethanthegenericalgorithmsusedinthistext. Nevertheless,anymodelwillopposethecostsofdebtwithitsbenefits,-aspinoffofthe originalMiller/Modiglianithesis. THEPROPOSEDIDEALANDITSINHERENTPROBLEMS Wecansolvethemarginalbenefitsfunction,∆ ∆∆ ∆Taxbenefits=∆ ∆∆ ∆Bankruptcycosts,. byfindingaleveloflong-termdebtthatmakesthefunctionequaltozero.Naturally,long- termdebtwouldbeavariableinbankruptcycosts,andwecansatisfytheequationby manuallyinputtingeachunitofdebtfromzerotothecapitallimit.Whenthefunction, (TaxBenefits)-(BankruptcyCosts)nolongerincreases,anoptimumisfound. Alternatively,wecanusealinearprogrammingmodulelikeExcels“Solver”tomaximize thefunction.Thegraphicaldisplaywouldbethus: Figure3-2 DEBT $COST TB CostofBankruptcy X 49 Inthis“ideal”depiction,theoptimalamountofdebtisatpoint“X”becausethedifference betweentaxbenefitsandbankruptcycostsareattheirgreatestpoint.Themarginal benefitsfunction,∆ ∆∆ ∆Taxbenefits=∆ ∆∆ ∆Bankruptcycostsisoperablewhentheslopesofeach curveareequal.Fromtheperspectiveofcalculus,thefunction,(TaxBenefits)- (BankruptcyCosts)ismaximizedwhenitsfirstderivativeissettozero.Bylinkingtax benefitstobankruptcycoststhroughthevariable,“Long-termdebt”,weproducea differentiatedfunction.Whenwesubtractthisoptimalamountofdebtfromagiven amountofcapital,weproducetheoptimalamountofequity. Amongtheseveralproblemsinsuchamodelare: • 1.Context-Optimizationvariablesarenotformulatedintermsofcorporatepotential, andyetcapitalallocationdependsonexpectations,judgment,andwisdomgarnered fromhistoricalobservations.Nooptimizationvariableiscognizant,forexample,ofa marketingstrategythatwillmakeafirmanindustry“front-runner”.Anotherexample wouldbetheexpectationofincreasinginterestrates;anoptimizationprogrammight lockafirmintoalevelofdebtthatisappropriateatalowerrate,butexcessiveifrates areexpectedtorise. • 2.Artificiality-Besidestheconstructionofhypotheticalvariables(amountofloss), somevariablesmustbeheldconstantwheninreality,theywouldchangedynamically withthelevelofdebt. • 3.Theassumptionofappropriatecapital.-Fewcompanieshavetheabilitytomatch theamounttheycanraisewithactualcapitalrequirements.Anyoptimizationmodel workswithinagivencapitalconstraintwhichmaynotbetheoptimalone.For example,ifunderperformingitspeersisconsidered“normal”performancefora company,aprogrammightoptimizedebtunderthepremisethatthefirmisearninga tenpercentreturnoncapitalandnotthefifteenpercentofitscompetitors.However,a trueoptimalproportionofdebttoequitywouldbringthereturnoncapital(ROC)up toindustrystandards.Inthiscase,eithertheamountofcapitalisexcessive,or 50 productionproblemslimittheamountofnetincome-aproblemthatisnotpartofthe optimizationfunction. CAPITALSTRUCTUREANDTHECOSTOFCAPITAL Bothstudentsandinvestorsgetthemisconceptionthattherelationshipbetweenthe costofcapitalandcapitalstructureisbasedonaffordability-thatthepriceofdebt (interest)determineshowmuchdebtafirmcanincur.Usingthisfaultylineoflogic,a companycanhaveafortypercentdebttoequityratiowheninterestratesaredownanda twentypercentproportionwhentheyrise.Infact,firmsdotakeadvantageofinterestrate cyclestoincurmoredebtbutdonotmateriallychangetheiroptimalcapitalstructures; lowerratesbecomeanincentivetotakemoreriskbutarebalancedbythediminished earningsoutlookthatusuallyaccompaniesaFederalReserveratecut. Capitalstructurehasaninteractivecauseandeffectrelationshipwiththecostof capital.Someofthefactorsthatregulatetheproportionofdebttoequityaffectthecostof capital.Ontheotherhand,themajorportionofthecostofcapitalisconfiguredaway fromtheinternaldynamicsofthefirmandisdictatedbyaconfluenceoffactorsinthe greatereconomy.Ineffect,anaggregateofthedemandformoneywilldeterminethe absolutelevelofinterestrateswhilecompetitiveforceswithinthecompany(salesstability, earnings,andassets)willdeterminethelimitsoftolerance-comparativelyfavorablerates, orratesthatareattheuppermostlevels. Thedecisiontousedebtisnotsoaffectedbythepriceofdebt,asbytheasset structureofthecompany-theproportionoffixedassets,andhowwelltheassetscanserve ascollateralforaloan.Forexample,realestatemayserveasbettercollateralforabond issuethanamorevaluablecommoditylikegoldsimplybecauseitislessvolatile.Thefixed incomemarketisbasedondependability,whichisactualized,bysteadyearningsand stableprices;creditorsdemandfromfirms’assetstructures,thesameperformancethat theyexpectwhenreceivinginterestpayments.Thus,largefirmswithsteadyincomesand salablecollateralcanseekandincurmoredebtintheircapitalstructures. 51 Threecharacteristicsofafirm’sassetstructureareespeciallysignificant:the stabilityofsales,operatingleverageandcapitalintensity.Thethirdcharacteristic,capital intensityismerelytheinverseofthefundamental,assetturnover,whichis:Sales/Assets. Whenweturnthisaroundto:Assets/Sales,weobtainacomparativeratioofthedegreeof fixedassetsinacompany.Riskiscreatedbythepropensitytogeneratefixedcosts,and highercapitalintensitieswillrequirejustthat-morecapitaltopayforoutdated machinery,managementsalariesandvarious“overheads”.Firmsdesiretomatchthe timingofcash-flowsfromprojectswiththetypeoffundingbecausesuchastrategy increasesreturnandreducesrisk.Consequently,firmswithhighercapitalintensitieswill haveprojectsthatencouragelong-runprofitabilityandrequireextendedfunding.By itself,capitalintensitywillaffectbothoperatingleverageandsalesstabilityandisvery dependentonthetypeofindustry.However,withinthoseconfines,thedecisiontofund withequityordebtstillexists,andisasmuchdependentontheriskofphysicalcollateral asitisonoperatingrisk-whichoftengo”handinhand”;firmswho,dealwithriskier assetswilloftenhavethemostoperatingrisk. ByMiller/ModiglianipropositionII,itisdoubtfulthatthefirmcanobtainalevelof riskthatpushesthecostofdebtabovethecostofequity.Therefore,bypricealone,the optimalcapitalstructureismadeupofalldebt.Theinvestorcanverifythisconceptby observingthedemandforanystockthathasajunkbondstatus;thepriceofthestock deterioratesasthecostofitsdebtsky-rockets,andthestockbecomesatoolforspeculators. Theriskofholdingonethousanddollarsworthofstockismuchgreaterthanholdingaone thousanddollarbondofthesamefirmbecausethebondneedstoberepaidorthestock willbecomeworthless. Inthelongrun,ahigherprobabilityofbankruptcywillraisethecostofcapital,a directcorrelationthatismaintainedatlowerlevelsaswell.Ineffect,thecostofcapital respondstotheriskofbankruptcyinthecapitalstructureandwilladjusttoreflectthe stabilityofthecompany.Capitalstructure,however,respondsslightlytochangesinthe 52 costofcapital,mostlyadjustingthroughtheamountofcapitalraised.Wheninterestrates arehigh,forexample,thecompanymaynotchangeitstargetstructuretoalowerlevelof debt,butmayadjusttheamountoffundingforallprojectsandmoveslightlypastitstarget withmoreequityfunding.Whenratesarelower,thefirmbeginstofundwithmoredebt, movingpastitsoptimaltargetfromtheotherdirection.Onlywhentherehasbeena systemicshiftthatwillraiseorlowertheinterestratethroughoutanentirebusinesscycle willtheoptimaltargetchangesubstantially.Suchpatternsoccurredinthelate1970swith massiveinflationanddoubledigitinterestratesandagainwithnearrecordlowratesin 2002-2003. Thestudent/investorshouldrecognizethatnolevelofinterestratewillensure steadyrepaymentofaloanbecauserateschangefrequentlywiththelevelofGDPgrowth. Forafirmwhofundswithallequity,takingadvantageof“cheapinterest”forashort amountoftimewillmerelycausethefirmtosuffertheconsequencesofapoordecision; shareholderswillsellthestockbecausetoomuchriskisincurred.Ontheotherhand,firms whoseassetstructureshavechangedandaremoreamenabletousingdebt,willreceivean initialboostupwardsastaxadvantageseclipsethecostofbankruptcy.Forcompanieswho usedebtonaregularbasis,changesinratesgiveincentivetoraisemorecapitalbutnotto makemajorchangesincapitalstructure.Ineffect,afavorablechangeinthecostofcapital isanimpetustomovewellpasttheoptimaltargetoreventemporarilymoveawayfromit altogetherbecausetheriskofdoingsoisrewardedwithalowercostofcapital.However, firmsrealizethatsuchconditionsaretemporaryandusuallymovebacktowardthe optimaltargetassoonaspossible. Thefollowingchartdelineatesdifferentlevelsofdebttoequityfromtheperspective ofboththecostofbankruptcyandthecostofcapital.Thecostofdebtisindicatedwith taxsavingsandriseswithbankruptcycosts.Thecostofequityisabovethecostofdebt andalsoriseswiththeprobabilityofbankruptcy.However,asindicatedbyasterisks,the costofbankruptcyisnotalwaysperfectlyalignedwithrisksandcallsforexecutive 53 judgment.Thecalculationsarebasedonalevelofcapitalof$1000andataxrateofthirty (0.3)percent.Thecostofbankruptcyisatypicalalgorithm. Table3-1 Percent D/E TB Costof Bankruptcy TB-Cost Interest Rate% AfterTax Costof Debt Costof Equity% 10 30 20 10 5.5 3.85 7 20 60 30 30 6 4.2 9 30 90 40 50 6.5 4.55 9.5 40 120 55 65* 6.8 4.76 9.5 50 150 90 60 7.4 5.18 10.5 60 180 130 50 8.4 5.88 12 70 210 180 30 9.75 6.85 14 80 240 170 70* 11.75 8.225 17 90 270 1250 -980 14 9.8 20 100 300 1350 -1050 17 11.8 24 Thiscostofbankruptcyfunctionisabimodaldistribution,producingtwoinflectionpoints: oneatfortypercentdebttoequityandanotherateightypercent.Thejudgmentofthe analystisparamountanditisobviousthatthemostcosteffectivepathisatfortypercent. Manybankruptcyalgorithmsareonlyaccurateacrossanarrowrangebecausetheyare basedonaveragehistoricaldistributions.Thechallengetoanymathematicianisto produceanalgorithmthatcanmaintainaccuracyoveritsentirerangeandyetbeflexible enoughtosolveforcapitalstructureinputs. THECONCEPTOFAWEIGHTEDAVERAGECOSTOFCAPITAL Thestandarddefinitionof“capital”comprisesseveralcomponentpartsincluding equity,long-termdebt,preferredstockandsomeshort-termdebtandcapitalleases.When thepercentageofeachcomponentinthecapitalstructureismultipliedbyitsspecificcost andthensummedtogether,themeasurementthatisformediscalledtheweightedaverage costofcapitalorWACC.Inthechartabove,forexample,attwentypercentdebttoequity, 54 wemultiplytheaftertaxcostofdebt(0.042)bythepercentageofdebt(0.2)andsumitwith theproductofthecostofequity(.09)anditsrespectivepercentage(0.8).Thetotal expressionis(0.2)(0.042)+(0.8)(0.09)=0.0804or8.04percent. Inthelongrun,theWACCwillfollowtheprobabilityofdefaulteitherupordown, butminimizeswhenthecapitalstructureisoptimal-atapointwheretaxbenefitsand bankruptcycostsareattheirgreatestdistance.Intheshortrun,theremaybeeccentric movementbecausethemarketdoesnotalwayspriceriskcorrectly.Whenthegovernment setsinterestrates,itdoessotopreservesystemicequilibrium,abalancebetweengrowth andinflation.However,afirmdoesnotsetpolicytoconformtoFederalReservedecisions. Itfundscapitalprojectstomakebackitsinvestmentasrapidlyaspossible.Therefore, therewillbeperiodswhenthefirmisusingmoreequitywhileinterestratesfordebtare decreasingandviceversa.Overatwoorthreeyearperiod,theWACCisareliable indicatorofmovementtowardanoptimalcapitalstructure;anydecreasewouldbeviewed asfavorable.However,itmayrisefromyeartoyearsimplybecausetherehasbeena systemicshifttowardcomparativelyhighercapitalcosts. Byusingtheamountofdebtthatkeepsthecostofbankruptcyatarelative minimum,butmaximizestaxbenefits,thegreatestamountoftheleastexpensivecapital componentisused.Forfirmswhofundonlywithequity,theprobabilityofdefaultis minimizedwiththesameactionsthatwillbooststructuralintegrityanddecreasethecostof equity-stabilityofsalesandincomeinthedomainofhigherreturns.Thus,anunlevered firmmustminimizeitsWACCwithouttheluxuryofsubstitutinglowercostdebtfor equity.However,byminimizingbankruptcycosts-keepingsharestoaminimumand loweringtheprobabilityofdefault-itwillalsominimizeitsWACC,producinganoptimal capitalstructure.Inmostcases,thecostofcapitalwillbeuniformlyhigherforallequity companies,whichtheyneedtoovercomebybothincreasingandstabilizingtheamountof salesandincome. 55 Figure3-3 EARNINGSANDCAPITALSTRUCTURE Inthechapteronleveragewementionedthattheamountoffixedassetsinan industrywilldeterminethestabilityofearningsandthatfinancialleveragecanbe increasedwhenoperatingleverageislower.However,theamountofearningsisalsoa majorfactorindeterminingcapitalstructure.Whenearningsarelarge,theycanbe retained,andoutsidesourcesoffunding(debtandnewsharesofequity)canbeavoided.In effect,largeamountsofearningstendtobeunstablebecauseoftherisk-returntradeoff, butenableafirmtobuildstockholders’equitythroughretainedearnings. Whenearningsareretainedandbecomepartofstockholders’equity,theyincurthe costofequity.Thefulldevelopmentofthecostofequityislefttoanotherchapter,butletit bestatedthatthecostofequityisacomparativecostverymuchrelatedtothereturnon equity,whichistheratio,NetIncome/Stockholders’Equity.Thus,ifthecostofequityis particularlyhigh,itwillcostthefirmmoretoretainearningsthantodistributethemas dividendsorbuybacksharesofstock.Inthisregard,dividendpolicyisveryintegrated withcapitalstructurebecauseitdeterminesretentionandtheultimateWACC. Debt/Equity Dollar Value WACC TB-CostofBankruptcy 56 Anothereffectofearningsoncapitalstructureistheminimizationoftheprobability ofdefault.Mostdefaultalgorithmsexplicitlydefinesomeaspectofearningsasavariable thatdiminishesthisprobability.Inthemarginalbenefitsequation,asmallerprobabilityof defaultallowsmoredebtfunding,moreshares,orsomecombinationofboth-capital appreciationthatleadstoassetgrowth.Inevitably,topreservetaxbenefits,theproportion ofequitygrowsthroughretainingthesameearningsthatdiminishedtheprobabilityof default;themarginalbenefitsfunctioncangrowwhenequityisaddedtodebtbutdoesnot replaceit.Onlywhenthecostofequityisconsideredtoohighwillsharebuybacksand specialdividendsneedtobeconsideredinlieuofretention. Thebusinesscycledetermineswhencertainindustrieswillhavefavorablesalesand earnings.Consequently,anentiresectorwillexhibitsimilarpatternsofmarginalbenefits andleveragethatwillbedependentontheprobabilityofdefault.Inadownturn,hightax benefitswillbeaccompaniedbyanevenhigherprobabilityofdefaultbecausefirmsare generatinglessincome;theymustparedowntheirdebtaboveallelse.Inexpansions,more incomelowerstheprobabilityofdefaultallowingmoretaxbenefits.Theeffectofan optimalcapitalstructure,however,istoreducethenegativeeffectsofthebusinesscycle andtoaccentuatethepositive.Thecostofcapitaliskeptlowenoughduringadownturnso thatcompetitiveprogresscanbemade.Consequently,duringanexpansion,thecompany willhavethefinancialflexibilitytotakeonmorerisk,andsometimeseventheluxuryof movingawayfromthe“safe”environsofanoptimalcapitalstructure. CAPITALSTRUCTURELOGIC Capitalstructuretheorydisplaysapatternofalternatingrisksandreturnsthatare encapsulatedinseveralmeasurements.Eachcategoryofriskstemsfromtheprevious categoryandhasacorrespondingsetofmeasurements.Inthefollowingchart,togettothe highestlevel,capitalallocation,theotherriskcategoriesmustbeproperlyassessed. 57 Table3-2 HIERARCHALRISK RISK/RETURN MEASUREMENT 1)CAPITAL ALLOCATION 1)COSTOFCAPITAL a.Proportionofdebtto equity a.Costofdebt b.Projectanalysis b.Costofequity c.Capitalbudgeting c.Costofvariousother components d.CapitalRequirements d.WACC e.Capitalmarket conditions e.ROE,ROC,Economic Profit 2)BANKRUPTCYRISK 2)COSTOF BANKRUPTCY a.Probabilityofdefault a.Measurementsand risksarethesame. b.Amountofloss 3)FINANCIALRISK 3)LEVERAGERATIOS a.Interestexpense a.Financialleverageratio b.Numberofequity shares b.Equitymultiplier= Assets/Equity c.TaxBenefits cChangeinNetIncome/ ChangeinOperating Income d.Stabilityofnetincome andoperatingincome d.TIE(TimesInterest Earned) 4)ECONOMICRISK 4)OPERATINGRISK a.Sales a.OperatingLeverage b.Fixedandvariable costs b.CapitalIntensity c.OperatingIncome c.OperatingMargin CHANGESINCAPITALSTRUCTUREANDSTOCKPRICES “Likeamonkeythrowingdarts”,hasbeenacommondescriptionoftherandom variationthatanalystsfacewhenpickingindividualstocks.Indeed,anysystemthat 58 purportstocomprehendsomeofthemismatchesinriskandreturnthroughouttheyears, includingcapitalstructuralism,appearstobeablindattempttorationalizechaosbygiving itmeaning.However,theonetothreeyeartimeframeofcapitalstructureanalysisdoes giveitsomeperspective.Althoughhistoricalpatternsseemtorepeatthemselves“witha newtwist”,capitalstructuralismdoesnottrytoforecaststockprices;itonlyattemptsto identifyanenvironmentconducivetoincreasingthem.Infact,thesamedynamicsthat provedsuccessfulforIBMinthe1960shavebeenprovensuccessfulforGoogleinthenew millennium;bothcompanieshaveworkeddiligentlytooptimizetheirrespectivecapital structures.Whilethe“devilisinthedetails”,thesefirmshaveimprovedstockpriceby increasingtheflowofearningstowardshareholders-mostlythroughcapitalgains.In effect,theabilitytokeepearningsexpectationshighhasbeentheprimeingredientinstock priceappreciation.Bystavingoffbankruptcycostsandkeepingthecostofcapitallow enough,earningshavebeenmagnifiedincomparisontothecompetition.Neithercompany worriedaboutthe“clienteleeffect”ofkeepingtheirsharepricestoohighandnotsplitting them;thepremiumwasplacedonminimizingthenumberofsharesoutstanding. Mostimprovementinstockpriceshappensconcurrentlywithearnings improvements.Ifthiswerenotso,“investingafterthefact”wouldbeprofitableand simple.Althoughsomemoneyismadeoffof“momentum”insomemarkets,WallStreet triestoseparateinvestorsintotwocategories:thosewhomakegoodjudgmentswellahead oftime,andtheneveryoneelse.Infact,mostfirmswillreachanoptimalcapitalstructure sometimeduringabusinesscycleanddosowhentheirentiresectorisdominatingthe market.Ifthehousingsectorisdominant,forexample,thepaintcompanieswillnotbefar behind.Itisacycleofmatchingtheopportunitiescreatedbytherelationshipbetween interestratesandequities,withthestructureofthefirmsthatcanmosttakeadvantageof it Asanexampleofaverytypicalscenario,considerthehypotheticalXYZCompany. Atthetopofabusinesscycle,theyhavenowincreaseddebttoequityfortwoyearsina 59 row,andtheyarewellpasttheiroptimalcapitalstructurewithtoomuchdebt.Their earningshavebeentepid,butnowtheysensethattheirinvestmentinChinesefurnitureis goingtobea“cashcow”andpayoff.Inthethirdyearoftheirinvestmentcycle,sales jumptwenty-fivepercent.Consumershavemoneytospend.Increasedearningsmovethe stockupthirtypercentandthefirmdecreasesitsdebttoequityratiobypayingoffsomeof itsloansandretainingearnings.Investorswhoenteredthisgameearlyenoughreceivethe spoilsofvictory.Aftertwoyearsofdiminishingitsdebt,thefirmmaybepastitsoptimum intheotherdirection,andinvestorsbeginignoringthestockbecausereturnsarenotas substantial.Atthispoint,thefirmneedstoregroupandfundnewprojectswithahigher risk-whichmayagainmeanincreasingitsproportionofdebt. Dartthrowingmonkeysnotwithstanding,someofthevolatilityinthemarketis attributabletobusinesscyclefluctuations;thosefirmswiththegreatestoperatingriskcan onlyperformforabriefamountoftimewhentheirparticularsectorisfavored.The performanceofsuchfirmsmayentailaneightypercentriseinthestockpriceovertwoor threeyears,followedbyweakorevennegativeresults.Lowertotalleverage(operating leveragemultipliedbyfinancialleverage)willbuffersomeoftheeffectofthebusiness cycle,butevenlow-riskfirmswillhavebriefperiodsofwildprofitabilityfollowedbybelow averageresults. Howdoesthebusinesscyclesuddenlymatchafirm’sstructurewiththepatternof availablefunding?Eachfirmhasanoptimalproportionofdebttoequitythatrespondsto thelevelofinterestratesandtheirrelationshiptoequity.Firmswhohavelowoperating risk,forexample,mayusemorefinancialleverageandfundprojectswheninterestrates arelowatthebeginningofarecovery.Atthispoint,thefirm’searningswillbegintofar outpacethecostofcapital,anditsstockpricewillsoar.However,thistypeofopportunity canonlyberealizedwhenthefirmismovingtowarditsoptimalcapitalstructure;the distancethefirmneedstotravelactuallyaccentuatesboththeriskandthereturn. FOUR“POSTULATES” 60 Intheworldofequities,hardandfastrules“breakliketwigs”andsowecallthese observations“postulates”withtheknowledgethateachwillbebrokenatsometimeor other: • 1)Allvariablesheldequal,moreearningstendtodecreasetheproportionofdebtto equity(D/E),becausethefirmwillpayoffsomeloansand/orincreaseretention. • 2)Companieswhosimultaneouslyandsubstantiallyincreasebothdebtandearnings maypayoutinincometaxesmuchmorethantheyhavedeductedininteresttax savings.Itisnocoincidencethatcompaniestimetheirdebtissueswhenearningsare lower. • 3)Thegreateststockreturnsoccurwhenearningsareacceleratingupwardwhilethe costofcapitalisacceleratingdownward. • 4)Smallblocksofdebtareprohibitivelyexpensivebecausethereareeconomiesofscale whenbondsareissued.Firmswhogarnerlargeloansusuallydosowithstrategic purpose.Alternatively,firmswhoincreasedebtbysmallamountsonaconstantbasis mayhavecostoverrunsorproblemswithremainingsolvent. SHARELIMITATIONS Theeffectoffinancialleverageistoreducethepotentialnumberofshares outstandingbyfundingwithdebtinsteadofequity.Thefirmreceivesatradeoffbetween thepotentialamountandvariabilityofearningspersharebecausenetincomeandmarket pricearenotdilutedbymoresharesoutstanding.However,givenachoice,mostfirms wouldfundwithsufficientretainedearningsaslongasdividendgrowthwereadequate. Thefactthatafirmneedstomakeachoicebetweenthe“lesseroftwoevils”(debtormore shares)isindicativeoftheinadequacyofinternallygeneratedfunds. Thisinsufficiencyisinnowayapejorative.Manyindustriesdonothavethe earningscapacitytodocontinualinternalfunding.Thesearewell-managedandprofitable companieswhohappentobeinanindustrythathavehistoricallylowmargins.Infact, theirabilitytoincreasetheirstockpricerestswholeheartedlyonmanagingacapital 61 structurethathashigheramountsofdebtbecauseoperatingleverageissolow;large increasesinsaleswilltranslateintosmallincreasesinoperatingincome.Thepremiumis placedonkeepingshareissuestoaminimum,andevenlimitingretainedearningsby payingasteadilygrowingdividend.Thistypeofcontrolovercapitalstructureallowsthese firmstocompeteinmarketsthathaveplayerswithprofitmarginsfivetoseventimesas much.Infact,anyquickstatisticalsurveywillfindthatindustrieswithmoredebttendto havelessstockpricevolatility-whichseemstobeananomaly-untiloneconsidersthat firmswithlessoperatingriskcanincurmorefinancialrisk. Firmswhoarefundedwithanall-equitystructurefacea‘doubleedgedsword”.On theonehand,theyareusuallyveryprofitable-periodically-andfundtheirprojects internallyfromretainedearnings.Ontheotherhand,theneedtocompeteandreplacea highleveloffixedassetsrequiresaconstantsourceoffundingwhichfurtherrequiresthese firmstoissuesharesofstock.Sincedebtisunwarrantedgiventhelevelofoperatingrisk, thesefirmswilldilutetheirEPSandmarketpricewithmoresharesoutstanding.Thus,the morestableandlargeistheiroperatingincome,thefewersharesneedtobeissued.Those “diamondsintherough”thatarefortunateenoughtohaveahighoperatingleveragewith stablesalescanfundalloftheirneedswithinternallygeneratedretainedearnings. However,thesecompaniesareusuallysmall,andwhenWallStreetrequiresthemtogrow, therewillbesometradeoffmadebetweenstabilityandthemethodoffinancing,i.e.,more sharesoutstanding. ADAPTEDMEASUREMENTS Thistextaccentuatesthetradeoffbetweenlong-termdebtandcommonequity.It considersallothersourcesofcapitaltobeadjunctsthatattempttolowerthecostofcapital. Whileothersourcesoffundingmayslightlychangetheriskprofileoftheentirefirm,the crucialcomponentsaretheamountoflong-termdebtandcommonequitybecausethese requirethemostexpenseandobligation.Ourdefinitionofcapitalmayincludeallof stockholders’equityandallliabilitiesforthepurposeoftheoreticalillustration:indeed, 62 anycorporationmustitemizeeverysourceofcapitalwhenitimplementsaproject. However,whenwecomparecompaniesandmakeinvestments,weexcludepreferredstock andinterestbearingdebtoflessthanoneyear’smaturity,becausewedesireastrict, categoricalriskmeasurementthatisremovedfromcorporateeffortstominimizethecost ofcapital. Forboththeinvestorandfinancialmanagement,thefocusneedstobeplacedon evaluatingthefirmthroughitslong-termcapitalobligationsbecausethesuccessofthe companyrestsontheirviability.Byeliminatingothervariables(preferredstockandother interestbearingdebt),weinnowaydiscounttheirimportance:infact,short-termdebtis viewedasamajorelementintheneedtofundwithlong-termdebt.However,wedowant indicativemeasurementstofocusonthetradeoffbetweenlong-termdebtandcommon equityandsoweadaptmeasurementstofitthisurgency.Forexample,insteadofalluding totheproportionofdebttoequity(D/E),weconcentrateonlong-termdebttocapital (LTD/CAP)whichismoresensitivetochange(mathematically)andbetterelucidatesthe tradeoffbetweencapitalobligations.Moreover,weusetheratio,returnoncapital(ROC) morethantheratio,returnonequity(ROE),simplybecauseinourmorenarrowdefinition ofcapital,thefigureismoreresistanttofalseinterpretation.Anotherexampleappliesto theweightedaveragecostofcapital(WACC).Bynarrowlydefiningcapital,weeliminate someoftheriskadjustingeffectsofothersourceslikeshort-termdebt,andformacostthat isdependentonlong-termdebtandcommonequity;itmaybeahigherfigurethanthe actual,butbettergaugestheriskofthesetwocomponents. EXPLICITVERSUSIMPLICITCOSTS Wearealreadyfamiliarwithsomeoftheexplicitcostsofcapitalstructure-those paidinanactualexchangeofcash.Costsaremadeupoffixedandvariablevarietiesthat togethermakeupthetotalcostwhenoperatingincomeissubtractedfromsales.Someof thesecostsinclude:wages,rent,machinerymaintenance,materialsandofficesupplies. 63 Interestexpenseisindeedaprominentexplicitcostthatneedstobepaidregularly. However,nolessimportantarewhataretermed,“implicitcosts”-coststhathaveaneffect onthepriceofthestockbutaredifficulttoenumeratebecausetheydonotrepresenta physicalasset.Ineffect,sincecapitalstructureanalysisinvolvesmakingchoicesbetween competingactions,manyofitsdecisionsarebasedontheseinherent“implicit“costs. Oneofthemostfamiliarimplicitcostsisdilution.IftheXYZCompanyhas100 sharesofstock,eachwithearningsofonedollar,increasingthenumberofsharesto110, willhaveanimplicitcostof(1-(100/110))x(110)=10dollars.Theactionofincreasing outstandingequitybytenpercenthadtheneteffectofreducingEPS,whichhasaneffect onthepriceofthestock,butdoesnotreducethefundamentalsonthebalancesheet. Anotherimplicitcoststemsfromdelayingactions.Forexample,Icanchoosenotto installpollutioncontrolequipmentandincurasmallfine,orIcanspendtoomuchfora systemthatwillbebothlessexpensiveandobsoleteinafewyears.The“fine”willbecome partofthebalancesheet,butthedecisiontodelayandsavemoneyhasnocorresponding entry. Sincecapitalstructureanalysisencompassesdecisionsaboutchoicesamong alternativeactions,theprimaryimplicitcostistermedan“opportunity”cost,againorloss thatoccurswhenwechooseoneactionoveranother.Thus,anopportunitycostimplies thatwearecomparingthecostoftwodifferentactions.Forexample,ifbondsarepaying sixpercentandstocksarepayingninepercent,myopportunitylossisthreepercentifI choosebondsoverstocks.Acomparativereturnonequity(ROE)ofcompetitorsinan industryexhibitsmanyofthecharacteristicsthatweterm,“thecostofequity”;thereisno physical,“upfront”cost,butifmyfirmunderperformsitspeersinROE,thereissome adjustmentmadetothestockwhichisdifficulttopredictorenumerate.Infact,anytime thatananalystresearchesindustryaverages,sometypeofcomparativeparadigm,an opportunitycostsotospeak,isbeingformedinhisorhermind.Thus,weoftenformthese costsunconsciously. 64 IMPLICITCOSTSOFDEBT Naturally,interestexpenseistheexplicitcostofdebtandtaxdeductibilityis acknowledgedwhenweuseittoformthecostofcapital.However,severalotherimplicit costsexistwhendebtisincurred. • 1)Theinterestrateneedstobecomparedtonotonlycompetitors’rates,buttothe risk-freerateofthetenyeartreasuryandtheaveragesintheequitymarketsaswell.If itistoohigh,debtneedstobecurbedandmoreearningsneedtoberetained. • 2)The‘real”costofinterestmaynotonlyincludetaxdeductibilitybutinflationas well.Highinflationhasavaryingeffectonfirmswithgoodcreditbecauseassets appreciatewhileloansarepaidoffindepreciateddollars.Thecosttosomefirmsis excruciating. • 3)Thecostofbankruptcyisalwaysimplicitunlessthefirmisactuallybankrupt.The firmneedstodeterminewhichassetscanbesecuredascollateralinadditiontothe immeasurableeffectonthestockofchangesintheprobabilityofdefault. • 4)Theeffectonthecostofequitymustbedetermined.Ifmoreleverageraisesthe returnthatinvestorsrequiretoinvestinafirm’sstock,canthefirmbeprofitable enoughtowarranttheincreaseindebt?Willdemandforthestockactuallydecrease,if somethresholdamountofreturnisnotsurpassed? • 5)Theimplicitcostofpossibleassetimpairmentmustbeexamined.Somerestrictive covenantsinbondindenturesrestricttheuseofassetsandputotherrestrictionsonthe actionsofmanagement. • 6.Theimplicitcostofimpairingfuturefinancialflexibilitymustbeexamined.Evenif interestratesdecline,thereissomecosttorefinancingaloan.Similarly,nofirmwants tobeladenwithdebtatthetopofamarketbecausethisisthepointofgreatest earningsopportunitiesformostcompanies. THEIMPLICITCOSTOFEQUITY 65 Thecostofequityistherateofreturnthatinvestorswillrequiretoinvestinafirm’s stock.Itisusedsynonymouslywiththeterm“requiredrateofreturn”andisreferredtoas an“opportunitycost”becauseitcomparestherateofreturnoffirmswithsimilarriskina leastsquarestypecorrelation.Whenequityisissued,theonlyupfrontcostswillbe “flotationcosts”whichareapaymentorapercentageoftheproceedstotheunderwriting firm.Whenequityisbuiltthroughretainedearnings,therequiredrateisappliedtoall retainedearnings(notjustthecurrentyear’s)aswellasalloutstandingstockthathasbeen issuedbythecompany.Thus,the“costofequity”isalmostentirelyimplicitandshiftsin valuefromyeartoyeardependingontherateinvestorswillrequire.Ina“bull”market, thisratenaturallyrises,whileina“bear”market,itdeclines. Anotherimplicitcostarisesinthetimingofastockissue.Ifafirmis“maxed”out onitscreditandisnotearningenoughtoraisecapitalthroughretention,itmustmeetits fundingneedsthroughissuingstock.However,thepricereceivedwillbediminished becauseinvestorswillnotfindthestockattractive;ifthefirmhasatargetlevelofcapital requirements,itwillneedtoissuemanysharestoachieveit.Thus,thereisathreshold pointwherefundingshouldbedelayedoravoidedbecauseitdiminishesthemarketprice ofthestocktoodrastically.Alternatively,afirmcanissuestockwhenthepriceishigh, receivingthemostcapitalpershareissued.Thislattertacticraisesadequatecapital,but maybeundertakenwhenthecostofequityisveryhigh-suchasattheendofabusiness cycle.Whenthecostofequity(therequiredrateofreturn)isexceptionallyhigh,thefirm mayhavetroublecoveringitwithadequateearnings.Theresultisoftenalarge adjustmentdownwardbecauseperformancedoesnotmeettheover-hypedexpectations. Thefinalstrategyismostpreferredbyinsidersandlargeinvestors:issuestockwhenthe priceislowenoughtoappreciatesubstantially.Whenthemarkethasnotfactoredin expectedearningsfromprojectsthatitknowsnothingabout,theriskandcostofequityis low.However,thecompany’scapitalstructuremustbeviableenoughnottodependona stockissueforitstotalfunding;anyCFOknowsthevalueofadiversifiedmixofsourcesof 66 funding.Anytimethatafirmdependstoomuchonasinglesource,thereismoreriskofa highercostofcapital. Tosummarizetheimplicitcostsofequity,wecanputthemintooneofthree categories: • 1)Theimplicitcostofdilution-ThefirmmustconsidertheeffectonbothEPSand marketprice,aswellasfuturedividendobligations. • 2)Theimplicitcostofthe“requiredrateofreturn”-Investorswillnotdemandthe stockofacompanythatunderperformsitspeers.Afirmthatdoessowillhaveavery high“opportunitycost”andbeunabletocoveritwithenoughearnings. • 3)Theimplicitcostoftiming.Raisinglargeamountsofcapitalwithanequityissuehas manyrepercussions.Theonlytimeitseemsjustifiediswhenamajormergeroccursin afavorableeconomicenvironment.Thusequityissuesshouldberelegatedto “executivecurrency”ratherthanexistasamajorsourceoffunding. THEMOMENTOFTRUTH Ultimately,mostinvestorswanttotimethemarketsothattheyareintheearly stagesofalargepayoff.Moreoftenthannot,thatscenariooccurstoanyinvestorwhois welldiversifiedandstaysinthemarketlongenough,despiteitsvolatilechanges.Itisnota frequentoccurrence.Ifthestudent/investorobservesthecorporatesideofanyinvestment, heorshewillunderstanditasashiftincapitalstructurewhereacceleratedearningsbegin topropelthecombinationsofassets,debtandequityinaparticulardirection.The uncertaintyisderivedfromthetimingofthatprospectandwhetheritwilloccuratall. Naturally,itistheprerogativeofmanagementtokeeplagtimebetweeninvestmentand payofftoaminimum.Someindustries,however(likepharmaceuticals),willhavealong lagtimebutreturnmoreoncethepayoffoccurs.Infact,morethanoneinvestorhaslefta firmonlytofindthatmorepatiencewouldhaveledtoprofitability.Inevitably,investment screensaredesignedtofailbecausethemarketwillchangeandmakesurethattheydo. 67 But-thereareafewsignalsthatarerelatedtobothcapitalstructureandearlyinvestment success. • ExecutiveTrades.Whenacompany’sownexecutivesarebuyingstockandaredoing sofromaleveragedposition,thetidemaystartshiftingtomoreequityfinancing. • Althoughfinancialstatementsoccur“afterthefact”,lookforquarterlyimprovement incapitalturnover,%∆ ∆∆ ∆Sales/%∆ ∆∆ ∆Capitalor%∆ ∆∆ ∆Sales/%∆ ∆∆ ∆Long-termdebt • Investinginacompanywhoisincreasingitsproportionoflong-termdebttocapitalis riskierthaninvestinginacompanywhoisbuildingequity.However,thereturncanbe greaterifthefirmknowshowtousedebtstrategically.Theriskshouldbe accompaniedbysomeconfirmationfromanalyststhatearningsaregoingtoimprove. • Lookforashifttoasmallerproportionoflong-termdebttocapitalaswellasashiftto alowerfinancialleverageratio(EBIT/(EBIT-InterestExpense)).Atfirst,thetwo ratiosmaybe“outofsync”;whenearningsincrease,thefinancialleverageratiowill begintodropinharmonywiththeotherratio. • Knowthebusinesscycle.Forexample,expectinglargegainsfromthehousingsectorat thetopofthemarketmaybewishfulthinking.Earningsacceleratewhenasectoris receivinghighdemandatthesametimethatitscapitalcostsarelow. (BacktoTableofContents) 68 APPENDIX:THENETOPERATINGINCOMEAPPROACHTOSTOCKVALUATION Student/investorsareadvisedtotakethemostconservativeapproachtovaluinga stock.Whilemoreleveragemayescalatethepriceofafirm’sstockinaworldwithout bankruptcy,realitydictatesthatreturnsmustbeevaluatedinthedomainofrisk. Academiciansdevelopedtwomethodstocontrastopposingviews.Thefirstmethodwas calledthe“netoperatingincome”methodandpostulatedthattheextrareturnfrom leveragewasbalancedoutbytheextrarisk,contributingnoadditionalvaluetothefirm. Thesecondapproachwasthe“netincome”method,whichproclaimedthatafirm’svalue wasanextensionofitsdegreeofleverage,andtherelationshipbetweenitsinterestrateand thecostofcapital.Bothapproachesweredevelopedinahypotheticalworldof“perfect competition”-notaxes,bankruptcycosts,ordifferentratesofinterestbetweenfirmsand individuals. Whilebothmethodsvalueacompanyasthesumofitsbondsanditsstock,thenet operatingincomeapproachdeterminesthevalueofthecompanyastheratioofcapitalized operatingincome:thatis-operatingincomedividedbythecostofcapital.Itthen subtractsthevalueofthefirm’sbondstodeterminethevalueofitsstock.Inmathematical notation,thevalueofthestockis:(X/CostofCapital)-Bonds,whereXisequalto operatingincome.Ontheotherhand,thenetincomeapproachtovaluingafirm’sstock subtractsinterestexpensefromoperatingincomeandthendividesthisdifferencebythe costofcapital.Inmathematicalnotationitis:(X-(InterestRate)(Bonds))/Costof Capital.Itthentakesthisvalueofthestockandaddsthevalueofitsbondstodetermine thevalueofthecompany.Asanexampleofthetwoapproaches,considerafirmthatis capitalizedat$10000withD/Eof0%,50%andthen100%.Operatingincomeis$1000 andiscapitalizedat10%.Theinterestrateonbondsis5%. 69 Table3-3 NetOperating IncomeMethod D/EPercent 0 50 100 NetOperating income 1000 1000 1000 CapitalizationRate 10% 10% 10% TotalMarket Valueofthe Company 10000 10000 10000 MarketValueof Bonds 0 5000 10000 MarketValueof theStock 10000 5000 0 Table3-4 NetIncome Method D/EPercent 0 50 100 NetOperating Income 1000 1000 1000 InterestExpenseat 5%onBonds 0 250 500 NetIncome 1000 750 500 Capitalizationrate 10% 10% 10% MarketValueof theStock 10000 7500 5000 MarketValueof Bonds 0 5000 10000 Totalmarket Valueofthe Company 10000 12500 15000 Bycapitalizingoperatingincomeanddeductinginterest,thenetincomeapproach addsasubstantialamounttoboththemarketvalueofthecompanyanditsstock. However,thenetincomeapproachassumesthatthereisnorisk;allincreasesinEPSare immediatelytransferredintothepriceofthestock.Alternatively,thenetoperatingincome 70 approachassumesthattheriskofleverageperfectlybalancesthepotentialeffectonEPS andthatriskandreturncanceleachotherout.Inthenetoperatingincomeapproach,the valueofthecompanyiscalculatedfirst,andtheamountofbondsissubtractedto determinethestockprice.Inthenetincomemethod,thevalueofthestockiscalculated first,andtheamountofbondsisaddedtothisfiguretodeterminethevalueofthe company.Thestudent/investorwillobservethatiftheinterestrateisthesameasthecost ofcapital,thereisnodifferencebetweenthetwomethods;indeedthenetincomemethod rewardsmanagementforkeepingbothratesaslowaspossible. Miller/Modiglianiarguedthattheonlycorrectapproachwasthenetoperating incomemethod.TheirPropositionIarguedthatinaworldwithouttaxes,nogainwould begarneredfromleveragebecausetheinterestratewouldalwaysapproachthe capitalizationrate.Thusinaperfectlycompetitiveeconomicenvironmentwhere individualsandfirmscanlendatthesamerate(nobankruptcy),therewouldbenobenefit fromtheproportionofdebttoequityinthecapitalstructure.Thesameamountof earningswouldflowtotheshareholdersregardlessofhowthefirmwasfunded (BacktoTableofContents) 71 4 THECOSTOFDEBT Whilewedefinecapitalaslong-termdebtandequity,wecannotneglectthe importanceofshort-termdebtandallcurrentliabilities.Infact,nexttoproportional capitalallocationitself,thestrategicuseof“workingcapital”formsthebackboneof sustainableprofitability.Althoughmostindustriesgothroughsomesectorandmarket volatility,itisworkingcapitalmanagementthatguidesacompanythroughthetroughof businesscyclesandgrantsittheflexibilitytotakeadvantageofthepeaks.Thereare, however,severalcharacteristicsthatmakeeachtypeofdebtunique.Despitecreating productivesynergy,thedifferencebetweenlong-termandshort-termcreditissubstantial enoughtocausetheeccentricpricingofrisk.Iftheclassicinvertedyieldcurveisasure signofimminenteconomictrouble,itaccentuatestheproblemofmatchingcostwithrisk. Whilethetwoareinseparable,theyarenotthesame,andcosthasatendencytoadjustto riskratherthantheopposite.Hereinliesamajorproblemincapitalstructure.Partofour solutionistofindaninterfacebetweencostandriskthatreconcilestemporaryinequalities. Weproposethatinterfaceinthischapter,anddefineitasthecostofbankruptcy. THEPROBLEMOFSHORT-TERMCREDIT Thesignificanceofshort-termdebtespeciallybegsthequestion,“Whynotdefine capitalstructureinthetraditionalmodeoftheproportion,debt/assets?”Whileitistrue thatsomecompanieswillfundlong-termfixedassetswithrevolvingcreditfromabank,the pricingandquantificationofriskbecomesobscured.Forexample,onecompanymaytake outshort-termloanstotakeadvantageoftradecreditdiscountsbecausetheinterestonthe loanislessthanthediscount.Anothercompanymaybemuchlarger,buysinvolumeand nevertakesdiscountsatall;infact,itmaytypicallyextendpaymentperiodswellpastdue andthevendortacitlyacceptsthisbehaviorbecausetheaccountissoprofitable.Which companyhasthelowerrisk?Theanswerisambiguousbecausewearecomparingan 72 opportunitycostofdebttoabusinesscondition.Thefirstcompanyismanagingitscosts, whilethesecondcompanyistakingadvantageofitsabsolutesizeandpower;thecost savingsfrom“floating’anon-paymentmaybejustasgreatasthefastidiousmanagement ofcredit. Althoughsomeacademicliteraturearguesthatshort-termdebtisonlyafactorin totalriskanddoesnotcontributesignificantlytomarketrisk,itmustbeobservedasan adjunct.Asbothalternativefinancing,andasasignalforapotentialincreaseinearnings, ariseinshort-termcreditcanhelpoptimizelong-termdebtinsomesituations.While short-termdebtmayhavenoeffectonmarketriskatsomepointsinthebusinesscycle,at othertimesitmaybeadeterminingfactor-forexample,asasubstituteforlong-termdebt attheendofacycleRatherthancommittothehigherinterestpaymentsofalargedebt issue,afirmmaywanttowaitouttheuncertaintyinanticipationoflowerrates.Short- termcreditencouragesthisfinancialflexibility. Whilelong-termdebtandequitycanbegaugedintermsofcostandobligation, short-termcreditislessamenabletoriskanalysis-onashareholderlevel.Thedeceptive qualityofshort-termdebtencouragestheconfusionofprofitabilitywithinsolvency.For example,aninventorybuildupcaneitherbeanticipatoryofsalesincreasesorasignalthat demandistoolow,dependingonthetiming.However,accountspayablewoulddisplaythe samelargebalanceregardlessoftheboom-bustcondition.Infact,asmorepurchasesare made,currentliabilitiesoftenrisedramatically,butthesameconditionarisesduringa shortageofcash. Inessence,thereisanon-linearrelationshipbetweenthecostofshort-termdebtand itsrisk.Unlikelong-termdebt,whichalmostperfectlycorrelatesprice(interest),andrisk, short-termdebtgetspricedintermsoftherisktothecreditorwithlessreferencetothe long-termviabilityofthecompany.Sinceshort-termdebtisnormallylessexpensivethan long-termdebt,itistemptingforafirmtocontinuallyfundlong-termprojectswithshort- termloanstokeepcapitalcostsataminimum.Suchastrategyisfraughtwithtwomajor 73 risks:1)Theloansneedtobefrequentlyrenewed,andthecompanymaynothavethecash flowatthetimeitisneeded.2)Interestratesvary,andashort-termloanexposesthefirm tobothapotentialriseinrates,andincomevolatility.Tocircumventtheserisks,most companieswillmatchthecash-flowsfromoperationstothematurityoftheirdebt,andend upfundinglong-termprojectswithlong-termloans.Ontheotherhand,firmswithmore seasonaldemandschedules-farmers,skilodgesandgolfcoursesforexample-wouldbe moreamenabletoshort-termcredit.Thepayoffwouldbemorecertainandbasedon historicalrepetition. INTERESTEXPENSEINEQUALITIES Anotheranomalyoccurswheninterestexpenseattheendoftheyeardoesnot reflecttheactivityintheshort-termcreditmarketduringtheyear.Interestexpensemay escalateduringthesecondandthirdquartersbutloansgetpaidoffjustintimeforthe annualincomestatement.Theresultisaninterestexpensethatdoesnotreflectthegreater amountofriskduringtheperiod.Infact,a“rogue”financialexecutivecanfinancewith short-termdebt,exposingthefirmtotheriskofdefault,buttimethecash-flowswith “luck”andpayofftheloansontime.Verylittleofthisactivitywillbereflectedinannual reportsanditappearsthattheexecutivesuccessfullyfundedlong-termprojectsand loweredthecostofcapitalsimultaneously.Moreover,ifafirmchoosestofinanceinventory withavendor,interestexpensecaneasilybesubsumedinto“costofgoodssold”,andthe higherriskwillnotbereflectedontheincomestatement.Othertimes,firmswill“netout” theirinterestexpensewithinterestgainedinsellingsecurities;suchobfuscationis sometimesaredherring:transparencycanbeconfoundedwithbotheuphemismsand extensivecircumvention.Thus,investorswhoneedaprecisebreak-downofinterest expenseattributabletolong-termdebtmustreadthefinancialnotesinafirm’s10Kor annualstatements.Eachmaturingissuewillbematchedwiththedateofmaturityandan interestrate.Forexample,inacompanywithaninterestexpenseof9millionand100 millioninlong-termdebt,asamplebreakdownisasfollows: 74 Table4-1 DEBTAMOUNT MATURITY INTERESTRATE 33.33(MILLION) 2009 5.5% 33.33 2011 7.5% 33.33 2013 10.5 Theamountofinterestexpenseattributabletolong-termdebtistheamountofeach maturitymultipliedbyitsrespectiveinterestrateandthensummedtogether: 33.33(0.055)+33.33(0.075)+33.33(0.105)=8.082 8.082millionininterestexpenseisattributabletolong-termdebtandjust$917,933 (0.917933)wasattributabletoshort-termdebt.Toconcludethetreatmentofshort-term debtwithinthecapitalstructure: • 1)Capitalstructureasdefinedinthistextisacombinationoflong-termdebtand stockholders’equity.Asshort-termcredithasamajoreffectonthesecomponents,itis anadjunctforcetotheriskofcapital. • 2)Withoutacomponentbreakdownofinterestexpense,theremaybesome computationalerrorifallinterestisattributedtolong-termdebt. • 3)Short-termcredithasanon-linearrelationshipwithrisk.Ifitwereincludedinthe costofcapital,anyattemptsatminimizationofthatcostwouldbeskewed. RISK,RETURNANDTHESIGNIFICANCEOFSHORT-TERMCREDIT Thenatureoftherisk-returnconflictforshort-termcreditisthetimingofcash- flows.Asafirm’scurrentratio(currentassets/currentliabilities)declines,itapproaches whatisknownas“technicalinsolvency”suchthatitcannotmeetshort-termobligations. Atthesametime,itmaybegearingupforasuccessfulperiodofrevenuegenerationby buildinguptradecredit,engagingvendorsthroughpurchasesandcreatinginventory.A balancesheetwillshowthatcurrentsaleshavenotproducedadequateaccountsreceivable orcashforthatmatter,whiledebtispilinguponthe“accountspayable”sideoftheledger. 75 Thisisoftenthetimewheninsidersbeginaccumulatingstockbecauseitmaybedepressed, butisreadytosoarwhenrevenuesimprove.Theaverageinvestorperceivesacrisisrather thananopportunityandfailstoinvest. Anexaminationoftwoliquidityratiosandtwocash-flowequationswillrevealthe ambivalentanalyticalnatureofshort-termcredit.Wehavealreadyobservedthe“current ratio”whichalsohasamodificationcalledthe“acidtest”or“quick”ratio,currentassets- inventories/currentliabilities.Theseclassicmeasurementsofsolvencydonotalways indicateprofitability,however.Ahigherratiosimplymeansthatshort-termobligations canmoreeasilybepaid;itisrarelyindicativeofstockmovement.Thereasonisfoundin twootherequations,freecash-flowandthecapitalrequirementsequation.Thebasic unleveredfreecash-flowequationis: (EBIT)(1-T)+DepreciationandAmortization-CapitalExpenditures-∆NWA Table4-2 EBIT EarningsBeforeInterestandTaxes T TaxRate ∆NWA TheChangeinNetWorkingCapital (definedasthechangeincurrentassets- currentliabilities) Thestudent/investorshouldnoticethatasNWAincreases,cash-flowdecreasesandthat morecurrentliabilitiesincreasecash-flow.Thesecondequationisthecapital requirementsequation,whichgivesaroughestimateofoutsidecapitalrequirementsinline withthesalesforecast.Itis:(Assets/Sales)(∆Sales)-(Liabilities/Sales)(∆Sales)- RetainedEarnings. Boththeassetsandliabilitiesvariablesarethosethatincreasespontaneouslywithsales whichincludesmostlythecurrenttype.AsLiabilities/Salesincrease,theneedforoutside capitaldiminishes.Tradecredit,forexampleisaninternallygeneratedsourceoffunding. 76 Thus,whenshort-termcreditincreasesandtheoutlookforearningsispositive,the potentialreturnisfargreaterthantheriskofinsolvency. Thestandardmethodforevaluatingshort-termdebtisinthecontextofcash-flow. Whentheratio,“Assets/Capital”risesatthesametimethatcash-flowincreasesasa percentage(tenpercentinthelastperiodandtwentypercentinthisperiod,forexample), thereturnfromrevenueswillmostlikelybegreaterthantheriskofincreasingshort-term liabilities.Theprimaryconcernisthatoperatingincomeisacceleratingenoughtolower theriskofdefault. Dependingontheleveragesituation,theratio,Assets/Capital,followsaloosechain oflogic.Thedifferencebetweennumeratoranddenominatoriscurrentliabilities.Amore positiveoutlookfortheissueoflong-termdebtoccurswhenthisratioisincreasingbecause businessactivityissteppedupandtheturnovertimeforthelargerinvestmentmaybe shorter.Moreover,anincreaseincurrentliabilitiesmayshowtheuseofalternative sourcesofcapitaltokeepthecostofcapitalataminimum.Assets/Capitalincreaseswill alsocontributetoanincreaseinthereturnoncapital(ROC),asitisoneofthethreemajor componentsofthatratio.First,areturnonassets(ROA)isformedbymultiplyingprofit margin(NetIncome/Sales)byassetturnover(Sales/Assets).TheresultingROAfigureis thenmultipliedbyAssets/Capitaltoformareturnoncapital(ROC).Essentially,therise incurrentliabilitiessupportsinvestmentinlong-termdebtandequity,becausethose componentsarepurchasedatahighercost.Anytimethatshort-termcreditcanbe substitutedforcapitalwithoutunduerisk,thefirmmovestowardanoptimalcapital structure. THECORPORATECOSTOFDEBT Banksandratingsagencieswilldetermineacorporation’spotentialdefaultby analyzingitsvariousleverageratiosanditsfutureprospects.Ariskpremiumwillbe 77 attachedtoanappropriaterisk-freerate(atenyeartreasuryyieldforexample)andthis willbethecompany’sinterestrateinputforitscostofdebt.Thus,anytimeanewloanis negotiatedatadifferentrate,acompany’scostofdebtchanges.Anewratewillchangethe marketpriceofafirm’sdebt,dependingontheamountofdecreaseorincreaseintherate, andalsoonthecompany’spriorinterestexpenseobligations.Wheneconomicconditions dictateaninterestratechangeforanentireindustry,anewprimerateforexample,those firmswhodonotincurnewdebtwillstillhaveachangeintheirrespectivecostsofdebt. Thecostofdebtisanopportunitycostandiscomparedtobothcompetitor’sratesandthe ratethefirmwouldactuallypayifitchosetoincurnewdebt. Sinceincurringdebtimpliesataxdeduction,thenewrateofinterestismultiplied bythereciprocaloftheeffectivetaxrate(1–taxrate)andthisfigureisfurthermultiplied bythemarketpriceofthefirm’sdebt.Thetheoreticalunderpinningsofthisprocessare discussedintheappendixentitled,“TheRealCostofCapitalandWhattheInvestorNeeds toKnow”. Fortheinvestor,itissimplyoverly“researchintensive”toconfigureafirm’snew costofdebteachtimeitoccurs.Thecomputationalandinformationalrequirementsare notjustifiedbytheperformancegainsinaccuratelymeasuringrisk.Aballparkestimate canbeformedbyequatingeachinterestratethatafirmpaysonitsexistingdebtwiththe correspondingproportionalmaturityinthefirm’sdebtstructureandthenforminga weightedaggregate.Suchexpediencywillderiveaninterestexpensethatequatesafirm’s bookvalueofitsdebtwithitsmarketvalue.This”nominal“costofdebtfailstogauge immediatechangesinthecostofcapital,butcanbeusedasproxyinotherrelationalvalues suchasthefinancialleverageratioandtheTIE(timesinterestearned). 78 THENOMINALCOSTOFDEBTANDTHECOSTOFBANKRUPTCY Whencalculatingthecostofcapital,theanalystusestherealcostofdebtwhichis composedoftheinterestrateonthedebtandataxdeduction;itisthiscostwhich comprisesthedebtcomponentinvaluationmodelsandcapitalbudgeting.However,in ordertofunctionallycalculatetheoptimalcapitalstructure,weneedtoincorporateseveral othercostsandinherentrisksintothemodelandweaccomplishthisbyforminga comprehensive“costofbankruptcy”. Themainattractioninusingdebttofinancecapitalneedsisitstaxdeductibility. Interestexpenseisfullytaxdeductible,whichallowscompaniestogrowatafasterrate thantheywouldiffinancedsolelybyequity.Thetradeoff,aspointedoutinthechapteron leverage,isthatearningspersharemayincreaseinvariability.Ineffect,thegovernment givesanadvantagetofirmswithloweralbeitsteadiercash-flowsbysubsidizinggrowth throughtaxbreaks.Thosecompanieswithmorevolatileearnings,whowouldbeingreater dangerofdefault,simplycannotcompeteonthisbasis.Naturally,several“taxstrategies” emergewhenafirmcanuseleveragetoitsadvantage;themixofcash-flow,deferredtaxes andtaxgainsorlossesbecomesparamount. Thecostofdebtissimply,i(1-t),wherei=theinterestrateandT=theeffectivetax rate.Thus,thenominalcostofdebtisthesummationofeachproportionalmaturity multipliedbyitscorrespondinginterestrate,andmultipliedagainbythecurrenteffective taxrate.Theeffectivetaxrateappliesbecauseitistherateatwhichcurrentdeductions areconsidered.Consequently,eachmaturitylevelofdebthasthepotentialofcostinga differentamountindifferentyearsbecauseofachangingeffectivetaxrate.Aprecise enumerationofthecostofdebtmaycontaintenorfifteenseparatematuritiesbutisquite simpletocalculateinaspreadsheet.Ifonecalculatestheproportionofeachmaturityasa percentageoftotaldebt,andthenmultipliesbythecorrespondinginterestrate,anaverage interestratecanbeobtained.Forexample,afirmhas30(million)of7%debt,50of8% debtand20of6%debtwithaneffectivetaxrateof30%.Totaldebtis100million,andso 79 theproportionsare0.3,0.5,and0.2respectively.Theaverageinterestratecalculationis 0.3(.07)+0.5(.08)+0.2(.06)=7.3%.Inowhavethechoiceofmultiplyingthisrateby(1- taxrate)toachievearateof.0511or5.11percent.Whenthisfigureismultipliedbythe 100millionintotaldebt,thecostofdebtis5.11million.Alternatively,Icouldalsogo throughtheentirecalculationof(30)(.7)(.07)+(50)(.7)(.08)+(20)(.7)(.06)=1.47+2.8+ 0.84=5.11.Whilethetypicalshareholderdoesnotnormallyneedthetypeofprecisionthat isusedforcapitalbudgeting,moreinformationhelpsformbetterdecisions. Whydoesthegovernmentmaketheuseofdebttaxdeductible?Therearemany sidestothiscontroversialquestion.Perhapsthebiggestreasonistoencouragesmall businessesandthosewhohavesmallcash-flowstostaysolvent.Asteadyincomethatis amenabletodebtfinancingencouragessteadyrevenues(forthegovernmenttoo!)and steadyemployment.Whileitseemstogiveanunfairadvantagetocompanieswhocanmost afforddebtbothfinanciallyandstructurally,manyotheradvantagesareincurredwhen equityfinancingisimplementedinstead:considerthelackofinterestrateriskandcredit crunchesthatall-equityfundedfirmscanembrace.Also,firmsthatfundwithequitymay reapmorebenefitsingoodeconomictimesbecauseincomeflowsdirectlytothe shareholders. Thecostofdebtisbothacomponentofthecostofcapitalandarisk-adjusted precursortoanoptimalproportionofdebttoequity.Ifthis“symbiotic”relationship seemsambiguous,itisbecause“risk”and“cost”arenotalwayscompatible.Considerthe resultoflargecutsinthefederalfundsrate;asinterestratesdecline,theaveragefirmwill domoredebtfinancingbutnotenoughtoradicallychangecapitalstructure.Theabilityto usedebtfinancingisderivedfrombusinessriskandimplicitoperatingleverage.;nolevelof interestratewillaltertheabilitytopayinterestinatimelymanner.Infact,astheriskof defaultbecomesgreater,banksandunderwriterswillchargegreaterinterestratesto compensateforgreaterrisk,evenasthis“cost”iscircumventedwithgreatertax deductibility.Atextremelevelsofdebt,creditorswillsimply“turnoffthespigot”and 80 interestrateswillstabilizeatahighlevel.Thereinliestheproblemofusingthecostofdebt todetermineanoptimalcapitalstructure.Thecostofequityisincreasedbytheuseof leverage,butthecostofthatleverageneversurpassesit.Evenattheaforementioned extremelevelsofinterest,taxdeductibilitywillensurethatthecostofequityisgreater; thereexistsa“riskpremium”thatcompensatesshareholdersforadditionalriskof uncertainincome.Ineffect,theMillerModiglianipropositionIIbecomesoperative becausetheoptimalproportionofdebtinataxedeconomywillbeonehundredpercent-if talliedbycostaloneThisextreme“cornersolution”wasproposedundertheassumptionof azeroprobabilityofbankruptcy,whichfurtheremphasizesthesignificanceofdefault.An optimalsolutiontofindingtherightcapitalproportionexistsonlywhenthecostofdebtis reconciledwiththecostofbankruptcy. THECOSTOFBANKRUPTCY Forafirmwhousesleverage,thekeytoanoptimalcapitalstructureistouseas muchdebtassafelypossible.Theword“safely”isaconnotativeterm,whichneeds objectification,anditisnoaccidentthatcompaniesspendmillionsinriskmanagementto gainaprecisedefinition.Inessence,thecorrectproportionofdebtwilldeterminethe optimalproportionofequity,andsincedebtislessexpensivethanequity,thecostofcapital willbeminimizedatanygivenlevelofoperatingincome. Evenwithinanindustry,eachfirmisultimatelystructuredinauniquewaywith differentpatternsofriskandreturn,dependenton“niche”orspecialty.Toforma “generic”costofbankruptcyistoassumeadauntingtask,butinroadsintoasolutioncan bemadewhenwelookforcommonfactorsinthecalculation.Twoofthosefactorsare:1) Theprobabilityofdefaultand2)Theamountofloss. THEPROBABILITYOFDEFAULT Whileteamsofactuariessetlossliabilitiesintheinsuranceindustry,mostinvestors areatleastfamiliarwiththeFICOscore-thecreditratingwhichneverseemsfactually accurate,buthauntseveryonewhoevergotturneddownforaloanThereinliesthe 81 problem.Economicconditionschangeandusingthesamemethodologywithdifferent inputsmayleadtoinaccurateprobabilities.Infact,ascomplicatedasdefaultanalysishas become,mostanalystssettlefortheproverbial“ballpark”figure.Logit,probitand multipleregressionanalysisarevaluabletools,butprecisepredictionofnear-term probabilitiesremainselusive.Althoughspecificconditionslikehousefiresorhomeowner- liabilitylawsuitscanbedeterminedbyprobabilitydistributions,typesofvariablesrelating totheeconomyaremorevolatile.Lawsuitsandfireshaveahistoricfrequencythatcanbe extrapolatedintoalossschedule;thereisnosuchscheduleavailableforperiodsofhyper- inflationorvariousspeculative“bubbles”.Ineffect,theextremesofsolvencyanddefault aremuchmorepredictablethanthechanceoftransitioningtoahigherriskcategory becausethelatterisdependentonshiftingeconomicvariables. BothMoody’sandStandardandPoor’shavebeenratinglong-termdebtissuesfor manyyears.Thisratingisofprimaryimportancetoboththeratedcompanyandto investors.Infact,bondratingssetthestageforcapitalstructurebecausetheyhelp determinethecostofdebtinbothnominalanddefaultforms.Thenominalcostisaffected becausebetterbondratingsleadtofavorableinterestrates,whilethedefaultformis affectedbecausethoseratingsareconfiguredintermsofsolvency-theprobabilityof default. Whiletherearemanycommercialalgorithmsavailable,mostratingsystemsare veryspecificandapplytoaparticularindustryoreconomicoutlook.However,forthe purposesofcapitalstructure,thegenericratingsystemspromulgatedbyvariousacademics haveheldupovermanyyears.Infact,theyarerobustfortheveryreasonthatcommercial bankscannotusethemtosetinterestrates:theyofferageneralpurposeriskanalysis withoutthespecificityneededtogaugeaparticulareconomicenvironment.Testedagainst commercialalgorithms,theacademicgenericshaveeclipsedtheminaccuracyovercertain periods,butlacktheabilitytotransitiontoachangedeconomy. 82 EdwardAltman’s“ZScore”isperhapsthebestknownofthese.Theproductof multiplediscriminantanalysis,itisstill94%accurateinpredictingbankruptciesoneyear away,andabout72-80%accurateinpredictingbankruptciestwoyearsaway.Forthe student/investor,itoffersacommongroundwithotheralgorithmsbecauseitbasesits analysisonanarrayofperformancevariablesasapercentageofassets;asummationis performedonasset-relatedcomponentsandthenamathematicalprocedureisperformed onthesum.InthecaseofAltman’sZScore,theprocedureinvolvesdeterminingwhether thesumislargeenoughorincreasing.Inmostotheralgorithms,logitoperationsare performedwherethelogarithmof“theodds”areobtained.Usually,theanalystwill merelyplugindecimalratiosintoaweightedcoefficientequationandtallyascore,whilea spreadsheetdeterminestheprobability. Thealgorithmanditscomponentsareasfollows: 1.2(X1)+1.4(X2)+3.3(X3)+0.6(X4)+0.999(X5) Allentriesaremadeindecimalform,butthelastfigure,X5,maybeaninteger. Table4-3 ALTMAN'SZSCORE (X1)=WorkingCapital/TotalAssets (X2)=RetainedEarnings/TotalAssets (X3)=EarningsBeforeInterestAndTaxes/TotalAssets (X4)=MarketValueOfEquity/Liabilities (X5)=Sales/TotalAssets Sometimesthemarketwillbeinflated,andoneneedstoformaZscorefrombookvalues. Inthatcase,themodifiedZScoreis0.71(X1)+0.847(X2)+3.1(X3+0.42(X4)+0.998(X5) ThisZScorewillproduceafigurefromzerotothreeormore,wherezeroisbankrupt,1.81 isatroubledfirm,andoverthreeisconsidered“safe”.Progressfromquartertoquarter canbediscernedbyformingarunningtotalonaspreadsheetandanylargeincreasesmay signalstockpriceappreciation. 83 COMMERCIALRATINGSYSTEMS TheMoody’sandStandardandPoor’ssystemsareupdatedregularlyanduse probitanalysisamongothertechniquestocreateratings.Inmosteconomies,theyare accuratebutrequireasubscriptionfromusers.Investorsarethoroughlyfamiliarwiththe letteredratingsystemof“AAA”standingforhighinvestmentgradeandasmall probabilityofdefault,downtoa“D”ratingwhichindicatesdefaultprobabilityofone hundredpercent.AccordingtoMoody’s,thesixmostsignificantvariablesinprobit analysisare: • 1)(EBIT+1/3Rent)/(InterestExpense+1/3Rent+(PreferredDividends/0.65)) • 2)AdjustedDebt/AdjustedBookEquity • 3)CashandEquivalents/TotalAssets • 4)FiveYearRevenueVolatility • 5)RetainedEarnings/AdjustedDebt • 6)AssetGrowth Again,weseemanysimilartypevariablesasinAltman’sZscore:operatingincomemust coverimmediateobligations,salesmustbeadequate,theremustbeliquidity(cash,working capital)andperformancemustbematchedagainstassets.Formanyyears,Standardand Poor’susedacombinationofprobabilityalgorithmsandqualitativejudgmenttorate companies-whichisentirelylegitimatebecausemanyeconomicvariablescanonlybe measuredaftertheirqualitativeanalogsoccur.Thefollowinglistentailssomeofthe variablesusedinthisjudgment: 84 Table4-4 VARIABLESUSEDINDETERMININGDEFAULTPROBABILITY 1)Debt/Asset 2)TimesInterestEarned(TIE) 3)TimesFixedChargesCovered 4)CurrentRatio 5)"QuickRatio" 6)MortgageProvisions(Collateral) 7)CallProvisionsOnBonds 8)OtherRestrictiveCovenants 9)SinkingFundProvision(AccountForRetirementOfDebt) 10)RegulatoryClimate 11)Anti-TrustLegislation 12)OverseasOperations,Diversification 13)EnvironmentalFactors 14)ResourceAvailability,VendorStability 15)LaborRelations 16)PercentageBreakdownOfSales/Customer-Diversification Itisquitedifficulttoendowavariablelike“laborrelations”withanaccurateprobability thatenhancescreditprediction,whichiswhyanyattemptatprecisionmustbeupdated frequently.Thestudent/investorshouldrecognizethatsettinganoptimalcapitalstructure isaprobabilisticventuresubjectto:1)thetechnicalaccuracyofthealgorithm;2)the effectoftheeconomyontheprobabilityofdefault;3)thecurrencyofdata.Any determinationofanoptimalcapitalstructurewhileusingdefaultprobabilitiescouldnot realisticallydelineatebetween30percentdebtand31percentdebtbutshouldbeableto alerttheinvestortoexcessivelevels-suchasfiveormorepercentaboveorbelowthe target. TYPESOFBANKRUPTCY Oneproblemthatoccurswithgenericapplicationofbankruptcyalgorithmsisthe lackofuniformityamonglegaldispositions.Largecompaniesaretypicallyencouragedto filechapterelevenandreorganizeratherthanliquidate;theyhaveassetsandobligations 85 thataregreatenoughtocontinueoperations-underdifferentcircumstances.Government entities,creditorsandeventhelocalcommunityhaveastakeinseeingtheoperation continueinadifferentcapacity-eitherdownsized,boughtout,orbrokenup.Thus,while recoveryofassetsmaybeaslowas20percentinaliquidation,theyare95-100percent recoverableinareorganization;claimsaresetbythe“doctrineoffairness”andare recognizedbylegalandcontractualpriority. Sincethedecisiontoreorganizeorliquidaterestswithtrustees,theyneedtoanalyze twobasicissues:1)Isthefirmworthmoreasanongoingenterpriseorassoldassets?,and 2)Canearningscoverfixedchargesinthefuture?.Obviously,thetypeofindustryis significantbecausetheassetstructurewilldeterminewhetheritismarketable.Firmswho havemoreintellectualpropertyandyethaveagoodhistorictrackrecordofmeeting obligations,willbemoreofareorganizationtargetthanafirmwhotradesinan appreciatingcommodity;onefirmrequiresmanagerialforesight,theotherprospersonthe marketpriceofanasset.Althoughthescienceofprobabilitycanfactorinqualitative decisions,mostmodelswillnotdifferentiatebetweentypesofbankruptcybecauseofthe legalambiguitiesinvolved. THEAMOUNTOFLOSS Amodelofdefaultwilldeterminetheamountcreditorsreceivebyreconcilingthe amountofliabilitieswiththeassetstructure.Shareholdersusuallyreceivenothing. Indeed,partofthetensionbetweenbondholdersandshareholdersstemsfromthe establishmentofa“riskpremium”whichisderivedfromcontractualpriority;anytimea firmincurslong-termdebt,thestockholdersforfeitaclaimonthefirm’sassets.Onthe otherhand,bondholdersareentitledtoaseriesoffixedpaymentandreturnofprinciple andnothingelse;anyfutureearningsgainedfromtheinvestmentindebtaccruesto shareholders.Inessence,shareholdersclaimtheintrinsicvalueofthecompany,whichis thepresentvalueoffutureearnings.Bondholdershaveaclaimonanextrinsicvalue, 86 whichmaybemarketableassetsorprioritizedearningsflow-inotherwords-interest paymentsandassetsthatcanbeeasilysold. Therearenosetformulasorseriesofcalculationsthatenabletheanalystto determineagenericamountoflossinabankruptcy.Eachcaseisuniquebasedonlegal type,liquidity,structureofdebtandabsolutesize.However,financialtheoristsarewell awareofthesignificanceofpotentialbankruptcytocapitalstructure,andevenadaptittoa modifiedversionofMiller/Modigliani’shypothesis:V(l)=V(u)+TB-PV.Thatis,the valueofaleveredfirmisequaltothesummationofthevalueofanunleveredfirm,the productofthetaxrateandtheamountofbonds,andthedifferenceofthepresentvalueof bankruptcycosts.Thecontroversyentailsthedefinitionandenumerationofbankruptcy costs. Ifwestatethatastockisworthlessuponbankruptcy,weareimplyingthatthecost ofbankruptcyistheentirelossofmarketvalue,andyetimplicitinthatvaluearethe prioritizedclaimsofcreditors.Whenweconsidertheseclaims,thetruemarketvalue assumessomemonetarymultipleabovethem.Ineffect,weareseparatingtheintrinsic claimsonearningsflowthattheshareholderspossess,fromtheextrinsicvalueofsalable assetsthatcreditorspossess.Inbankruptcy,eachisaseparateentity.Duringnormal operations,the“extrinsic”assetsprovidetheframeworkforearnings;intrinsicvalueis basedontheprospectoffutureearnings.However,oncebankruptcycommences,the confluenceisnolongerviable.Unlessthefirmisreorganized,theassetsdonotproduce presentorfutureincomeandthestockisworthless.Creditorshaveaclaimtopartofthose extrinsicassetswhileshareholderslayclaimtowhatisleftover. Thus,itwillbehelpfultoputbankruptcycostsintoperspectivebyrelatingthevalue ofvariousassetswiththemarketvalueofequity.Boththepriorityofclaimsandthetype ofassetsdeterminetheamountofloss.Notethatshareholdershaveclaimsduringongoing operations,butwillforfeitthegreaterpercentageofthemduringbankruptcy. 87 Table4-5 ASSETCLASS PRIORITY TYPES COLLATERALIZED a.SeniorMortgageDebt a.Marketable b.AdministrativeFees b.Salvage c.AccruedLiabilities dNotespayable e.Debentures INTANGIBLE NONE-Byprior agreement a.Patents somecreditorsmayhave claims b.Trademarks onmarketablepatents c.Goodwill UNCLAIMED CommonStockholders DependentonPrior Claims Figure4-1 Theintrinsicvalueofthecompanyisequaltomarketvalueaslongasthecompanyis solvent.Oncebankruptcyproceedingsbegin,priorityisgiventocreditors,and shareholdersforfeittheirclamsonassets. Assetsaremadeupoftangibleassetsandintangibleassets.Tangibleassetsmayor maynothaveamarketvalue,andshareholdersmaylayclaimtosomeofthem. Analogously,intellectualpropertylikepatents,classifiedas“intangibleassets”,canhave Intrinsicvalueofthestock(ongoingoperations) BookValue Market Value TangibleAssets+Intangible TrueintrinsicValueatBankruptcy AmountofLoss ExtrinsicValue 88 marketvalueandbeclaimedbyeithercreditorsorshareholders.However,themajorityof intangiblesaremadeupbythedesignatedassetclass“goodwill”,whichisprimarilythe excessmarketvaluepaidforacquisitions.Atbankruptcy,itisvirtuallyworthless,unlessit isusedasabargainingtoolforadministrators;trusteescanpointtoacompany’shistoric responsibilitytocreditorsandtheindustry.Priorityofclaimsgoestoseniormortgage holders,followedbyadministratorswhomayget20-25%ofmonetarygainsfromasset sales.Subordinateddebtholdersmayreceiveanywherefrom20to70%dependingonthe claimandmarketabilityofassets.Theleewayforclaimsisquitelarge,andsome subordinateddebtholdersreceivenothingatall. Totheshareholder,theamountofassetsacreditorreceivesisimpertinent;heorshe isleftwithaworthlessinvestment.However,toviewtheamountoflossastheentire marketvalueofsharesisimproper.Stockholders,ineffect,agreedtopayariskpremium whentheydecidedtoincurlong-termdebtinanefforttoboostearningsandshareprice. Thispremiumwasadoubleedgedswordofunrealizedpotential:1)debtmightincrease thevolatilityofEPSbeforeitboostsshareprice,and,2)notonlyweretangibleassetsgiven uptocreditors,thepotentialtoraisetheprobabilityofdefaultwasgreater.Aslongas operationswereongoing,shareholders“rented’acertainportionoftangibleassetsto producefutureincome.Oncethecontractwasneartermination,thoseassetsrevertedto thecreditors.Anyunclaimedassetsbelongtothecommonstockholdersandthoseare designatedbytheratio,“tangiblebookvaluepershare”.Itistheamountashareholder wouldtheoreticallyreceiveuponliquidation.Therealityofmostbankruptcyproceedings, however,dictatesthatclaimsbyemployees,pensionfundsandevenauctioneerswill eliminatethepossibilityofsuchcompensation. Toconstructaworkingmodelofthisconcept,weneedthecalculationfortangible bookvaluepershare,whichis(loosely):(TotalAssets-IntangibleAssets(including goodwill)-UnamortizedDebt)/(NumberofSharesOutstanding).Next,weformaratio 89 betweentangiblebookvaluepershareandthemarketpriceofthestockandsubtractthis fractionfrom“1”. (1-(TangibleBookValuepershare/Pricepershare)).Lastly,wemultiplythisvalueby themarketvalueofthecompany,whichisdeterminedbytheproductofoutstanding sharesandpricepershare.Theentireexpressionisasfollows: AmountofShareholderLoss=(1-(Tang.book.Val.persh/Mkt.Val.persh))x(#sharesx MktVal.persh) Implicitinthislossisallmarketvalueaboveafirm’sextrinsicvaluenetofliabilities,which rephrasestheterm,“tangiblebookvalue”. Essentially,theamountoflosswillbesomefunctionofthesethreevariablesanditis therelationshipamongthethreethatcandeterminetheproperamountofdebttoemploy. Whenthesevariablesinteractwiththeprobabilityofdefault-whichgaugesthepotentialto coverinterestexpenses-acostofbankruptcycanbeestimatedBymultiplyingthetwo constructs:(ProbabilityofDefault)x(AmountofLoss)wecanestablishacomparative proxy.Althoughthetruecostofbankruptcymustbeconfiguredforeachuniqueentity, ourstructurallydependentmodelcanbeusedincapitalstructuredevelopment. Despiteitsutility,thecostofbankruptcymodelhassomeseriousshortcomings.The amountoflossisprobabilistic,dependingoninteractionwiththeeconomy,competitors andtheequitymarket.Forexample,aspeculativerunupinthestockwouldaffectthe amountofloss,perhapsevenseparatingitfromitstrueassetvalue.Secondly,the probabilityofdefaultneedstoformanexponentialcurve,penalizingfirmsatanincreasing ratefortakingontoomuchdebt.Alinearformatwouldallowfirmstoreapthetax benefitsofmoredebtwithoutviolatingtheconstraintsofbankruptcy.Sincemost algorithmsareformedfromprobabilitydistributionsthatemphasizecentraltendency, findingadefaultprobabilitycalculationthatisalsopreciseattheextremeswouldbe difficult.Anyoptimizationmodelmustallowfortheprimacyofall-equityfundingifthe assetstructureisnotamenabletofinancingwithdebt.Lastly,theamountofshares 90 outstandingisavariablethatisdeterminedoutsideofthemodel.However,atheoretically viablemodelwouldallowthatnumbertobedependentonnecessaryfunding-thelevelsof equity,debtandcapitalalreadyinthecapitalstructure.Toobservethisdiscrepancy,note thefollowing“paradox”thatoccurswhenastockhasbeen“beatendown”: Table4-6 THRIVINGCOMPANY NEARBANKRUPTCOMPANY #SHARES=100 #SHARES=100 TANGIBLEBOOKVALUE/SH=4 TANGIBLEBOOKVALUE/SH=4 MARKETPRICE/SH=20 MARKETPRICE/SH=7 LOSS=(1-4/20)x(100x20)=1600 LOSS=(1-4/7)x(100x7)=300 Thedifferenceinshareprice($20vs.$7)allowedtheamountoflosstobemuchlessforthe nearbankruptcompany;thatsmallernumbermayopenthedoorformoredebtfinancing inanoptimizationmodel-theoneelementthatwouldcertainlypushtheprobabilityof defaultupward. THEROLEOFDEBTINCAPITALSTRUCTUREOPTIMIZATION Debtisthefulcrumbetweenhighertaxbenefitsontheonehandandthechanceof lossontheother.Theoptimalamountofdebtwilloptimizeequityaswellbutitmustfirst bereconciledwiththeproductionofincome.Ifmoredebtviolatestheparameterssetforth inthecostofbankruptcy,futurecash-flowsarejeopardizedandthevalueofthecompany deteriorates.Toseehowthisworks,wewillrevisittheadaptationoftheMiller/Modigliani “valueequation”:V(l)= V(u) +TB-PVwhichstatesthatthevalueofaleveragedfirmis equaltothevalueofanunleveredfirmplustheproductofitstaxrateandbondsand minusthepresentvalueofitscostofbankruptcy.Thusthedistanceinvaluebetweenthe unleveredfirmandtheleveragedfirmismaximizedwhenthedistancebetweenTBandPV isalsoatamaximum(V(l)–V(u))=(TB-PV).Themaximumofafunctionsuchas(TB- PV)occurswhen∆TB=∆PV.Whenthemarginaltaxbenefitsequalthemarginal bankruptcycost,thefunctionTB-PVwillbeamaximum.Mathematically,whatwehave 91 doneissetthefirstderivativetozero;thatis∆TB-∆PV=0Thefunctionwillincrease untilthechangeintaxbenefitsequalsthechangeinthecostofbankruptcy,andthenit shoulddecrease,indicatingthattoomuchdebthasbeenincurred.Therefore,bothTBand PVmustbeafunctionofdebt;whenwechangethatvariable(debt),bothTBandPVwill changeaccordingly.Inthatregard,stockholders’equitywillbesetbysubtractingthe optimumlevelofdebtfromagivenlevelofcapital.Thispresumptionthatcapitalissetat thecorrectlevelisalimitingconstraintoftheoptimizationmodel;weknowfrom observationthatfirmsdonotalwaysraisethecorrectamountofcapitalandthatthe properamountistheoutgrowthofprojectanalysisandcapitalbudgeting-producinga positivenetpresentvalueoncorporateprojects. Asafirmincreasesleverage,itstaxbenefitsincrease,butthecostofbankruptcy alsoincreases.Atsmallerlevelsofdebt,thetaxbenefitswillexceedthecostofbankruptcy. Asmoredebtisadded,theprobabilityofdefaultincreasesatamuchfasterratethantax benefits,creatingadisproportionateincreaseinbankruptcycosts.Eventually,ifenough debtisadded,bankruptcycostswillexceedtaxbenefits,whichisasignalforthefirmto takeactionandlowerleverage. Doesthismodelminimizethecostofcapital?Inthelongrun,themarketwill correctlypriceriskbycharginghigherinterestratesforriskierleverage.However,inthe shortrun,variousmispricingsoccurthatwilltemporarilyallowtoomuchortoolittledebt. Byencouragingtheuseoflowerpriceddebttotheextentwhereitdoesnotmaterially underminethepriceofthestock,norviolatedefaultconstraints,thismodelwillminimize thecostofcapital.Ifequity,intheformofretainedearnings,wereusedinsteadofdebt, theleftsideoftheequation(TB)wouldremainstable.Inthemeantime,anyappreciation ofthestockwouldbecounteredbyalowerdefaultprobability,whichwilllowerthecostof bankruptcy.Theresultisahighermarginalbenefitsequation,andmovementtowardan optimalcapitalstructure.Hadthepatternoffundingwithretainedearningscontinued, theappreciationinthestockwouldbegreatenough,andthedecreaseindefault 92 probabilitysmallenough,toactuallybeginincreasingthecostofbankruptcy.Atthis point,thecompanyneedsaninfusionoflowercostdebtbecausethecostofretained earningsistoohigh. THEOPTIMALAMOUNTOFDEBT Fromtheperspectiveofatheoreticalidealwheretheintrinsicvalueofthestockis equaltomarketvalue,wealsoneedanidealdefaultprobability.Besidesdisplayingthe aforementionedshape,apracticalmodelisconsummatedwhenvariablesinthedefault algorithmaremadedependentonchangesindebt.Forexample,achangeinlong-term debtwillaffectinterestexpensewhichwillaffecttheamountoftaxes,andultimately-net income.Sincethemodelimpliesthatthetaxrateandtheamountoflossareconstants (determinedasgiven),onlythechangeinlong-termdebtanddefaultprobabilityarethe dynamicfactors.Infact,theequationcanreduceto: ∆Long-termdebt=(∆ProbabilityofDefault)x(AmountofLoss/TaxRate) Whenwesolvefor∆Long-termdebt,weadd(orsubtract)thisamountfromtheactual long-termdebtthatthefirmhadonitsbooks,andthisfigurewillrepresenttheoptimal amountofdebt.Subtractingtheoptimalamountofdebtfromthetotalamountofcapital invested,willyieldtheoptimalamountofequity. Thefirm’scapitalinvestmentisanotherconstraintthatisinputtedasgiven;wemay secondguesstheamountofcapitalasexcessiveorinadequate,butweneedtouseitasa realisticconstraint.Asinvestors,wecancontrolourownflowofmoneyintoastockbased onnumericalaggregatesliketheamountofdebt.However,theamountofcapitalisana priorifigure.Ifearningsarelarge,whatlookedlike“excessive”capitalturnsinto “adequate”capital.Thejudgmentofadministration,theanalysisofvariousprojects,and thetimeframeofmanagementareallvariablesinthedecisiontoraisecapital.Thus,while theamountofcapitalisimperativetocapitalstructure,itisagivenconstraintin optimization,outsidethepurviewofthemodel.Whileourmodelcandelivera“realistic” optimum,itcannotdelivera“true”optimumbecauseitrelinquishescontrolofthecapital 93 functiontooutsidefactors.Whenweoptimizedebt,wedosoinrelationtothevariablesof defaultprobability-assets,income,andotherliabilities.,butthemostonerousconstraintis theamountofcapital. LONG-TERMDEBTANDTHEAMOUNTOFLOSS Thepeculiaritiesoftheamountoflossmeshwiththeadditionoflong-termdebt: • 1.Growthstocksoftenhavehigherearningsthanaveragebutmorevolatility.Manyof themarefundedwithequitybecausetheinstabilityofcash-flowkeepsthemawayfrom thecreditmarket.Thehighermarketvaluethatisimplicitwithequityfundingwillbe anoutgrowthofmoresharesissuedwhenearningsareup.Atthisjuncture,thecostof bankruptcywouldbeprohibitivelyhighforadebtissueevenwithoutahigher probabilityofdefault. • 2.Higherstockpricesencouragetheissuanceofequity.Inthismodel,ahigherstock pricediscouragestheuseofdebt,becausemoredebtwouldraisethechanceofdefault andbemagnifiedbythehigherprice.ThusthemodelmirrorstheP/Echaracteristics ofdebt;firmswithalotofleveragetendtohavestockswithalowerpriceearnings ratio. • 3Moretangibleassetswoulddecreasethecostofbankruptcyandallowmoredebtto beused.Fromacreditor’sstandpoint,thisisarealisticassumptionbecausetangible assetsprovidecollateralforloans.Abusinessthatisfoundedupon“managerial expertise”isusuallylesscreditworthythanonethathassalableassets. GRAPHICDEPICTION Togainabetterunderstandingoftheoptimum,itissignificanttonotethatthe expressionTBisastraightline.Ontheotherhand,thecostofbankruptcywillbeacurve thatincreasesatanincreasingratetoreflectmoreriskathigherlevelsofdebt.Sincethe amountoflossisaconstant,mostofthecurvatureisderivedfromtheprobabilityof default.WhenthefunctionTBminusthecostofbankruptcyisatamaximum,therateof 94 change(slopes)willbethesameandtheamountofdistancebetweenthelineswillalsobeat amaximum. Figure4-2 AtpointO,anoptimumoccurswhichindicatesthatthedistancebetweentaxbenefitsand bankruptcycostsisamaximum.Themarginaltaxbenefitsequalthemarginalcostof bankruptcybecausetheslopesarethesame.AtpointD,anoptimalamountofdebtis incurred.Wherethetwolinescross,debtissoexcessivethatitmustbereducedtosavethe company. Itshouldbenosurprisethatthedecisiontousedebtisfoundedonclassiccost/ benefitsanalysis.Whiletheprobabilityofdefaultcurvecanassumemyriadshapesand includemanycomplexvariables,moststrategieswillbelinearenoughtoallowthe assumptionofsomeunderstandablerelationshipbetweenbasicfundamentals.Whenthe modeliscircumvented,itisusuallybecausemanagementisgamblingonsomeunrealized streamofincomethatisoutsidethenecessaryparametersoftaxincentives,assets,market valueanddefault. (BacktoTableofContents) $Cost Debt TB Costof Bankruptcy O D 95 APPENDIX:MAKINGSURETHATWHATYOUSEEISWHATYOUGET Mostcompaniesandbusinesspeopleareinherentlyhonest,butthereisagrowing factionintheinvestmentcommunitythatoperatesontheedgeofthelaw.Armedwith competentteamsoflawyers,thesecompaniesconcentratemoreonwhattheycanlegally garnishthanonservicingthecustomerorinvestor.Oftenthecrygoesupthat“Weneedto doitbecausethecompetitiondoesit.”,whichmaycertainlybetrue.Whilethisisnota courseonbusinessethicsorforensicaccounting,theobfuscationofdebtispertinentto capitalstructure;decisionsneedtobebasedontransparentfinancialstatements.The complexityofthemodernbalancesheetwarrantsanarrayoffootnotestoaccompanyit. Makesureyoureadthesenotescarefully.Theyshouldnotactastheaccountingversionof “fineprint”,butasanefforttoeducatetheinvestor,andfullydisclosethecomplicationsin thebalancesheet. Enronwerethesuprememastersofobfuscation,butmanyoftheirshenaniganshad tipoffsthatwouldhavealertedprudentanalysts.Enronwasextensivelyinvolvedin “entitystructuring”.Whilemanycompanieshaveseveralincorporatedsub-unitswhose performanceisriskyenoughtoseparatethemfromthelargerentity,Enronattemptedto usethemtoshieldtheextentofliabilitiesfrominvestors.Forexample,GeneralMotorsis intheautobusiness,butkeepstheirfinancialunitseparatedfromproducingautos.A distributioncompanymayhaveitsownincorporated,“inhouse”transportationcompany thatservesonlythedistributionbusiness.Itiskeptseparatetomaintaincore competencies,controlexpensesandcreatetaxadvantages.However,Enronpiledsub-unit uponsub-unitandhadmorethanmostanalystscouldrealisticallytrack.Onelookattheir last10Kshowedpageafterpageofsub-unitsallovertheglobe.Investorsandanalystsdid notquestionthepracticebecauseEnronwasa“cashcow”thatapparentlymademoney andincreaseditsstockprice. ThisauthorfirstencounteredEnronbecausetheywereusersofacertainbrandof riskmanagementsoftwarethatdidValue-at-Riskcalculationsonderivativeslikeoil 96 futures.Afteralittleresearch,Ispiedafinancialstatementitemthatwastermed, “Obligationsincurredinriskmanagementactivities“orsomethingtothateffect,andI wascurious.Atfirstglance,itappearedthatEnronwasplacingmoneylostinhedging futuresintoalong-termdebtcategory,whereitcouldconvenientlybepaidbackassoonas cash-flowimproved.IclosedtheInternetwindowandshuddered. Whilethefuturesmarketishighlyregulated,anycredit-worthyindividualcan engageinaforwardcontract,swappingfixedratesforvariablerates,cash-flowsinone currencyforcash-flowsinanother,adinfinitum.Infact,mostbanksprofitfromthese swapseverydaybutdonotdetailthemonfinancialstatements.Theincomeislegally claimed,buttheinvestorknowsnothingabouttheotherpartiesinthecontractor especiallyabouttheriskinvolved.Asthischapterwasbeingprinted,theentiredomestic bankingsystemhasbeenunderscrutinyformortgagelendingpracticesandthefurther securitizationofsuchmortgages,creatingderivativesthatneedtobetracedbacktothe originator. Onacorporatelevel,abalancesheetmust“balance”whichallowstheinvestorto play“detective”iftherearemisappropriationsbetweenshortandlong-termdebt,assets andequity.Whenmightthelargestshareholdersrefusetomovetowardamoreoptimal capitalstructure?Whenissuingmoresharestakescontrolawayfromthem.Ifcorporate controlisanissue,thenafewshareholdersatthetopcanprofitbyissuingdebtinsteadof equity,evenifitmeansanominaldecreaseinthestockprice.Alargeramountofdebtwill alsoimmunizeacompanyagainstatakeover,becausemostcompanieshavesecond thoughtsaboutassumingtheobligationsofanother.Althoughitmayseemthatsenior managementisdefendingthestockagainstalowershareprice,theunstatedreasonisto maintainpower. Sometimestakingondebtisanemotionalissue.Theterm,“agencyfriction”was developedoutoffrustrationthatmanagementwaslookingoutaftertheirowninterestsand nottheshareholders’.Acaseinpoint:ACEOemphasizescurrentsalesandoperating 97 expensestotheexclusionoffundingfutureprojectswithlowercostdebt.Althoughinterest ratesareatanalltimelow,andcompanyincomeisrocksteady,theCEOisleeryabout “toomuch”debt.Hisreferenceisa“creditcrunch”incidentsixyearsagothatalmost bankruptedthecompany.However,thisisanewphaseinthebusinesscycle.Theold paradigmdoesnotwork.Hegoesaheadandfundsprojectswithequity,evenasthestock pricediminishes.Afewshareholderswillgripe,butmostwillbe“outoftheloop”.The reason?Thereisnoindicationinfinancialstatementsofwhatthe“proper”capitaltarget shouldbe.Shareholderswillbeleftwonderingwhythecurrentadequacyofsalesand incomecouldnotbufferthestockfromdeclining.Ineffect,thecapitalstructuralist recognizesthatthecostofcapitalroseevenasincomeremainedsteady;thestockhadto deterioratebecausenotenoughriskwastaken. Anextremecaseof“balancesheetmagic”iscalled“insubstancedefeasance”.A companywillattempttoclean-upitsbalancesheetbyconvertingitsdebttohigherinterest rate,lowerfacevaluebonds.Suchamove,doneinanticipationofhigherearnings,will wipeoutasubstantialamountofdebt,whileallottingthedifferencebetweenthetwo amountstonetincome-awin-win,but“borrowingfromPetertopayPaul”,nonetheless. Inbrokereddefeasance,abrokerbuysafirm’soutstandingbondsandtradesthemtothe companyfornewlyissuedshares.Thebrokersubsequentlysellstheshares.Ineithercase, debtisnotbeingremovedthroughincreasedoperatingincome,butismysteriously disappearingoffthebalancesheet-anotherreasontoreadfootnotes.Whilethese techniquesarenotimplementedtoobfuscatefinancialviabilityortoconfuseinvestors,they needtobemanifest,asshouldthepracticeofrefinancingloans. Sometimesitappearsthatacompanynegotiatedafantasticdealoninterestrates, wheninfact,itissuedzerocouponbonds-securitieswhichwillbepaidoffinalumpsum whentheymature.Suchanissueisadvantageoustomostcompanies,becauseoftax savingsandnoregularinterestpayments.Long-termprojectsthatmightnotpayoffuntil yearslatercanbeimplementedwithoutexcessiveinterestexpense.Nevertheless,itis 98 importantfortheinvestortotreattheamortizationoftheloanasinterestexpenseinan optimizationmodel.Theriskofnon-paymentwhentheissuecomesdueisviableandneeds tobedistributedoverthelifeoftheloan.Theseindicationsarealsoavailableinthe footnotes. Investorsshouldbeawareofthevariouscategoriesoflong-termdebt,becausesome companieswillstatethesubunitwithoutreferringtoitas“longtermdebt”.Capitalleases, anyleasetoowncontractsoranyliabilitylastingoveroneyearshouldbereferredtounder theheading“long-termdebt”.Iftheinvestorhastomatchthenumberofyearsthata contractlastswiththeliabilitytype,thatistoomuchworkandaphonecallneedstobe madetoinvestorrelations.Clarityistothebenefitofbothmanagementandinvestors. Lastly,thereissometimesanunintentionalmisstatementofdebtobligationsbecause ofconfusionornaivetéofthepreparer.Tokeepandestablishgoodrelationshipswith clients,accountingfirmswillbegenerallycompliantandgooutoftheirwaytomakethe customerlookgood-legally.Theincreasedcomplexityofcorporatefinancehasledtoboth oversightsandpurposefulmanipulationsCallingabadhedgea“riskmanagementloss”is notillegalandmaynotevenbepurposefullydeceptive,butwhenitiscouchedinlanguage thatimpliessuchalossisnormalpartofbusiness,whenitisnot,aredflagshouldbe waved.Thisauthoroncewentthrougha10Kforawellknownsteelcompanyandcould notfindanyinterestexpenseintheincomestatement-despiteitshavinglong-termdebton thebooks.Miraculously,thecompanydecidedtonetitoutwithinterestgainedfrom investments.Evidently,noonethoughtthatinterestexpensewasmeaningfulenoughto declare. (BacktoTableofContents) 99 5 THECOSTOFEQUITY Thereisacostthathoversaboveeverymajorbusinessdeallikeapeskyflyata summerpicnicthatnoonewantstoacknowledge.ItisrarelymentionedintheWallStreet Journal.Noaccountingbalancesheetitemizesit,andyetitisomnipresent.Infact,itisa costsoprominentthatitnotonlyaffectseverymultinationalbusinessfromCiscotoUS Steel,itisanintegralpartofeverymomandpopstoreaswell.Itisthecostofequity. Thecostofequityisnotanaccountingcostbutaneconomicone.Itrepresentsthe “opportunity”oftakingoneactionoversomealternativeaction.Ifweconsiderthatthe foundationofstockownershipisbasedupontheflowofcorporateearningstoward shareholders,thisopportunitycostissimpletofathom:acompanycaneitherreinvest earningsorreturnthemtostockholders.Ifthecompanyreinvestsearnings,itneedsto makeareturnatleastasgreatasinvestorswouldmakeinalternativeinvestmentsin companieswithsimilarrisks.Therefore,ifallsteelcompaniesreturn15%andsteel companyZreturnsonly10%,steelcompanyZhasanopportunitycostof15%butunder performstheindustryby5%-giventheconstraintthatallsteelcompanieshavethesame risk. Whilethisexamplerepresentsanoversimplification,thereadershouldunderstand thatthecostofequityisacomparativecostthatisimplicitineveryprofit-makingvehicle: eachdollarofaddedincomecoststheshareholderssomeamountofunrealizedbut potentialloss.Someofthefactorsthataffectthatpotentialloss(risk)arethefollowing:1. Dividendyield.2.Thepriceofthestock.3.Alternativeinvestments-intheindustry,in thebondmarket,orinthestockmarketingeneral.4.Currentincome.5.Future prospects.Thislistisbynomeansexhaustiveasmyriadotherfactorsaffectthecostof equity.Infact,acomprehensivelookatthissubjectwouldincludeamodelwithhundreds 100 ofvariables.Forourpurposes,weneedtocreateonlythebestestimatepossible,giventhe limitedamountofinformationavailabletoinvestors. MODIFICATIONS Ultimately,weneedtobringthetheoreticalconcept,“thecostofequity”intothe practicalrealmofcapitalstructureoptimization,whichrequiressomemodification. Traditionaluseofthecostofequityforstockvaluationandcapitalbudgetingneedstobe expanded.Thefollowingissuesneedtobereconciled: • 1)Thecostofequityisbasedonmarketvaluesandnotbookvalues.Theproper methodofcalculatingnotjustthecostofequity,butthecostofanysourceofcapitalis tousemarketvalues.Sincethecostofanycapitalcomponentisgaugedbyitsrequired return,themarketmethodallowstheanalysttobalancethisequivalencebetweencost andreturn.Considerafirmwhofloatsabondissueof100millionat8%interest. Essentially,theybeginbypaying8millionayearininterestexpense.Nowsuppose inflationtakeshold,andinterestratesareraisedto10%,andthepriceofthis company’sbondsfallsto80million.Ifweusethe8%rateastherequiredreturnas wellastheinputintothecostofdebt,thenthecompanycanbuybackitsownlower pricedbondsinthemarketandearnayieldof10%.Thebookvaluedesignatesdebtof 100millionandinterestpaymentsof8million,butmarketvalueshaveshiftedboththe requiredreturnondebtanditsconsequentcostupwards.Usingthelowerinterest valueof8%astherequiredreturnondebtwouldleadtoadebtladencapitalstructure becauseitwouldunderpriceitscost. • But-Theobjectiveofcapitalstructureistoseekanequilibriumbetweenmarketrisk andcompanyriskthatmayrequiretechniquesthatcomparemarkettobookvalues. Simplystated,bookvaluesaremorecontrollableandactionable.Sincethemarket valueofastockisbasedonprojectedfuturegrowthinearnings,anycomparison betweennetincomeandstockpriceisobfuscatedbyextrapolation.Ifearningsareto bedistributedtoshareholders,andshareholderspayanimplicitpriceforsuch 101 earnings,bothvariables,netincomeandequity,needtobecalibratedwithinthesame timeframe;presentvaluecalculationsonlydiscountfutureearnings.Sincenetincome isabookvaluecomponent,whichlaterbecomesthebookvalueitem,“retained earnings”,oncedividends(ifany)arepaid,itisconsistenttousethebookvalueof equitywhenmakingdirectcomparisons-atleastwhengaugingnear-term performance.Therefore,capitalstructureanalysismultipliesamarketderived percentagecostofequitybythebookvalueofequity,andthencomparesittonet income.Thiscompositevalueisneitherthecostofequity,arequiredreturn,noran opportunitycost,butreflectsmovementtowardsanoptimalcapitalstructure.Infact, theconcepthasbeentrademarkedbythefirm.SternStewart,Inc.andiscalledEVA ® or“economicvalueadded” 2 .Theinvestorfriendlyversionisfurtherdevelopedinthe chaptercalled,“TheCapitalDynamic”. • 2.Everyfirmhasanoptimalproportionofdebttoequitythatincreasesreturnsand minimizesrisk. • But-Thetargetproportionisalwaysmovinganditisquestionablewhetheran optimumcanbereached.Anothertenetofcapitalstructureisthatthestockpriceis maximizedwhenafirm’scapitalstructureisatanoptimum.Thus,thestockpricerises andfallsbasedontheproportionofdebttoequity,butitneverremainsconstant. Multiplyingtheestimatedcostofequitypercentagebythemarketvalueofthestock yieldsarealisticcostofequityforonlyaslongasthestockremainsstable-whichmay notbeanylongerthanafewminutes!Thisisacircularargumentwhichassumesthat thepremiseistrue;whenthecostofequityisdependentonashiftingstockprice,a concreteminimumcostofcapitalcannotbeobtained.Mosttheoreticalmodelsignore thefactthatsharescannotbeissuedataconstantpriceandthatthetruemathematical optimumisfleeting.Whileacompanyneedstostruggletominimizeitscostofequity,it 2 EVAistheregisteredtrademarkofSternStewart,Inc. 102 mayneverrealizehowcloseitistoitsgoal,becausethedeterminantvariable(stock price)keepschanging. • 3)Thecostofcommonstockisvaluedlikethecostofretainedearnings,exceptfor flotationcostsonnewissues. • But-theremaybenoactualcashoutlayonanyofthecomponentsWhenan investmenthouseunderwritesanissueofstock,itchargestheissuingcompanya substantialfeewhichisdeemeda“flotationcost”;itmaywellbejustapercentageof thecashreceivedfromtheissue.Therefore,thecostofequityfornewcommonstockis higherthanforbothretainedearnings,andpastissuesofcommonstockbythisfactor. However,theamountofpastissuesandretainedearningshasnoimmediatecashoutlay (disregardingadministrationfees)andmayrepresentinvestmentinsomeassetthatwas amortizedlongago.Nevertheless,thesecomponentsrepresentasubstantialinvestment thatmustbeenumerated.Apreciseitemizationwouldrequireseparatevaluationsfor newcommonissuesandthecategoryofretainedearningsandpastissues,allofwhich ascendanddescendatdifferentrates.Withoutacashoutlay,andwithinherent volatilityinthemeasurement,theneedforstalwartprecisionisinsubstantial;itismore prudenttoseekoutamodelthatyieldsthebestestimategiventhecurrentlevelofrisk. Fortunately,suchacomprehensivetoolexists:the“capitalassetpricingmodel” comparesthechangesinequityofanindividualfirmtothegeneralmarket,andeach componentpartofthefirm’sequityisimplicitinthecomparison.Withan “opportunitycost”thatshiftsfrequently,thevariablesoftrendandcomparative magnitudearemostsignificant.Whatweloseinprecision,wegaininexpediencyand utility.Sincethemarketcanbevolatile,itismorepracticaltousethismeasurementin near-termcomparisonswhenseekinganoptimalcapitalstructure.Moreover, observationoflongtermchangesinthecostofequitymayindicatehowthecompanyis reactingtomacroeconomictrends. • 4)Thedefinitionofstockholders’equityincludespreferredstock. 103 • But-preferredstockisnotvaluedliketheotherequitycomponents.Preferredstockis ahybridsecurity,combiningaregulardividendwithasetissuedprice,butwithoutthe projectedgrowthcharacteristicsofcommonequity;themarketforpreferredstockis limitedandconsistslargelyofcorporateownershipbecauseofthetaxbreaksinvolved. However,itisvaluedasacomponentofthecostofcapital,anditscostisasfollows: (PreferredDividend/NetIssuingPrice)x(ProportionofPreferredEquityinthe CapitalStructure).Manycompaniescarrynopreferredstockatall.Iftheinvestor ignoresitinhisorhercalculations,thecostofequitywillbehigherthanitactuallyis becausepreferredstocklowerstheriskofequity.Theregularityofpaymentsgivesit someoftheriskcharacteristicsofbonds,whileitmaintainstheliquidityofstock.Thus, makingcomparisonsbetweentwocompaniesthathavepreferredstock,andpricingit atthecostofcommonequity,willnotbeasaccurateasifpreferredstockwereitemized asacomponentcost.Inaquickcomparison,itisaconservativeerrorthatwillhave fewtacitrepercussions.However,inathoroughanalysisorincapitalbudgeting problems,preferredstockneedstobefullyaccountedfor.Likeshort-termdebt,itacts asanadjuncttoothersourcesofcapital.Inthiscase,itkeepsthecostofequityless risky;lesssharesofcommonequityneedtobeissuedwhensomefundingisdone throughpreferredstock THEREINVESTMENTRATEANDOPPORTUNITYCOSTS Someconfusionmayexistbetweenthereinvestmentrateonearningsflowingto shareholdersandthereturnthatwouldbereceivedoncomparativelyriskyinvestments. Whenthemarketisinperfectequilibrium(whichisrare),theseratesareequal.If CompanyAcangetgreaterreturnsthanothercompanieswiththesamerisk,investors flocktoCompanyAandbidupitsprice.Thisupwardpressureonthestockpricewill continueuntilthepointwhereCompanyAnolongergarnersanextraordinaryreturn,but isinlinewithcompaniesofsimilarrisk.Infact,anyinvestorwhopurchasesarisingstock thatisbeatingthemarketislikelychoosingonethatisnotinequilibriumwithit.This 104 “riskarbitrage”isfundamentaltoallspeculation.Similarly,anyreinvestmentratethatis lowerthancompanieswithsimilarriskwillendupdepreciatingitsstockprice.The “opportunitycost”istheopportunityofnotinvestinginthepresentcompany,butinvesting insimilarlyriskycompanies,Thus,ifCompanyAmakes25%onreinvesting shareholders’earnings,itsreinvestmentrateishigh,butitsopportunitycostmaybelowif comparativelyriskyfirmsaremakingonly10%.Iftheopportunitycostexceedsthe reinvestmentrate,itistimetosellthestockand“taketheopportunity”toinvestinother companies. Academicsoftencallthereinvestmentrateofreturn,“theexpectedrateofreturn”, andtherateofreturnoncompaniesofsimilarrisk,“therequiredrateofreturn”.For purposesofcapitalstructure,wealwaysrefertotheopportunitycost,“therequiredrateof return”asthetruecostofequity.Weneverassumethatthemarketisinequilibrium,nor doweassumethatreinvestmentratesaresimilaramongcompaniesofcomparativerisk. Sincecompanyfundamentalstendtodeterminereinvestmentratesandmarketcompetition determinesrequiredrates,wesubmittothemarketasthegreaterofthetwoforces. However,theforcesmayoverlapeachother,workinconcert,oragainsteachotherand onemaydominatetheother.Indeed,thestudentwhohascomparedthedifferencebetween “value”investingand“growth’investing,canobservethatbothmethodsarepartofthe samephenomenon-alackofequilibriumbetweencompanyandmarketratesofreturn thatworkonthesamecontinuumtocorrectthemselves. EVALUATINGTHECOSTOFEQUITY:METHODOLOGY Theattempttoenumerateatheoreticalcostrequiresatoleranceforuncertainty.To thosewhoworkintheaccountingorengineeringfields,basingdecisionsona “guesstimate”,maytakealeapoffaith.However,therewillbeenoughcorroboration fromseveralsourcestocreateourown“inhouse”tolerancelevels.Anymeasurementof changerequiressomerelationalvaluesarounditinorderforittobemeaningful.In 105 capitalstructureanalysis,thoserelationalvaluealwaysprovideasolidfoundationoverthe long-term,becauseafirmcannotbeinbusinesswithoutadheringtothem. • 1)TheRiskPremiumMethod:Yearsago,wheninformationwasnotsoreadily available,analysts(withgreeneyeshadesintact)wouldaddfromtwotofour percentagepointsontoafirm’sbondyieldandcallit“thecostofequity”.Realizing thatriskierlowerratedebtwouldcontributetomoreequityrisk,analystssetarisk premiumthatcombinedthetimingofthebusinesscyclewiththeratingofthefirm’s bonds.“TripleA”(AAA)ratingswouldjustifyjusttwopercentagepointsadded,while lowerratedbondsatthetopofthebusinesscyclewouldwarranttackingonfourpoints. Althoughthismethodissubjective,itdoesgivethemodernanalystsomeguidance:if heorshecalculatesaneightpointdifferencebetweenequityanddebtcosts,thefigureis mostlikelytoohighandneedstoberecalculated. • 2)TheRuleofThumbMethod:Theruleofthumbmethodisamodificationofmarket valuationmethodsthatusethepresentvalueofdividendsreceivedinthefutureto obtainthe“fairvalue”ofastock.Inthecasewherethedividendgrowthisconstant, theGordonModelemerges:P=D1/(K–G).Sincethedividendgrowsataconstant rate,presentvaluediscountingisimplicitinthecalculation.Theterminologyisas follows: Table5-1 SYMBOL EXPLANATION P Currentpriceofthestock D1 Dividendreceivedinthenextperiod K Costofequity ROE ReturnonEquity(netincome/equity) RETENTIONRATIO 1-(DividendsPaid/NetIncome) G Growthratethatdividendsareexpected tofollow 106 Thegrowthrate“G”iscontroversialbecauseitispresumptiveandextrapolatesinto thefuture.Forthesakeofsimplicity,weequateitwiththeretentionratiomultipliedby thereturnonequity(RetentionRatioxROE).Inactuality,thegrowthratemaybea long-termaverageofthisfigure.Forthetimebeing,wewillassumethatG= (RetentionxROE),andproceedwithderiving“theruleofthumb”method. Ifwesubstitute“X”fortheretentionratioandusethesymbol,“E”forearnings,the modelisnowP=(1-X)E/K-(ROE)X.,Furthermore,ifweassumethatROEisequal tothecostofequity,“K”,thenthefollowingequalityismet:(1-X)E/K-(K)X=(1-X)E /(1-X)K=E/K.SincewenowhaveanequalitywhereP=E/K,werearrangeand formK=E/P.Obviously,thisinverseofthe“ageold”financialindicator,“P/E”,is merelythereturnonthemarketpriceofthestock.ThefactthatP/Escantellusso muchaboutastock’sbehaviorispartlyattributabletothisrelationship.Used judiciously,theP/Eisanindicatorofgrowth,andwhen“E/P”approaches“ROE”,the stockmaybeconsideredabargain.However,liketheP/Eratio,when“E/P”isused asaproxyforthecostofequity,itmustbegaugedwithareferencetothemarket averageForexample,growthstocksdonotqualifyforconsiderationusingthismethod. ManygrowthstockshaveveryhighP/Es;aP/Eof29basedona$20sharepriceand 0.70foranEPSwouldassumearidiculouslylowE/Pof3.5%.Thus,forstockswith muchanticipatedgrowth,themethodissimplyinadequate.Forstocksthathavean averageP/Ethatisclosetothemarket,themethodoffersaveryquickestimationofthe percentagecostofequity.Infact,analystscanscanthisnumberrapidlyandcombineit withotherinformationtoscreenforprospectiveinvestments. • 3)ValuationMethods:TheGordonModelisonlyoneofmanydividenddiscount modelsthatequatethepriceofthestockwiththepresentvalueofcashflowsreceivedin thefuture.Manyofthesecanbequiteelaborateandrequireforeknowledgeofthepath ofearningsfarintothenextbusinesscycle.Withastablemultinationalfirm,the estimationoffutureincomeisnotthatformidableatask,butwithuntestedorvolatile 107 companies,themethodcomprisesa“randomwalk”.WhyusetheGordonModelifitis onlyapplicabletodividendpayingcompanieswherethedividendisgrowingsteadily? Theassumptionofsteadygrowthwillnotradicallyaltertheviabilityofthecalculation inthenearterm.Withoutexcessiveextrapolation,weassumethatthenextdividend willbegrowingatthesamerateasthepresentdividend,aprobabilitythatavoids judgmentabouttransitions.Consequently,weusethecostofequityfornearterm comparisonsbetweencompanies,andnottoextrapolatethe“fairvalue”ofastock.We havealreadybeenintroducedtotheGordonModel,P=D1/(K–G).Thesimple adjustmentthatwemakeistosolvefor“K”,whichbecomes,(D1/P)+G=K.A modifieddividendyieldusingthenextexpecteddividendratherthanthecurrent,and summedwiththegrowthratewillproduceacostofequity.Themajorproblemwith usingtheGordonModelisthatitissodependentonthefundamentalsofthecompany. And–obviously–likeanyotherdividenddiscountmodel,itcannotbeusedwithfirms whopaynodividend.Ultimately,itisabetterstatementofthereinvestmentrate,“the expectedrateofreturn”thananequitybased,comparative,“requiredrateofreturn”. However,becauseitmesheswellwiththeearningsandfundingcharacteristicsofa company,ithasmanymoreusesbesidesdeterminingacostofequity. • 4)TheCapitalAssetPricingModel(CAPM):Thismodelisthepreferredchoicefor capitalstructureanalysis.Despiteyearsofacademicandprofessionaldisparagement, theconceptcontinuestothrivelikeaDarwinian“missinglink”.Withahighvolumeof literaturetobothsupportandexcoriateit,theCAPMseekstobetheessenceofrisk comparisonbutfallsfarshortofanyclaimstopreciseaccuracy.Whatitdoesbeyond disputeistocompareanysecuritytoamarketindexinalinearfashion.Itthen multipliesaderivedcomparativenumberbythedifferencebetweentheindexandwhat istermed,the“risk-free”rate,andthenaddsthatfigurebacktotherisk-freerate.In effect,itcomparestwosetsofnumbers:1.Theindividualsecuritywiththemarket.2. Themarketwithwhatisthoughttobearisk-freeyield(usuallythetenyeartreasury). 108 Itisthis“requiredreturn”thatformsthecostofequity,andtheproverbial“rateof returnoncompanieswithsimilarrisk”.Noothermethodbetterincorporatestheforce ofthemarketwithacomparisonbetweenalternatives,i.e.,theindividualsecurityand therisk-freerateofreturn. Thekeytothismodelisthecomparisonfactor.TheCAPMusesregression techniques(seethechapteronstatistics),andderivesacomparisonfactornamed “Beta”,whichisessentially,∆Y/∆X,thefamiliar“riseoverrun”-onagraph.In fact,theCAPMisbasicallyastraightlineintheformofY=A+B(x)withsome modification,andthereinliesitsdownside.Mostfinancialvariableshaveanongoing multinomialrelationshipthatconstantlychangesandformsacurve.Inoneweek,for example,therelationshipbetweenassetsandearningsmightbelinear,butinthenext weektheymightassumeapolynomialrelationshipofthefifthdegree.Whiletrue linearitybreedscertainty,itisrarelyencounteredinfinanceanditspresumptioncan leadtopoordecisionmaking-witnesstheimpliedchaosofwageandpricecontrolsin theearly1970stoobservejustoneexample. TheCAPMneedstobeperceivedasavaluablebutpotentiallymisleadingtool. Whenusedcorrectly,eachmethodofdeterminingthecostofequitywilltendto corroboratetheotherinthedomainofchange-albeitatadifferentabsolutelevel;that is-ifthechangeinthecostofequitywhileusingthe“riskpremium”methodis20%, therewillbeacorrespondingincreaseusingtheCAPMmethod,buteachrespective methodmayyieldadifferentvalue.Thecostofequityisameasureofcomparativerisk andnotanaccountingstandard.Usedinisolation,however,theCAPMissusceptibleto periodicinstabilitythatwillleadtoincorrectdecisions.Asmuchinformationabout leverage,thebusinesscycle(whethertheFedhasraisedorloweredrates)andother methodsmustsupporttheuseoftheCAPM.Ultimately,performanceinoneyearwill becomparedtoperformanceinthepreviousyear,andtheemphasismustbeplacedon consistency.Aslongasthesamemethodologyisusedinthecomparison,theforces 109 actingonthestockwillbeverysimilarandwillbemeasuredinsimilarways.Itiswhen twodifferentmethodologiesareusedthattroublearises,becausethecomponentsofthe measurementswillbesodifferent.TheE/Pmethod,forexamplewillmeasuremarket pricemuchmorethanthe“riskpremium”methodwhichwillmeasurethecomparative riskondebt.Thus,theabsolutevaluesofthesecostsmaybetotallydifferent,butwhen usedproperlyinthecontextofcapitalstructure,eachpointstooptimality;the dynamicsofchangedefinetheirutility. • 5)ConsensusMethods:Thelogicbehindcombiningmethodsintoa“fusion”costof equitystemsfromtheoppositionoffundamentalforces.Forexample,combiningthe returnonequity(ROE)withamarketbasedmethodlikethe“E/P”,utilizesthe differencebetweenbookandmarketvalueswhichisanimplicitfactorinthecostof equity.Similarly,combiningtheCAPMmethodwiththeGordonmodelgivesweightto bothmarketcomparisonsandinternallygeneratedcompanyfundamentals.The judicioususeofthesecombinationsgivestheanalystanadvantagewhengauging changesinrisk,andinsomespecificsituations,formsamoreaccuratemeasureofthe costofequity. Themostfrequentuseofthecostofequityisnotininvestmentanalysis,butin capitalbudgeting.Whenacorporationoutlinesthecapitalneedsforfutureprojects, accuracyandprecisionareapremium.Moremistakesaremadefromunderestimating theneedforcapitalinputsthananyother-partlybecausethepoliticalnecessityof “tellingthetruth”aboutaprojectcanactuallyundermineit.Whilemanycompanies usesophisticatedMonteCarloanalysistoforecastapotentialcostofequity,historical datawithconsensustechniquesmayworkforothers. Onesuchmethodistoregresseachofthemethodsagainsttheperformanceofthe stockoverafiveyearperiodinonemultipleregression.Thedependent“Y”variableis thefiveyearperformanceofthestock,whilethemethodsactasthesequenceofX variables.Thecoefficientsoftheregressionwillbetheweightsofeachmethodinthe 110 consensus.Whilethismethodislinear,itwillallowtheanalysttocitethemost historicallyaccuratemethodandtobetterunderstandtheforcesbehindthestock’s changes. THECAPMASABUILDINGBLOCKFORCAPITALSTRUCTUREANALYSIS TheCAPMisintegraltoanareaoffinancethatcommandstheutmostrespect- modernportfoliotheory,thecombinationofsecuritiesthatwilldiversifyawayriskand maximizereturns.Itsuseincapitalstructureanalysis,however,isnotsocomprehensive. Infactonlypartofitspotentialisrealizedinderivingacostofequity.Whileitistypicalof aneclecticsystemlikecapitalstructuralismtousewhatispractical,thestudent/investoris encouragedtostudythismodelfromtheperspectiveofcreatingaworkingportfolio-one thatmaximizesthemeanreturnwhileminimizingthestandarddeviationthereof. Anydisciplinethatisasprobabilisticasfinancemustadapttoexistingconditions, sometimestradingutilityforstructuredprecision,butneverthelessencouraging experimentation.Forexample,manyprofessionallyacceptedstatisticaltechniquesassume “normality”wherenoneexists.Thatis:theydescribeaprobabilitydistributionwitha centraltendencyaroundameanandanareathatisdefinedbyasetnumberofstandard deviations.Therealityoffinancialdataisthatitisactually“heteroskedastic”-definedby extremeswithheavierweightsattheends.Theareadefinedbyahandfulof“bigwinners” andalotof“losers”isheavilyskewed.Indeed,iffinanceworkedlikephysicswhichis mathematicallypreciseandpredictable(onanon-theoreticallevel),thenmuchofthe returnwouldbeextractedfromit.Thedistributionwouldbelikeasavingsaccountata bank-highlypredictablebutwithlittlereturn-almostaflatline.Byitsverysimplicity andflexibility,theCAPMcanbeusedinseveraldifferentways–asaninvestmenttool,for portfoliomanagement,orforcapitalbudgeting.However,theinvestormustrealizethatit describesonlya“bestestimate”,givenanarrowrangeofvariables,forabriefmomentin time. 111 Ifoneconceptcancharacterizecapitalstructuretheoryitisthateverydecision requiresabalancebetweencompetingalternatives-atradeoff.Themarket-enmasse- willmakethisdecisionwhenitfavorsinvestmentgradebondsovertheequitiesmarket- theclassicflighttoqualitythatoccursinadownturn.Thefinancialexecutivemakesthis tradeoffwhenheorshedecidestoraisedividendsandretainfewerearnings.The individualinvestormakesthisdecisionwheninvestinginadebtladencompanyratherthan ahigh-flyingcashgeneratorthathasjustpeaked.Essentially,allofthesetradeoffshavea singleelementincommonwiththeCAPM:whenevermoreriskisengaged,agreater returnisexpected. Thecapitalmarketlinedisplaysthistradeoffgraphically,usingthemarketrateof returnandthestandarddeviationofthemarketasmeasuresofrisk. Figure5-1 Thegreencurveisacombinationofstocksthatencompassesthemarket.AtpointB,there isalmostzerorisk(nostandarddeviation),butnoticethatboththereturnandstandard deviationarelessthantheyareatpointAwherethegreatestreturnwiththesmallest standarddeviationisachieved.PointBistermed“therisk-freerateofreturn”.Whileno securityistruly“risk-free”,becauseofinflationandthenominalprobabilityof Market Return Standard Deviation B A C 112 catastrophe,thispointrepresentsU.S.treasuriesandhasthelowestamountofinvestment riskassociatedwithit.Ontheotherhand,pointArepresentsthe“efficientfrontier”which iswherefundmanagersandinvestorswanttobe.Speculators(andunscrupulousfund mangers)willchoosepointCbecauseithasthehighestreturnassociatedwithit.However, mostprofessionalswouldeschewthispointbecausethechanceoflossistoohigh. Pointsabovethelinehaveahigherlevelofreturngivenanylevelofriskassociated withthem.Thesepointsareonlytemporarilyachievableandtherewillbeforcesthatput downwardpressureontheirprices.Analogously,pointsbelowthelinewouldhavemore riskthanwarrantedbytheirreturns,andhaveupwardpricingpressureonthem.Infact, realitydictatesthatinmanycasesthereturnswillnotsubstantiatetheriskincurred,anda portfoliocanlanguishformanyyearswithoutmovingtowardtheline.However,the investorshouldrealizethatsomeoptimalportfolioexistsevenduringrecessions,andthat therisk/returnlinewouldchangeinproportiontothechangeinthemarket-inclining duringabullmarketandflatteningduringabearmarket.Ineffect,theslopeoftheline changeswhentherisk/returncharacteristicsofthemarketchange.Inbullmarkets,itis typicaltogainlargereturnswithlittlerisk,andsothecapitalmarketslineinclines.Ina bearmarket,anyamountofriskincurredseemstoyieldnegativeresults,andsotheline flattens. ThecapitalmarketslinedisplaysthetheoryofriskbehindtheCAPM,butitshould notbeconfusedwiththeCAPMitself.Thecapitalassetpricingmodelisactuallya compendiumofthecapitalmarketsline,andtheregressionlineofanindividualsecurity againstthemarket.Thatregressionlineiscalledthe“characteristicline”. 113 Figure5-2 CompanyX %Stock Increase Market%Increase Thislineisusuallydevelopedoverfiveyearsofreturns,usingmonthlychangesinthestock priceandthemarket,foratotalof60datapoints.Themonthlyreturnandnumberofdata pointsmakesthecomparativenumericchange,“beta”,lesssusceptibletoerrorthrough volatility.ThemarketdataisusuallyderivedfromawidelyencompassingindexliketheS &P500,butmuchdebatehascenteredaroundwhichindexifany,cantrulyinteractwith thismodelbecauseeachhasdiscrepanciesthatbiasitsdistribution. The“Y”interceptor“alpha”componenthasalitanyofitsown.Portfoliomanagers arequicktopointouttherelevancyofalphatotheparticularsituationofthecompany.A comparativelylargeorsmallalphacanindicateeverythingfromindustrydominanceand protectionismtoadecayingcompanywithapredilectionforbankruptcy.Whatthe investorneedstoknowisthatinterpretationofalphaleansheavilyoninvestment experiencewiththecharacteristicline,andthatextremeYinterceptspointtoless associationbetweenanindividualcompanyandthemarket. ThepropermethodtocalculatetheCAPMisasfollows: • 1.Calculatethecharacteristiclineasstated.The“Y”dataarethepercentagechanges instockpriceoversixtymonthsfortheindividualcompany.(N+1)entriesofthetype 114 ((MarchPrice/FebruaryPrice)-1)willproduceadecimalpercentage.Sixty-oneof theseentrieswillproducesixtyworkingdatapoints.The“X”dataarethemarket changesoverthesameperiod.Ifneeded,refertothechapteronstatistics.Oncethe studentgetsusedtodoingtheseregressionsinaspreadsheetlikeExcel,theprocesswill beswiftandmechanical:datacanbeinputtedandregressionsprocessedinlessthan thirtyseconds.Oncearegressionlineisformed,thecoefficientofXisthe“beta” component.WhathappenstotheYintercept,“alpha”?Whilealphaisnotentirelyout ofthepurviewofcapitalstructure,itisusedtogaugetherelevancyofbeta.A comparativelyextreme“alpha”,incombinationwithalowcorrelationcoefficient(R) wouldindicatethattheassociationofthefirmwithothercompaniesofsimilarriskis weak. • 2.Therisk-freerateistheaverageyieldforthetenyearTreasurybondovertheperiod oftheregression-usuallyfiveyears.Insubsequentchapters,weshowhowtocreatea “FederalReservebias”,byusingashorterspan.However,thepropermethodistouse theaverageoverfiveyears,becausethebetacomponentshouldcorrespondtothesame lengthoftimeastheothercomponents. • 3.Theaveragemarketreturnoverthefiveyearperiodisinputted • 4.Thethreecomponentsarearrangedasfollows:Riskfreerate+(Beta)(Marketrate- Riskfreerate). Thedifference,(Marketrate-Riskfreerate)isespeciallysignificant.Thisdifferenceis calledthemarketriskpremiumanddelineatestheexcessreturnofthespecificmarketover theriskfreerate.AlthoughtheCAPMisnotexclusivetothestockmarket,andcanbe appliedtootherassets,themarketriskpremiumisespeciallysignificantinthatarena.In essence,thedifferenceservesasagoodindicatorofhowstrongtheequitiesmarketis comparedtothefixedincomemarket;alargeriskpremiumwillpeakandthenshrink wheninterestratesareraisedtocombatinflation.Infact,thebusinesscycleisoftenan expressionofthesizeandaccelerationoftheriskpremium.Atthebeginningofarecovery, 115 theriskpremiumbeginstowiden;fundmanagerssellbondsandbuystocks.Although sellingbondswillraiseyields,thestockmarketacceleratesatamuchfasterpace.Atthe peakoftheequitiesmarket,demandforequityisatitshighestpointwhichraisestherisk premium.Eventually,inflationcreepsin,ratesareraisedandbusinesseshaveaharder timestayingafloat.Theequitiesmarketdeclinesmuchfasterthanratescanbelowered, andtheriskpremiumisloweraswell.Analogoustothedifferencebetweentheratesof changeofearningsandthecostofcapital,theaccelerationdifferencebetweentherisk-free rateandthemarketratedeterminesperformanceinthecapitalmarkets.Whileearnings andthecostofcapitalworkonamicroeconomiclevel,thecomponentsoftheriskpremium workonamacrolevel.Thissensitivitytotheinteractionbetweencreditandequity marketsisonereasonthattheCAPMisagoodtoolforgaugingthecostofequity:capital structureisdependentonthebusinesscycleandtheinherentchangesofriskinvarious sourcesofcapital. PartofthevalueoftheCAPMisthatitissoflexible.Itcanbeusedonanyasset (commodities,realestate,junkbonds),butitalsodefinesdifferenttypesofrisk.Thebeta componentiscalledsystematicriskandisnondiversifiable.However,the“Y”interceptin theregressioniscalledthe“alpha”componentandmaybecombinedwithotherassetsto diversifyawayrisk.ThemainpurposeoftheCAPMwastoenumeratetheserisksmore clearlyandcreateaninvestmentpatternthatallowedtheinvestortotakeabroaderlookat thereasonswhyastockroseordeclined.Thepremiumwasinfindingthecombinationof securitieswiththegreatestreturnperunitofrisk. HarryMarkowitzpioneeredmodernportfoliotheorynearlyfiftyyearsago(asof 2008).AlthoughtheCAPMdevelopedintoanecessaryadjunct,theconnectionbetween portfoliodiversificationandcapitalstructure(aportfolioofcapitalfunding)wasnotso readilyapparent.TheflexibilitythattheCAPMengenders,however,mayhaveledtosome ofitslackofacceptanceintheacademiccommunity.Itcombinedflexibilitywithlinear constraints,whichisanoddmathematicalcombination:itneedstomaintainastraight 116 line,butitdoessobyreconcilingvariablesthatwouldnotordinarilyhavealinear relationship.InfactsomeprofessionalrejectionoccurredafterthefamousBlack,Jensen and.Scholesstudyofthirty-fiveyearsofstockperformance.Theyfoundthatalthoughthe actualrelationshipbetweenbetaandaveragemonthlyreturnsmatchedthetheoretical model,therewereseveralperiodswheretherelationshipwasinverted:thatis-a downwardslopingcurvewasencounteredthatdescribedlessreturnwithmorerisk.Since thevalidityoftheCAPMiscontingentonarisk-returntradeoff,WallStreetquickly rejectedtheconceptandbecamemoreenamoredwithtechnicalanalysis. Figure5-3 Black,JensenandScholesStudy Return Beta(35Years) Actual Theoretical Figure5-4 Return Beta(1957-1965) Actual Theoretical 117 Naturally,suchacomprehensiveandimperfectmodelhasstimulatedmany attemptsatimprovement.Theoriginalmodelcanbeattributedtoaunifiedeffortby Sharpe,Lintner,TreynorandMossin,allofwhomtriedtoelaborateonitthroughother techniqueslikethe“singleindexmodel”or“multipleindexmodel”,whicharetermed “market’modelsmoreconducivetodisplayingtheactualbehaviorofasecurity.The originalmodelalsohadabevyofrulesthatmadeitdifficulttoapplyinarealisticsituation. Theseempiricalconstraintswereasfollows: • 1.Investorsareriskaverse.Theyprefermorereturnandlessriskonaconstantbasis. • 2.Firmscanlendandborrowatthesamerate. • 3.Notaxesortransactioncostsexist. • 4.Themarketportfolio(index)chosenistheappropriateone. • 5.Investorsaremoreconcernedaboutdomesticcurrencyreturnsthanexchangerates. • 6.Betasarestable Everyoneoftheseruleswillbeoverturnedinaninvestmentscenario,anddifferent combinationswillbeeitherinplaceorsuspendedatanygiventime.Whileadherenceto theserulesremovessomeoftheconceptualliabilitythemodelwouldhaveifastockdidnot behaveasexpected,italsoopensupseveralmoreavenuesoftheoreticalapproach.What aretheeffectsofhightaxesonthemodel?WhatifthedollarisslippingandIwantto applythemodelinEuros?Doesthemodelworkbetterforastockwitharocksteady beta?Theseareessentiallyinvestmentquestionsandoutsidethepurviewofcapital structure.However,liketheMiller/Modiglianimodelwhichalsohasstringentrules, experimentationisencouragedbecausetherulesareimpossiblenottobreak.Whenever theorystraitjacketsbehavior,itistemptingtoinventsomewayaroundit. ANESTIMATEOFRISKANDNOTPREDICTION PartoftheinconsistencyoftheCAPMhasstemmedfromitsusebyfundmanagers whoexpectittobeapredictivetool.Inthiscase,variationisconfusedwithperformance: 118 variationwiththemarketisexpectedtopresentaforecastofexpectedbehavior,i.e.,ifthe marketgoesup17%,thenabetaof1willdothesame.However,theequivalenceof variationmaynotleadtothe“causeandeffect”relationshiprequiredbyprediction,and eachstockhasaunique“alpha”componentthatisnotsoeasilyquantifiable.Infact,the myriadcomponentsofalphahaveasmuchtodowithastock’sactualperformanceas beta;alargealphasignifiesastockthatwillriseevenasthemarketisstagnant. WhenusingtheCAPMasaproxyforthecostofequity,oneperiodofequityriskis comparedtoanother,andtherequiredrateisnotusedasagaugeofabsoluteperformance. Sincewearemeasuringriskandnotcombiningportfolios,askepticmayask,“Whatifthe risk-returnlinesuddenlybeginsinvertingforonecomparativeyearandnottheother?”. Inotherwords,theanalystwouldwrongfullyinterpretanincreaseinriskasadecreasein risk.Obviouslysuchaprospectcanbeembarrassingandfinanciallydisastrous.Forthis reason,theanalystmuststaywelldiversifiedinhisorhermeasurementactivities- resortingtoothercostofequitymeasurementsaswellasrelatingittotheproportional increasesindebt.Forexample,iftheproportionofdebttoequityrises,andyetourCAPM measurementofthecostofequitydiminishes,weknowthatsomethingmaybeawry, becausemoreleverageissupposedtoincreasethatcost.Attimes,afirmcansubstantially increaseearnings,butdecreasethecostofequitysimultaneously.Sinceearningsandthe costofequityarehighlycorrelated,examinationofthe“alpha”componentintheoriginal regressionmayholdthekeytothemystery.Analphathatincreasesinarisingmarket, evenwhilebetashrinks,willimprovestockperformancebutreducemarketrisk. Therefore,itistotheinvestor’sadvantagetointer-relatethethreecomponentsofthe CAPM,evenifusingsomeothermethodofanalysis.Thereturnonthemarket,therisk- freerate,andbetainteractinwaysthatmakethemarketmorecomprehensible.Ifwe disassociateriskfromabsoluteperformance(marketreturns),wecanextractthe componentsofthecostofequityandaddthemtocapitalstructureanalysisforamore comprehensiveperspective. 119 Besidesthesometimes-periodicdysfunctionoftheCAPM,theanalystmustcontend withwhatistermed,“lookaheadbias”-thefactthatdataisonlyavailablesubsequentto companyperformance.Inotherwords,theearningsdatamaydescribeatotallydifferent companythantheonethatiscurrentlyfunctioning.Whenaneconomyundergoesa fundamentalshiftintoadifferentphaseofthebusinesscycle,datalagstheperformanceof thegivensector.Theinvestor,whoeyes“ahotstock”,makesthemistakeofbuyinga companywhoseearningswilldomuchworsethancurrentlyexpected.Unfortunately,even quartertoquarterobservationbyanalystswillfailbecausethereisnoguaranteeof continuation.Whencapitalstructureanalysisfocusesonproportionalcapitalchanges,itis becausetheyaremoreindicativeofatrendthanpastearnings;itisalmostprohibitively expensivetomakemajorchangesincapitalstructureandthentryto“undo”them.For example,fewcompanieswillmergeinconsecutiveyearsorissuelargeamountsofstockon topofeachotherbecauseoffearofdilution.Ascoveredinthechapteronthecostofdebt, mostproportionalrisesindebtwilloccurinsubsequentyearsbecausethereisalagtime beforetheinvestmentbeginstopayoff.Thus,anyanticipatorychangesinearningsare derivedfromcapitalshiftsratherthanobservedmomentumfromquartertoquarter. Momentumisafunctionofbeinginafavorableleveragestatethatistrending. ANOMALIESPERTAININGTOTHEUSEOFTHECOSTOFEQUITYINCAPITAL STRUCTUREANALYSIS Inaprevioussection,itwasstatedthatusingbookvaluestodeterminethetotalcost ofequity(percentagecostmultipliedbyequity-eithermarketorbookvalues)was consideredimproper.Marketvaluesreflectthetruetotalcostofequityalthoughsome textbooksusebookvaluesforthesakeofsimplicity.Oneofthemaintoolsincapital structureanalysisistheproductionofa“hybridcostofequity”whichusestheCAPM derivedrequiredrateofreturn,andthenmultipliesitbythebookvalueofstockholders’ equity.Ineffect,wecombinemarketriskwithchangesinequitythatare“organic”-either 120 increasesinretainedearningsornewstockissues.Themorerandomandspeculative marketvaluesarefilteredout. Anothertacticistomodifytheriskpremium.Incertainmarkets,theriskpremium willnarrowenoughsothatthereisabsolutelynoreasontobeinthemarketotherthan speculation:therisk-freereturnisactuallylargerthanthereturnfromthemarketindex. Underthesecircumstances,thecostofequitymightevenbemeasuredasnegative!To calculateausableriskpremium,capitalstructuralistswillharboranassumptionthathas historicalcredibility,butmaynotbecurrentlyoperable:themarketriskpremiumisgiven a“floor”ofatleastfivepercent.Astockmarketwithverylowriskpremiumsissimplynot sustainableforalongperiod.Ifaninvestorcangetalargerandmorestablereturnfrom investinginacertificateofdepositwhichhaszerovolatility,thereisnoreasontoinvestin stocksexceptforspeculativepurposes.Researchershaveestimatedtheaveragerisk premiumtobebetweenfourandonehalftosixpercentanditwillbethisnumberthatwill beinputtedintothemodelwhentheactualpremiumdipsbelowit.Essentially,weare imposinganartificialequilibriumfactorontothemarkettomakethecostofequitya workablepercentagefigure.However,whenthemarketriskpremiumgoeswellabovefive percent,weusethatfiguretoemphasizeincreasingrisk.Thus,whenderivedfromthe CAPM,thecostofequityhasanupwardbiasbecauseitismeaninglessasagaugeofriskat extremelylowlevelsoftheriskpremium. ADAPTIVEEXPECTATIONSVS.RATIONALEXPECTATIONS Previoussectionsencouragedthestudent/investortodoasixtymonthregressionon bothanindividualstockandthemarketwhileusingtheaveragemarketreturnsandten yearyieldsasinputsintotheCAPM.The“rationalexpectationshypothesis”which professesthatinvestorstakeallinformationintoconsiderationwhenmakingadecision, wouldencompasssuchalongregression;anyrationaldecisionwouldconsiderthe possibilityofreversiontothemean.Ontheotherhand,an“adaptiveexpectations hypothesis”wouldgivemoreweighttorecenteventswhenmakingadecision.For 121 example,anyaccelerationofinterestratesinthenear-termwouldmakeusbelievethat inflationwasbecominganobstacle. Bothofthesehypotheseshavemerit,dependingonthetimeframeoftheforecast. Whencapitalbudgetingforalongproject,ananalystwhochoosesfive-yearaveragesmay beclosertodeterminingtheactualcostofequity,thanonewhoprojectsseveralcurrent scenarios.However,capitalstructureisconstantlychangingbecausetheoptimumtargetis inastateofcontinuousflux.Adaptingthecostofequitytoreflectcurrentrisk-freeyields andmarketreturnswhilemaintainingthelong-termbetaisacombinationthatwillbias themeasurementinfavorofthecurrentmarket.Iftheinvestorchoosestoinputtheactual yearlyfiguresformarketreturnandtenyearbondyield,theresultwillbeamuchgreater orlessrequiredrateorreturn;anymarketsurgeorinterestratehikewillbeadetermining factorinthis“adaptive”costofequity.Thelong-termregressionwithlong-termaverages willbelessvolatilebutlessanaccuratereflectionofthecurrentsituation.Again,weare doingperiodiccomparisonsofthecostofequity:aslongaswedouniformapplicationsfor eachyear,theresultswillbecomparable.Forexample,allowingthefirstyeartobeafive- yearaveragewhileusingcurrentmarketreturnsandinterestratesinthesecondyear,will violatethisuniformity.EvenifwechoosetheE/P“ruleofthumb”method,wecanget functionalresultsaslongasthemethodisappliedequallytoeachcomparativeyear.That is:wecannevermixonemethodwithanotherorespeciallyusedifferentaverageswhen comparingdifferentyears.Whileeachmethodmayyieldadifferentabsolutepercentage, ourobjectiveistodeterminewhichyearhasthelesserorgreateramountofequityrisk-a functionofchangeinthederivedcost. Howaboutbeta?Shouldwemodifyit?Whilebetashaveproventobeunstable, modifyingbetainaoneyearregression,forexample,wouldbecounterproductive.A short-termbetawouldsimplyreflecttheexternalriskofreactiontothemarketforthat yearonly;ifwehavealreadyusedthecurrentmarketrateforthatyear,themodification ofbetawouldleadtoredundancyandvolatility.EvenintheBlack,JensenandScholes 122 studythatobservedtheinstabilityofbetaoveralongperiod,thetheoreticalthirty-five yearratewascloselycorrelatedwithactualreturns.Infact,reversiontothemeanis typicalforbetawhichseemstoexoneratetherationalexpectationshypothesis-atleastfor thismeasurement.Therealdifferenceinthisareaisconfusionbetweentheconceptsof performanceandrisk.Afundmanagermayusetherequiredrateofreturn,ex-post,asa measureofperformanceandawhollyshort-termCAPMwillyieldvolatileresultsthat exhibittheincreaseinstockpriceforthatyear;butanystockchartcantellusthesame thing.Whenwedoalong-termregressionandmodifytheCAPMwithcurrentmarket inputs,wecancomparethisfiguretothefigurethatusestheproperfiveyear(long-term) averagesandobtainameasureofrisk.Ifwearecurrentlybelowthelong-termfigure,we canlookforariseintherequiredreturnandvice-versa.Thus,adaptiveexpectationscan notonlygaugetheriskofastock,buttheycandefineourpositioninthebusinesscycleas well;anydecliningriskpremiumwithalargerrisk-freeratewouldindicateamarkettop. Whilesuchobservationsmustbetakenincontext,i.e.,sometimesmarketsareincoherent andlackevenrandomlogic,thereisoftenadiscerniblepattern. THEDECOMPOSITIONOFBETA (Thereaderisreferredtothechapteronstatisticsandmathematicsiftheconceptof covarianceisconfusing) AnybetacoefficientcanbebrokendownintotheratioCOV(y,x)/Varianceofx.In astockbeta,“x”isequaltothedistributionofmarketreturnchanges,while“y”isequalto thedistributionofanindividualsecurity’schanges.Essentially,thexcoefficientina regressionlinewhichweterm,“beta”ismadeupofthecovarianceofxandydividedby thevarianceofx.Anycovariancecanbeobtainedfromaregressionbymultiplyingthe correlationcoefficient,“R”,bytheproductofthestandarddeviationsofxandy.The expressionissimplifiedasRσ(x)σ(y) . Thus,whenweobtainbetafromthecharacteristic line,anylargeorsmallcorrelationbecomesanimplicitpartofrisk,andthedegreeofrisk dependsonallthreefactors,R,σ(x)andσ(y). 123 Twosimplerulestoapplytocovarianceare: • 1.ThecovarianceofasumCOV(x+y)withanotherelement,“m”,(COV(x+y,m))isthe sumofthecovarianceofeachelementoftheoriginalsum(x+y)withtheotherelement. ThisisequaltoCOV(x,m)+COV(y,m). • 2.Whenaconstantismultipliedbyoneelementinacovariance,itcanbefactoredout andmultipliedbytheentirecovariance-example-COV(cx,m)=cCOV(x,m) Themostsimplewaytoenvisionbetaisthus,COV(Returnonthestockofacompany, Returnonamarketindex)/Varianceofthereturnonthemarketindex.Thesame expressionreducesto:(Rσ(x)σ(y) . )/σ(x) 2 . Inthemid1980s,theresearchteamofMandelkerandRheedidmuchtoprovidethe financiallogicbehindthedecompositionofbeta.Theytheorizedthatsincestockpricesare thepresentvalueoffutureearnings,thereisanetincomefactorineachstockpricechange. Moreover,thereissomereturnonequitythatisalsoimplicitinstockpricechangesand thatthisconstantcouldbefactoredoutandthenmultipliedbytheseriesofearnings changes.Accordingtorulenumbertwoabove,COV(ROEx%∆NetIncome,Returnon themarket)/VarianceoftheReturnontheMarket=ROExCOV(%∆NetIncome,R market)/VarianceofRmarket.“Rmarket”ismerelythereturnonthemarketandisnot areferencetothecorrelationcoefficientintheregression. Ifthereaderwillreviewthechapteronleverage,heorshewillrememberthattotal leveragemultipliedbythepercentagechangeinsalesyieldedthepercentchangeinnet income.Thechangeinsalesandthechangeinoperatingincome(EBIT)canceloutand yieldthechangeinnetincome:(%∆Sales)x(%∆NetIncome/%∆EBIT)x(%∆EBIT/ %∆Sales)=(%∆NetIncome)Inastablecompanywithcontrollableleverage,thischange innetincomewasderivedfromusingtotalleverageasapredictivetool.Ifwesubstitute DFLandDOLforthedegreesoffinancialandoperatingleveragesrespectively,thechange innetincome,%∆NetIncome,isequalto(DFL)x(DOL)x(%∆Sales). 124 Wecansubstitutethisexpressionforthechangeinnetincomeintheoriginal covariancetoyield:ROExCOV((DFL)x(DOL)x(%∆Sales),Rmarket)/VarianceofR market.Wenextdoanexchangeofconstants.BothDOLandDFLareconstantsasis ROE.Wefactortheleverageconstantsoutoftheequation,whilere-multiplyingROEback intotheequationandeliminatingdecimalpercentages.Inthisnewmodel,wehavetwo newsubscripts,p=previousyearandc=currentyear.Thus,ROE=NetIncome(p)/ Equity(p).Theentireexpressionisnow: (DOL)(DFL)COV[(NetIncome(p))x(Sales(c))/(Equity(p))x(Sales(p)),Rmarket]/ VarianceofRmarket.Toobtainthisexpression,wereasonedthat%∆Salesisequalto (Sales(c))/(Sales(p))-1,andwhenmultipliedbytheROEconstant,NetIncome(p)/ Equity(p),itisequalto[(NetIncome(p))x(Sales(c))]/[(Equity(p))x(Sales(p))],which becomesthefirstterminthecovariance.Weeliminateddecimalpercentagesbyusingthe ratioofcurrentyear/previousyearandthensubtractingone(“1”). Forthesecondtime,weagainfactoroutROE.Thisfinalexpressionis: (DFL(c))x(DOL(c))x(ROE(p))xCOV(Sales(c)/Sales(p),Rmarket/VarianceofR market Ineffect,itisthecurrenttotalleveragemultipliedbylastperiod’sROE,andagain multipliedbythecovarianceoftheperiodicratioofsaleswiththemarketindex.This expressionisthendividedbythevarianceofthemarketindex.Noticethatsalesisno longerapercentchangebutthereturnonthemarketandvarianceofthemarketremain decimalpercentchanges. MandelkerandRheeusedtheoreticalassumptions,thefirstandforemostbeingthat thereturnonastockcanbedecomposedintoareturnonequityandaseriesofnetincome changes.Thesecondmajorassumption,whichisbackedupbyleveragetheory,wasthat thepercentagechangeinnetincomecouldbederivedbymultiplyingtotalleveragebythe percentchangeinsales.Thatisamathematicalfactattributabletothenatureofthe componentsintheequation.Thefirstassumption,however,ismorecontroversialbecause 125 itassumesthatastockisworthitsintrinsicvalueinnetincomerelativetotheamountof equity-withoutbeingdiscountedataspecificcost,whichisthecostofequity.Thatisa disposablecircularargument,becausetheequationdefinesseveralimportantrelationships amongthecomponents. Fromacapitalstructureperspective,thedecompositionidentifiesthekeyelements ofrisk.Althoughtheargumentforaprecisedecompositionremainselusive,Mandelker andRheegaveusaseriesofsignificantvariablesandconnectedtheminalogicalmanner. Whileitmayseempretentioustoattributethereturnonastocktowhollyknown components,bothROEandtheincreaseinnetincomehavegreaterlong-runcorrelation withstockincreasesthanmostanyotherelementwiththeexceptionofearningspershare. Producingavalidbetameasurementthen,mightjustrequirealong-termregressionthat containstherelationshipofROEandnetincomeincreasesasprimecomponents. Moreover,theuseofbothtotalleverageandthegrowthrateofsales,Sales(c)/Sales(p) alsocontributetotheoverallinduction.Steadysalescanallowmorefinancialleverageto beused,thuslimitingtheamountofequityissued.Byusingthedegreeoffinancial leverage(DFL),MandelkerandRheemadeequityanimplicitvariableinthe decomposition.NotethattheROEwasthepreviousperiod’sandthattheleverageis current;theassumptionofimpliedchangesinequityisclear. Therefore,onapragmaticbasis,wecanusetheseidentifiedvariablestogaugerisk. Weneednotsubmitthemtoactualbetaequationsbecausetheyexistinthedomainof probability;someelementwillalwaysbe“outofsync”withtheactualresultsofa regression.Thevariablesare: • 1)∆NetIncomeanditsstandarddeviation • 2)ROE,thereturnonequity • 3)∆Salesanditsstandarddeviation • 4)Thedegreeofoperatingleverage • 5)Thedegreeoffinancialleverage 126 • 6)∆Returnonthemarketanditsstandarddeviation • 7)Thevarianceofthereturnonthemarket Thereadershouldobservethatthecostofequityisrelatedtothechangesinnetincome andexistingequity,butisneitherthereturnonequity(ROE)northepercentageamountof growthinnetincome.Itcombinesmarketforceswithinternalcorporatedynamicsto produceariskfactorentirelyseparatefromincomefundamentals. Whilethedecisiontoaddorsubtractequityiscrucialtoformingatotalcostof equity-thatis:thepercentagecostmultipliedbycommonstockholders’equity-we developthosecalculationsinotherchapters.Theessentialobservationis:thesubtle connectionbetweenasustainablereturnandthecostofequity.Thekeywordis “sustainable”.Withinourdecompositionofbetawereseveraltypesofstandarddeviations whichstandasthegenericstatisticalmeasurementsofrisk.Thesmallerthestandard deviation,themoresustainableisaflowofincome.Equityriskisacompositeofthese otherrisksandwhencomparedtoactualnetincome,addsadifferentdimensiontothat figure;netincomeisaderiveddeductionofaccountingcostslikeinterestandtaxesfrom operatingincome.Ontheotherhand,thecostofequityisanopportunitycost,animplicit economiccostderivedfromseveralriskfactorsthatincorporatemarketvariancewithan individualcompany’sperformance.Muchofthepracticalapplicationofcapitalstructure theorycomesfromtheconflictbetweenthesetwoelements-theconcrete,financial statement-profitabilityofnetincome,withthetheoreticalriskfactorsofthecostofequity. Together,theirrelationshippointstotheconceptofeconomicprofit,andthedevelopment ofvariantssuchas“EconomicValueAdded® 3 ,and“thecapitaldynamic”.These relationshipsareex-postindicatorsofperformancebutex-anteestimatesoftheamountof equityfinancingafirmcanviablydo.Whentheobjectiveistoincreaseperiodiceconomic profit,balancingthecorrectamountofequitywiththeriskofanychangesinrequired 3 Theacronym“EVA”andtheterm“economicvalueaddedareregisteredtrademarksofSternStewart,Inc. 127 returniscounterpoisedwiththeamountofnetincome.Thesethreeelementsarealwaysin conflict:tocreatemoreincometendstorequiremoreequitywhichcausesmoreriskwhen thecostofthatequitygetstoohigh. BALANCINGLEVERAGE,RISK,ANDTHECAPM Onemisconceptionthatbothstudentsandfinancialprofessionalsharboristhatthe optimalcapitalstructureisafunctionofcostratherthanrisk.Ineffect,theybelievein “shopping”aroundforthelowestpricedcapitalandthencombiningitwiththeright weightsofequityanddebttoachieveanoptimum.Thus,besidesflotationcosts,thecostof equitywouldbethecostofdilutingearningspersharebyaspecificamount.Toillustrate thisconcept,consideracompanywithearningsof$100and100sharesoutstanding.The priceofthestockis$20pershareandEPSwouldbe$1.Ifthecompanyneededtoexpand andissued25moresharesforatotalof125sharesoutstanding,atotalof$500incapital wouldberaised.However,becauseofflotationcostsof-let’ssay2%-andadilution valueof($1-(100/125))x(125shares)=25,thecostofequityis(2%x500)+25=35.This myopicapproachtopricingequityturnsitintoanaccountingcost.Retainedearnings wouldbesimilarlypricedbytheadministrativecostdifferencesbetweenkeepingitinan interestbearingaccountandmanagingcash. Costisafunctionofrisk.Whenmeasurableriskisatanoptimum,thecostof capitalisminimized.But-anoptimallevelofriskmaynotbetheleastriskorthegreatest riskbutalevelthatisperceivedbytheinvestingpublictobethebestforthefirm.The subjectiveconnotationof“best”isobjectifiedbymeasurementsfromcreditagencies,and furthersubstantiatedbythepresentandprojectedcash-flowofthecompany. Whileequitybearsitsowncostsandrisks,thosefirmswhochoosetofinancewith debthavetheaddedabilitytoaffectthecostofequitybyincreasingordecreasingitsrisk. Purelogicwoulddictatethatifacompanychangesitsdefaultprobability,thentheriskof owningthestockshouldgoupordown.Afirmthatfinanceswith“junk”bondsshould haveaveryhighcostofequityaswell.Indeed,thereisarelationshipbetweenthecostof 128 bothdebtandequitythatwetreatasdeterministicforpracticalreasons.The Miller/ModiglianipropositionIIstatedthatmoreleverageincreasestheriskandcostof equity.Itessentiallystatedthattherequiredrateofreturnonequitywasafunctionofthe amountofleverage,theinterestrateandthecorporatetaxrate.Whilethe Miller/Modiglianipropositionswerepostulatedinthedomainofmanyconstraints,the developmentoftheCAPMallowedthefreedomtomeasureanestimatedamountofchange intheriskofequityduetochangesinleverage.Theeffectofleverageonthecostofequity wasnowbeingcomparablyquantified. BETAANDLEVERAGE RobertHamadatooktherelationshipbetweenbetaandcapitalstructuretonew heights.HisseminalarticleintheMay1972editionoftheJournalofFinancewasentitled “TheEffectoftheFirm’sCapitalStructureontheSystematicRiskofCommonStocks”. Essentially,Hamadadevelopedthestandardonthetheory:notonlywastheprincipleset forththatmoredebtraisedequityrisk,Hamadatoldthereaderbyhowmuchandbywhat mechanism.Infact,theCAPMcouldbeusedtodrawthesameconclusionsas Miller/Modigliani,whichfurthersubstantiatedtheirresearch:inaworldwithouttaxes,or bankruptcycosts,capitalstructurewasimmaterial,butwhentaxeswereadded,the optimumproportionwasonehundredpercentdebt.Theseextremecornersolutions eliminatedtheconceptthatcapitalcostsweredeterminedoutsideoftheconfinesofthe corporation;theywerenotdictatedbythemacroeconomy,onlyaffectedbyit.Therefore, acombinationoftaxpolicy,andmanagingdefaultriskwoulddeterminetheamountof leverageandhaveaneffectonthecostofequity. Formanyyears,acommonobservationwasthatmoreleverageinapubliclytraded companyledtogreatervolatilityinthestock.Worldeventsappearedtoaffectdebtladen companiesmuchmorethanunleveredones,whichwascorrectlyattributedtothelink betweenexchangerates,inflationandinterestrates:leveragedcompanieswereatthe 129 mercyofa“creditorchain”-frombankstoforeignsuppliers,andtheincreasednumberof variablesthataffectedtheflowofearningsmadethestockmoresensitivetochanges. Hamadagaveanalystsanimprecise“ballpark”figureforchangesinequityrisk duetoleveragechanges.Again,thefoundationislinearbecausethebestestimateof extremevariationineitherdirectionisastraightline.Essentially,betahadtwostates:an unleveredstateinwhichriskwouldflowmostlyfromoperationsbecausethecompanyhad nodebt,andsecondly,aleveragedstatethatreflectedchangesinthelevelofdebt. Throughasimpletransformationofbeta,theanalystcouldobservehowmuchriskwas addedtothestockafteracertainamountofdebtwasincreased,andevenobservetherisk inastockthatwasattributabletooperations(businessrisk)alone.Giventherestrictions oftheCAPM,amathematicalproofcouldnowbesetupthatindicatedhowleverage increasedbeta.Moreover,itbecameapparentthattheoptimalcapitalstructurewas foundedasmuchupontheeffectofleverageonthecostofequity,asitwasonprevailing interestrates.Althoughbetaisaffectedbymyriadothervariablesandaprecise determinantiselusive,thisnewviewofleverageemboldenedthefieldofriskmanagement. A“ballpark”figurewascertainlybetterthannoneatall,andnowanalystscould experimentwithcombinationsofdifferentcapitalsourcesandlevelsthatpurportedly minimizedrisk. ASIMPLEEQUALIZER TomakeHamada’sanalysisworkable,onlyfourinputswerenecessary.Theywere: • 1)Thecurrentbetaofthestockwhichwasavailablethroughregression • 2)Themarketvalueofdebt • 3)Themarketvalueofequity • 4)Thecurrenteffectivecorporatetaxrate Thesevariableswerecombinedinthefollowingexpression: (1+[(1-taxrate)xMarketvalueofdebt/Marketvalueofequity]) Thus,theunleveredbetabecameafunctionofthecurrentbetadividedbythisexpression: 130 Unleveredbeta=Currentbeta/(1+[(1-taxrate)xMarketvalueofdebt/Marketvalueof equity]) Theleveredbetabecame: (Unleveredbeta)x(1+[(1-taxrate)xMarketvalueofdebt/Marketvalueofequity]) Achangeinthebetaduetomoreleveragebecame: (Originalbeta)/(1+[(1-taxrate)xOldMarketvalueofdebt/Oldmarketvalueofequity]) x (1+[(1-taxrate)xNewMarketvalueofdebt/Newmarketvalueofequity]) Iftheriskofequityisdependentupontheamountofdebt,doesitfollowthatthe costofdebtdeterminescapitalstructure?Onlysofarthatinterestratesarea“perfect” transmissionmechanismforbothgovernmentpolicyandtheprobabilityofdefault.While thecostofdebtisacontributingfactorthroughtheeffectofinterestratechangesinthe macroeconomy,the“real”costofdebtisthecostofbankruptcywhichisattributableto theinterfacebetweenoperatingincomeandtheamountofinterestexpense,aswellashow assetsarestructuredinthefirm.Bankschargeinterestbasedoncreditworthinesswhich stemsfromtheabilitytocoveraloan.Thesteadycash-flowthatisimplicitinsuch coverageisanoutgrowthofoperatingriskandtheamountofdebtalreadyincurred.Since moredebtincreasestheprobabilityofdefault,itraisesthecostofequity.Whilemoredebt mayleadtohigherinterestpayments,higherinterestratesalonewillnotleadtomore equityrisk. Therelationshipbetweenbetaandtheprobabilityofdefaulthasnotbeenwell developed.Onewouldexpectinordinatelyhighorlowbetastoforeshadowtheinabilityto makepaymentsoninterest.However,betaisnotameasureofperformancebutof volatility.Whilemoredebtmayincreasebothbetaandtheprobabilityofdefault simultaneously,afirmthatisnearbankruptcywhenthemarketsurgesmayhaveastock thatlanguishes.Theinabilitytomakepaymentsmaystemmorefromalackofincome generationthanitdoesfromtoomuchleverage.AsinMandelkerandRhee’s 131 decompositionofbeta,salesthatarelanguidcouldcounterbalancetheeffectofhigher leverage,makingbetanon-reactive.Inthatcase,the“alpha”variableintheregression wouldtakeoverandbecomeverylow.Evenwhenbetaisup,thestockwouldstillunder performthemarketandthecorrelation,“R”,wouldbelowaswell. Theeffectofleverageonbetaismostobservablewhenweunleverafirm’sbeta- thatis-findingouthowmuchequityriskithaswithoutanydebt.Firmsthathavea comparativelyhigh-unleveredbeta,likemanytechstocks,canaffordlittledebtbeforethe stockbecomesprohibitivelyrisky.Analogously,manyhighdebtcompaniesthathave steadyincomes,likeutilities,willhavelowunleveredbetas;thesecompaniesseelittlerisein theprobabilityofdefaultevenwhentheyincreasedebt,mostlybecausetheyareheavily regulated.Thus,manyfirmsthathavemoreleveragethanwarrantedbytheirbetaswill havenaturallyriskierstocksandbecomethefodderforspeculativebettingonWallStreet. Ontheotherhand,astockwithtoolowabetawhosefirmisnotoutproducingtheir respectiveindustrymayunderperformthemarketbecausetoolittleriskisbeingtaken. Thehappymedium,ofcourse,isthefirmwithalowbetastockwhoisindeedout producingtheirsector;thesefirmswillhavelargedifferencesbetween“expected”and “required”ratesofreturns.Theywillalsoofferthemostreturnfortheleastamountof risk. ThestrategicimplicationsofHamada’sresearch(andmanyothersinthatsame period)shouldbeapparent.Anyfirmthatiswealthyenoughtoaffordthelowerinterest debtfinancingthatisofferedatthebeginningofarecovery,canincreasebetaatthesame timethatthemarketispickingup.Whentheinvestmentindebtbeginstopayoff,cash- flowsimprove,attractingmoreequitycapitalintothecompany.Bythetimethenext downturnoccurs,theproportionofdebttoequitywillhavedecreased,diminishingsomeof thehighercostofequitythatwouldhaveoccurredbecauseofhigherinterestratesanda surgingmarket. CONCLUDINGCOMMENTS 132 Anymodelthatpurportstobeascomprehensiveasthecapitalassetpricingmodel (CAPM)needstobeusedwithjudiciouscaution.Whenitisaccurate,itcanresolveseveral issues: • 1)Itcanmeasuretheadditionalriskcausedbyincurringmorefinancialleverage. • 2)Itcanmeasureoperatingriskandanyadditionalriskderivedfromthecombiningof assets. • 3)Itestablishestherequiredrateofreturnonafirm’scommonstock. • 4)Itcanbecomparedtothe“expected’rateofreturnandusedasa“buy/sell”signal. • 5)Itestablishestheriskofastockincomparisontotheoverallmarket. • 6)Itestablishesacostofequityforthecompany. • 7)Itcanbeusedtogaugethebusinesscycleaswellasthedifferencesbetweenrisksin theequityandfixedincomemarkets. • 8)Itcanbeusedasatoolinportfoliomanagementtodiversifyawayrisk. Thewatchwordsinanyanalysissystemarealwayscorroborationandbalance.Without understandingandseekingoutalternativeanalyses,theinvestoris“puttingallofhisorher eggsinonebasket”.Thatisariskyproposition,byanystandard. (BacktoTableofContents) 133 APPENDIX:THEMATHEMATICALRELATIONSHIPBETWEENLEVEREDAND UNLEVEREDCOMPANIESINTERMSOFTHECAPM ThecapitalassetpricingmodelcorroboratestheMiller/Modiglianipropositions.In thecaseofnobankruptcyandnotaxes,capitalstructureisimpertinent;firmsthatfund withequityhaveanequalamountofriskasfirmswhofundwithdebt.However,when taxesareapplied,theoptimalstructureisoneofalldebt;taxdeductionsoninterestgivea distinctadvantagetoleveragedfirmsaslongasthereisnoprobabilityofbankruptcy.In thissection,weexaminetheMiller/Modiglianiargumentsfromthestandpointofthe CAPM.First,weestablishthedifferencesbetweenbetainbothleveragedandunlevered companies,andconcludethattheyarederivedfromtheamountofleverage.Secondly,we usetheseconstructstoexhibitthedifferenceinthecasesofbothtaxesandnotaxes. 1)EQUATINGTHELEVEREDFIRMWITHTHEUNLEVEREDFIRM Studentstendtointerpretthisproofasreferringtotwodifferentfirms.Itismore enlighteningtothinkofthecomparisonasthesamefirmintwodifferentsituations-with debtorwithoutit.Thus,thenetoperatingincome(NOI)willbethesameforeach situation,butthereturnonequity(ROE)willbedifferentbecauseintheleveredstate,the firmhastopayinterest,butintheunleveredstate,itdoesnot.Anotherassumptionisthat betaisstatedintermsofROEratherthanthestockprice.Sincethestockpricewilltendto mirrorROEoverthelong-run,makingbetaadeterministicfunctionofnetincomeand equitywilldisplayacauseandeffectrelationship.Thecertaintythatweimposehelpsto exhibitpossibletrends,althoughwefullyrealizethatbetaandstockpricehaveamore volatileandprobabilisticrelationship. 134 Table5-2 LEVEREDFIRM(L) ROE(L)=ReturnonEquityfortheLeveredFirm NOI=NetOperatingIncome B(L)=AmountofDebtfortheLeveredfirm S(L)=ValueofStockoftheLeveredFirm r=InterestRateonDebt.Therefore,rB(L)isInterestExpense. ROE(L)=NOI-rB(L)/S(L) Table5-3 UNLEVEREDFIRM(U) ROE(U)=ReturnonEquityfortheUnleveredFirm NOI=NetOperatingIncome S(U)=TheValueofTheStockoftheUnleveredFirm.S(U)=V(U)=Valueofthe UnleveredCompany ROE(U)=NOI/S(U)=NOI/V(U) STEP1)NotethatNOI=(S(U))(ROE(U)).ThiscanbesubstitutedintotheNOIofthe leveredversionofthecompanysothatunleveredandleveredcanbeequated. STEP2)Theleveredfirmisnow[(S(U))(ROE(U))/S(L)]-[r(B(L)/S(L)].Wemerely broketheexpressionintotwoparts,givingitthecommondenominatorofS(L). STEP3)Determinethebetaoftheunleveredfirm.Rememberthatbetaisacovariance dividedbythemarketvariance.WeareusingtheunleveredROEratherthanthestock priceandso:betaisCOV(NOI/S(U),Rmarket)/VarianceofRmarket.Thisisthesame asCOV(ROE(U),Rmarket)/VarianceofRmarket.RememberthatRmarketisthe returnonanappropriatemarketindex. STEP4)Determinethebetaoftheleveredfirm.Thisismorecomplexbecausewenow expresstheleveredfirmintermsoftheunleveredfirm(stepsoneandtwoabove).First,we usetheadditivelawforcovariancetocombinetheentireexpressionintoonecovariance. Next,weeliminatetheconstantsbecausetheyhaveavalueofzerointhecovariance. Thirdly,wemultiplythevariableswithinthecovariancebythetermS(U)/S(L)toequate 135 theunleveredbetawiththeleveredbeta.Fourthly,wefactoroutthissameterm(S(U)/ S(L))fromthecovariance. a)Usingtheadditiveproperty:BETA(L)=COV(((NOI-rB(L))/S(L)),Rmarket/ VarianceRmarket=(1/VarianceRmarket)(COV((NOI/S(L),Rmarket)))(COV(rB(L) /S(L)),Rmarket).Wehavebrokentheexpressionintotwoseparatecovariances. b)Eliminateoneofthecovariancesbecauseitisaconstant.WeeliminateCOV((rB(L) S(L),Rmarket).NowBETA(L)=(1/VarianceRmarket)(COV(NOI/S(L),Rmarket) c)IfwedivideNOIbyS(U)andthenmultiplythetermbyS(U)/S(L),thetermNOI/S(L) willremainthesame.ThatisBETA(L)=(1/VarianceRmarket)(COV((NOI/S(U)x (S(U)/SL)),Rmarket) d)Bythelawofcovariancefactoring,wecanfactoroutS(U)/S(L)sothatnowBETA(L)= S(U)/S(L)(COV(NOI/S(U),Rmarket))(1/VarianceRmarket) Noticethatinpartd,theleveragedbetaisthesameastheunleveredbetaifwe multipletheunleveredbetabyafactorofS(U)/S(L).Theexpression(COV(NOI/S(U),R market))(1/VarianceRmarket)isthebetaoftheunleveredfirmthatwestatedinstep3. Thus,merelybymultiplyingtheunleveredbetabyafactoroftheratiosofmarketvalues (S(U)/S(L)),wecandeterminethebetaoftheleveredcompany.Iftheleveredbetaisa given,wedivideitbythesametermtoyieldanunleveredbeta. Bothbetasareequalbyaconstant(S(U)/S(L))thatreflectthedifferencein financialleverage.Sincetheleveredfirmissuesdebtinsteadofstocktoraisethesame amountofcapital,thissmalleramountofstockwillbethedeterminingfactor.Thereader willrememberthatfortheunleveledfirm,S(U)=V(U).Thatis:thevalueoftheunlevered firmwastotallydependentontheamountofitsstock,andsowearecomparingawhole value(S(U))toapartialvalue,(S(L)).Thisfigurewilltakeonsomeintegervaluegreater than1,makingtheleveredfirm’sbetaalwaysgreaterthantheunleveredfirms-bya factoroftheamountofleverage. 136 2.USINGTHECAPMTOCORROBORATETHEMILLER/MODIGLIANI PROPOSITIONS Toreviewthelastsection,recallthatwehaveestablishedtwosubstitutions: • A)BETA(L)=S(U)/S(L)(BETA(U)) • B)SinceNOI=(ROE(U))(S(U)),thenROE(L)=[(ROE(U)(S(U))/S(L)]-[rB(L)/S(L)] whichessentiallystatesthatthedifferenceinthetwoROEsisbecauseinterestexpense (rB(L))issubtractedfromoneandnottheother. Wefurtherestablishtwopremises,equatingthemarketreturnswiththeindividual companyreturnsthroughthemechanismofbeta. • Premise1:(E(ROEofafirm)-r)/Betaofafirm=E(ROEmarket)-r.Inthis expression,“E”istheexpectedreturn,while“r”isequaltotherisk-freerate.Inthis instance“E”isnotmultipliedbyROEbutistheexpectedreturnthereof.Thisassumes thattheexpectedreturnofthereturnonequityofanyfirm,dividedbythefirm’sbeta, isequaltotheexpectedvalueofthereturnonthemarket. • Premise2:(E(ROE(U))-r)/BETA(U)=(E(ROE(L))-r)/BETA(L).Thisstatesthat theexpectedvalueofanysecuritydividedbyitsbetaisequaltotheexpectedvalueof anyothersecuritydividedbyitsbeta. PROPOSITIONI InpropositionI,Miller/Modiglianibasicallystatedthatcapitalstructuredidnot matterinaworldwithouttaxesorbankruptcy. STEP1:WesubstituteforROE(L)andBETA(L)inpremise2.Again(ROE(U))(S(U))is substitutedforNOI,while(S(U)/S(L))(BETA(U))issubstitutedforBETA(L). (E(ROE(U)-r)/BETA(U)= [(E(ROE(U))(S(U))/S(L)-rB(L)/S(L)-r]/(S(U)/S(L))(BETA(U)) STEP2:Weeliminatetheterm,S(U)/S(L),ontherightsideoftheequationby algebraicallymultiplyingbyitsinverse,S(L)/S(U).Weobtain: (E(ROE(U)-r)/RETA(U)=[E(ROE(U)-rB(L)/S(U)-rS(L)/S(U)]/BETA(U) 137 STEP3:Weeliminatetermsandmultiplyitthrough: -r=-rB(L)/S(U)-rS(L)/S(U) STEP4:Wemultiplyby-1andthenfactor: r((B(L)/S(U))+(S(L)/S(U))=roralternatively,S(U)=S(L)+B(L).Sincethevalueofthe unleveredfirmisequaltoitsstock(V(U)=S(U))andthevalueoftheleveredfirmisequal toitsstockandbonds(S(L)+B(L))=V(L),thenV(U)=V(L).Capitalstructuremakesno differenceinthevaluation. PROPOSITIONII InpropositionII,Miller/Modiglianiarguedthatinaworldoftaxes,theoptimal capitalstructurewouldbemadeupofonehundredpercentdebt.Inthetaxcase,wego throughthesamesetofequationsexceptthatNOI(andthesubstitutedvariables)are multipliedbyafactorof(1-taxrate). STEP1:ROE(L)=[(ROE(U))(S(U)(1-taxrate))/S(L)]-[(1-taxrate)rB(L)/S(L)] STEP2:Multiplyingtheequationsthroughyields: -r=[(-(1-taxrate)rB(L))/S(U)]-[rS(L)/S(U)] STEP3.Reducingtheexpressionandeliminating“r”: S(U)=(1-taxrate)B(L)=S(L)oralternatively, S(U)+(taxrate)B(L)=B(L)+S(L) STEP4:Applyinglogic:SinceS(U)=V(U)andsinceV(L)=B(L)+S(L)then V(L)=V(U)+(taxrate)B(L) Thevalueofaleveragedfirmisequaltothevalueofanunleveredfirmplustheproductof thetaxrateandtheamountofbonds.Eitherahighertaxrate,ormoredebt,orboth,will increasethevalueofthefirm,andsothevalueismaximizedatonehundredpercentdebt. (BacktoTableofContents) 138 6 THECAPITALDYNAMICANDOTHERTOOLS Thischapterwillprovidethestudent/investorwithsomepracticaltoolsfordetecting movementtowardanoptimalcapitalstructure.Whilemathematicaloptimizationremains controversialandinexact,thesemeasurementsexploitcorrelationandprobability;their raisondetreَstemsfromcreatingdistancebetweennetincomeandthecostofequity,and notfromthetime-dependentrequirementsofmaximizingafunction. Fortunately,thefinancialcommunityprizesthetransitionalprogresstowardagoal morethanitsactualachievement.Ifthisstatementseemsparadoxical,considertherisk factorsinvolvedwhenacompanyisalongwayfromreachinganygoal-sales,earnings stockprice,etc.Atfirst,expectationsareincreasedbyanalystsoratargetwouldnothave beenoriginallyset.Next,whenacompanyisagreatdistancefromthetarget,therisksof notachievingitaregreaterthanwhentheachievementis“adonedeal”,andthefirmis veryclosetoitsobjective;atthispoint,severaldifficult“hurdles”mustbenegotiated. Finally,WallStreetrewardsthosecompaniesthathavesuccessfullyovercomeobstaclesat apointwherethegreaterinvestmentcommunityisunawareoftheiroccurrence-before theknowledgeofprofitabilitybecomescommonplace.Infact,whileitappearsthatthereis substantialdownsideriskfromnotmeetingobjectives,theupsidereturnoccurs“earlyin thegame”whentheriskisgreatest. THEPERCENTAGETRAP Althoughitismoremathematicallysoundtouselogarithmsratherthanpercentages asthemeasurementofanincrease,thistextsubmitstothepercentageimperative. Percentagegainsarethestandardinbusinessprimarilybecausetheyareunderstandable inacomparativesense;theyaresomewhatakintoanAmericanusingtheEnglishconcept of“foot”ratherthanthemetricsystem.However,theirowndependenceonreference allowstheirbasestoshift,creatingshort-termhyperboleratherthanscientificexactitude. 139 Considerasanexample,anickelincreaseonanickelinvestment;suchanincreasecanbe toutedasa“onehundredpercentgain!”,whiletenmilliondollarsonabilliondollar investmentis“only”onepercent.Indeed,whenaninvestmentgoesupfrom80to100,itis a25percentincrease,butifitgoesfrom100to80,itisonlya20percentdecrease,simply becausethereferencebasehasshifted.Incapitalstructureanalysis,wedonotsell investmentsortoutearningsincreases.Infact,percentagegainsarerelevantbecause absolutesizematterslessthanthemovementitself;withoutaneedforreference, percentagegainsareasimplemethodofenumeratingchange.Dissociatedfromemotional content,atwentypercentgaininthecostofequity,forexample,willhavemuchless meaningtotheaveragebusinesspersonthatatwentypercentgaininthecostofgoodssold. Forthatveryreason,percentagechangesinthese“backgroundfundamentals”– opportunitycosts-becomethebuildingblocksforchangesinmoreprominent measurements. THEWEIGHTEDAVERAGECOSTOFCAPITAL Introducedbrieflyinthechapteroncapitalstructure,theweightedaveragecostof capitalorWACC,formsthebackboneofcapitalstructuremeasurement.Althoughitdoes notalwaysadhereperfectlytoitstheoreticalunderpinnings,WACCchangescantellas muchaboutthedirectionofthecompanyasearnings.Infact,whenusedintandemwith earningsmeasurementsandinformationabouttheeconomy,theWACCcanpredict slowdowns,peaksandplateausaswellasanyeconomicbarometer. Theproportionalcomponentcostofeachsourceofcapitalinaggregatemakesup theWACC,anditisessentialthatthiscostbeminimized.Thosecomponentcostsmay includelong-termloans,bonds,retainedearnings,preferredstock,capitalleases,and commonstock-bothnewissuesandpaidincapital.Onatechnicallevel,suchaswould occurincapitalbudgeting,eachcapitalitemneedstobespecifiedasaseparatecost.For genericestimation,however,wehavegroupedlong-termdebtandstockholders’equity togethertoformourdefinitionof“capital”:aninvestorwhoneedstocompareseveral 140 companiesinrapidsuccessionneeds“bestestimates”ratherthantechnicalprecision.The libertiesthataretakenbecomeapracticalextensionofturningatheoretical“opportunity cost”intoapracticaldecision-makingtool:thecostofequityisremovedfromthe contingenciesofaccounting,andsotheriskshavetobedefinedbymodelinterpretation. Someoftheriskwillalwaysremainsubjective,althoughwetrytoeliminateitbydefining asmanyvariablesaspossible. Therefore,acorporationwhopreciselyitemizesitscostofcapitalneedstoensure implementationandreturn,notjustriskandreturn;theimplementationofanyprojecthas massivelogisticsproblemsonarelativelevel.Abank,forexample,mustdoacomplex itemizationofcosteffectivenesswhenitinstallsautomatictellers.Theinvestor,however,is moreconcernedabouttheriskofinstallingthem:thereislessneedfordetailedknowledge, butmoreofaneedforcomparingalternatives. Oncethedifferencesinmethodologyarediscarded,thedecisionvehiclewillbe similar.Corporationswillderiveaninternalrateofreturnandcompareitwiththe discountedcash-flowofaproject.Investorswillobservethereturnfromacompany,and compareittowhatcompaniesofsimilarriskwouldyield.Together,eachentityusesthe weightedaveragecostofcapitaltoformthecomparison. Onapragmaticlevel,theminimumcostofcapitalisalsotheoretical.Thetaxeffects ofinterestdeductibilitymandatetheuseofdebtuptothepointwhenacompany’srisk manifestlybeginstoundermineitsstockprice.Mostcompaniesareriskaverseenoughto performatalevelsubstantiallybelowthatthreshold.Thus,themajorconcernforthe investoristoobservethechangesintheWACCandnotworryaboutatheoreticalabsolute. Forexample,attimeswhentheFederalReserveraisesrates,theWACCmayrisealthough thefirmhasminimizeditscapitalcosts;sinceacompetitor’sWACCrisesaswell,the increaseisentirelyrelative.Toresolvesuchproblemsofinterpretation,werelatethe WACCtoothercapitalstructurevariables,andalmostcompletelydispensewiththe observationofitsindividualmovement. 141 THETHEORETICALSHAPEOFTHEWEIGHTEDAVERAGECOSTOFCAPITAL CURVE Combiningthelinearityofthecostofequitywiththeupwardtrendingofthecostof debtformsaconvexcurve.Atsmalllevelsofdebt,theWACCcurveslopesdownwardto reflecttheflatterslopeandgreaterproportionoftheequitycurve;moredebtactually decreasesthecostofcapital.OncetheWACChitsaminimum,debtbecomesmore expensive,anditbeginstocurveupwardstoreflectthesteeperslope,andgreater proportionofthedebtcurve. Figure6-1 Theshapeofthecurveiswhollydependentonthespreadbetweentherespective costsandrisksofequityanddebtateachlevel.Capitalstructureoptimizesatthe inflectionpointandthisisalsowherethestockpricetheoreticallymaximizes.Arealistic renditionwillobservethesometimesabsurddivergencebetweenbondandequitymarkets: attimes,yieldcurvesbecomeinvertedandshort-terminterestratesarehigherthanlong- termrates.Theriskpremiumbetweenthemarketsmayevaporateforlongperiodsand equitieswillhavegreatvolatilitywithonlyaminimumofreturn.Atothertimes,therisk premiummovesintheoppositedirectionasinvestorssellbondsandbuystocks.Thus,over Costof Capital Debt/Equity CostofEquity CostofDebt WACC Optimum D/E 142 thelongterm,thecostofcapitalmirrorstheinherentrisksofeachcomponentsourceand formsahypotheticalcompendium.Whileshort-termvolatilitypreventstheanalystfrom pinpointinganabsoluteminimumcostofcapital,severalrelationalvariablesallowthe examinationofadirectionalflow,anditisinthis“vectoring”context,thatinvestment decisionscanbemade. Tounderstandtheneedforacombinedcapitalstructuremeasurement,considerthe theoreticalrelationshipbetweendebtandequity;whenmoredebtisincurred,thecostof equityrises.However,intherareoccurrenceofFederalReserveratecuts,firmswhocan stillaffordalotofdebthavetheopportunitytodecreaseboththecostsofdebtandequity simultaneously.Afirmwithalowprobabilityofdefaultcanmostbenefit.Oncethe marketisbackinequilibrium,thecostofequityrisesasmoredebtisincurred,andthe hypotheticalrelationshipisreadilyrestored.Inthemeantime,theWACCmayrisewhenit issupposedtofallandviceversa.Tocorroborateitsmovement,theanalystusesboth earningsandchangesintheproportionsofcapitalsources.WhentheWACCismultiplied bytheamountofcapitalandthensubtractedfromearnings,acomparisoncostifsformed. Thechangeinthiscomparisoncostwillconfirmordenymovementinthecostofcapital. ACCELERATIONRATES ArehigherearningsandadownwardtrendintheWACCasignaltoinvest?While suchacrosscurrentmayleadtoanoptimizedcapitalstructure,theinformationistoo spottytomakeadecisivejudgment.Whatisofprimaryimportanceistheaccelerationof earningsincomparisontotheWACC.Duringarecovery,forexample,bothearningsand theWACCmayrisetogetherbecausethemarketgetsbetter,andtheFederalReserve raisesratestocontaininflation;therateofincreaseinearningsfaroutpacestherateofthe WACCandstocksconsequentlyrise.Whenmarketconditionschange,thecostofcapital becomeshigherorlowerateachlevelofdebttoequity,changingtheoptimummix.In effect,theWACCcurveshiftsupordown,butwilllagtherateofearningsineither direction.Itisthisdisparitythatcreatesopportunityintheequitymarkets;corporations 143 canshifttheirmixesofdebttoequitytotakeadvantageofthedistancebetweenearnings andtheWACC. THEWACCANDRISK Themarketdoesnotalwayspriceriskefficiently.Whilesomeseeopportunityin thevariouseconomic“bubbles”thatarise,thereisusuallyacompensatorydownsidethat createsforcesintheoppositedirection.Whendebtisespeciallyinexpensive,thetendency to“gooverboard”isjustifiedbyalowerWACC.Aslongasthecostofdebtislowerthan othersourcesofcapital,theWACCtellstheanalysttoloaduponitbecauseitistheleast expensive.Toproperlygaugerisk,however,acompanymustconsidertheprobabilityof defaultfirstandforemost.Withoutthecontextofpotentialbankruptcy,theWACCwill optimizeatacapitalstructureofalldebt,simplybecausetherearenoupwardboundson it.Sinceinterestrateschangefrequently,andwillrarelybeabovethecostofequityfora solventcompany,theWACCcannotbeminimizedonthebasisofitsownparameters.For example,afirmmayloaduponalargeamountofzerocouponbonds,andtheWACC woulddiminish,notfactoringinthepotentialdistressofpaymentinthefuture.Thus,the WACCisanindicatorofrisk,butitisnotdefinitive.Itmayshowpositiveornegative trends,butisnotthefinalarbiterofcorporateaction.Forthatdecision,thefirmneedsto balancethetaxadvantagesofdebtwiththeprobabilityofdefault.Onlywhenafirm balancesthestabilityandamountofearnings,thetypeofassets,andtheabilitytomake promptpayments,willanoptimalcapitalstructurearise. THEMECHANICSOFTHEWACC:RISKADJUSTMENT Whenthecomponentproportionsofallsourcesofcapitalaremultipliedbytheir respectivecosts,aweightedaveragecostofcapitalisformed.Aftercalculatingindividual costs,thecruxoftheequationrevolvesaroundtheproportionsandthedefinitionof “capital”.Corporationsandlargeinvestorsneedacompleteitemizationofeachsourceof funding;theywillincludebothpreferredstockandshort-termnotesinthecapital denominatoraswellaslong-termdebt,commonequityandretainedearnings.Thesmaller 144 investor,however,maywanttomakequickcomparisonsandusejustthelatterthree components.Consequently,thedenominatorwillbesmaller,andbothlong-termdebt,and commonequitywillbealargerproportion.Sincethesmallerinvestorismoreconcerned withrisk,andlesswithimplementationofaproject(oracontrollingstake)expeditingthe WACCwillcreateutility;long-termdebtandequityarethemajorriskfactorsandthe contractionofthedenominatorwillrecognizethis.However,this“riskadjusted”WACC isnotthetrueWACC,andwillleadtogrosserrorsifusedinthecontextofcapital budgeting;mostfirmsdiversifyawayriskbyfundingfromasmanydifferentsourcesas possibletoavoidthisexacttypeof“corner”solution. Tocontrastthetwomethods,observethefollowingbalancesheetwhereeach componentisitemizedbyapre-taxcost.Thereaderisreferredtopreviouschaptersfor themethodologyofdeterminingindividualcost. Table6-1 TYPEOF CAPITAL AMOUNT PERCENTAGE COST ASSETS 1000000 100 STOCKHOLDERS' EQUITY 700000 70 PreferredStock 70000 7 8% RetainedEarnings 140000 14 10% CommonStock 490000 49 12% DEBT 300000 30 Long-term 250000 25 7% CurrentLiabilities Notes 50000 5 4% OtherCurrent Liab. 0 0 CORPORATEMETHOD Forthesakeofillustration,othercurrentliabilitiesareassumedtobezero.Hadtheybeen enumerated,itwouldbepropertoexcludethemfromthecapitaldenominatorbecausethey 145 areasourceofinternalandnotexternalfunding.Inthatcase,allothersourcesoffunding wouldriseinproportion.Forexample,hadothercurrentliabilitiesbeen25000,andnotes 25000,thentheproportionofcommonequitywouldbe(490000/975000)or50.25%and not49%.Notes,however,wouldgodownto(25000/975000)or2.56%asacomponent proportion. Assumingataxrateof30%,debtiscalculatedas(interestrate)(1-taxrate).Theother costsaremultipliedbytheirproportionsandthensummed(costsareinparentheses) .07(.08)+0.14(0.1)+0.49(0.12)+0.25(0.7x0.07)+0.05(0.7x0.04)=0.09205or9.2%. Thefullcapitalbudgetingtypeofanalysisyieldsa9.2%WACC. RISKADJUSTEDMETHOD Wechangethecapitalbasetothesumofretainedearnings,commonequityand long-termdebt.Thecomponentproportionsoftheseelementswillrise: Table6-2 TYPE AMOUNT PERCENTAGE COST RetainedEarnings 140000 15.9 10% CommonEquity 490000 55.68 12% Long-termDebt 225000 28.4 7% TotalCapital 880000 100 Thecalculationisasfollows: 0.159(0.1)+0.5568(0.12)+0.284(0.07)=0.1026or10.26percent.Thus,theriskadjusted WACCissubstantiallyhigher.Suchabiaswillinflatetheimportanceoflong-termdebt andcommonequityinthecapitalstructurewhichisitsintendedpurpose. THEMARGINALCOSTOFCAPITAL Fewmoderncorporationscanattainanoptimalcapitalstructureandremainthere foranysignificantlengthoftime.Notonlywillchangesintheeconomyalsochangethe targetmixofdebttoequity,butthefirmcanbepenalizedfortakingtoolittlerisk.In 146 effect,thecompanywillmoveawayfromitsoptimalstructurefromtimetotimebytaking oninordinateamountsofrisk,andthenideally,movebacktowardthetargetasrapidlyas possible;suchrapidityimpliesareturnfortheextrariskthatthefirmincurred. Moreover,itiscosteffectivefromanadministrativeperspectivetoraisecapitalinlarge incrementsbecausesuchinflowsareoftenpurposefulandacttofocusmanagementonan objective.However,largecapitalinfusions,whetherindebtornewstock,maytakea longertimetointegrate,creatingstagnationanduncertainty.Thechallengeatthispointis perhapsoneofthemostdifficultinbusiness:tobegingeneratingaprofitwhenthe infrastructureisnewanduntested. Thesolutiontothisproblemistobeginmovementbacktothetargetcapital structurewhichcanbetakeninincrements;mostinvestorswilldemandequitywhenthey seesalesandprofitsgrowing.Theprocesscantakeuptothreeyearsormorebutwillcome tofruitionwithconsistentperformance.Inthemeantime,managementneedstobeaware ofitstargetcapitalstructureandthatahigherWACCwillbeimpliedwhenthefirmmoves awayfromit.Forexample,ifthetargetstructurecallsfor35%debtand65%equity,and thefirmhas500millionindebt,then500/.35=1429canberaisedintotalcapital. KnowledgeofthislimitcanhelpaCFOgaugethefeasibilityofcapitalprojects;when combinedwiththeamountofearnings,someofwhichwillsupplycapitalthroughretained earnings,theCFOcansetdividendpolicy,andcapitalbudgets,andthenplantofinda sourceoffundingifneeded.But-itismucheasiertomoveofftargetandthenbackagain thantobeconstantlyconstrainedbytheamountofearningsandcapital.Movingoffthe targetwillraisethemarginalcostofcapital-thatis-eachadditionaldollarofcapitalraised willbemoreexpensive,butwillprovidegreaterpotentialreturniftheriskisnegotiated well.Ineffect,managementhastoactinatemporarycapacitynottomaximizethepriceof thestock.Itisatthisjuncturewhenshort-termperformanceissacrificedforlong-term gainsthatcareersaremadeorbroken. 147 Forinvestors,themostimportantbreakpointcomeswhenretainedearningsare exhaustedanddebtisalsohigh.Itisatthispointwhennewstockisissued,andthe investormustdealwithatleastfivediminishingcharacteristics: • 1)NewissuescandiluteEPS. • 2)Newissuescandilutemarketprice. • 3)Newissuescanraisethecostofequitythroughflotationcosts. • 4)Newissuescandilutecontrolfromexistingshareholders. • 5)Newissuescanentailanongoingobligationofdividendpayments. Realizingthatexcesscapitalisbeingraised(orthatretainedearningsareinadequateto meetthetarget),theinvestorneedstohavefaithinfutureprojectsenoughtowarrant remainingashareholder. DECISIONMAKINGANDTHEMARGINALCOSTOFCAPITAL Considerafirmwiththefollowinglimitations: Table6-3 FUNDAMENTAL AMOUNT OptimalTarget 40/60Debt/Equity NextDividend $1.10 GrowthRate 10% SharesOutstanding 68.18(Million) NetIncome 300(Million) Retention 75% Thisexamplebringsseveralissuestotheforefront:itshowstherelationship betweenthevariablesaswellastheadvantageofknowingthetargetcapitalstructure.The firstquestiontoaskis“Howmuchfundingcanbedoneonthebasisofretainedearnings alone?”Hadtheretentionratenotbeengiven,itcouldbederivedbymultiplyingthe sharesoutstandingbythedividend,andthensubtractingthatfigurefromnetincome:(300 148 -(68.18x1.10)=225.Thisamountofretainedearningsisinadditiontotheamount alreadyretainedinstockholders’equity.Todeterminethetotalamountofadditional fundingthatcanbedone,wedividethe225bythetargetproportionofequity,0.6.Thus, thetotalamountofadditionalfundingis225/.06=375.Outofthisamount,(375-225) or150millionwouldbeinnewdebt. Ratherthanmeetstringentrequirementsoncapitalfunding,mostfirmscancreate strategicmovementtowardtheoptimaltargetonaperpetualbasis.Movingpastthetarget inanygivenyearwillrequireacountermovementintheoppositedirection.Sucha strategyallowsflexibilityincapitalfundingbecauseopportunitieswillarisethatmay requireinflowsgreaterthantheamountoffinancingthatiscondonedbytheoptimal target.Toobservethepotentialdilemmaofbeingconstrainedbythemarginalcostof capital,considerascenariowhere390andnot375millionwasrequired.Theoptionsare asfollows: • 1.Cutcosts,rationcapitalandunderfundsomeprojects.Thisscenariowillleadto eventualearningsdisastersandshouldbeavoided. • 2.Fundtheshortfallwithdebt.Thisstrategywillmovethefirmoffitstarget,andmay reducethesharepriceofthestock,whileraisingWACC. • 3.Cutdividendgrowthto7or8%.Thisstrategywillalsoleadtoadiminishedstock price.Bycuttingdividends,notonlyareexpectationslowered,butmoreearningsare retained,andthefirmstillmovesoffitstargetstructure. • 4.Issuestock.Only60%x15millionhastoberaisedinnewlyissuedstock.This relativelysmallamountof9millionwillkeepthecompanynearitstargetlevel. However,thecostofcapitalgoesupbecauseflotationcostshavebeenincurred,which willimplyanewlevelofoptimaltarget.Intermsofbankruptcycosts,theamountof lossrises,andnewtaxadvantageswillbeneededtobalanceit.Atthisjuncture,the student/investorshouldnoticethathadnetincomebeengreater,theextracapitalcould havebeenraisedthroughretainedearnings.Infact,moredebtwouldhavebeenraised 149 aswellbecausetheincreaseinnetincomewouldhaveloweredtheprobabilityof default. Whyissuestockinsteadofraisingdebt?Obviously,bothstrategieswillmovethecompany awayfromitstarget,butraisingasmallamountofstockislessdisruptive.Themarketwill rewardacompanyforissuinglargeamountsofdebtbecausegeneratingprofitswill consistentlymovethecompanybacktowarditstarget.Suchalargeissuehasstrategic value-acquisitions,largeprojects,evenleveragedbuyouts.However,alargestockissue canunderminethemarketpricethroughdilution.Mergersareuncertain,andlesspeople wanttoownastockunlesstheyaresureofapayoff.Thus,smallissuesofbalancingequity canbothpreservestockpriceandfundcapitalshortfalls.Similarly,smallissuesofdebt arelookedonas“movesofdesperation”,tokeepthecompanysolvent;thetaxadvantages willbelessthantheincreaseinbankruptcycosts. Inaperfectworld,firmswouldbeattheiroptimaltargetsandcontinuallyfundat thatlevel.However,thoseconstraintsarehardlyrealistic,andafirmmayneeda compensatoryamountofcapitalfromonesourcetomovebacktowardthetargetlevel.For example,iftheoptimaltargetcallsfora50-50mixofdebttoequityandthefirmisat60 percentdebt,itwillneedmuchmoreequitythandebttomakeuptheshortfall.This balancingactismuchmoredifficultthanitappears:notonlywillthefirmneedtolookat alternativesources,conditionsintheindustrymaychangethatwillabruptlychangethe optimalmixaswell.Fortunately,thereisananalysissystemthatcanbothdetectchanges inthecostofcapitalaswellasmovementtowardtheoptimum-thetheoryofeconomic profit. ECONOMICPROFIT,EVA ® ANDTHECAPITALDYNAMIC:UTILIZINGTHE OPPORTUNITYCOST Whileaccountingprofitsaremadebysubtractingcostsfromrevenues,economic profitsareconfigureddifferently.Anopportunitycost,theamountgivenupbypursuing onecourseofactionoveranother,issubtractedfromtheearningsderivedfromthe 150 originalcourseofaction.Thus,itisacomparisoncost.IfIamastockbrokerandIgiveup acareerinmedicine,Ineedtosubtractthecostandincomeofbeingadoctorfrommyown. Inthecaseofeconomicprofit,ifonefirmoutperformsthoseofsimilarrisk,itseconomic profitisgrowingbecausetheopportunitycost(whatisgivenup)issosmall.Therefore, economicprofittheoryoccupiesahypothetical“middleground”betweencost/benefits analysisandrisk/returnanalysisanduseselementsofeach. Thefirm,SternStewart,pioneeredtheconceptofeconomicprofitinapractical “handson”accountingenvironment,callingit“EVA ® ”oreconomicvalueadded 4 .Its uniqueapproachenabledmanymajorcorporationslikeGE,AT&TandCoca-Cola,to notonlybuildcapitalbuttocompensateemployeesbasedonimprovementinthe measurement.However,EVAcalculationscanbeverycomplex,requiringknowledgeof taxlawandaccountingskills,andthemeasurementnevercompetedwiththemore simplistic“P/E”asafavoriteofmutualfundsandindividualinvestors. ELEMENTSINANEVACALCULATION Usersneedtoderiveafigurecalled“NOPAT”,whichisanacronymfor“net operatingprofitaftertaxes.Initssimplestform,wetakeEBIT(earningsbeforeinterest andtaxes)anddeductjusttaxesfromit,leavinginterestuntouched.Inmostcompanies, severalotherdeductionswillbemadeatthistimeaswell,andderivingacoherentNOPAT willrequireknowledgeofitemizeddeductions,andcorporatetaxlaw;EVAbecomesa seriousmanagementtoolwhenusedproperly. ThesecondvariableinEVAcalculationsistheWACC(corporateversion).Without interestdeductionsinNOPAT,thetaxadvantagesofdebtbecomeimplicitintheWACC, andweendupcomparingthosefiguresaftermultiplyingWACCbytheamountofcapital. ThecomparativeopportunitycostistheproductofWACCandcapital,becausethatisthe figurethatwouldbemadeonalternativeinvestments.ThusWACCbecomessimilarto 4 EVAistheregisteredtrademarkofSternStewart,Inc. 151 ROC(returnoncapital)exceptthatitisappliedtocompaniesofsimilarrisk-throughthe costofequitycomponent. ThisnextexamplegivesastepbysteprenditionofasimpleEVAcalculation: Table6-4 FUNDAMENTAL AMOUNT OperatingIncome(EBIT) 145 TaxRate 31.03% InterestRateonDebt 8% CostofEquity 10% Capital 700 Debt 200 Stockholders'Equity 500 STEP1.Thisstepassumesnoothertaxdeductionsexceptinterest.DeriveNOPAT= (OperatingIncome)-[(TaxRate)(OperatingIncome)]=145-(0.3103)(145)=100 STEP2.DeriveWACC=[(InterestRatex(1-TaxRate)]x(ComponentPercentageof Debt)+[(CostofEquity)x(ComponentPercentageofEquity)]=[(.08)(0.6897)(0.2857)]+ [(0.1)(0.71249)]=0.87193or8.72%. STEP3.CalculateEVA.=NOPAT-[(WACC)x(Capital)]= 100-[(0.872)x(700)]=100-61.0355=38.9645 THECAPITALDYNAMIC Withoutbeingversedinmanageabledeductions,aninvestorcanformthissame resultusingjustthreevariables:netincome,thecostofequity,andthevalueof stockholders’equity.Thisadaptationallowstheinvestortoquicklyextractthesefigures fromfinancialstatementsandthendetermineonlyonecost-thecostofequity.The conceptofeconomicprofitremainsrelevant,andtheinvestorcanfocusonobtainingan accuratecostofequity;thesimplerframeworkallowslessroomforerror.Ineffect,the equationbecomes:NetIncome-[(CostofEquity)x(Stockholder’sEquity)].Thetax 152 advantagesofdeductedinterestexpensebecomeimplicitinnetincome,andtheinvestor willnothavetoitemizethedifferentinterestrateswithcorrespondingdebtmaturities. However,managerialcontrolislessapparentinthissimplerstructure,becausedeductions arenotitemizedasinNOPAT;thecapitaldynamicbecomesaninvestorfriendlyversionof economicprofit. Intheaboveexample,netincomeisderivedas((145)-(16)]x(0.6897)=88.97. Interestexpenseof16iscalculatedas8%of200.The0.6897isafigurefor(1-TaxRate). Nextwederiveaproductofthecostofequity,10%,andstockholders’equity,500: (0.1)(500)=50.Finallywesubtract50from88.97andderive38.97,whichisthesame figureasforEVA. THERELATIONSHIPBETWEENEVAANDTHECAPITALDYNAMIC EVAandthecapitaldynamicwillbethesamefigurewhentheinterestrateon currentdebtmatchestheactualinterestexpensethatispaidout.Theinvestorforcesthe bookvalueofdebttoequalitsmarketvaluewhichgivesthecapitaldynamiclessresilience asapredictorthanEVA.However,neitherdoestheinvestorneedtobeprivytothelatest negotiationoverinterestrates(theriskpremiumthatisattachedtotherisk-freerate),nor doesheorsheneedtobemiredinintricatecalculations.Thecapitaldynamicoffersthe investoralegitimatecomparisonbetweentwomainsourcesofrisk:netincomegrowthand thesizeandvolatilityofequity.WhileEVAmoreaccuratelyreflectsthecurrentcostof capitalandtheWACC,thecapitaldynamicbetterreflectsthecurrentinvestmentoutlook becauseincomegrowthisinherentinthecalculation.Analogously,thetypicalinvestoris muchlessconcernedwithaccuratecapitalbudgetingwhichwouldbeamajorconcernof thecorporateEVApractitioner. ECONOMICPROFITANDCORRELATION 153 DespitetheprodigiouseffortsofSternStewarttoeducateinvestors,EVAanalysisis notaspopularassomeothersystemssuchasthe“PEG”ratio(price-earningsgrowth). Nevertheless,itishighlycorrelatedwithstockperformance,andnotjustbecausethe earningscomponentaccountsforsomuchofthevariation.TheWACCcapturesthe interfacebetweencorporateriskandthestateoftheeconomy,whilethecapitalcomponent measuresproportionandimplicitlyencompassesmarkettobookvalue Sincedefaultprobabilityisnotanexplicitvariableinthefunction,anyeconomic profitformulawillnotoptimizecapitalstructuremathematically-thatis-ina deterministicfashion.However,thecorrelationvalueissogreatthatimprovementsinthe measurementcanbereadasmovementtowardthetargetmix.Inthecapitaldynamic, thereishighcorrelationbetweenallthreevariables;theeffectivenessofthefunctionis derivedbyobservingoneormorevariablesdecliningwhiletheotherrises,orbyexamining thegrowthratesofeachcomponent.Ineffect,eachvariableisaffectedbytheother:net incomeincreasesequitythroughretainedearnings;higherinterestratesthatareimplicitin thecostofequitymaydiminishnetincome;higherequitydiminishesbeta,whichdecreases thecostofequity. TodisplaythecorrelationbetweenEVAandstockprice,SternStewart trademarkedanotherconceptcalledMVA ® or“marketvalueadded.” 5 Essentially,MVA isthedifferencebetweenmarketvaluesofcapitalandtheirbookvalues.Whenallfuture EVAsarediscountedintothepresentatthecostofcapital,theresultis“MVA”.While suchextrapolationmaybedebatable(asitisinanyvaluationmodel),thesimilarities betweenstockpriceandEVAcanbeobserved.AsresearchedbytheteamofFamaand French,markettobookvalueishighlycorrelatedwithstockprice,thecostofequity,and especially,equityrisk.ThusitisnottoofarreachingtomakeaconnectionbetweenEVA andstockprice. 5 MVA ® istheregisteredtrademarkofSternStewart,Inc. 154 MANAGEMENTOFECONOMICPROFIT Implicitinthecapitaldynamicisthespecterofdebt;itisneveroutwardly acknowledged,andyetithasthegreatesteffectonallthreevariables.Thetaxdeductibility ofdebtbringspotentialincomebutthepossibilityofdefault,andthesefirmsmustmanage creditadeptly.Ontheotherhand,foracompanywhofundsonlywithequity,the prioritiesaresimilarbutthefocusisdifferent;thesefirmsmustperformmanyofthesame actionsasdebtladenfirms,butemphasizesalesandearningsgrowthandstability-two characteristicsthatareofteninconflict.Thesefirmsdependonequitymanagementand changesinoperatingrisktokeepahigheconomicprofit. 155 Table6-5 ALLEQUITY FIRMS NETINCOME COSTOF EQUITY STOCKHOLDERS' EQUITY 1Match marketing strategieswith demandtrendsin thebusinesscycle. . 1.Betais regulatedby operatingrisk Worktokeepsales asstableas possiblethrough diversificationand focusonfixed costs. 1.Focuson retainedearnings andpayspecial dividendsifneeded. 2.Taxstrategies 2.Buybackstock ontheopenmarket. 3.Diversifyamong products, customers, acquisitions 3.Managetheissue ofoutstanding sharesthrough smallincrements. 4.Worktolower operatingriskby focusingonfixed costs. 4.Fundcapital shortfallsinthe short-termcredit market. 5.Bewaryof expansionatthetop ofabusinesscycle 156 Table6-6 DEBT CARRYING FIRMS NETINCOME COSTOF EQUITY STOCKHOLDERS' EQUITY 1.Debtladen firmsmust concentrateonthe sameincome generating strategiesasall equityfirms. 1.Betais decreasedby increasingthe proportionof equityinthe capitalmix. 1.Likeallequity companies, leveragedfirms mustpaystrict attentionto retainedearnings andnewissues. 2.Anindebted firmmustbalance theamountof interest,andtax deductionwiththe increased variabilityofnet income. 2.Likeallequity companies,these debtladenfirms mustworkfora highmeanandlow standarddeviation insalesand income.Suchan effortwilllower betarisk. 2.Withless dependenceon equityfunding, thesecompaniescan payasteadier dividend. 3.Bothlong-term andshort-termdebt willworktolimit theamountof sharesoutstanding, butthefirmmust monitorthe probabilityof default. 4.Sharebuybacks canbeimplemented withdebt Leveragedfirmshavegreaterfinancialflexibilityandtheabilitytomaneuver throughlessprofitablephasesinthebusinesscycle.Theycangrantasteadydividendand “brace”thecompanythroughleverage,whenthefirmbecomesatakeovertarget. However,thepriceextractedisthegreaterriskofdefault,andtheurgencyofmeeting 157 earningstargets.Whileinvestorsdonotexpectall-equitycompaniestobe“bigearners” duringsomephasesofthebusinesscycle,leveragedcompanieshavebothacommitmentto shareholders-and-creditors.Theyareputintothepositionof“groworfail”quite frequentlybecauseoftheaddednecessityofmeetingbothinterestanddividendpayments. Thisdoublecommitmentcreatesthephenomenonofgrowingbeyondtheconfinesof “normal”growthoftheindustry;manyofthesecompaniesmustexpandintounrelated territorieslikefinance(GMCapital)orevenautorepair(Wal-Mart,K-Mart,Sears). COMPONENTMOVEMENTSOFTHECAPITALDYNAMIC Thecomponentmovementsofthecapitaldynamicfollowasequentialchainoflogic onwhichtheinvestorwillfocusi.e.,therateofchangeforearningsmustbegreaterthan forstockholders’equityorthecapitaldynamicwillfall.Whileeffortsmaybedirectedat keepingequitygrowthataminimum,theinvestormustrealizethatequitygrowthis actuallypermissiblewheninterestratesarehigh,andsothecontextofeachchangeismost significant.Onlywhenthemeasurementistakenasacohesivewhole,willtheperformance ofthecomponentpartsbecomprehensible.Asanexample,considerthecorrelation betweennetincomegrowthandthecostofequity:netincomeshouldactuallydecreasethe costofequitybysteppinguptheproportionofretainedearnings,andloweringbeta.And yet,bothnetincomeandthecostofequitywilloftenrisetogether.Someofthatpositive correlationhastodowithincreasingsalesjustasthemarketisrising,butalotofitis relatedtotheperformanceofthebusinesscycle;theFederalReservewillraiserateswhen themarketisacceleratingtocombatinflation.Thegreatereffectistoincreaseboththe levelandthesizeoftheriskpremium(differencebetweenthemarketandrisk-freerates). Theresultisalargercostofequity.Thus,eachcomponenthasseveralcountervailing effectsthatmayoccurwhenthemarketisinequilibrium.Theassumptionofcoherency, howeverisoftenfalseintheshort-term;anyrandomvariablecanputpressureona component(especiallythecostofequity)thatmakesitbehaveeccentrically. 158 Thefollowingscenarioshavethecapitaldynamiclaidoutascomponentdriven changes.Thearrowsindicatewhetherthecomponentisrisingorfalling.Inreductionist terms,thestockmarketissimplyanaggregationofthesethreecomponentchanges. • 1)Earningsareadequate,andthecompanybeginstopayoffitsdebt.Thescenariofor thecapitaldynamicwouldbe:(NetIncome↑ ↑↑ ↑)-[(CostofEquity↓ ↓↓ ↓)(Stockholders’ Equity↑ ↑↑ ↑)].Netincomeincreasesretainedearningswhichdecreasestheproportionof debttoequityandreducesbetainthecostofequity. • 2)Thefirmfundscapitalrequirementswithmoreleverage.Thescenariowouldbe: (NetIncome↓ ↓↓ ↓)-[(CostofEquity↑ ↑↑ ↑)(Stockholders’Equity↓ ↓↓ ↓).Inthiscase,net incomemaybedecreasedbyhigherinterestpayments,whilestockholders’equitymay havelessretainedearningsandbenefitfromlessofaneedtoissueshares.Thecostof equitywouldrisebecausethegreaterproportionofdebttoequityincreasesbeta. • 3)Marketforcesaretakingoverandinflationisbeingcurbedbyinterestratehikes. Thescenariomightbe:(NetIncome↑ ↑↑ ↑)-[(CostofEquity↑ ↑↑ ↑)(Stockholders’Equity↑ ↑↑ ↑)]. Thefirmmaybeatthetopofthemarket,andperformancewilldependontherateof accelerationofthethreefactors. • 4)Amarketdownturnarisesasinvestorsfleetohighqualitybonds.Suchascenario putsnegativeforceonallthreecomponents:(NetIncome↓ ↓↓ ↓)-[(CostofEquity↓ ↓↓ ↓) (Stockholders’Equity↓ ↓↓ ↓)]Thereislittledemandforanycompany’sequitywhilesales andearningsarefalling.Thecostofequityfallsbecausethelevelandsizeoftherisk premiumdeclines;themarketisdescendingandtheFedhasloweredrates. • 5)Theinitialphaseofarecovery:(NetIncome↑ ↑↑ ↑)-[(CostofEquity↓ ↓↓ ↓)(Stockholders’ Equity↓ ↓↓ ↓)]Thismaybethebesttimetoinvestbecauseprofitsarerecoveringbut pricingpressureonthestockissolow. • 6)Astockbuybackpurchasedwithleverage.Thatfortunatescenariolookslike this::(NetIncome↑ ↑↑ ↑)-[(CostofEquity↑ ↑↑ ↑)(Stockholders’Equity↓ ↓↓ ↓)].Thecostofequity risesbecauseleverageforcesbetatoascend.Netincomerisesbecausetheamountof 159 debtisnotsubstantialenoughtoraisethefinancialleverageratio(EBIT/(EBIT- Interest).Stockholders’equityistargetedfordecline. Themovementinstockpriceisalmostalwaysconcurrentwiththechangeinthecapital dynamic,althoughsometimesmarketinefficienciesariseanditfollowsit.Changesinthe componentsareamatterofdegreeandtheaccelerationofeachisassignificantasits absolutelevelanditsdirection. THECOMPARATIVECAPITALDYNAMIC Althoughsomeperformancecanbegaugedbyincreasesintheactualsizeofthe capitaldynamicorEVA,itisanabsolutemeasurementthatisafunctionofcorporatesize. Toplacecompaniesonamorecomparativebasis,itwillbenecessarytocreatearatio betweennetincomeandthetotalcostofequitythatequalizesqualitativegains.The comparativecapitaldynamicismerelyNetIncome/[(%CostofEquity)(Stockholders’ Equity)]andisaveryapplicablemeasurementforcompaniesinthesameindustry.Itcan beusedwithmorecautionforcompaniesindifferentsectorsaslongastheanalyst recognizesthatstockperformanceisrelatedtoeclipsingthegainswithinanindustry.For example,ifacomparativecapitaldynamicof2.5ishighforaspecificindustry,that valuationwillcarrymoreweightthanif2.5weretypical.Someindustrieswillnaturallydo lessequityfinancing,andtheyneedtobecomparedwithsimilarcompanies. 160 APPENDIX:THEEFFICIENCYOFEVAVERSUSROE ManyCFOswillconcentrateonthereturnonequity(ROE)measurementasa gaugeofcorporateperformance.Thebasicmeasurementistheproductofprofitmargin (Sales/NetIncome),assetturnover(Assets/Sales),andtheequitymultiplier(Assets/ Stockholders’Equity).Ineffect,theseindicatorsdonotmeasureprogresstowardan optimalcapitalstructureasefficientlyasEVA.WithbothROEanditsdistantcousin ROC(returnoncapital),debtcanbedirectlysubstitutedforprofitabilityandtheratios willstillrise.Observethefollowingtable: Table6-7 ProfitMargin AssetTurn. Equity Multiplier ROE Year1 0.07 1.2 1.8 0.1512or 15.12% Year2 0.06 1.1 2.4 0.1584or 15.84% Year2hasahigherROEwithmoredebtbutalowerprofitmarginandassetturnover. Technically,thefirmcangothroughadownturnbut“redeem”itselfwithmoreleverage. Bysubstitutingdebtforequityinthecorrectproportion,ROEwasabletorisedespitethe lowerprofitabilityratios.Ontheotherhand,hadthefirmusedEVAasitsmeasurementof progress,asmallerincreaseinnetincomemayhavereflectedthelowerprofitabilityratios andbalancedthedecreaseinequity.Moreleveragewouldhaveincreasedthepercentage costofequityaswell. (BacktoTableofContents) 161 APPENDIX:THEREALCOSTOFCAPITALANDWHATTHEINVESTOR NEEDSTOKNOW Thecostofcapitalisdefinedasthereturnabusinesscouldmakeifitchosean alternativeinvestmentwithsimilarrisk.Allofthecomponentcostsofcapital–debt, commonequityandpreferredstock-areconsideredopportunitycostsandaredetermined intherealmofmarketvalueswhenconfiguringtheweightedaveragecostofcapital (WACC).However,theinvestorisleftinthedarkwhendeterminingsomeofthesevalues. Whilethecostofcapitalmaybethemostimportantfundamentalindetermininga company’sdirection,theprocessofcalculatingoneofitscomponentcosts,thecostofdebt, isprivilegedinformation.Justasattorneysandclientshaveaprivaterelationship,sotoo docreditorsanddebtors.Yournext-doorneighborneednotknowyourmortgagerate, andneitherdocompetitorsinanybusinessneedtoknowothercompetitors’borrowing ratesuntilthoseratesbecomepublicknowledge. Therealcostofdebtisthenextinterestratethatacorporationcanincurafter analyzingitsriskcriteria(leverageratios)andinterfacingthisanalysiswiththemarket(a potentialcreditor):adefaultpremiumisderivedandisaddedtotherisk-freerate(an appropriateTreasuryyield).Whenmultipliedbythetaxratereciprocal(1–TaxRate) andthenappliedtothemarketvalueofafirm’sdebt,acostofdebtisformed.Ifthe corporationdecidestoincurdebtatthisrate,interestexpensewillbetalliedand transcribedtothenextfinancialstatement.However,nomatterhowhighorlowpast interestrateshavebeen,thecostofdebtisconfiguredatthenewrate.Ifinterestrateshave been10%andforsomecataclysmicreasonsgodownto5%inonemonth,thenewcostof debtisconfiguredat5%.Whathappenstopastinterestratesandtheircollectiveeffecton corporatedebt?Theeffectofongoinginterestpaymentsisreflectedinthemarketvalueof 162 thedebt.Justaswithanybond,themarketpriceofdebttendstorisewheninterestrates arefallingandvice-versa. Wheredoesthisleavetheinvestor?Ifthe“nominal”costofdebtisderivedby averagingeachdebtmaturitywitheachcorrespondinginterestrate,theamountofinterest paymentswillbetheonethatmakesbookandmarketvaluesequal.Amuchlessaccurate costofdebtwouldbeformedbyapplyinginterestexpensetothebookvalueofafirm’s long-termdebt.Theseaccountingversionsofthecostofdebt,despitetheirdefianceofthe definitionoftrue“opportunitycosts”,maybetterservetheinvestoringaugingrisk.Inthe followingexample,acompanydecidesnottoincurnewdebtatalowerinterestrate becauseitconsidersitselfoverleveraged.The“new”percentagecostofdebtismuchlower thantheinterestrateitactuallypaysbecausetherisk-freeratehasdecreased. Whiletheaverageinvestorcancopyamodelthatsimulatesthemarketvalueof debt,heorshehasnoaccesstothenegotiationsthatdeterminetherealinterestrate,nor wouldsuch“transparency”becosteffective.Asshowninthechapteronleverage,Chapter Two,thefinancialleverageratiohaspredictivevaluebasedonpastinterestexpensewhich isanexplicitcomponentofboththecapitaldynamic(EVA)andtherisk-adjustedWACC. DETERMININGAMARKETBASEDWACC:ANEXAMPLE TheXYZCompanyisnegotiatingwithitsunderwritertoconfigureanewinterest rateforaprospectivebondissue.AlthoughalowerFederalFundsratecollectively decreasedinterestratesbyapproximately2%,XYZisbeyonditsoptimaltarget proportionandanticipatesadownturn.Thus,theywilleschewlower-ratedebtandare unabletorefinancebecauseitsbondshaveano“call”provision.Themaincriteriathatthe underwritersuseisXYZ’sinterestcoverageratiowhichdeterminestheirdefaultspread. 163 XYZhasearningsbeforetaxes(EBT)of500(million)andinterestpaymentsof250,giving themacoverageratioof2.Onthefollowingchart,thatcoverageratioyieldsadefault spreadof2%overtheTreasuryyield. InterestCoverage>than InterestCoverage<than DefaultSpreadPercentage 0.8 1.249999 5% 1.25 1.499999 4.25% 1.5 1.749999 3.25% 1.75 1.999999 2.5% 2 2.499999 2% 2.5 2.999999 1.5% 3 4.249999 1.25% Toconfigureamarket-derivedWACC,weneedtoapplythemostup-to-dateinterestrate tothemarketvalueofafirm’sdebt.WewillconsiderXYZ’spositionovertwoyears.The firstyear(Year1),themarketvalueofthedebtisthesameasthebookvalue.Inthe secondyear(Year2),weapplyamodeltodeterminethemarketvalueofXYZ’sdebt. Besidesthismarketderiveddifferenceincapitalproportions,theonlyotherdifferences betweentheyearsaretherisk-freerateandconsequentnewcostofequity,andthenew interestrate. Themodelforestimatingmarketvalueisasfollows: MarketValueofDebt=InterestExpense[(1–(1/(1+NewRate) AverageDebtMaturity ))/New Rate] +BookValueofDebt/(1+NewRate) AverageDebtMaturity . 164 Thefollowingchartsdelineatethechangesinthecostofcapitalovertwosubsequentyears. ThechangeinthecostofequityisaconsequenceofthelowerTreasuryrate,whilethe marketvalueofchangesincapitalisderivedfromtheeffectofthenewinterestrateonthe priceofthefirm’sdebt. XYZYEAR1 VALUE Risk-FreeRate 6% CAPMCostofEquity 10% InterestRate 8% MarketValueofEquity(SharesxPrice) andPercentageinCapitalStructure 7200=0.6973=69.73% =7200/(7200+3125)=69.73% MarketValueofDebtandPercentagein CapitalStructure 3125=0.3027=30.27% =3125/(7200+3125)=30.27% BookValueofEquityandPercentagein CapitalStructure 4000=0.5615=56.15% =4000/(4000+3125)=56.15% BookValueofDebtandPercentagein CapitalStructure 3125=0.4385=43.85% =3125/(4000+3125)=43.85% TaxRate 0.3=30% CostofDebt(Book) (1-0.3)(0.08)=0.056=5.6% CostofDebt(Market) (1-0.3)(0.08)=0.056=5.6% AverageMaturityofDebt 10Years WACC(Book) (0.4385)(0.056)+(0.5615)(0.1)=8.07% WACC(Market) (0.3027)(0.056)+(0.6973)(0.1)=8.668% 165 Becauseinterestpaymentsareequaltotheproductoftheinterestrateandthebookvalue ofdebt,themarketvalueofdebtisequaltoitsbookvalue: 250[(1–(1/(1+0.08) 10 )/0.08]+[3125/(1+0.08) 10 ]=3125 Forthenextyear(Year2),wemusttakethenewinterestrateandapplyittowardthe marketvalueofdebt.SincethenewTreasury(risk-free)yieldis4%,thenewinterestrate (asdeterminedbythedefaultspreadof2%)is6%.Themarketvalueofdebtis: 250[(1–(1/(1+0.06) 10 )/0.06]+[3125/(1+0.06) 10 ]=3585.Theincreaseinmarketvalue occursbecauseXYZ’spriceisbidupbyinvestorswhowantthehighercouponrateonits debtratherthanthenewlowerrate.Noticealsothatinterestexpenseremainsat250 Millionbecausenonewdebthasbeenincurred–justachangeinmarketinterestrates. XYZYEAR2 VALUE Risk-FreeRate 4% CAPMCostofEquity 8% InterestRate 6% MarketValueofEquity(SharesxPrice) andPercentageinCapitalStructure 7200=0.6676=66.76% =7200/(7200+3585)=66.76% MarketValueofDebtandPercentagein CapitalStructure 3585=0.3324=33.24% =3585/(7200+3585)=33.24% BookValueofEquityandPercentagein CapitalStructure 4000=0.5615=56.15% =4000/(4000+3125)=56.15% BookValueofDebtandPercentagein 3125=0.4385=43.85% 166 CapitalStructure =3125/(4000+3125)=43.85% TaxRate 0.3=30% CostofDebt(Book) (1-0.3)(0.08)=0.056=5.6% CostofDebt(Market) (1-0.3)(0.06)=0.042=4.2% AverageMaturityofDebt 10Years WACC(Book) (0.4385)(0.056)+(0.5615)(0.08)=6.94% WACC(Market) (0.3324)(0.042)+(0.6676)(0.08)=6.74% Theaccountingversionofthecostofdebtdidnotrecognizethenewinterestrate andmaintaineditsrelationshipbetweeninterestandprinciple(0.08)(3125)=250.Onthe otherhand,themarketvalueversionofthecostofdebtcreatedamoreexaggeratedchange intheWACC.Academictraditionacknowledgesthemarketcostofdebtastheonlytrue opportunitycostandcreatesconsistencywiththecostofequity.However,the computationalrigorsofthemeasurementbringitawayfromthepurviewoftheindividual investorandintothecorporateboardroom.Indeed,thereadercanassessthecomplicated effectofchangingmarketvaluesofequityonthecostofcapitalandcanviewthetemporal characteroftheWACC.Sincetheinvestorneedstomakerapidcomparisonsbetween companies,itisrecommendedthatheorsheusetheriskadjustedversionoftheWACC; therelationshipbetweeninterestandprinciplewhencombinedwithearningsinformation, isjustasforwardlookingandperhapsmorestablethanthemorevolatilemarketversion oftheWACC. DISTORTIONANDACTIONABILITY 167 Forthefinancialprofessional,amarket-basedcostofcapitalcanonlybeminimized byadheringtotherisk-orientedcriteriathatlowerstheinterestrateprofferedbycreditors, andbymatchingthelifeofcorporateassetswiththematurityofliabilities.Themarket valueofequitymayfluctuatespasmodicallywhichdirectsstrategicattentiontotheonly controllablevariablesinthecostofequity–operatingriskandtheproportionofdebtto equitywhichbothaffectbeta. Fortheinvestor,amarket-basedcostofcapitalissimplynotactionableandheor shemustbecontentwithsubstitutingthejuxtapositionofinterestexpensewiththebook valueoflong-termdebt.However,interestexpenseandlong-termdebtarebothmajorrisk factorswhenmakinginvestmentdecisions.Infact,applyingtherealcostofdebttobook valueswillcreateadistortionbecausethecostofdebtwillnothavepriorinterestpayment obligationsasanexplicitvariable.Forexample,intheabovescenario,usingbookvalues, XYZ’scostofdebtwouldhavedecreasedto4.2%from5.6%,understatingtheongoing couponrateof8%.Whilenotwhollyintegratingopportunitycosts,thecapitaldynamic betterquantifiesriskfortheinvestorthandoesanEVAthatusesthemarket-basedcostof debt.Assumingthatpriorinterestexpensequantifiestheinterestratebeingpaidisnot theoreticallycorrectintermsofaweightedaveragecostofcapital.Suchanassumption, however,allowstheinvestortogaugeriskfromfinancialstatementswithouttheneedtobe privytocreditornegotiationsandmayevenbemoreprofitablethandiscerningamarket- basedWACCthatfluctuateswithvolatilestockprices. (BacktoTableofContents) 168 7 FUNDAMENTALSANDCAPITALSTRUCTURE “Fundamental”isawordwithdiversemeanings.Inareligiouscontext,itusually connotesaliteralinterpretationofasacredtext.Infinance,itreferstoconcrete performancemeasurements-sales,netincome,andassets.Inthecontextofcapital structure,weareawareoftheliteral,concreteaspectofbalancesheetitems,andthenwe turnthemupsidedown:onlyinthedomainofchangedothesefigureshaveanygreat significancefortheinvestor. Foroverfortyyears,therehasbeensomeequivocationinacademiaaboutthe properteachingoffundamentals.Sinceearningsforecastsarebasedonfundamentals,one wouldcertainlyseetheimportanceofteachingfundamentalsinanybusinesscurriculum. Indeedmostbusinessstudentsneedtotakeatleastacoupleofaccountingcoursesinorder tograduate.However,amongacademics,thereisarighteousadherencetowhatistermed the“semi-strong”formoftheefficientmarketshypothesis.Inbrief,thatdoctrine proclaimsthatnopubliclydisclosedinformationcancorrectlyforecastthefuturepathofa stock;themarketisso“efficient”atpricingasecuritythatthepricereactstoinformation beforeitisannounced.Thisdivergencebetweenthegreatamountoffinancialdata available,andtheinherentinabilitytoutilizeit,hasfrustratedmanyastudentand professionalalike. Tothecapitalstructuralist,theamountandriskofcapitalinflowsdeterminesthe fundamentalsthemselves,andthereforebecomesthebuildingblockofearningsforecasts. Whilemostexecutivesandanalystsareoptimisticaboutsalesandearningsincreases,few willconsidertheirsource.Thenebulousworldofcapitalproportionsandallocationentails riskandinterpretation,andnotjustthestraightforwardpronouncementofariseinsales. However,capitalstructureanalysiscanbeclassifiedasamoresophisticatedformof fundamentalanalysis-albeitonethatanticipatesfundamentalsratherthanreactingto 169 them.Whileatypicalfundamentalanalystmightlookatindustrymarginstoanticipate revenues,thestructuralistlooksformoreindirectnumbersthatwouldsignalan environmentconducivetoearningsincreases-asmallerinterestexpense,lowerfederal fundsratesoranincreaseinlong-termdebt,forexample.Similarly,afundamentalanalyst wouldbeconcernedwithastock’sintrinsicvalue,thepresentvalueofdiscountedfuture earningsascomparedtotheactualmarketvalue.Alternatively,acapitalstructuralistwill beasmuchconcernedwhetherthenumberofsharesoutstandingistheproperamount;the structuralistseesearningsinthedomainofimmediatechangesinrisk.Toquantifythis risk,heorshestepsoutoftherealmofrelationalfundamentalsandusesstatistical techniquestocomparethemean,skew,andstandarddeviationsofadistribution.Only whenriskandreturnareoptimizedthroughactionablechangesintheproportionofdebt toequity,willheorshebesatisfied. Iffundamentalsbythemselveslackpredictivevalue,theymaybethemost significanttoolintheeducationofafinancialexecutive.Whiletheefficientmarketmay anticipatefundamentalsanddiscounttheirvalue,itistheexecutiveswhowillimplement thechangesthataffectthoseratiosforbetterorworse;thesearethechangesthatportenda riseorfallinstockprice-anewmarketingstrategy,alowercostofdebt,orabroader customerbase.Thus,inthemostultra-efficientmarket,thestockwillriseconcurrently withsalesandearnings,andnotafterthefact.Althoughmomentuminthesefigurescan spurevenmorecapitalinfusioninlaterperiods,thattypeofariseisa“hitormiss” proposition,basedasmuchoncontinuedsectordominationasonimmediateearnings history.Whilegreaterearningscanhelpmakethecostofcapitallessexpensive,future prospectsaregearedtofundingprojectswithahighnetpresentvalue(seethechapter entitled,“CapitalStructure”)andnotonthegenerationofpastearnings.Inessence,an idealcombinationoflowcostcapitalsources,someofwhichmayberetainedearnings, mustbecoordinatedwiththecapitalbudgetingprocess. 170 Notwithstandingtheirvalueasapredictivetool,fundamentalsarelikea“yard stick”:theyenableexecutivestocompareandcontrastthevariablesthatneedchanging.It isinthisrevelatoryrolethatastructuralistusesfundamentals-togaugecapitalinflows anddeterminewhethertheyarebeingchanneledefficiently. DUPONTANALYSIS TheDuPontCompanydevelopedanextensive,albeitdeterministicsystemthat linkedleveragewithsales,profitsandeventheeffectivecorporatetaxrate.Asameasured result,theDuPontsystemsetthefoundationforfurtheranalysis;itcreatedthe underpinningsforexaminingtheprobabilitiesandchangesbehindeachinput.For example,themeasurementEBIT/Assetsmustexceedtheinterestrateifleverageistobe justified.However,noaccountingsystemcanprojectanoperatingincome,apriori,that willleadtothisresult.Ultimately,thedifferencebetweenfinanceandaccountingis establishedbythebalancebetweenchoosingtheactionwiththebestprobabilisticoutcome andthenmeasuringitsresult. Thedecompositionofafirm’sreturnonequity(ROE)isamultipurposeexercise withfourdistinctrewardsforthecapitalstructuralist. • 1.Whencomparingcompanies,ROEisanaccurate,generalindicatorofcurrent .performance. • 2.ThecomponentsofROEwillindicatewhatrisksneedtobeaddressed,andwhich elementshavepotentialforimprovement. • 3.ThecomponentpartsofROEpointtowardacompany’scapitalstructure,and delineateafirmsoperatingandfinancialleverages,howtheyarebalanced,andwhether theyneedtobechanged. 171 • 4.EachcomponentofROEisaffecteddifferentlybyeachphaseofthebusinesscycle andanyimbalanceinthecomponentswillaffectperformance. SomeinstructorswillrefertoROEas:earningsbeforeinterestandtaxes(EBIT)/ Stockholders’Equity,whichthisauthorbelievesnegatestheroleofdebtinreducing stockholders’equity.Sincecapitalstructureanalysisisfoundeduponthiscrucial difference,wesubmittotheDuPontAnalysissystemandendupwithROE=NetIncome/ Stockholders’Equity. Eachfundamentalisextractedfromafinancialstatementandthendividedby anotherfundamentaltoformacomponentpartofROE.Theresultingfivecomponent partsarethenmultipliedtogether. Table7-1 COMPONENT PARTSOFROE FUNDAMENTAL LOCATION RATIO DESCRIPTION 1)Earnings BeforeInterestand Taxes(EBIT) IncomeStatement 1)EBIT/Sales OperatingMargin 2)Sales IncomeStatement 2)Sales/Assets AssetTurnover 3)Assets BalanceSheet 4)Stockholders' Equity BalanceSheet 3)Assets/Equity EquityMultiplier 5)Earningsbefore Taxes IncomeStatement 4)EBT/EBIT CostofDebt (Financial Leverage) 6)NetIncome IncomeStatement 5)NetIncome/ EBT TaxRetention Theshortenedformofthisequationis:(NetIncome/Sales)x(Sales/Assets)x(Assets/ Equity)=ReturnonEquity,butthefullarrayofcomponentsbetterdescribesthedynamics oftheequation. Aprimeobservationthatwillbereadilyapparentwhenoneworkswiththeseratios istheirtendencytobeconsistentwithincertainboundaries-especiallyassetturnover, 172 profitmarginandassetstoequity-thethreecomponentsoftheshortenedformofROE. Thecharacteristicsofeachindustrydictatelimitationsonthecomponents,andthe companieswithintheindustrywillsharesimilarlevels.Forexample,Wal-Martmight struggletoobtaina4%profitmarginandnevermatchMicrosoft’s20%,butneitherwill MicrosoftmatchWal-Marts’largeassetturnoverorassettoequityratio.Thetwo companiescanobtainthesameROE,however,byeitheremphasizingwhattheydobest,or workingontheirrespectiveweaknesses.But-neverwillacompanywithatypicalasset turnoverof0.8becomeacompanywitharatioof“2”,ifitisinfact,thesamecompany. Thus,thesizeofthecomponentsofDuPontanalysisconveystherisksofworkingina particularindustryandcanbeincreasedordecreasedwithinthelimitationsproscribedby theindustry.Again,aswithleverage,theamountoffixedassetsandtechnologywithinthe industrywilldeterminetheboundariesofcapitalinflows.Fixedcostsinadepartment storechain,forexample,aretotallydifferentfromthoserequiredinasemi-conductor company,andeachwillhaverespectivelydifferentassetturnovers,profitmarginsand equitymultipliers. Todisplaythedifferenceinratiosthatanindustrycanimpose,wewillcontrasttwo companies:Barra,adeveloperofriskmanagementsoftwarewithhighresearchand developmentcosts,andWal-Mart,theworldrenownedconsumerretailer. Table7-2 1999 COMPANY NET INCOME SALES ASSETS EQUITY BARRA 23.4 187 169 101.3 WAL-MART 5575 165013 70349 27872 173 Table7-3 1999ROE COMPANY ProfitMargin Asset Turnover Asset/Equity ROE BARRA 23.4/187=0.1251 187/169=1.1 169/101.3=1.67 23.09% WAL-MART 5575/165013= 0.0337 165013/ 70349=2.35 70349/27872 =2.524 20.00% Bymerelyobservingthesethreeratios,ananalystcanlearnmuchaboutafirm.Inretail, theadvantagecomesfromassetturnoverandthesafeuseofcredit-asset/equity.In specialtysoftwaredevelopment,theadvantageismorequalitativeandstemsfromthe higherpricessuchproductsmaygarner.Eachcompanyhasdifferentinherentrisks. WhileBarrahasnocreditrisk,theyareexposedtocompetitionthatcanunder-pricethem. Ontheotherhand,Wal-Marthasalargecreditriskbutcushionsitwithalargeasset turnover;morecash-flowdiminishestheprobabilityofdefault.Thereturnonequityis similarbutthewayeachcompanyarrivedthereistotallydifferent. Thestudent/investorwillfinddifferentpatternsamongtheindustriesthatbalance eachotherinthecontextofROE.Forexample-highprofitcompaniesusuallyhave pricingpowerbutlowerassetturnoveranddiminisheduseofcredit.Thereason?Higher profitmarginssometimesentailmorevariationinsalesand/ortheuseofretainedearnings toraisecapital.Withoutthenecessary”trackrecord”ofsteadyearningsandloanhistory, creditorsarelesslikelytoloanatfavorablerates.Moreover,pricingpowerimpliesthe creationofunique“niche”productsthatcannotbeeasilyduplicated.Hence,higherprofit margincompaniestendtobeincapableof“churningout”productatamomentsnotice becauseoftechnicalormarketinglimitations,andwillconsequentlyhavealowerasset turnover.Ontheotherhand,highassettoequitytypecompanieswillalmostalwayshave higherassetturnoversbecausealargeturnoverdiminishesriskandprovidegreater securityfortheircredit.Theyareoftenfromolder,moreestablishedindustriesthatchurn 174 outsimplebutverynecessaryproducts-whethertheyaretiresorbakerybread.However, theirownproductivityinterfereswiththepricethattheycancharge,andthesecompanies mayhaveadifficulttimeincreasingprofitmargin.Infact,inmanyindustries,increasing theprofitmarginbyeventwopercentagepointsonaconsistentbasiswouldbeconsidered anamountthatwoulddominatetheindustry,andwarrantasoaringstockprice.Thus, thesefirmscanconcentratewholeheartedlyonthissinglelack.Infact,ifthemarketwill pricestocks“efficiently”,itwilldosobycreatingimplicitbenchmarksintheseareas;once afirmimprovesonadeficiency,itisrewardedbyahigherstockprice.InWal-Mart’s case,evenaonepercentagepointgaininprofitmarginwouldhaveearnedthecompany approximately1.6Billion.Analogously,Barraneededtobothlimititsamountofequity inproportiontoitsassetsandincreasesales.Sincestockpriceishighlycorrelatedwith improvementsinROE,bothofthesecompaniescouldhaveshiftedcomponentpartstofind thefactorsthatproducedanoptimum.However,theneedforbalanceisevenmore imperativebecausedecreasingonevariablehasatendencytoincreasetheother;without “synergy”,therecklessimprovementinonevariabletothedetrimentofanothercanlead to“shocks”inthesystemthatneedtobereconciledinfutureyears. Toestablishabenchmarkcomponentratio,themosthelpfultoolistoresearchthe industryaveragesoverafiveyearperiod.Thethreelargestcompetitorswillusuallysupply enoughdatatoestablishavalidaverage,andsomewebsiteswillhavedonethework already.Indeferencetotheefficientmarketshypothesis,itisdoubtfulwhetherinvesting onthebasisofimprovedratioscanbeatthemarketforanylengthoftimebecausegainsin sharepricehappenconcurrentlywiththeratioimprovement.Theonebigadvantagethat companyinsidersenjoyisthattheyknowwhichratiosareimprovingandbyhowmuch, becausetheystrategicallysetouttochangethem.Thus,theobservedconfidencethat managementhasinitsownstockisoneofthefewwaysthattheaverageinvestorcan indirectlyprofitfromfundamentals.Buyingopportunitiescarrymoreweightthansales becausemanyinsiderswillsellstockfortaxpurposes.Forthecapitalstructureanalyst, 175 theprimarymissionistocoordinatea“map”oftheterritory-decidingwhetherthe investmentenvironmentisconducivetoearningsandobservinghowthesectorismeshing withthebusinesscycle.Inthisregard,theanalystwantstofindasituationwhereearnings willacceleratemuchfasterthanthecostofcapital;some“insiders”aresimplyfervent optimistsandweneedtocorroboratetheirenthusiasm. COMPARINGROECOMPONENTS ThefollowingchartsdisplaythethreebasicROEcomponentsfortendifferent companies.Mostofthesefirmsareindifferentindustrieswhicharereflectedinboththe sizeandstabilityofthecomponents.AnotherpotentialadvantageofROEanalysisisthata firmcanuseitsmoststableROEcomponentasaplanningtool.Ifafirmhaslargesalesin termsofassets,butlowprofitmargins,afirmcanplaninventorylevelsandsalesdistricts aroundassetturnover.Similarly,iftheassettoequityratioisstable,afirmmaywantto performnearitstargetcapitalstructureatalltimesandnottaketheriskofbrief directionalmovementsawayfromit.Moreover,keepingonecomponentverystableallows theothercomponentstovary(inathreecomponentROE).Amongthoseothertwo components,onewillrepresentthe“weakestlink”toahealthyROE;thechallengetoany firmistoimprovethisweaklinkwithoutdamagingtheperformanceofothercomponents. Forexample,ifafirmdisposesofassets(net),itprobablyisnotgrowing.However,sucha moveintheshort-runmightartificiallypumpupassetturnover.Theobjectiveistoexceed theindustrystandardforthatcomponent,i.e.,aonepercentriseisprofitmarginmaybe verysubstantial.Finally,therewillbeonecomponentthatgivesthefirmthemajorityofits ROEstrength.Thiscomponentwillrepresentitscompetitiveadvantageandmayormay notbethesameasthe“stability”component.Forexample,acompanylikeCSX(see charts)hasahugeadvantageinitsassettoequityratio;apparently,cash-flowisstable enoughtofunditsheavyneedforcapitalwithdebt.ThatisCSX’scompetitiveadvantage. Theymaylowerthisratioandstillremaincompetitive,buttheyneedtobalanceany changewithreinforcementfromothercomponents.TheexhibitofROEcomponents 176 displaysthecomponentnumberoverfiveyearsandfollowsbyrankingeachcompanyfor whichcomponentisstable,weakand/orcompetitive. Table7-4 DILLARDS (DDS) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 3.9 1.7 1.9 1.1 0.8 Asset Turnover 1.19 0.95 1.1 1.19 1.15 Asset/Equity 1.99 2.88 2.8 2.74 2.65 Stable Component Asset Turnover Weak Component Profit Margin Competitive Component Asset/Equity 177 Table7-5 ST.JUDE MED(STJ) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 5.5 12.7 2.2 11 12.8 Asset Turnover 0.68 0.73 0.72 0.77 0.83 Asset/Equity 1.48 1.72 1.46 1.63 1.38 Stable Component Asset Turnover Weak Component Asset Turnover Competitive Component Profit Margin Table7-6 USTOBACCO (UST) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 31.3 32 31 28.6 29.4 Asset Turnover 1.7 1.56 1.49 0.94 0.83 Asset/Equity 1.89 1.95 0.37 6.07 3.47 Stable Component Profit Margin Weak Component Asset Turnover Competitive Component Profit Margin 178 Table7-7 ECOLAB (ECL) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 8.2 8.2 8.5 9.2 8 Asset Turnover 1.16 1.25 1.31 1.32 0.93 Asset/Equity 2.57 2.14 1.7 2.26 2.87 Stable Component Profit Margin Weak Component Asset Turnover Competitive Component Asset/ Equity Table7-8 INT. RECTIFIER (IRF) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 0 3 3.7 9.7 9 Asset Turnover 0.71 0.75 0.77 0.73 0.56 Asset/Equity 1.62 1.93 1.77 2.59 2.07 Stable Component Asset Turnover Weak Component Asset Turnover Competitive Component Asset/Equity 179 Table7-9 MOLEX (MOLX) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 10.8 11.2 10.4 10 8.6 Asset Turnover 0.94 0.99 0.9 0.99 1.07 Asset/Equity 1.32 1.3 1.27 1.32 1.25 Stable Component Asset/ Equity Weak Component Asset Turnover Competitive Component Profit Margin Table7-10 NATURE'S SUN.(NATR) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 7.2 7.9 6.2 5.4 5.2 Asset Turnover 2.93 2.85 2.7 2.67 2.44 Asset/Equity 1.43 1.4 1.38 1.35 1.37 Stable Component Asset/Equity Weak Component Asset/Equity Competitive Component Asset Turnover 180 Table7-11 CSX(CSX) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 7.5 5.4 0.5 6.9 3.6 Asset Turnover 0.53 0.48 0.52 0.4 0.39 Asset/Equity 3.46 4.49 3.6 3.41 2.94 Stable Component Asset Turnover Weak Component Asset Turnover Competitive Component Asset/ Equity Table7-12 ARCH.DAN. MID.(ADM) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 2.7 2.5 2 2.3 1.9 Asset Turnover 1.22 1.17 1.02 0.96 1.4 Asset/Equity 1.88 2.13 2.25 2.36 2.26 Stable Component Asset/Equity Weak Component Profit Margin Competitive Component Asset/Equity 181 Table7-13 ARGOSY GAMING (AGY) YEAR/ COMPONENT 1997 1998 1999 2000 2001 ProfitMargin 0 1.3 6.1 6.7 8.4 Asset Turnover 0.61 0.9 1.05 1.29 0.598 Asset/Equity 17.13 13.81 9.74 5.16 7.37 Stable Component NONE Weak Component Asset Turnover Competitive Component Asset/Equity *TheAsset/Equityratioisassetsdividedbycommonstockholders’equityandnotthe fullstockholders’equitythatmayincludepreferredstock. MODIFYINGANDENHANCINGDUPONTANALYSIS Thecapitalstructuralistneedsadetailedviewofafirm’sdebtstructure;heorshe willfurtherdecomposetheassetstoequityratio(commonlycalledtheequitymultiplier) andalsousethefullarrayofcomponents(five)inthemodel.Theasset/equityratiois decomposedasfollows:ASSET/EQUITY=(LTD/Equity)x(Asset/Short-termdebt)x (Short-termdebt/LTD).“LTD”isanacronymforlong-termdebt,while“short-term debt”issynonymouswithcurrentliabilities(forthesakeofthisanalysis). Often,fundingwithmoreshort-termdebtandlesslong-termdebtwillaccomplish oneofthreeobjectives:1)itmayreduceoverallinterestexpenseasshort-termratesare lowerthanlong-termratesinanormalmarket;2)itmayindicatemoretradecreditwhich isoftentreatedasaninterestfreeloan,andadditionally,morebusinessactivity;3)ifused judiciously,itcanreduceexposuretointerestraterisk.Thislastitemneedsqualification: 182 whilefundinglong-termfixedassetswithshort-termdebtcanleadtoinstability,funding withshort-termcreditwheninterestratesareanticipatedtodropmayreduceexposureto long-termcommitments-aslongassuchamoveistemporary. Althoughanincreaseincurrentliabilitiesmayaffectshort-termsolvencyandthe probabilityofdefault,itisbothasourceofinternalfinancingandfreecash-flow,andisan importanttoolinmanagingcapitalstructure.Byfollowingthechangesineachcomponent, thestudent/investorcanobservethebehaviorofthecompleteassettoequityratioina profitableyear:(LTD↓ ↓↓ ↓/Equity↑ ↑↑ ↑)x(Asset↑ ↑↑ ↑/Short-termdebt↑ ↑↑ ↑)x(Short-termdebt↑ ↑↑ ↑/ LTD↓ ↓↓ ↓)=Asset/Equity. Table7-14 COMPONENT DIRECTIONOF CHANGE REASON LTD DOWN Lessinterestexpense,risk ASSETS UP Growthinsalesand earnings EQUITY UP Moreretainedearnings SHORT-TERMDEBT UP Morebusinessactivity Ifthereaderhasbeenfollowingthetext,heorshewillrealizethatincertainsituationsthe oppositechangescanaffecttheROEmeasurementinapositivemanner.Thisissimplyan exampleandnotmeanttoconveyonesetpositivepattern:whentaxbenefitssoarwhilethe probabilityofdefaultdoesnot,theresultwillbeanincreaseinstockprice.Ourobjective istocitethemechanismsthatmakethathappen.Sincemanycompanieswillfurtherraise long-termdebtinveryprofitableyears,itisthesizeofthelong-termdebttoequityratio thatissignificant,andnotwhetherlong-termdebtisactuallydecreased. Ofthefourremainingratiosinthefullmodel,twoareimperativetocapital structureanalysis:theratio,EBT/EBITdistinguishesthecostofdebt,whileNetIncome/ EBTdeterminestheeffectoftaxesonshareholderearnings.Thestudent/investorshould 183 beawarethatearningsbeforeinterestandtaxes(EBIT)minusinterestexpensewillequal earningsbeforetaxes(EBT),andthatinterestisataxdeductibleexpense.Thus,Net Income/EBTincreasesasinterestincreases,butthecostofdebtisareciprocalthat actuallydecreasesasinterestrises.Toputthesefiguresoncommonground,laterinthe chapterwewillreversethecostofdebtratioandturnitintoEBIT/EBT;thisfigurewill increaseasinterestexpenseincreases.Thismore“understandable”formistermedthe “financialleverageratio”oralternately,thedegreeoffinancialleverage”andisthesame ratiowehaveusedinpreviouschapters.ItisEBIT/EBIT-InterestExpense. Thetworemainingratios,operatingmargin(EBIT/Sales)andassetturnover (Sales/Assets),arenolessimportantthantheaforementionedcomponents,buttheyare lessdirectlymanageable.Theyaremoresensitivetobusinesscyclefluctuations,thelevelof technology,ageoftheindustry,andcompetitivepressure.Indeed,itwouldnotbe “steppingoutofbounds”tosaythatthegoalofROEstrategyistooptimizetherisk/return characteristicsofthesetworatiosbycontrollingtheotherthree.Forexample,financingan acquisitionwithleveragemayraisetheassettoequityratio,andsincethetargetcompany hasalargeassetturnover,operatingriskwouldbediminished.Inthatcase,thedeciding pointwouldbethefearofpotentiallydecreasingtheprofitmargin,andthatissuewould havetobeaddressed.Thus,eachdecisionmadeaboutanyoneratiowillaffectallofthe othersandfinancialmanagementmustpursueaproperbalance,understandingthe repercussionsofeachaction. Theentirefivecomponentmodelisconfiguredasfollows: (EBIT/Sales)x(Sales/Assets)x(EBT/EBIT)x(NetIncome/EBT)x(Assets/ Stockholders’Equity) Thereadercannoticethatthefirstfourcomponentsreducetoareturnonassets(ROA) becauseallbutthetwofundamentals,netincomeandassets,canceleachotherout.When wemultiply(NetIncome/Assets)by(Assets/Equity),wegetthefullROEeffect,(Net Income/Equity). 184 Tocorporatemanagement,athoroughsensitivityanalysisinvolvingallthe componentsisimperative.Tothecapitalstructuralist,itisthechangesineachratiothat aresignificant.WhilefinancialexecutivesoftenattempttoextractthelargestROE possible,theinvestormustgaugetheriskofinteractivechangetowardanoptimalcapital structure.Inahandfulofcases,morelong-termdebtisneeded,andthetaxbenefitswill enhancegreatersalesandearnings.However,inthevastmajorityofcapitaltransitions, profitswillbeextractedfromlessuseoflong-termdebtincomparisonwithequity.While debtisbeingpaidofffromenhancedprofits,interestexpenseisbettercoveredbyoperating incomeandtheprobabilityofdefaultdecreases.Moreearningsareretained,andthe proportionofequityincreases.Whentheeconomy“heats”upandthecostofretaining earningsbecomesexpensive,thecyclewillbeginagain;thefirmwillloaduponlowinterest loans,takeonnewprojects,andsometimesmovepastthepointwheretheproportionof debtisoptimal. Inreality,theremaybeaninterimwhenoperatingmarginandassetturnoverdo notincreaseenoughtoparedowndebt,andacorporationwillmakeseveraladjustmentsto theotherratiostokeepROEfromdiminishing;thisiswheretalentinfinancial managementisrealized.Althoughnew,profitableprojectsmustflowinto”thepipeline”to keepthecompanycompetitive,thecapitalstructureorientedratioscantemporarilymake upthedifference.Eventually,ifsalesarenotgeneratedfromnewprojects,thecompany facesassetcutbacksand/ordownsizing.However,enhancinganROEwithanincreased equitymultiplierevenwhileoperatingmarginsandassetturnoveraredepleted,isarisky strategywithalargepayoffiftheoutlookisfavorable. Alessriskystrategyistousethethreeleverageratios(EBT/EBIT,NetIncome/ EBT,andAsset/Equitytokeepthetwoprofitratios(EBIT/Sales,Sales/Assets),ashigh andstableaspossible.Infact,the“ideal”ROEhasbeenobservedmanytimeswhena dominantcompanyinadominantsectorcomestofruition. 185 Table7-15 THEIDEALROE COMPONENT RATIO DIRECTIONOF CHANGE Operatingmargin EBIT/Sales UP AssetTurnover Sales/Assets UP CostofDebt EBT/EBIT UP(Indicatesdecreased costs) TaxRetention NetIncome/EBT UP(Indicatesreduced taxes) LTD/Equity LTD/Equity DOWN(Moreretained earnings) Asset/Short-termdebt Asset/Short-termdebt NEUTRAL(Increase both) Short-termdebt/LTD Short-termdebt/LTD UP(Morebusiness activity) Athighlevelsofearnings,thereisapenaltyforretainingearningsandnot distributingthemasdividends,andalsoapenaltyforissuingtoomuchequityintheform ofnewlyissuedsharesofstock.OurDuPontmodelisnotsophisticatedenoughtodeal withtheseconcerns-yet.Sofar,wehavenarrowedourperspectivetosevenratios,three ofwhich(LTD/Equity,Assets/Short-termdebt,Short-termdebt/LTD),area decomposedversionoftheequitymultiplier,(Assets/Equity).Thetaxretentionratio,Net Income/EBT,issubjecttoinfluencesfromsalesandoperatingincomeforwhichthereis littletacticalcontrol.Byconcentratingonthecostofdebtratio,EBT/EBIT,weimplicitly affecttaxretention;interestexpenseisacommonfactortobothearningsbeforetaxesand thecostofdebt. Whiletheabsolutesizeoftheratiosisimportanttotheinvestor,moreconcernis placedwiththeinteractivedynamicsbetweenthem-whichcomponentsareincreasingor decreasingandhowwilltheyaffecttheothercomponents.Asanexample,considera scenarioinwhichafirmunderperformstheindustryinassetturnover.Woulditbe expedienttolowerassetstoequitytoamorepermanentlytolerablelevel?Thatquestion 186 wouldbeselfevidentiftheequitymultiplierconsistentlyexceedstheindustrystandard, butwouldbeadifficultchoiceotherwise.Inessence,WallStreethasneverplayedagame ofabsoluteswhenitcomestoeitherfundamentalsornear-termstockpriceincreases;an undercovered,smallcapfirmcanoutpacealarge-capDowcomponent.Thechallengeof goingfrom0.30pershareto0.70pershareinEPSisoftenrewardedmorethanahigh performingstockthatgoesfromanEPSofthreetofourdollarspershare.Therewards aremorederivedfrombeatinghigh-riskoddsanddoingtheunexpected,thanfrom consistentperformance-intheshort-run.Anditisthattimeframe-the“near-term”or short-run-inwhichreturnonequitychangesoccur.Obviously,themicromanagementof theseratioscanleadfinancialexecutivesintoatrap,andwithoutastrategicoverlayof projectanalysisandlowcostfunding,financialengineeringofadesirableoutcomewould beimpossible.Therefore,ROEcomponentsrepresentanobjective,butnotameanstoan end.Theyrepresentperformancegoalsbutdonotclarifythedistinctionbetweenriskand returnenoughtoactastoolswithwhichtomanageafirm. WhattheROEcomponentslackastoolsofmanagerialfinesse,theymorethanmake upassignalsforanalystsandinvestors.Wemust,however,modifythemtoencompass capitalstructurechanges.Inparticular,thecostofdebt(EBT/EBIT)goesupwhenthe actualcostofdebt,asmeasuredbyinterestexpense,goesdown.Secondly,increasesin equitydonotautomaticallytranslateintoacompanymovingtowardamoreoptimal capitalstructure;retainedearningscanbeexcessiveandnewstockcanbeissuedatan inopportunetime.Thirdly,theassettoshort-termdebtratioisambiguous,andabetter ratiocanunifythethreeelementsofAsset/Equityintoanunderstandablewhole. THERETURNONCAPITALRATIO Throughasimplemodification,thecapitalstructuralistcreatesa“returnon capital”orROCratio.Capitalisdefinedasthesumoflong-termdebtandstockholders’ equity.Byreplacingthefundamental,“Equity”,with“capital”,thedynamicswillbe mathematicallychanged.Intheratio,long-termdebttocapital(LTD/Capital),adding 187 long-termdebtincreasesboththenumeratoranddenominator,creatingalessvolatileand morestatisticallysignificantmeasurement.Ontheotherhand,theratio,AssetstoCapital, becomesmorevolatileandsensitivetochangebecausewearemakingitsmallerthanAsset /Equityandnotchangingbothnumeratoranddenominatorintheexpression.Any increaseinassetstocapitalimpliesarelativeincreaseinshort-termdebtbydefault.Tosee howthisfitsintothenewexpression,observethefollowing: ASSETS/CAPITAL=(LTD/Capital)x(Assets/Short-termdebt)x(Short-termdebt/ LTD).AssetstoCapitalbecomesmoreofariskmanagementtoolbecauseitcanbeapplied toseveralleveragesituationswhereshorttermdebtincreasesinproportiontolong-term debt.Moreover,long-termdebttocapitalwillshowsignificantchanges,whereasthe formerlong-termdebttoequityratiomightnot. Thecostofdebtratio,EBT/EBIT,isinvertedtoEBIT/(EBIT-InterestExpense), whichwillshowincreasesofinterestexpenseinrelationtooperatingincome.Thisisthe muchreferenced“financialleverageratio”whichisverymuchliketheinterestcoverage ratioinbondratings.Increasesinthisratiowillusuallysignifythatthedefaultprobability forthefirmhasrisen.However,sincedefaultprobabilitieshavecountervailing components,theremaybesomeotherelementthatbufferstheriskofanincrease. Withthesethreecomponents,LTD/Capital,EBIT/EBTandAsset/Capital,we havetheelementsofriskreduction.Reducingallthreecomponentswilllowerrisk,butnot necessarilyincreasereturn.Itiswhentheseelementsinterfacewiththeprofitcomponents, operatingmarginandassetturnover,andbecomeintegratedwiththedynamicsofthe greatereconomy,thatriskisreducedinthedomainoflargerreturns.Infact,theratio asset/capitalisaveryneutralelementintermsofabsoluteriskbecauseitcanbeusedasa capitalsubstitutewhenthecostofdebtisveryhigh.Thus,itisdependentonotherfactors forinterpretation-thebusinesscycle,theotherleveragefactors,theamountofabsolute debt.However,increasingtheratio,Assets/Capital,willalwayscontributetothereturn 188 oncapitalonasuperficiallevel;aslongasraisingitdoesnottacitlycontributetoa depletedoperatingmarginorassetturnover,itwillincreasebothriskandreturn. Infullform,thecompletedreturnoncapitalequationis: (EBIT/Sales)x((Sales/Assets)x(EBT/EBIT)x(NetIncome/EBT)x(LTD/Capital)x (Assets/Short-termdebt)x(Short-termdebt/LTD).Fromaninvestor’sperspective,itis prudenttolookattheshortenedformbecausemanycomparisonsneedtobedonetofind the“right“company.Webringthisbacktobasicswith: (NetIncome/Sales)x(Sales/Assets)x(Assets/Capital)=ROC(ReturnonCapital). Backinthe1990s,adherencetotheprincipleofincreasingallthreecomponents yieldedasituationweaffectionatelycalled,a“trifecta”.Thenormalsuccessratefora stockincreasewasabout73%accordingtoourdata,typicalfora“bull”market.When allthreeofthesecomponentswereraisedsimultaneouslyinanygivenyear,thesuccessrate hoveredataround90%.Althoughstockswereinflatedandmomentumwasrampantin thatdecade,an“afterthefact”increaseofthatmagnitudewasphenomenal. (BacktoTableofContents) 189 8 CAPITALSTRUCTUREANDTHEBUSINESSCYCLE Thegeniusofmanagementistoplaceitself“intherightplaceattherighttime”. Whileatalentedmanagercantakethehelminarecessionandslowlyguideacompanyinto recovery,evenamediocremanagercanprosperifheorsheisprescientenoughto anticipatethedynamicsofanindustrywhichis“suddenly”favoredbytheeconomy.That foresight,whencoupledwithevenaslipshodleadershipability,maybeenoughtogarner hugebonusesattheendoftheyear.Despitetheprotestationsof“moretalented” underlings,themanagerwhocanoutforeseethecompetition(inthiscase-otherpotential managers)willbetheonewhoisrewarded. Apersonalstory-MynephewRandywasawhizatcomputersandhelpeddesign manyInternetsitesearlyinthegame.Heday-tradedstocksatatimewhensuchapassion wasunique.However,hefailedto“foresee”shiftsintheeconomywhichsuccessful managerswillanticipate.Whenstockstumbledattheendof2000,hegothit.Hisfamous line?“Butthefundamentalsweregreatonthatcompany!”Corporatefundamentalsare alwayssecondarytotheactionsoftheentireeconomywhichthemarketanticipates.Huge earningscanbemade,butiftheyaregarneredinaninflationaryperiod,forexample,the marketwilltotallydiscountthem.Therefore,thebusinesscycleistheultimatesourceofall stockgainsand“structures”capitalstructure,sotospeak. COMMONELEMENTSOFBUSINESSCYCLES Attimes,theeconomyfollowsadiscerniblepattern,andalthougheachmarketis differentfromthelast,distinctsimilaritiesemerge.Whileeconomistscannotpredictthe peaksandtroughsofthispatternwithprecision,theyhaveisolatedcommoncharacteristics ofmostmarkets: 190 • 1)Fourphasesappearvalid(someeconomistswouldargueformoreorless)and smallersub-cyclesaresometimesprevalent.Theyare:1.Recession/contraction2. Recovery3.Expansion4.Plateau • 2)AsimilarpatternofinterestratechangesoccursoverthecycleastheFederal Reserverespondstoboththeneedforinvestmentandthepotentialblightofinflation. • 3)Companiesprosperatdifferenttimesovertheentirecycle,separatingthemselves intoindustrialsectorsthathavesimilarcash-flowpatternsandborrowinghabits. Thedangerofbusinesscycleanalysiscomesfromexpectations;wenaturallyassume thatpastpatternswillbeduplicatedandbegintoextrapolateintothefuture.However, eachmarketisdifferentinsomerisk-takingaspect-legislatively,tax-wise,intheamounts ofinflationortheratesofforeignexchange.Phasesrarelymakesmoothtransitionsfrom onetothenext,andleadingindicatorssuchasM2orthestockmarketcanevenbelagging insomecycles.Sometimesentiresectorswillbeleftoutofarecoveryandexpansion becauseconditionsinthatindustryhavechangedsincethelastcycle.Afteraboutthreeto fiveyearsintoarecovery/expansion,prospectsmayagainseemdimandtheFederal Reservemayengineerwhatistermeda“softlanding”or“growthrecession”-aperiodof lowGDPgrowthwhichallowsa“bull”markettocontinueforafewmoreyears,contrary tothebestpundits’predictions.Fortunately,thestudent/investordoesnotneedtomake accuratepredictionsinordertomakemoney.Heorshecanconcentrateonthecapital structuredecisionsofthebestsectorsandseekoutpatternsofleveragethatwillreplicate throughoutthecycle. THEYIELDCURVEANDINTERESTRATEBEHAVIOR Studentstendtothinkoftheyieldtomaturitycurveasasupplyanddemandcurve whichisnotacorrectassumption.Infact,itismuchmoreagaugeofinvestorexpectations aboutthedirectionoffutureratesthanademandgraphforloanablefunds.Thereasonfor thisconfusionisthatthedemand(andprice)foradebtissuemovesintheopposite directionfromchangesintheinterestrate.Alogicalquestionmightbe:“Whywould 191 anyonepaymoreforaninvestmentwithalowerrate?”Theansweris:“Theydon’t.”.As interestratesareraisedbytheFederalReserve,thepriceofexistingissueswithlowerrates goesdown.Investorscanmakemoreonanissuewithhigherratesandselloffanybonds thathavetheolderandlowerrates.Thus,wheninterestratesspikeabovethedebtissue withlowerrates,thereislessdemandfortheoldissueandthepricegoesdown.Atalower price,thelowerinterestratenowyieldsmorereturn,givingitparitywithnewissues– whenheldtomaturity.However,whentheissueissoldbeforematurity,andinterestrates haverisen,thepricewillbebelowpar.Theprocessisdynamicandnotstatic. Withtheyieldcurve,theyieldwhichrelatespricetointerestratesisafunctionof thetimetomaturity.Duringthebusinesscycle,astheFederalReserveraisesorlowers rates,thecurvechangesshapetoreflectinvestorsentimentandtheprevailingrateforeach levelofmaturity.Ifinvestorsexpectanormalspateofinflationtooccur,long-termrates willbeaboveshort-termrates,becauseinvestorsneedtobecompensatedfortheriskof holdinganinvestmentthatismoreexposedtochangesininterestratesandinflation. Analogously,financialinstitutionsmakeprofitsbyborrowingatlowershort-termrates andlendingatlong-termrates,andtheascendingyieldcurveisuniversallyacceptedby economistsas“normal”. Asecondincorrectassumptionisthatafirm’schoosingthelowestinterestrate, basedonthematurityofaloan,willminimizecapitalcosts.Asweshallsee,justthe oppositeisoftentruebecauseratesreflectinvestorexpectationsasmuchasthepriceof debt.Financialprofessionalsanalyzeyieldcurvebehaviorthroughwhatistermed“the segmentedmarketshypothesis”.Thistheoryimpliesthatthesupplyanddemandfor loanablefundsisderivedfromthecash-flowpatternsofabusiness.Companieswithlarge amountsoffixedassetsdemandtheuseoflong-termfunds,whilecompanieswithmore currentassetswillborrowshort-term.Banks,forexample,havemanyshort-term liabilitiesandtheywillmatchthosematuritiesbyinvestinginshort-termsecurities,mostly treasurybills.Astheeconomyheatsup,theFederalReservebeginsraisingratesatthe 192 sametimethatloandemandisalsohigh.Tomeetbothrequiredreserveratiosandhigher loandemand,bankswillselltheseshort-termsecurities,floodingthemarketandpushing uptheyieldsonthem. Academics,ontheotherhand,generallyanalyzeyieldcurvebehaviorthroughwhat istermed,“theexpectationshypothesis”.Long-termratesareviewedasthesumofshort- termratescombinedintoalongermaturity.Therefore,theshapeofthecurveisa reflectionoffutureexpectationswhichwilldeterminebasicsupplyanddemand.Iflong- termratesaretemporarilylowerthanshort-termrates(thedreadedinvertedyieldcurve), itisbecausetheeconomyisinbadenoughshapefortheFederalReservetobeginlowering rates. Inessence,thesetheoriesarenotmutuallyexclusive,andbothappeartoexplain yieldcurvebehavior.Thesegmentedmarketshypothesisproclaimsthatshortandlong- termdebtcannotbesubstitutedforeachotherwhichisbackedupbyempiricalevidence. BraniffAirlinesinthe1980sisacaseinpoint:whenthiscompanybeganusingshort-term debtasasubstituteforlong-termdebt,itwentbankrupt.Thevolatilityofshort-termrates began“eating”thecompany’sprofits.Althoughmanycompanieswillconvertshort-term debtintolong-termonceitreachesaspecifiedlevel,theoppositedoesnotoccurbecause firmsdonotwanttopayvariableinterestrates;theuncertaintyofthelevelofdefaultrises. Therefore,demandforloanablefundsisafunctionofboththefinancialstructureofa companyandexpectationsaboutthedirectionofinterestrates.Whilealargecompany cannotdelayallofitsfundinguntilinterestratesarelower,itmayrationcapitalby limitingoutlaystocurrentprojectsandforegoanynewprojectsuntiladownturn transpiresandtheFedlowersrates. GRAPHSTHATUNITETHETWOTHEORIES Tomatchcorporatebehaviorwiththeshapeoftheyieldcurveineachphaseofthe cycle,thesegraphsservetouniteboththesegmentedmarketsandexpectationshypotheses. 193 Figure8-1 1)Phase:Contraction/Recession 2)ExpectationsTheory:Short-termratesareabovelong-termratesandlong-termrates areexpectedtodecline.Companiesdonotwanttolockinaloanatahigherrateandwait forinterestratestodecline.However,thisis“self-fulfillingprophecy”,becauserecessions beginwheninvestmentandloandemanddeclines,andbecomeworsewhenfirmsdelay buildingtheirinventories. 3)SegmentedMarketsHypothesis:Bankshaveraisedshort-termratesbyselling securitiestomeetloandemandinpriorperiods.Consumerspendingisnowdeclining,and companieswithsteadydemandandshortproductcycleswilldowell.Withhigherdebt levels,industrialcompaniesmaynotbedemandingloanstomaintainfixedassets,andthe priceoflongtermdebtisdeclining.Thelackofcapitalexpendituresisaleadingindicator of“”trouble”downtheroad. 4)CompanyBehavior:Mostcompaniesretrenchandattempttodepleteexisting inventories.Asstatedabove,firmsdelaymajorpurchasesofplantandequipmentbecause theyareuncertainaboutthedirectionoftheirrespectivesectors.Anycompanywith steadydemandwillgenerallyoutperformtheeconomy,evenifprofitmarginsare squeezed.Thus,“sin”companieslikealcoholandtobacco,healthcaremaintenancefirms, YIELD TIME 194 andconsumerstapleslikebathroomtissueetc.willatleastmaintainthevalueofmost portfolios. Figure8-2 YIELD TIME 1.Phase:Recovery 2.ExpectationsTheory:Theyieldcurvetendstobeflat(itmaybeslightlyascendingor descending),butratesarenowlowerandcompaniesbegintoborrowattheselowerrates forlargecapitalexpenditures.TheexpectationisthattheFedisthroughwithcuttingrates andthatnowwouldbeagoodtimeto“lock”themin.Refinancingloansisagaincommon forfinancialinstitutions. 3.SegmentedMarketsHypothesis:Companieswithmorefixedassetsandsteady operatingleveragecanbestbenefitfromlowerinterestratesonlong-termdebt.These firmstendtousealotofdebtandhavelargeperiodicexpenditureswhichwouldlower capitalcostsifboughtwithlowercostloans.Long-termdebtisnowindemand.Lower riskcompanieshavethegreatestaccelerationofEPSbeyondthecostofcapital. 4.CompanyBehavior:Firmswillextendinexpensiveloansamongthemselvesandto consumers.Housing,banksandutilitieswillbenefitfromlowercostborrowing,andthe slightlyhigherpricetheychargefortheirservices.Thesesectorsaredeemed“interest sensitive”. 195 Figure8-3 YIELD TIME 1.Phase:Expansion 2.ExpectationsTheory:Companiesexpectinterestratestoriseastheeconomy“heats”up TheFeddoesindeedraiserates,perhapsseveraltimesbeforeinflationappears.The economicimperativeistoborrowandbuy“now”,beforeinterestratesriseandinflation takeshold.Essentially,thisscenariorepresentsacollectiveraceforimmediate consumption. 3.SegmentedMarketsHypothesis:Long-termdebtisstillinhighdemand.Companies withhigheroperatingleveragebegintodowell.Atthispoint,theequitymarketsarein fullswingwhichrepresentsthemainsourceoffinancingformanyofthesefirms.Higher GDPgrowthandhigherconsumerspendingoffsetsthehigherrisk. 4.CompanyBehavior:Consumerspendinghasbeenbolsteredbyagreatemployment marketandloansareextendedfor“consumerdurables”-autos,boats,appliances-so called“bigticket”items.Intermediateindustrialgoods(usedtoproduceconsumerend products)andtransportationcompaniesbegintodowell. 196 Figure8-4 YIELD TIME 1.Phase:Plateau 2.ExpectationsTheory:“Mixedsignals”aregivenoffbyaneconomythatisunsureof whichdirectionitisgoing.SomethinktheFedwillraiseratestocurbinflation,while othersseediminishingreturnsandhopeforaratecut.Inessence,theflatyieldcurve mimicsanearlyrecoveryandyetinterestratesaretoohighforanotherexpansion.The financialsectorbeginstosuccumbtointerestrateworriesandlobbiesforacut. 3.SegmentedMarketsHypothesis:Atthemarketpeak,companieswithhighoperating leverageandlittledebtdothebest.Consumersarestillspendingeventhoughthereisless industrialactivity.Thetechsectorcanbenefitatthistimebecausetheirfinancingis independent(superficially)ofhighinterestloans.However,theremaybemorevolatilityin themarket,andsomeofthesestockswill“sky-rocket”,onlytofacetheconsequencesofa highbetaduringthelatterpartofthephase. 4.CompanyBehavior:Bothpersonalincomeandinterestratesarerelativelyhigh. Capitalgoodshavebenefitedfromreplacementneeds,butmostcompaniestakea“wait andsee”approachasmergerandacquisitionactivityslowsdown.Inthismarket,itis “winnertakeall”andsomesectorsareheavilyfavoredoverothers,causingvolatilityas investorsareunsureofwheretoputtheirmoney. 197 Thestudent/investorshouldrealizethatthecharacteristicofaphaseisnotderived fromitspredictedpatterns,butdefinedbyitsanomalies.Canamarketgofromrecovery toplateauandmissanexpansion?Canbanksdobetterwheninterestratesarehigh?Can stocksdowellinthefirstpartofarecession?Asimprobableasthescenariosbehindthese questionssound,theincreasedcomplexityoftheeconomy,especiallythebehaviorof foreignmarkets,canturnthemplausible.Banks,forexample,increasinglycherish“non- interestincome”thatdiversifiesthemawayfromdependenceontheFed.Borrowingat lowratesinJapan,inordertoinvestinChina,makescompanieslesssensitivetotheyield curve.Inevitably,theremaycomeatimewhenthestockmarketnolongerreflectsriskin theUnitedStatesalone,andintheperiod2003-2008,wewitnessedanotheranomaly:gold andstockswereperfectlycorrelated.But-ifwecombinedourownyieldcurvewiththatof othercountriesandweighteditbyGDP,thatcurvewouldreflecteconomicconditions throughoutthedevelopedworld.Therealutilitybehindtheshapeoftheyieldcurveisthat itgraphicallyrepresentsthebusinesscycleateachpointintime.Althoughitfluctuatesand shiftstoreflectimmediateconditions,noothermeasurementsopreciselycapturesinvestor behaviorandexpectations.Ifweweretoconcludethatcapitalstructuredecisionsarean outgrowthoftherelationshipbetweenlongandshort-terminterestrates,wewouldbevery closetothetruth.Thoseratesaffecttheequitymarketsandthetypeofassetsthatgenerate income.Thelengthoftimethattheymaintainapattern,theleveltheyareat,thedistance betweenshortandlong,aswellastheirinherentvolatility,alldeterminetheamountand timingofcashflows-andalsodeterminehowthosecash-flowsarefunded. STRATEGICCONSIDERATIONS • 1)Themarketprecedesactualbusinessactivitybyapproximatelysixmonths.One reasonthatinsidersprofitistheirabilitytoanticipateincreasedactivityintheir respectivebusinesses.Thisisalsothereasonwhy“chasingprofits”byinvestinginthe mostprofitablecompanies,isdoomedtofailure.Bythetimeaninvestmentismade,the yieldcurveshiftsandbeginstofavoranothersector.Theonehighlyrecommended 198 strategyistoinvestinageneralmarketindexaboutsixmonth’safterarecessionis officiallyannounced.Historydictatesthatamarketbeginstorecoveratthispoint,and thatthegreatestgainwillbeatthebeginningofarecovery.Nooneexceptmarket professionalswillbewatchingtheindex,andthemarketseemstorespondtothislevel ofanonymity. • 2)Afavorableleveragestatewillmoveafirmtowardanoptimalcapitalstructureand ahigherstockprice.Unfortunately,asinterestratesriseandtheyieldcurveshifts,the confluenceofidealincomeandcapitalcostconditionswillonlybetemporary.Thespan canbefromsixmonthstotwoyearsbeforeastockbecomes“overbought”.Whenan investorhasseenacompanydoubleitsstockprice,itusuallysignifiesthatthesectoris abouttowash,althoughitdoesnotmeanacompanywillmakelessthanoptimalcapital structuredecisionsorlosevalue.Ifearningsdeceleratecomparedtothecostofcapital, investorsbeginlookingtoothersectors,butthatdoesnotmeanthatthecompanyisnot agoodlong-termprospect. • 3)Twootheraphorismshavehistoricalmerit:1.Thegreatestgainsinthemarket occurbeforetheFederalReserveraisestheinterestratethreetimesinsuccession.At thispoint,themarketbecomesmoresectororiented,respondingtoindividualcash- flow/capitalcostcircumstances.2.Ifafterthreetofouryearsofrecovery,andshort- termratesexceedlong-term,itmaybetimetoshiftmoneyoutofthemarket.Atthis point,theFedwilltrytoincreasethemoneysupplywithopenmarketoperationsand otherlendingfacilities,tryingtoengineerwhatistermed,”asoftlanding”.While enormousspeculativeopportunitiesexist,prudentinvestorswillcurbtradingbecause thepotentialriskoutweighsthereturns.Iftheinvertedyieldcurveisatlowenougha level,itdoeshavethepotentialtoflattenoutandbeginascendingtocontinuethebull market.However,suchamovewouldsignifythattheFedconsidersinflationtobeat anacceptablelevelandwouldpenalizecreditorswithdeflateddollars.Sincedebtlevels 199 arehigher,losseswouldbehighunlessinflationwasactuallyheldincheck.The averageinvestorbestnottemptfate. • 4)Politicalconsiderationsmaytrumpeconomicdiscretion.Sincethepublicequates economicbehaviorwiththeprevailingpresidentialadministration,theFederalReserve triesnottoraiseratesinthelasttwoyearsofanincumbent’sreign.Ifinflationgetsout ofhand,thepostponementofaratehikemaymakeitworse,forcingtheFedtotake evenmoredrasticmeasures.Historically,stocksgenerallydotwiceaswellinthelast twoyearsofaPresident’stermthaninthefirsttwoyears. • 5)Majortrendsareconfirmedbysequentialmovementinthesamedirectionofall threeDowcomponentindices,utilities,industrialsandtransports.Atfirst,electric utilitiesleadtheway,upordown.Secondly,theDowIndustrialsfollowtheutilities. Lastly,thetransportationindex,followstheDowindustrials.Unlessatrendoccursin thatspecificorder,itisnotconsidered“confirmed”,andmayleadtoalesspredictable businesscycleandachaoticmarket.Fromasectorrotationstandpoint,theDow Theoryiscompletelyrational.Interestratesgettoohighforthe“interestsensitive” utilities,pushingthemdown.Fewerorderspushdowntheindustrialsbecause companiesdonotwanttocommittoprojectsathigherrates.Lastly,thetransports sufferbecausetheyarethemainserviceunitfortheindustrials. THEBUSINESSCYCLEANDTHECOSTOFEQUITY Thefirstmisconceptionthatinvestorshaveaboutbetaisthatcompanieswith greaterfinancialleveragehavehigherbetas.Infact,whileleverageincreasesbeta,the greaterproportionofbetaisderivedfromtherelationshipbetweensalesandmarket return.ToreiteratetheMandelkerandRheeequation,Beta=(DOL)(DFL)(ROE)[(COV %Sales,%Market)/Variance%Market)].Salesisaprominentpartofboththedegreeof operatingleverageandthecovariancecomponentoftheequation.Moreover,itisan assumptionofcapitalstructuretheorythatoperatingandfinancialleveragebalanceeach other;acompanywithmorefinancialleveragewillhavelessoperatingleverageandvice- 200 versa.Althoughsomesectorsemploymoreofbothtypes,withinanysector,themixtureof leveragewillbesimilarandbalanced. Theoretically,the“ideal”companywouldhavealowbetabecauseitwouldbewell diversifiedandbeabletoincreaseitsbetawithmoredebtand/oracquisitions.Playingthe cycle,itwouldtakeadvantageoflowinterestratesuntilthemarketpickedup,andrates wereincreased.Atthispoint,itsinvestmentinassetswouldbegintopayoff,andthe companywouldraiseitsreturnonequity(ROE),whilemaintaininghighdemandforits products.AstheFederalReserveraisesinterestratestostaveoffinflation,thefirmbegins topayoffsomeofitsolddebt,attractingequitythroughitshigherEPS.Simultaneously, thecompanybeginstoretainmoreearningsandlowersitslong-termdebttocapitalratio, decreasingbetajustasthebusinesscycletransitionstoaplateau.Duringthestagnant market,thefirmcontinuestodiversifywithacquisitions,loweringoperatingrisk,and tryingtobroadenitscustomerbaseforthenextprofitcycle. Thus,the“ideal”companyengineeredastrategythattookadvantageofthree cyclicalcharacteristics.First,ittookadvantageoflowerinterestratesandbegantakingon debtandraisingitsbetajustastheeconomywasimproving.Secondly,asinterestrates werecontinuallyraised,itlowereditsdebtratioandbeganrestructuringitscapital towardsanequitybase.Concurrently,EPSwasrisingandan“overheated”economy ensuredthatdemandforitsproductswasstable.Lastly,thecompanypreparedfora downturnintwodifferentways:1.Itjettisoneditshighinterestdebtandpositioneditself forgreatersolvency.2.Itbegantobroadenitscustomerbasebyinvestinginrisklowering acquisitionsthatwoulddiversifyitsoperations.Fromtheperspectiveofcapitalstructure, afteradownturn,thefirmusesleveragetoacceleratethechangeinEPSwellpasttherate ofchangeinthecostofequity;thecostofequitywasatacyclicallowbecausethemarket haddeclinedandtheFederalReservehadloweredinterestrates.Inthesecondphaseofits strategy,thefirmactuallybeginstolowerbetainresponsetohigherinterestrates.By paringdownitsproportionofdebttoequity,itnullifiestheriskofleverageathigherrates. 201 Ontheotherhand,itcandonothingaboutthesystemicriskofanoverheatedmarket,and soitattractsequityfundingwithitshigherEPSandsubstitutesitfordebt.Thethird phaseofthestrategylowersitsbetainresponsetoastagnatemarketbutalsopreparesfor thenextbusinesscyclebydiversifyingawaysomeofitsoperatingrisk.Inessence,thefirm isbothpreparedforadownturnandyet“cautiouslyoptimistic”aboutfutureprospects. Preposterousyousay?Afantasy?Whileeventhebestruncorporationscannotgo througheverycyclewithsuchmachinelikeprecision,manyfirmshireeconomiststoguide themthroughthevariouspitfallsandmissteps.Thebankingindustryinparticularis exposedtocyclicalrisk,whichcarriesovertoallthosewhoareinfluencedbytheprime rate-whichencompassesatleastsomeaspectofnearlyeverysectorintheeconomy. Again,thepremiumisplacedonforesightandnothindsightbecauseifacompanyadopts thesestrategiesasareactionaryresponse,itwillfinditself“outofsync”withchangesin thecostofcapital,i.e.,therateofearningsincreaseswillslowandbesuddenlyeclipsedby thecostofequity,whichwillbegintoaccelerate.Onthedownside,earningswillusually outpacethedecreaseintherisk-freerate,butevenifearningsarestable,ahigherbetawill createanoverreactiontoamarketdownturn.Atthispoint,weoftenseefirmswithtwenty percentearningsincreases-loseandnotgain-twentypercentinstockprice.Themarket simplyfactorsindiminishedfutureprospects. Betararelyperformsasexpectedintheshort-run.Infact,rightafterthetheoryof theCAPMwasproposed,WallStreetimmediatelyjumpedonthebandwagon.Investingin highbetastocksduringanupswingandtheninlowbetastocksduringadownturn, institutionalinvestors’attempttotimethemarketwasfutile.Performancewasspottyat best.However,therealculpritwasmethodology;betawasusedasatoolforpredicting stockprices,andnottogaugecomparativerisk.Itisalwayspossible(althoughnotlikely) forahighbetastocktoonlyreactviolentlywhenthemarketisdecliningandtomake meagergainsduringtheexpansionphaseofatypicalcycle;betaencompassescumulative volatilityandisnotstableintheshortterm.Suchanomaliescanbefurtherpunctuatedby 202 alowcorrelationcoefficientandahighbutvolatilealpha-thepartofaregressionthat dependsonfactorsoutsideofmarketinfluence.Forexample,iftariffpolicysuddenly favorsaparticularindustry,i.e.,steel,thenthatindustrymayprosperwiththeeffectofa greater“alpha”,andlessbeta;itisnolongerasdependentonthemarket.Theindustry mayhaveacollectivebetaofperhaps“1”,buthardlyreactsatallduringamarket downturn. THECAPITALASSETPRICINGMODELANDSENSITIVITYANALYSIS ThetrueworthofbetaisgaugedinrelationtotheothercomponentsoftheCAPM. TheCAPMisadynamicmodelthatchangesdailyasthemarketchanges.Periodically,the risk-freerateismanipulatedoutsidethesystembytheFederalReserve,butitchanges yields(notcouponrates)basedondemandfortreasuries.Themorevolatilestockmarket indexcanvaryfromnegativetwentypercenttopositivetwentypercentoverthecourseofa year,andthedifferencebetweenthatfigureandtherisk-freerate,knownasthemarket riskpremium,istheprimaryeconomicfactoraffectingthecostofequity.Ultimately,the singularrelationshipbetweenbetaandtheriskpremiumwilldeterminethecoherenceof themodel Inthefollowingexample,wewillobservetheeffectofaonepercentchangein interestratesontheCAPM,definingthreelevelsofbeta:lowmedium,andhigh. Table8-1 EQUILIBRIUM RiskFree % Beta Market RiskPrem. CAPM LOW 0.05 0.75 0.098 0.048 8.6% MEDIUM 0.05 1 0.098 0.048 9.8% HIGH 0.05 1.25 0.098 0.048 11% 203 Table8-2 INCREASE 1% RiskFree % Beta Market RiskPrem. CAPM LOW 0.06 0.75 0.098 0.038 8.85% MEDIUM 0.06 1 0.098 0.038 9.8% HIGH 0.06 1.25 0.098 0.038 10.75% Noticethatwiththehighbetastock(1.25),increasingtherisk-freerateactuallydecreases thecostofequity(from11%to10.75%).Thisisasystemicadvantagethatencompasses firmswhomustfinancewithmoreequity.Companiesthathavehighbetasmayhavea difficulttimeturningtothecreditmarketsfornecessaryfinancing.Wheninterestrates rise,thecostofcapitalgoesupforthosefirmswhouseleverage.Ontheotherhand,those whouseequityhaveacompetitiveadvantage,especiallyifthemarketignoresthehigher ratesandkeepsascending.Thehigherbetaallowsafirmtobothescalateearningswith lowercapitalcosts,andtemporarilyoutperformthemarket.However,asthemarket keepsrising,thesecompanieshaveamuchhighercostofcapital,andwheneitherearnings orthemarketdeclines,highbetastocksfallprecipitously. Thisnextexamplewillshowsensitivitytomarketchanges.Italsoshowstheimportanceof alwaysexaminingthecostofequityinthecontextofearnings.IfweexaminetheCAPMin isolation,amarketdeclineactuallyreducesthecostofequity,butthenwemustremember thatearningsarehighlycorrelatedwiththemarket;adeclineinmarketvalueimpliesthat earningsmaybedecreasingbyanevengreateramount. Table8-3 EQUILIBRIUM RiskFree % Beta Market RiskPrem. CAPM LOW 0.05 0.75 0.098 0.048 8.6% MEDIUM 0.05 1 0.098 0.048 9.8% HIGH 0.05 1.25 0.098 0.048 11% 204 Thisscenariocallsforathreepercentdropinmarketreturnto6.8%. Table8-4 MKT. DECREASE RiskFree % Beta Market RiskPrem. CAPM LOW 0.05 0.75 0.068 0.018 6.35% MEDIUM 0.05 1 0.068 0.018 6.8% HIGH 0.05 1.25 0.068 0.018 7.25% Evenwiththeintroductionofextraordinarycashflowduringthisperiod(beatingthe averagecompanyinearningsgains),itisdoubtfulwhetherahighbetastockcould maintainitsprice.Certainly,thecostofequitydroppedmorethanthatoflowbetastocks, butthestudent/investorshouldconsiderthesourceofthehigherbeta-usuallyhigher operatingriskattributabletovolatilesalesandearnings.Anyfirmthatfinanceswith equity,andhasahigherbeta,musthavehigheroperatingvolatilitybydefinition. Oneofthekeystounderstandingthecostofequityistherecognitionofhowmarket dependentitis.Althoughinterestratechangescanmovethemarket,theywillrepresent onlyasmallfractionofthecostifthemarketdoesnotrespond.Thenexttableshould convincereadersthatmarketreactionisparamount:wecuttheinterestrateinhalf, withoutamarketresponse. TheequationistheCAPMwithabetaofone:Risk-freerate+(Beta)(Marketrate-Risk- free). Table8-5 TheMarketgoesfrom20%to15%Return .05+(1)(0.2-.05)=0.2or20% .05+(1)(.15-.05)=0.15or15% 205 Table8-6 InterestRatesareCutinHalf(therisk-freegoesfrom5%to2.5%) .05+(1)(0.2-.05)=.2or20% .025+(1)(0.2-.025)=.2or20% Moreover,thereisnofactorinthemodelthatexpressestherelationshipbetweeninterest andmarketreturnexceptforthecomparativeriskpremium(marketreturn-risk-free rate).Thecorrelationbetweenratesandmarketisnotexplicitandcannotbepermanently quantifiedbecausevolatilityinbothcreditandequitymarketschangestherelationship. Ultimately,therelationshipisafunctionofyieldcurvebehavior,inflationandaggregate demandamongmanyfactors.Expressedasalogarithmiccurve,therelationshipisina stateofperpetualchangewhichdiminishesforecastingability,i.e.,themarketcreates inflation,whichraisesrateswhichlowersthemarket,whichlowersrates,whichincreases themarketwhichcausesinflation... Figure8-5 Themostconstructiveadvicethataninvestororcorporateprofessionalcanreceive istokeepbetaataminimum;theempiricalevidenceisunchallenged.Lowbetastocksare rewardedoutofproportiontotheirriskprofiles,whilehighbetastocksarepenalized.In statisticalterms,adownwardbiasexistsforhighbetastocks.However,thedangerof InterestRate MarketReturn 206 makingthisgeneralizationistolosetheupsidepotentialthatsomanyhigherbetastocks provideduringtheexpansionphaseofthebusinesscycle.Whilesomehighbetastocks offeranexceptionandachievesomedegreeofstabilitybytheirverysizealone(Intel,for example),theimperativeistoseekoutfirmswhocanraisebetaandtakeadvantageof upswings,butnothavethetypeofoperatingriskthatwoulddamagethestockduringa downturn. Oneofthebestexamplesofhowbetacanbedeceptive,iswithChapterOne’s introductoryillustration,FisherScientific(nowThermo-FisherScientific).Withasmall profitmarginof3to4percent,Fisherwascapableofincreasingsaleseveninadownturn. Thecompanywaswelldiversifiedwithalowoperatingleverage,buthadadebttocapital ratiothatsometimesapproachedseventypercentbecauseofnumerousleveragedbuyouts. Nevertheless,thecompanymaintainedabetaofabout0.6andwasabletoincreasetheir sharepricesevenfoldinabouteightyears.Howwasthatpossible?Besidesbalancing operatingandfinancialleverage,thisfirmrarelyreactedasmuchasthemarket.Noone everbraggedabouthowFisherwasbeatingWallStreetestimatesandthestockwasnot heavilytraded.Infact,salesandcash-flowwerecomparativelylarge,butthelowprofit marginsputthecompanyinthe“highturnover”class;assetgrowthwasleveragedthrough thestabilityofsales.The“mystery”behindFisherwasthatthestocksimplydidnotmove onthebasisofsalesorprofits,butoncetheword“acquisition”wasmentioned,ittookoff likeaguidedmissile.Eventually,itwouldagainsettleinforanothersoporificinterim, waitingforwordofthenextgrowthopportunity.Mathematically,theselargejumpswould notbeinterpretedasvolatilitybecausetheywereisolatedtospecificshortperiods;thus, thesmallbeta.Inotherwords,Fisherhadupsidepotentialbutwasprotectedfrom downsiderisk OnefamousstudywasconductedbytheformidableresearchteamofBlack,Jensen andScholes.TheyfoundthattheentireperiodfromApril1957toDecember,1965,was characterizedbyaskewedrisk-returntradeoff-higherbetastocksproducedlowerreturns 207 thanlowbetastocks.Infact,thesefindingsareoftenpointedoutasacondemnationofthe entireCAPM.Again,theanomalycanbeexplainedbytheextremityofthemovement;a stockwithahighbetacantriplethemarketsdownwardpathinonlyafewshortdays, whileitmaytakeayearormoreforittoachieveanewhighduringanexpansion.That jackrabbit-volatilityalmostalwaysrepresentsareturntorelativebookvalueandaway fromspeculativerisk.Anotherexampleofbeing“burnedbybeta”wasduringthetech stockspeculationofthe1990s.Thisperiodofferedrewardstohighbetainvestorsonlyif theywereadeptenoughtopartwiththeirinvestmentsbeforethey“fellthrough”.Very reputablefinancialprofessionalsthoughtthata“newera”haddawned,characterizedby permanentlyhighpricetoearningsratios,andvolatilestocksthatwould“gobackup”.By 2001,manyhighbetastocksthatcommandedahundreddollarsasharejustafewyears earlier,couldbeboughtfortendollarsandchange. Unfortunately,thevolatilityofcertaintechstockshasledtomorespeculationand lessinvestment.EvenawellruncompanylikeGoogleisnotaninvestmentgradestockas ofthiswriting(2008),becauseitisoverexposedtoadownturn.Ontheotherhand, Microsofthasbecomemoreinvestorworthyasitgivesoutspecialdividendsandtrades betweenamuchnarrowerrangethanitoncedid.Itsbetaisnolongerexceptionallyhigh andithasdiversifiedintovoicegeneratedprogrammingandvideogames.Therefore,a highbetashouldnotcondemnastock’sinvestmentpotential;awellmanagedhighbeta companycanloweritsbetaasittakesadvantageoflowercostequity.However,even duringanupswing,lowbetastockscanoutperformthehigherbetasbecauseofatendency toholdtheirgains. CIRCUMVENTINGTHEOPTIMALCAPITALSTRUCTURE Therelationshipbetweenshort-terminterestratesandlong-termratesdefinesthe businesscycle.Ifincomestreamsarecertainoveralongperiod,thecostofborrowing short-termwouldbesignificantlylessthanborrowinglong-term-givena“normal”yield curve.Thereasonforthelower“price”ofshort-termdebtisthatthelenderincursless 208 risk.Anypricinganomalyintheborrower’sfavorisshort-livedbecausetheloanwillbe frequentlyrenewedatthecurrentrate.However,theborrowerincursmuchgreater defaultrisk-loansmighthavetobepaidwhencash-flowisnegative-and-thereisgreater riskincurredbecauseofthepotentialforinterestratefluctuation.Mostbusinessesneedto planfarenoughaheadtobecompetitiveandfrequentratechangesontopoffrequent paymentswillincreasedefaultprobability.Thisdisparitybetweenthepriceofdebtand theriskofdebtskewstheabsolutecostofcapitalinthedirectionofgreaterrisk.In essence,thecompanyisforcedtochoosehighercostdebtbecauseitislessrisky.Sincea wellrunfirmmatchesitscash-flowwithitsfundingneeds,thereisalwaysthetemptationto uselessexpensivedebtwhenthefirmisinatemporary“holding”patternatthetopofthe businesscycle:itmaywishtodelaymajorprojectsbecauseinterestratesareexpectedto decline.Asbanksselloffshort-termsecuritiestomeetloandemand,thereisatendencyfor short-termratestorise,indicatingtheprospectofacontraction.Thus,afirmmay temporarilysubstitutetheincreasinglyexpensiveshort-termloansforlong-termdebt,in thehopesofavoidingbeing“lockedin”atahigherrate.Thistrade-offofmoreimmediate riskandexpensefortheprospectoflowerratesinthefutureisaluxurythathighbeta companiescannotafford.Lowerbetacompanies,whohavelessoperatingriskcanmakea strategicplaytolowercapitalcostsinthelong-runbyacceptingtheseimmediaterisks: theydonotfacetheprospectofgreatmarketvolatilitywhichwoulddamageahighbeta firmthatacceptedthesamerisks.Interestexpensewouldriseintheinterim,aswould defaultprobability,butthismovementawayfromanoptimalcapitalstructurewould positionthefirmforgreatergainsoncearecoverycommenced.Essentially,thehigher pricepaidfortheshort-termloansbuffersthefirmagainstuncertainty,actinglikean insurancepremium,andcounterbalancingtheriskofdefaultbecausetheactionisonly implementedtemporarily. THEGAMEOFCAPITALSTRUCTURE“GOTCHA” 209 Highbetacompanieswhopeaknearthetopofabusinesscycleprimarilyfinance withequity.Long-termdebtisbothprohibitivelyexpensiveandrisky,becausecash-flowis notstableenoughtowarrantit.Undernormalcircumstances,retainedearningsprovide sufficientequitygiventhelowereconomicprospectsthatoccurduringtheplateauphaseof abusinesscycle.However,thecostofequityisrelativelyhighatthisstage,pushedupbya marketthathassurgedthroughafewyearsofexpansion.Unlessearningsaremaintained andevenaccelerated,thecostofequitywillberisingandeclipsingtherateofearnings, sometimesvectoringoffintoadifferentdirectionaltogether.Withtheprospectofboth lowercapitalandconsumerspending,theforecastforastablestockpriceisdim. Whenahighbetastockisatitspeak,companyofficialssometimesfeel indestructible.Theirunbridledoptimismispunctuatedbyhugebonusesandastockprice thatissoaring.Opportunitiesforgrowtharesightedandthefirmmaybeginraising capitaltofundlargeprojects.Atthesametime,anexpandingmarketcreatesmany unsophisticatedinvestorswhohaveneverseenadownturn.ThemarketseemslikeaLotto ticketthatalwayspaysoff.Thus,the”perfectstorm”occursincapitalstructure.The unbridledoptimistsarematchedwiththeunsophisticatedinvestors;oneentitydemands plentyofequityandtheothersuppliesit.Infact,thehighbetafirmisminimizingcapital costsbyraisingequitywhenthestockpriceishigh:morefundswillberaisedwithfewer shares.Whateachplayerinthisscenariodoesnotrealizeisthattheyhaveuppedthe “ante”atthewrongtime,takingonmoreriskthaniswarrantedbytheeconomicoutlook. Naturally,theinvestmentbeginstoimplodeassoonasanyofthemajorindustry participantsmissesearningsexpectations.Therefore,aninvestorneedstoviewequity issuedlateinthebusinesscyclejustasacompanyviewsshort-termdebt.Thehigherrisk needstobecounterbalancedwiththepotentialforappreciation-except-inthisgame,the loandoesnotexpirelikeshort-termdebt.Oncestockisissued,itiskeptuntilsold;the investormustenterthedeallikeagambler,expectingto“dump”thesharesbeforegetting 210 “burned”.Inessence,evenwiththebestofintentions,bothpartiesenduplikepoker playersinaLasVegascasino. IDEALIZEDTRENDS Businesscyclebehaviordefiesexpectationsmorethanitconfirmsthem.Highbeta stocksmayrecoverearlierthanlowbetastocks.Temporarily,thecostofequitymay actuallydecreaseduringanupswing.Themarketmaygointoatailspinjustasthe expansionphaseisexpected.Exceptionstotheruledonotnegateitslogic,butoffersthe opportunitytoobservetheworkingsofotherforces.Often,thepoliticalmotivesofalobby canbeexhibitedinsomepieceoflegislationthatseemsinconsequentialbuthas repercussionsinthemarket;thelawsofsupplyanddemandarecircumventedbycreating anartificialscarcityforexample.Speculativeexcess,a“bubble”createdbygovernment actionorinaction,canoccur.Anexampleofsuchabubblewasthehousingspeculation thatoccurredintheearlymillenniumwhichwaspurportedtobeanoutgrowthofinterest ratesbeingtoolowfortoolong.Insuchascenario,theinvestingpublicwillputtoomuch capitalinoneareatothedetrimentofothers,andtheresultwillbeasector(housingor Reits)thatremainsprofitablebeyonditscapacitytogenerateincome.Thus,asectorwhose internaldynamicswouldnormallyletitprosperforonephaseonly,endedupbeing favoredbytheeconomyforthree.Thisimbalancecanonlyendupaffectingothersectors, andindeedwesawafalloutwithmortgagecompanies,banksandbrokerscausingasevere creditdebaclein2007.Byrecognizingtherationalitybehindthe“ideal”businesscycle,the investorcanthereforebealertedtothedangerofanexception-whetheritbeaspeculative bubbleoramorecomplicatedsupplyanddemandissuesuchasoccurredwiththepriceof oil. 211 Figure8-6 VARIABLE PHASEA PHASEB PHASEC InterestRates LOW MEDIUM HIGH MarketIndex LOW MEDIUM HIGH CostofEquity LOW MEDIUM HIGH RelativeBeta LOW MEDIUM HIGH RelativeOp.Leverage LOW MEDIUM HIGH EPSAcceleration INCREASING HIGH DECREASING PHASEA PHASEB PHASEC LOWBETASPROFIT BETASNEAR“1”PROFIT BETAS>“1”PROFIT 1.FinancialLeverage 1.FlexibleMixtureof Debt/Eqty 1.MarketSpendingPeaks 2.StabilityofDemand 2.MoreMarketSpending 2.EquityislessExpensivethan Debt(relatively) Themanagerialimperativeistofinditsown“efficientfrontier”withintheparametersof itsindustry;thatis-itshouldstrivealwaystoachievethegreatestreturnwithagivenlevel ofrisk.ThereadershouldobservethateconomicriskislessinphasesAandBanditisat thisjuncturethatcorporateriskcanbeincreased.OnceafirmentersphaseC,corporate riskisnolongeranoption:itshoulddecreasebeta,reducedebt,andattempttolower operatingriskthroughdiversification. Averyobviousexampleofanindustrychangingitsriskprofileislandlinetelecom. Theseformerlyregulatedcompanieshaveincreasedtheircollectiveriskbybranchingout intowireless,theInternetandeventelecomequipment.Withoutthelowbetaofa MarketCap. Time A B C 212 regulatedutility,thesefirmspeakmuchlaterinthebusinesscycle.Infact,AT&Twitha betaofaround“1”,onlyhadalongtermdebttocapitalratiorangingfrom15%to33% throughoutthe1990s.Thetypicalutilityhasmorefinancialleveragebutnotnearlythe greateroperatingriskthatAT&Thasincurredovertheyears.Extremeamountsofcapital pouredintothesefirmsrightatthemarketpeakofthelate90s,whichwasthewrongtime toincursuchrisk.Whenthemarketimploded,telecomsufferedmorethanmost,butcould haveavoidedsomeofthedamagebyexpandingintheearlierphasesofthecycle. SECTORROTATION Toflawlesslypredictwhichsectorwillbethenexttoprofitisapipedream.Most sectorswillbesegmentedbyperformanceandallfirmswillnotprosperatonce.Norwill itbepossibletopredicthowlongasectorwillprofitalthoughsixmonthsofaccelerated earningswillbeanindicatedminimumtobeconsideredas“outperforming”theeconomy. “Chasing”profitablesectors,however,ismuchlikechasingearnings:theinvestormight gainfrommomentum,butjustaslikelywilllosemoneybecausethelargeinvestorshave alreadyspotted“thenextbigthing”andmovedon.Therealvalueinsectorrotationisto recognizethatadiversifiedportfoliocanbeachievedbyconcentratingonfirmsthatbenefit fromtherecoveryandexpansionbutshouldalsohavesomedefensivestocksincaseofa contraction.Inaddition,stocksthatdowellduringaplateaushouldbechosenonthebasis oflowerrisksimplybecausethehigherriskstocksthatdowellduringthisphasewillbethe firsttofall. SECTORLOGIC Theconceptbehindsectorrotationisthatthecyclefavorsspecificindustrieswithinthe boundsofinterestratesanddemand.Forexample,wheninterestratesarelow,more applicantsqualifyforcreditandany“bigticket”itemrequiringaloanwillbefavored. Consequently,anyindustrythatdependsoninterestrates-autos,housing,banks,etc.-will befavoredaswell.Duringacontraction,whencreditriskisstillhigh,consumerswillnot betakingoutloansforpurchases,butwillstillbebuyingbreadandgoingtovisitthe 213 doctorifneeded.Thus,consumerstaplesandthehealthcareindustryarefavored.When therecoveryandexpansionphaseshit,goodsthatareusedintheproductionofothergoods -subassemblies,smallmotors,heatexpandersetc.,oftencalledintermediategoods,willbe indemandandsetthestageforthelateexpansion,earlyplateau.Itisatthisjuncturethat unusedcapacitydisappearsandmanufacturersbegintoexpandbypurchasing“capital goods”-thetoolsandmachinerythatconstitutethefinalproductoftheintermediate goods.Naturally,allofthisfreightmustbetransported,andthevarioustransportation stocksbegintodowell-railroads,trucking,shipping.Whiletheastutereaderwillnotice thatsectorrotationismerelyadetailedelaborationofthe“Dowtheory”,thepremiumfor economistsistoobservehowitisdifferent.Manysectorswill“bleed”overintoaphase whoseinterest/demandcharacteristicdoesnotfittheindustry.Additionally,most economistswilldebatewhichphasetheeconomyhasenteredaswellashowmanyphases orsubphasesactuallyexist.Asageneralguideline,thefollowingtableoffersarough estimateofatypicalbusinesscycle: Table8-7 PHASE Contraction RecoveryI RecoveryII Expansion Plateau GOODS 1. Healthcare 2. Consumer Staples 3.Foodand Beverage 1.Interest Sensitive (banks, homes, electric utilities) 1.Consumer Durables (autos, appliances, “bigticket” items) 2. Intermediate Goods 1 Intermediate Goods 2.Consumer Discretionary 3.Capital Goods 1.Capital Goods 2.Consumer Discretionary 3. Transportation Onepatternthatisworthnoting:bothconsumersandbusinessesfollowsimilarborrowing patterns.Theconsumerswiththebestcreditwillbepurchasingthehomesandautosthat stimulatetheeconomyinarecovery.Likewise,thelargestmanufacturerswiththeleast operatingriskwillexpandwiththegreatestfinancialleverage.Paradoxically,asinterest 214 ratesrise,creditisevenmoreobtainableandlesscreditworthycustomerswillbepicking uploanslaterinthebusinesscycle,albeitathigherrates.Thebasicreasonforthis anomalyisthatpersonalandbusinessincomesdonotriseuntillaterinthecycle,thus qualifyingcustomersforloans.However,theaxiomthatthemostcreditworthycustomers needtheleastamountofcreditistrue;wealthybusinessesandpeopleusedebtasamoney- makingtoolratherthanasanecessity.Individualsandcompanieswhoaretheleastcredit worthy“somehow”enduppayingmoreinterestforloansbecausetheydidnotpossessthe collateralearlyenoughinthe“game”whenrateswerelow.Thisrecipefordefaultcosts financialinstitutionsbillionsbutnoalternativesystemseemspractical.Thesuccessful firmswhofinancewithequitydosonotbecausetheycannotqualifyforloans,butbecause itisthemostcosteffectivemethod,helpingtomaximizestockprice.Ontheotherhand, firmswhohaveavolatilecashflowwillsometimestakeondebtwhencredittermseaseand willsuffertheconsequenceslateintheplateauphase.This“survivalofthefittest”scenario canbecombatedwithknowledgeofthebusinesscycleandappliedjudgment. Thereadershouldnoticethattheriskpremium,thedifferencebetweenstocksand therisk-freerate,alsomirrorssectorrotation.Attheprospectofacontraction,thereisa “flighttoquality”,ageneralmovementintolowrisk,highqualitydebtinstrumentslike treasuriesandAAAbonds.Thegreatercertaintyinthebondmarketattractscapitalaway fromstocks.Asthemarketexpands,investorstakeonmorerisk,inflatingtherisk premium,andstocksareagainfavored.Infact,atthetopofthemarkettherewillbe investmentinjunkbonds,IPOs(initialpublicoffering)andeveninfirmswithoutany earnings.Thishigherdemandfordebtinstrumentsintheinitialphasesoftherecovery, makesdebtlessexpensiveinthecapitalstructure,butasriskpremiumsrise,firmscan garnermorefundsfromanequityissuedespiteitshighercostbecausestocksarein demand.Thissituationpresentsanotheranomaly:whendebtisindemandbyinvestors,it isrelativelyinexpensivefortheissuingcompany,butwhenequityisdemanded,the companymustpayahigherprice;therisingmarketraisestheriskpremium,andthecost 215 ofequityWhensuchanequityissueissupportedbyhigherearnings,bothbetaandthe costofbankruptcyproceedtodrop-betabythedecreaseindebttoequityandbankruptcy costsbyanearningsgenerateddecreaseindefaultprobability. INDUSTRYRESPONSETOTHEBUSINESSCYCLE Thefollowingindustriesrespond(notalwayspositively)totherespectivephase: Table8-8Table8-9 1)CONTRACTION 2)RECOVERY Utilities ElectricPower ConsumerStaples PaperProducts,Forestry Tobacco Chemicals Food,Beverages Steel Publishing HouseholdFurnishings,Autos, Appliances Drugs CrudeOil Healthcare Banks Apparel SmallMachineTools Intermediateparts DefenseElectronics PollutionControl WasteManagement Table8-10 Table8-11 3)EXPANSION 4)PLATEAU CapitalGoods AllTypesofMining MachineTools OilRefineries GoldMining TelephoneSystems Tobacco CommunicationsEquipment Beverages SpecialtyChemicals Drugs Transportation Cosmetics Aviation OilEquipment Aerospace ComputerSystems FinancialServices 216 Table8-12 CROSSOVERSECTORS INDUSTRY PHASES Publishing Contraction,Recovery Beverages Contraction,Recovery Mining Expansion,Plateau Oil Recovery,Expansion,Plateau ElectricUtilities Contraction,Recovery Tobacco Expansion,Plateau,Contraction Drugs Contraction,Expansion DefenseElectronics Recovery,Plateau ECONOMICSIGNALS Economicindicatorsaresomixedthatitisverydifficulttoachieveconsensusamong leadingeconomists.Thedecisionstoinvestinacertainsectormustanticipateitseconomic milieu,andmuchmoneyismadeinmakingacorrectforecast.However,oncethemajority ofeconomistsagreeonthestateoftheeconomy,itisalmosttoolatetomoveintoasector becauseallthemoneyismadethroughearlyanticipation.Theaverageinvestorcanbenefit fromknowledgeofeconomicsignalsifonlyasaninstructionaltoolfor“whatnottodo”. Chartsandtablessetuptheillusionthatanticipationofaneconomicphaseiseffortless,but thesetablesdonotfunctionas“tealeaves”.Forexample,whywouldanyinvestorbuy stocksonmarginwheninterestratesarehighduringaplateauphase?Iftheinvestorhasa speculativebent,itismuchbettertotakeadvantageoflowinterestratesandspeculate duringtherecoveryphase.Nooneneedstobeafortunetellertoalmostguaranteethat interestrateswilldropduringacontraction.Thefollowingtablesgiveabriefoutlineof whatsignalstoexpectineachphase: 217 Table8-13 CONTRACTION RECOVERY 1.GDPDeclinesasaNegative Percentage 1.GDPPercentageChangeturns Positive. 2..12MonthAveragePercentage ChangeinFederalFundsRateturns Negative. 2.12MonthAverageofIndustrial ProductionChangebecomesPositive 3.ThepercentagechangeinM2turns fromNegativetoPositive. 3.NonFarmPayrollsIncrease. 4.InterestRatesDecline. 4.InitialUnemploymentClaims Decrease. Table8-14 EXPANSION PLATEAU 1.InterestRatesIncrease. 1.TheMovingAveragefortheRateof ChangeofIndustrialProductionturns Negative. 2.6MonthMovingAverageintheCPI (Inflation)turnsPositive 2.GDPDeclinesasaPercentageChange (2%insteadof4%forexample) 3.NonFarmPayrollsIncrease 3.UnemploymentClaimsIncrease 4.12MonthMovingAverageoftheRate ofChangeintheFederalFundsRate turnsPositive. 5.6MonthMovingAverageoftheRate ofChangefortheRealMonetaryBase turnsPositive 218 ThearbitrarinessofsomeofthesesignalsisapparentifoneviewsanindicatorliketheCPI inisolation.Insomeeconomiestherewillbeinflationandrecessionatthesametimeand sotheaccelerationintherateofchangewillcertainlynotbeanexpansionphaseindicator. Commonsenseandcoordinationofseveralfactors(somenotenumerated)willbethe watchwords.Again,thepremiumisplacedonanticipationofthephase,butinvestmenton thebasisofsuchindicatorsisinitselfariskyproposition. (BacktoTableofContents) 219 9 OPERATINGRISK Onemajoradvantageoftheinformationageisthatwecanobservethe transformationofacademictheoryintopracticalapplication.Beforetheubiquityofthe PC,datawasobtainedfromoldnewspapersandmicrofiche.Asquaintasthatmayseem, investorswouldplotchartsoutofstockguides,sometimeslosingtrackofoverallobjectives inpursuitofesotericdata.AcorrelationbetweenRandD(researchanddevelopment) expendituresandnear-termsales,forexample,wouldbeconsideredtime-consuming research,nottypicallyinitiatedbytheaverageinvestor.Ifacademictheorywasreadily observableinthemarketplaceatall,itwasrestrictedtosuchgeneraltruismsas“higher earningsleadtohigherstockprices”.Technologyhastransformeddata.Wecannow obtaindetailedinformationofbothhighquantityand(somewouldargue)quality. Althoughmorecommunicationhasalsocreatedtheexchangeofinformationwithdubious qualitythatmostfindobfuscatingandconfusing,thegreaternumberofsourceshas enabledustocomparedataandgroundourselvesmorefirmlyinthetruth. Theonearea,inwhichthereislittlemoretransparencythantherewasin1965,isin theaccountingforfixedandvariablecosts.TheSEC(SecuritiesandExchange Commission)doesnotrequirecorporationstobreakdowntheircostsintofixedand variablecategoriesinrequiredfinancialstatements.Althoughitisstandardpracticeto itemizecostsbetweenfixedandvariable,confusionexistswhenassetcategoriesoverlapand thetwocostsbecomeindistinguishable.Nevertheless,theFASB(FederalAccounting StandardsBoard)andtheSEChasnotmadesuchdisclosuremandatoryalthoughitwould immeasurablyaidtheinvestoringaugingrisk.Thedelineationofsuchcostsiscrucial, becausethegreatestamountofcorporateactivityrestsentirelyontheconceptofoperating riskanditsconcurrentmeasurement,operatingleverage. 220 Sincefixedcostsareoftenassociatedwiththetechnologicalinputsofabusiness- thatismachinery,testequipment,andautomation-manyobserversconsiderthemimplicit inthetypeofproductoneisselling,andalmostbeyondmanagement’scontrol.For example,ifmybusinessrequiresaneutronparticleemitter,andallcompetitorshaveone, thereislittlediscretion;Imustbuytheemitterandpaythefixedcostsformaintenance, etc.Thus,thetypeofbusinessisthegreatestdeterminantoffixedcosts.Thisoperating riskissometimesreferredtoas“economicrisk”;itismostaffectedbythebusinesscycle andthetimingofdemandforthatcompany’sproduct,buttheneedforlong-term commitmentofcapitalisalsoanimperative.Infact,theleveloftechnologycharacterizes theamountoffixedcostsinabusiness.Asemiconductorfirm,forexample,willcompete onthebasisofamuchhigherleveloftechnologythanadepartmentstore.Consequently,it willbedeemed“capitalintensive”andhaveahigherleveloffixedcosts.Almostwithout exception,operatingleverageprovidesthefoundationforstructuringabusiness; marketing,finance,eventheaveragelevelofemployeeeducationisdependentonoperating leverage,andespeciallytheamountsandconsistencyofitscomponentparts. FIXEDCOSTSANDECONOMICS Intheclassicbreakevenpointequation,Profits=Sales-VariableCosts-Fixed Costs.Salescanbefurtherdifferentiatedinto(QuantityxPrice),andvariablecostscanbe differentiatedinto(Quantityx(VariableCost/Unit)).Increasingprofitsisasimplematter ofeitherincreasingsalesordecreasingcosts.Sincefixedcostsreflectthelevelof automation,wouldreducingthemchangethenatureofthebusiness?Issuchreduction possible?Theanswertobothquestionsisaffirmative.Theobjectiveofmanymergersisto “share”fixedcosts,andthusreducetheircostperunitbyincreasingcapacity.However, thebasicproductionprocessstillrequiresthesameinputsandremainsunchanged.Fixed costsinsuchamergerwouldalsoremainstable,butthequantityofunitswouldincrease. Onelookatthebreakevenequationdictatesthatvariablecostswouldincreaseaswell;this 221 proportionalriseinvariablecoststofixedcostslowerstheoverallrisktothemerged company. Another,muchmoreinsidiousreductioninfixedcostsoccurswhenproduction processesquietlybecomeobsolete.Thegreaterthenoveltyandrelativecomplexityof technology,thesmallerwillbeitssupply.Ifsuchtechnologyoffersacompetitive advantage,itwillbeindemand,andcommandahigherpricemerelybecauseofitslimited source.Whenaprocessisgraduallyimprovedovertime,itsrelativecomplexitydecreases, becauseitcanbemoreeasilyduplicated,i.e.,moreindividualsjumpintoaprofitable market,andthesourcesofsupplyincrease.Theresultisalowerprice.Eventually,the supplyofthistechnologywillbegintoeclipsedemandforitbecause“newandimproved” processesreplaceit,andtheoldtechnologybecomesobsolete.Industriesbecome“capital intensive”becausethecompetitive“edge”goestothebusinesseswiththelatesttechnology, whichmustbepurchasedat“anycost”.Suchcompetitionbecomesaratherexpensive propositionforshareholderswhotendtoinvestinthesecompaniesatspecificpointsinthe businesscycleandthenselltheirshareswhenthecostofcapitalclimbsandfurther expansionisunwarranted.Thesecompanieshavealmostinvariablyhighoperating leverage. Highfixedcostsareabarriertoenteringamarket.Theyareafundamentalsign thataprocessrequiressomuchcapital,thatonlyafewenterprisesarewillingtotakethe riskandparticipate.And-usually-suchparticipationconnotesahigherprofitmargin andhighunitprices.Ifthereisonemajorflawtothebreakevenpointequation,itisits failuretocorrelatehigherfixedcostswithhigherprices.Whilepricesaredeterminedby themarket,higherfixedcostsraisethebreakevenpointforsales.Sincesalesareafunction ofbothpriceandquantity,atleastoneofthosevariablesneedstochange.However,when technologyischeapenedtothepointwhereitssupplyisvastandunending,notonlydo fixedcostsproportionatelydecrease,butthepricechargedfortheproductdecreasesas well.Thus,profitsaredrivenbychurningoutmoreofaproduct;revenuesincrease 222 becauseunitquantitiesincreaseandnotbecausepricesrise.Asmoreparticipantsentera market,vendorscutpricestomeetcompetition,whichwillmakeafirm’sinputsless expensive.Thechainofcostreductioncontinues:cuttingafirm’sfixedassetcostswill allowmoreparticipantstoenterthatfirm’sproductmarketalbeitatalowerprofitthan waspreviouslyacceptable.Atthispoint,thefirmcanstampoutaproductthatwas previouslydeemed,“hightech”,almostatwill.Infact,theveryeaseofproductionhas allowednumerouscompetitorsintothemarket,forcingcompaniestocompetewithmore “assetturnover”andlessprofitmargin.Eventually,ifdemandisstableandnotincreasing, moreofagoodcanbeproducedwithlessinvestment,butbythistime,themarketwillbe gluttedandpriceswillhavedeflated. This“negativeeconomicsynergy”isoneofthebailiwicksoftechnologicalchange. Itistheprimereasonthatcompaniesneedtoremainflexibleandwillingtochange.Ifa companyisfixatedonaspecificmethodof“doingthings”,orencumberedbyalimited numberofproductsandprocesses,itwillbedoomedbyalackofdiversification.No companycanaffordtowaitfor“itstimeinthesun”,i.e.,atimeinthebusinesscyclewhich favorsitssector. Thisneedforcompetitivestrategieshasbeenthoroughlyanalyzedanddocumented byMichaelPorter.Hisoutstandingresearchproducedthreegenericstrategiesthatmost companiesemphasizeandshare:1)Costleadership2)Productdifferentiationand3) Focusonnichemarkets.Eachstrategyusesfixedcostmanipulationtogainmarketshare. Inthefirst,costleadership,acompanycanchargelowerpricesbecauseitcutsvariable costsandspreadsfixedcostsovergreaterproduction.Anexamplewouldbeadepartment storethatrentsunusedspacetoanotherbusinesslikeMcDonald’sorStarbucks.A “partnership”arisesbetweenthetwocorporations,andthedepartmentstoreisspreading outthesamefixedcostsovermorerevenue.Thesecondstrategywouldbeexemplifiedbya semiconductorcompanywhocreatesanewchipforamedicalprosthesis,andtestsitwith thesamedevicesitusesforchipsinPCs.Notonlyisthecompanyspreadingfixedcosts 223 overmoreunits,itisdiversifyingitsproductline,creatingalessriskycashflowand changingthecharacterofitsoperatingleverage.Nichemarketsaresometimesspecialized enoughtorequirehigherfixedcostsi.e.,theresearchanddevelopmentinpharmaceutical companies,andnaturallyrequirehigherpricestocoverthosecosts.Asinthecaseof “negativesynergy”,thenichestatuscandeteriorateovertimeiftechnologybecomeseasily duplicable,andmanyparticipantsenterthesamemarket.Flexiblecompaniesmust ordinarilyadoptallthreestrategies,sometimesemphasizingonetotheexclusionofthe othersandsometimescreatingahybridofallthree. THECASEOFCOMPAQCOMPUTER Duringthe1990sthecomputerbusinessbecamesohotlycompetitivethatmany companiesfellbythewayside.Becauseoffrequent“pricewars”,computersbecamevery accessibletotheaverageperson.UserfriendlyoperatingsystemslikeWindows95from Microsoft,createdacommercialenvironmentfortheInternetthatstimulatedshopping, gameplaying,andsocialactivitiesaswell.However,suchaccessibilitycamewithapriceof itsown:manyofthesecompanieswerewellmanagedandyetsawtheirprofitmargins shrink.Fightingtokeeptheirheadsabovewater,consolidationbecameasurvival mechanismratherthananopportunity.Compaq,aPCmanufacturerwhoswallowedup theflounderingenterprisesystemmaker,DigitalEquipmentCorporation,battledto remainviable,onlylatertoseeitselfswallowedup.Thecomputerindustrywaschanging– inwhatseemedlikeananosecond. Whilecomputersarecertainlyhigh-tech,theycanbeeasilyassembledfromaseries ofmodules.Themodulesthemselvesarerelativelysophisticated,butonedoesnothaveto bean“electronicswhiz”toconnectseveralofthesemodulestogetherandforma computer.Increasingdemandforqualitycomputers(onesthatdidnotloseinformation whenyouturnedonthevacuumcleaner),withgreater“meantimebetweenfailures”, meantthatmachinerynolongerhadtobereplacedbecauseit“woreout”liketennisshoes. Infactasstoragespacebecamegreatenoughtoeclipseourabilitytouseit,therewasless 224 ofaneedtoreplacecomputerswithan“upgrade”.Foryears,computerspeedneededto keeppacewithevermoresophisticatedprogramming,andpartswerenotstandardized enoughtobeinterchangeable.Inotherwords,puttingonemoduleontopofanotherwas likeputtingChevypartsintoaFord:Itmightgetyoutothecornerstoreandback,butnot muchfarther.Whenoperatingsystemsbecoming“userfriendly”andprogramsbecame compatible,computersmarchedrightoutofthescientificlaboratoryandintotheliving room.Computerswerenotonlyaccessible,theywerearequirement. Frequentchangesintheproductionprocess,coupledwithanincreaseinmarket participantsledtoaninflexibilitywithinthelargestmanufacturers.Custombuilt computersbythelocalelectronicsstorecouldcompetewithhugePCmakerswhohadthe addedburdenofhighfixedassetsandcosts.Forecastingdemandbecameahitandmiss art,andunsoldinventoryforcedCompaqtoadoptanewdistributionmodelcalled,“the distributionallianceprogram”;computerswouldbemadetoorderandnottoforecast. Inevitably,computermakerswouldcompeteonthis“customstyle”basis,almostlikea servicecompany.ManufacturerslikeDellstoppedsellingcomputersinstores,and computergiant,Gateway,endedupclosingthemajorityofitsdistributioncenters. THENATUREOFCOSTSANDMARGINS Thetelltalesignofadeterioratingeconomicpredicamentissimplyfallingoperating margins,measuredbyOperatingIncome/Sales.Notethatoperatingincomeinthis instanceisbeforedepreciationandamortizationhasbeendeductedtoderiveEBIT (earningsbeforeinterestandtaxes).Thebreakevenpointequationbestestablishesthis relationship: 225 Table9-1 FIXEDCOSTS(FC)=100 SALES(S)=500 QUANTITY(Q)=100 PRICE(P)=5 VARIABLECOSTPERUNIT(VC/UNIT)=3 (PxQ)-Q(VC/U)-FC (100x5)-(100x3)-100=100 OPERATINGMARGIN=100/500=20% Aslongasfixedandvariablecostsarestable(thatisfixedcostsremainthesameand variablecostsareastablepercentageofsales)operatingmarginswillincreaseassales increase. Table9-2 FIXEDCOSTS(FC)=100 SALES(S)=1000 QUANTITY(Q)=200 PRICE(P)=5 VARIABLECOSTPERUNIT(VC/UNIT)=3 (PxQ)-Q(VC/U)-FC (200x5)-(200x3)-100=300 OPERATINGMARGIN=300/1000=30% Noticethatquantityistheonlyvariablethatchanged.Aslongasmoreunitsarebeing sold,variablecoststhatriseasapercentageofsales,donotadverselyaffectmargins.In thenexttable,amere16.66%riseinvariablecoststotallydilutestheeffectofaone hundredpercentincreaseinsales:Operatingmargindeclinesbackto20%. 226 Table9-3 FC=100 S=1000 Q=200 P=5 VC/U=3.5 (PxQ)-Q(VC/U)-FC (5x200)-(200x3.5)-100=200 OPERATINGMARGIN=200/1000=20% Overtime,operatingmarginsrisewhenthepercentchangeinsalesexceedsthe percentchangeintotalcost,regardlessofthemixturebetweenfixedandvariablecosts. Thetendencyinmostbusinessesisforfixedcoststobecomeasmallerpercentageoftotal costsoverthelifespanofaproduct.Such“parity”wouldnotbedestructiveofmarginsif onlyvariablecostsdidnotrisetocompensateforthereduction.Infact,variablecostsrise asapercentoftotalcostsandevenaccelerateasfixedcostsdecline:increasedcompetition “raisesthebar”andeachsaleinvolvesmoredistribution,advertising,packagingand marketingexpenses.Theconceptisobviouswhenonewalksintoanysupermarketand pricesgenericcornflakesvs.anamebrand.Thetwocerealsarenearlyidenticalin productioncosts.Fixedcostsoftheprocessaresimilar,althoughsome“quality”premiums arebuiltintothenamebrandthatwouldmakeitslightlymoreexpensive.Therealcost differencesareinsuchthingsasTVadvertising,breadthofdistribution,endorsementsetc. thatpushupthecosttoselltheproduct.“No-Brand”cornflakesneverhaveaprizeinside inordertokeepcostslow. Takentotheextreme,whentheincreaseintotalcostsisgreaterthantheincreasein sales,operatingmarginssuffer.Insomeindustries,fixedcostsandnotvariablecostsare theculprit.Theairlineindustry,forexample,hasneverbeenprofitableforanextended periodoftime.Theircollectiveinvestmentinfixedassetsislarge,buttheirproductive capacity,(seats)islimited.Whenaplane(afixedasset)carriesitscapacityof300people,it 227 cannotaddcapacityorspreadthefixedcostsofaplaneamongagreaternumberofunits. Thatisthenatureofthetechnology,andalsothereasonwhyseatsarecreatedtoconform tothedimensionsofarunwaymodel.Mostairlineshaverespondedbycuttingfrillssuch aspillows,peanuts,andmeals,inanefforttomakeupinvariablecostswhatitmustspend infixedcosts. InCompaq’scase,ithadfinishedadealtobuyDigitalEquipmentandinvested heavilyintheInternet(CMGIandAltaVista).Ultimately,Compaqwasplaguedbyan overcapacityandresortedtotheaforementionedchangeinitsdistributionmodel.While theseinvestmentswereimmediatelyextraneoustotheproductionofPCs,theyshowedthe capacityofCompaqmanagementtoadapttoasituationthatwasrapidlygettingoutof hand.The“higherups”couldseethewritingonthewall-theyneededtodiversifyanddo sorapidly.However,bothfixedandvariablecostswererisingfasterthansalable production,becausecompetitionhadloweredpricesandfloodedthemarketwithPCs.The onlywaythatCompaqcouldhavemaintaineditsmarginswouldbetoeitherexpand productionandsellatdrasticallyreducedprices,orcutitsvariablecostperunitenoughto bringupoperatingincome.NeitherpathwasfeasibleanditwasnaturalforCompaqto lookforamergerpartner,whichitfoundlaterinHewlettPackard. FIXEDCOSTSANDTHEBREAKEVENPOINTFORSALES Whilehigherfixedcostsactasabarriertoentry,theyalsoentailmoreriskbecause raisingthemraisesthebreakevenpointforsales.TheidealsituationforCompaqwould havebeentoeffortlesslyincreaseitscapacityandincurthecostincreasesthatwere normallycontingentonexpansion-bysellingmoreunits,albeitatalowerprice.However, Compaqneededtoincreasemarginsandcouldnotrelyonsalesalone.Suchastrategy wouldhaveraisedvariablecostsinproportiontofixedandmarginswouldhavestabilized ataverylowlevel.Nocompanycancutitsoperatingmargininhalfandcontinueto operateunscathed.Thus,themainproblemwithCompaq’scostswerenotariseinfixed costs,whichactedsolongasabarriertoentryandwerenowdiminished,butadeclinein 228 pricingpowerthatwasderivedfromanaccelerationinvariablecoststhatreplacedfixed costs. Table9-4 FC=100 FC=200 VC/U=3 VC/U=3 Q=50 Q=100 PROFIT=0 PROFIT=0 PRICE=5 PRICE=5 Notethatwithoutpricingpower,acompanywouldhavetoraiseitsunitquantitywhen fixedcostsrise.Acompanywhocanraisepriceswithoutcustomercomplaintisinafar superiorpositionthanthosewhomustdowhatthemarketdictates.Itisfareasiertoraise apricethantogearupproductionprocessesandchurnoutmoreunits,andyettheresultis thesamewhenfixedcostsareraised;salesmustrisetomeettheincrease. Theriskofraisingfixedcosts(orevenbuyingthestockofacompanywithhigh fixedcosts)stemsfromtherelationshipbetweensalesandoperatingincome.Whenhigher fixedcostsareincurred,asmallchangeinsaleswillcausealargechangeinoperating income.Whilethisscenariomaysoundlikeaterrificprofitmakingventure,suchprofits areonlycontingentuponthestabilityofsales,whichhasanalmostnegativecorrelation withveryhighoperatingleverage.Forexample,autilitycompanywhohashigherfixed costs(turbinesetc.)willalsohaveahighdegreeofsalesstabilitybecausetheirproduct (electricity)hasademandduringdownturns.Ontheotherhand,asemiconductor companywhomayhavethesameleveloffixedcostsfacesamuchmoredauntingtask becausethedemandforitsproductismuchmoreelastic:bothbusinessesandconsumers tendtodolesscapitalspendingtoupgradeequipment(orlifestyle)duringrecessions.In factthevolatilityofstockpricesissomewhatdependentontherelationshipbetweenthe amountoffixedassetsandthestabilityofdemand.Thosecompaniesfacedwithboth higherfixedcostsandunstabledemand,willtendtohavehigherbetas,andanunstable 229 stockpriceaswell.Thus,fixedcostsareadoubleedgedsword;innewerindustries,they actasabarriertoentryandpumpuppricesbecausethesupplyoftechnologyislow.But- whenpricesfallduetocompetition,risingvariablecostsdepleteanyprofit.Theywillalso destabilizeoperatingprofitbecausehighercapitalexpendituresareneededtomaintaina highleveloffixedassetswhichwillalsorequireahigherlevelofsales.Sinceitisrareto havesuchidealconditionsoccursimultaneously,thesecompaniestendtodoenormously wellwhentheirrespectivesectorsarefavoredandthenlagbehindthemarkettherestof thetime. COMPAQCOMPUTER:THERESTOFTHESTORY By1997,Compaqbegantoseemarginsdecline.Thecompanycouldnot compensateformarketpricedecreaseswithadditionalsalesvolume.Desperatelyseeking toincreasesalesandprofits,itacquiredUNIXoperatingDigitalEquipmentCorporationto gainafootholdintheenterprisesystemsmarket.Inwhatwastermedthelargestcomputer companyacquisitioninhistory,Compaqaddedittoitspreviouspurchase,Tandem Computer.Enterprisesystemswere“bigticket”itemsincomparisontothecommoditized PCindustry.Compaqsawthemasasteppingstonetoregainingoperatingmargins.To understandthegravityofthesituation,twomeasurementsneedtobeintroduced.Thefirst iscalledthe“capitalintensityratio”,whichmeasurestheimpliedamountofcapitalafirm needs.Itismerely,TotalAssets/TotalRevenuesanddiscerningstudentswillnoticethatit istheinverseofthe“assetturnover”ratio.Whenitiscomparedtobothyeartoyear changeandanindustryaverage,ittendstorevealtheamountoffixedassetsand consequentfixedcostsinacompany.Thesecondmeasurementisthenoncurrentasset ratiowhichismerely,(TotalAssets-Currentassets)/TotalAssets,andlikethefirst, measurestheamountoffixedassets,albeitinamoredirectmanner.Althoughthisratio canbeobfuscatedbylargeamountsofintangibleassetslike“goodwill”,itcanofferagood comparativesnapshot;itmustbereiteratedthattheinvestormustdeciphertheamountof fixedassetsinacompanyassuchinformationisnotdirectlyproffered. 230 Table9-5 COMPAQ YearofDigital Equip.Purchase YEAR 1996 1997 1998 CAPITAL INTENSITY 0.58125 0.5951 0.7395 NONCURRENT ASSETRATIO 0.12908 0.1787 0.3412 TheDigitalEquipmentpurchasewaspaidforwithoverninebilliondollarsincash andstockforwhatwas,atthetime,thelargestacquisitionincomputerhistory.The “higherups”knewtheyweretakingamajorrisk,butrightfullybelievedthatapathto highermarginsincludedmoreoperatingriskwithhigherfixedcosts;sucha“shotinthe arm”wouldreformulatethecompanyawayfromhighquantity,lowleveragePCsand towardanicheproduct-hugeenterprisesystemsthatcouldrunamajorcity.Notall computersystemsarecreatedequal.Withthepurchaseoflargeamountsoffixedassets, Compaqhadamonumentaltasktointegratethemefficiently.TheWindowsbasedPCs wererapidlybeingreplacedbytheproverbial800poundgorilla-systemsthatwereboth complexandmassive. Hadtheacquisitionbeenconsummatedfouryearsearlierin1994,Compaqmay havehadachanceofridingoutwhatwasabouttooccur:oneoftheworstmarket correctionsinmodernhistory.Techstockswerecompletelyslammed.Whatwas anticipatedtobea“normal”tentofifteenpercentdecline,increasedtofiftyandeven seventypercentformanystocksinthetech-heavyNASDAQ.By2001,manystocksthat hadsoldforeightyorninetydollarsasharecouldbepurchasedforaslittleastendollars. Compaq’svaliantattempttoreviveitsprofitmarginswasdoomedtofailure.Inessence, thehigherfixedcostsoftheDigitalEquipmentmergerentailedmorepotentialprofitbut moreoperatingrisk.Sincethebreakevenpointforsaleswasrisingatthesametimethat 231 theeconomywasindecline,thehigheroperatingleveragejeopardizedCompaq’sposition: Compaqneededtosellhighprofitenterprisesystemsjustastheeconomywasgongintoa tailspin.ThelowerleveragefromPCmanufacturingcouldnothelpweatherthestorm becausePCswereadiscretionaryitem,i.e.,theleveragewasnotlowenoughtomake consumersbuytheseproductswhen“moneywastight”.Ultimately,amergerwithHewlett PackardbecameCompaq’ssavinggrace.Theparadoxofseeingmarginsdecreasebecause oftechnologicalchange,andthenanattemptathigherleveragefailbecauseofchangesin thebusinesscycle,shouldnotbelost.WhathappenedtoCompaqwas“theperfectstorm” inoperatingrisk. OPERATINGLEVERAGEANDPREDICTION Tounderstandoperatingleverageandtheconceptofasmallchangeinsaleshaving amagnifiedeffectonoperatingincome,onemustrealizethattheratiohasmultipleuses.It notonlydisplaystheriskofhigherfixedcosts,butcanpredicttherelationshipbetween salesandoperatingincome.Thebasicratioisformedbysalesminusvariablecostsinthe numerator,oftencalled“thecontribution”,andthendividedbyEBIT(earningsbefore interestandtaxes)inthedenominator.Itshouldalsobeobservedthatthecomponentsof EBITaresalesminusvariablecostsminusfixedcosts(S-VC-FC).Thetotalratioisthus: S-VC/S-VC-FCorS-VC/EBIT.Simplystated,fixedcostsdeterminethemagnitude oftheratio,becausetheothervariablesoccupyboththenumeratorandthedenominator withthesamerelationship. Thegreat“mystery”ishowtheseconcretenumbersactuallytransforminto percentagechangevariableswhenwecompareoneyeartoanother.HowcanS-VC/ EBITturnintothebetaratio,%∆ ∆∆ ∆OperatingIncome/%∆ ∆∆ ∆Sales?Theexplanationis:if variablecostsremainthesamepercentageofsales,andfixedcostsremaintruly“fixed” (theydonotchange)thentheratioofS-VC/EBITattheendofthisyear,willmirror%∆ ∆∆ ∆ OperatingIncome/%∆ ∆∆ ∆Salesattheendofthenextyear,andthushaveapredictive relationshipwiththosevariables.Sincetheratioisacompendiumofperhapsthousandsof 232 internalbreakevenpoints,itisdebatablewhetheroperatingleverageforanentire companycanbepreciselymeasured.Thediscerningreaderwillnoticethattheconcrete formoftheratiocanneverdropbelow“1”andyetmanycompanieshaveprocesseswhere operatingincomepercentchangesarealwayslessthansalespercentchanges.Infact,just bydiscontinuingonebusinesssegmentandaddinganother,acompanycanradically changeitspercentages.Therefore,whatweobserveasinvestorswillrarelyconformtothe academicideal.Sincecostsareneverpublishedintheformatof“fixed”and“variable”, butmorelike“costofgoodssold”or“administrativeandsellingcosts”,theinvestorneeds todoaclever“endaround”andinferoperatingriskfromstatisticalrelationships.By observingthemeanandvarianceofoperatingmargins,sales,operatingincome,andcapital intensity,overafiveyearperiod,wecanbetterunderstandtherisksofanyfirm’scash flow.Someresearchreportsmaycontainspecificoperatingleverageinformationforan industryoracompany.Whenwefindsuchinformation,wecomparethestabilityofactual percentagestothisidealandtrytocoordinatetheleveragetrendwiththetypesof productsafirmissellinganditsplaceinthebusinesscycle.Thesevariablesaredifficultto predict,butahandfulof“astute”investorssawtheCompaqdebaclecomingandacted withforesight.Itisnotimpossible. CHARACTERISTICSOFOPERATINGLEVERAGE WhilehigheroperatingleverageultimatelyincreasesthevariabilityofEPS,justas financialleveragedoes,mostcompaniestrytominimizethisvariabilitybybalancingthe twotypesofleverage.Thus,ifacompany’soperatingleverageishighitwillhavea tendencytolowerfinancialleveragebyfinancingwithequity;moresharesissuedactually diminishtheriskofunstablecashflow.Conversely,companieswithlowoperatingleverage haveachanceofraisingEPSsimplybyfinancingwithmoredebt,whichtheycan presumablydoifoperatingincomeisstable.Creditorsfavorthosecompanieswithstable revenuesbecausetheriskofdefaultislower.Consequently,banksforexample,willcharge the“primerate”tothecustomerswhoaremostcapableofpayingoffandrenewingloans- 233 firmswithlargeandstableoperatingcharacteristics.Thecyclecontinuesbecausethese customerscanusemoredebtintheirrespectivecapitalstructureswheninterestratesare lower.Thusoperatingleverageisthemajorsourceofmostothertypesofriskincluding: financialleverage,politicalrisk,foreignrisk,inflationriskandevenrandomeventrisk. Thetypeofproductoftendeterminesthescopeofrisk.Theelasticityofdemand determinesthepercentchangeindemandbasedonapercentchangeinprice.Agricultural products,forexample,usuallyhavelowoperatingleverageandmustbeproducedin quantitytomakeaprofit.Sincepriceisdeterminedbyacompetitivemarketofmany buyersandsellers,demandisunresponsivetopriceandtheseproductsarealmostinelastic. Ontheonehand,mostpeoplewouldbuythesameamountofbreadifpriceswereto increasebytenpercent.Ontheotherhand,ifpricesquadrupled,astheywouldina famine,peoplewouldbuylessbread.Elasticityisnotcalculatedfornonnormalperiods, andthenearinelasticityofbreadholds.Analogously,productswithhighelasticityimplya higheroperatingleveragealthoughtherelationshipiscertainlynotcauseandeffect. Vacationresorts,plasmatelevisionsandhigh-techexerciseequipmentallhavehigh elasticityandwillseelargeshiftsindemandwhenpriceschange.Consequently,ahigher fixedcostperunitwillbeimplicitintheirdevelopment.Stocksthatpertaintothese productswilldowellinthebrieftimetheirrespectivesectorsarefavoredandthenlag behindothersectorsintheinterim.Diversificationcansmoothoutthisrisk,buteachfirm mustmaintainacorecompetency,i.e.,itwouldnotbodewellforamakerofexpensivegolf clubstostartsellingbatteries. Anothercharacteristicofoperatingleverageistheabilitytocombinedifferent operatingleveragesandchangeariskprofile.Intheaboveallusiontobatteriesandgolf clubs,thecombinationseemedcontrivedandawkward.However,duringthe1950sand 60s,suchcomboswerefrequentlytriedand“conglomerates”wereformedforthe expressedpurposeofdiversifyingawayrisk.Unfortunately,asprocessesoverlapped, inefficienciesoccurredthatdiminishedsalesandprofitforeachproduct.Corporations 234 wereunsureofwhatbusinesstheywerein.A“good”productfromanaveragecompany wasdiscardedinfavorofa“great”productfromagoodcompany,andmanyfirmsbegan torefocusduringthe1980sand90s.Such“reengineering”,oftenrequirednumerous dislocationsanddownsizingsbutincreasedtheoverallflexibilityofthecompany. “Outsourcing”and“partnerships”becamethebuzzwords,andinsteadofhavingagolf clubcompanyowningabatterycompany,wewillwalkintoWal-Martandseeakioskset uptoserveMcDonald’shamburgers.Insteadofplayingtenniswiththesalesmanager,a salesrepmayplayvideogameswithaprogrammerinIndia.Thediversityincashflows havebeenoffsetbydiversifyingtheprocesses,althoughfirmsstillattempttodiversifysales withinanarrowerframework,a“niche”. Finally,capitalbudgetingtimeframesareanimportantcharacteristicofoperating leverage.Whenaventurecapitalisttakestheriskofinvestinginahigh-techoperation,the payoffsneedtobemorerapidthaninalowoperatingleveragescenario.Onereasonthat Americahasfailedtoinvestin“alternativeenergysources”(asof2008),isbecausethe payoffisuncertain.Noinvestorwantstotieupcapitalforfiveyearsinthehopesof doublinghisorhermoney“somewheredowntheroad”.Thereasonthatamajorproject liketheinterstatehighwaysystemwascompletedissimplybecauseitwasundertakenby thegovernment,whocanassumetheriskofahugefixedcostoperation.Privateinvestors expectareturn,andwithhigherfixedcosts,thatreturnneedstobegreaterwithatleast somemodicumofcertainty.Whiletaxbreaksandincentivescanstimulatesomeaction, massiveprojectsrequiretheorganizationofmanyinvestors,andtoomanyotherprofitable investmentopportunitiesaboundforthattooccur. Moreover,investmentincapitalintensivefixedassetsinvolvestheriskof obsolescence.Considerthatfixedassetsofteninvolveprocesseswithhightechnologythat canrapidlybecome“newandimproved”andreplaceexistingtechnology.Withouta rapidpayoff,theinvestorrisksbeing“eatenalive”bythecompetitionwhomayalready havethenewtechnologyinplace.Thisbehaviorisreadilyobservableinwhichstocksget 235 “daytraded”andinwhatperiods.Whilethemantraof“buyandhold”servesawell thoughtoutportfolio,manytraderswantto“getinandgetout”.Thereasonissimple.Ifa pharmaceuticalfirm,forexample,ownsthepatentonamiraclecure,itwillbeashorttime beforeitisduplicatedwithadifferent,“designer”formulabysomeothercompany. Traderswanttogetinandthenleavethestockbeforethetechnologybecomesobsoleteor duplicated. OPERATINGTRENDSANDREVERSALS Thefollowingsectionsaregoingtodiscussevaluatingthesizeandstabilityof operatingcomponents.Whilestatisticsgiveusinsightintothequalityofsalesand operatingincomeatanymomentintime,theycannotpredictthefuture.Wecanobserve increasingrisk,butwecanneversaythatacompanyisaboutto“implode”.Infact,oneof thehardestdecisionstomakeistofindacompanywithoutstandingsizeandstabilityof operatingcomponents,andthenviewaperiodofdecliningmarginsasanarbiterof investment;trendspottingrequirescoordinationbetweencompany,products,industryand theoveralleconomy.Similarly,ifweseeincreasingmarginsoveralongperiod,weknow thatsuchtrendsstopwhenmorecompetitionentersthemarket.Sometimes,onehasto enterthemarketbeforeatrendevenstarts.Inaneedforcertainty,itistemptingto formulatetrendrulessuchasneverinvestinginacompanythathasathreeyeardeclinein marginsorlookingforareversalafterthreeyears.Suchrulescanblindaninvestortobig opportunities.Thetrendinoperatingcomponentsmustbeseeninthecontextofanoverall movementtowardatargetcapitalstructure;isolatingtrendsinoperatingmarginwillbe selfdefeatingbecausetheinvestorwillbe“chasingearnings”,andnoanalystcanpredict whensuchmomentumendsbyobservingoperatingcharacteristicsbythemselves. THEQUALITYOFANOPERATINGMARGIN Forlong-terminvestment,thesizeandstabilityofanoperatingmarginisoneofthe determiningfactors.Weapplyboththemeanvariancemethodandthecoefficientof variationtofiveyearsofoperatingmargindata.Whentwoinvestmentoptionsdifferin 236 thismeasurement,weusuallychoosetheonewiththelowestcoefficientofvariation.The reasonthatthisissuchatelltaleratioisthatmostfirmswillgooutoftheircollectiveways toensurethepreservationofthisfigure.Indeed,wesawhowCompaqevenmadeanine billiondollaracquisitiontodoso.Thisratiodefinesthecompany-itsmarket,itsfunding, anditsresponsetochangesinthebusinesscycle.Thefollowingtablespresentthe operatingmarginsoftwocompaniesalongwithameanvarianceandacoefficientof variationforeach: Table9-6 INTUIT YEAR 1997 1998 1999 2000 2001 Operating Inc. 107 70.5 254 200 265 Sales 599 593 848 1094 1261 Operating Mar. 17.86% 11.89% 29.95% 18.28% 21.02% Table9-7 Statistic Methodology Result Mean AverageofOperating Margin 16.26 StandardDeviation Squarerootofvariance 11.09 MeanVariance MeanminusStandard Dev. 5.17 CoefficientofVariation StandardDev.dividedby Mean 0.6819 237 Table9-8 H.J.Heinz YEAR 1997 1998 1999 2000 2001 Operating Inc. 1096 1834 1412 1575 1282 Sales 9357 9209 9300 9408 9430 Operating Mar. 11.71% 19.92% 15.18% 16.74% 13.59% Table9-9 Statistic Methodology Result Mean AverageofOperating Margin 15.43 StandardDeviation SquarerootofVariance 3.129 MeanVariance MeanminusStandard Dev. 12.14 CoefficientofVariation StandardDev.dividedby Mean 0.2028 Heinz’s’marginsareassteadyandslowasitscatsup,whichiswhythecompanyhas becomeaninstitution.Foraquickanddecisivesnapshotofoperatinghistory,thereisno betteranalysisthandoingacoefficientofvariationonoperatingmargindata.Foramuch moreminuteexamination,anacquisitionspecialistcoulduse60monthsofdata;moredata pointstranslatesintomoreaccuracy. ARISKYPROPOSITION:CONFIDENCEINTERVALS Anydecisionmakingbasedontheextrapolationsofdataisveryrisky.Asinvestors, ourinformationismuchmorelimitedthanthatofprofessionalanalysts.Ourfiveyear timeframescontainalimitednumberofdatapoints,andfuturetrendsrarelymirrorpast performance.Sowhymaketrendpredictions?Whytakethechanceandhaveto“biteour lips”whenwearewrong?Statisticallyderivingapredictionintervalisnotthesameas makingaprediction.Wearemerelyimplying,thatfromourverylimiteddata,thereisa 238 certainpercentagechanceofaneventoccurring.Weaddthisinformationtomanyother indicatorsbeforemakingadecision,andcollectively,weconsidersmallsamplepredictions tobetheleastreliable.Inourquestfora“ballpark”figure,wearelookingforasmany diverseindicatorsaspossible.Someofthesewillbecontradictory,andsowewillhaveto weighteachmeasurementonthebasisofreliability.Suchreliabilityisestablishedbythe numberofdatapointsandthepropermethodology.Additionally,somedataissoskewed thatnomeasurementismeaningful;inthatcasewemayuseanunreliableindicatoronlyto confirmareliableoneRegression,forexample,isproperlyappliedtodatathathasa “normaldistribution”.Wemayuseittoconfirminferencesaboutdistributionsthatare anythingbut“normal”,butweshouldnotmakeittheprimarycriteriafordecision making. Toformconfidenceintervalsthatexpressvolatility,wetakethefiveyearmeanand samplestandarddeviationsofoperatingmomentum,whichis%∆ ∆∆ ∆OperatingIncome/% ∆ ∆∆ ∆Sales.Toachievethisamountofdata,weneedsixyearsofconcretefigures,asweare determiningfiveyearsofpercentagechanges.Wethenformconfidenceintervalsby multiplyingthesamplestandarddeviationbyaTscore(seeStatisticsPrimer)andthen dividingbythesquarerootofthesamplesize(5).Thisfigureisthenaddedto(or subtractedfrom)themean,thusformingtheinterval.Ifourcurrentfigureliesoutsidethe interval,weknowthereisaspecificpercentagechancethatitwillmovebacktowardthe mean.Sincedataislimited,wetrytousethe99thpercentileasarulebecauseitwillbeso extreme,andweneedtotrustasmallsample.Afigurethatwouldviolatethatextremea constraint,wouldmostlikelyreverttothemean-butevenmoreimportantly,itsuggestsan upheavalintheproductionprocessandgivesasignalforfurtherinvestigation.The followingexampleisverytypicalofasmallsamplesize: 239 Table9-10 VARIABLE YEAR 2000 2001 2002 2003 2004 Operating Income% 20 29 17 14 19 Sales% 15 18 23 7 24 Operating Momentum 1.33 1.64 0.739 2 0.79 Thestudent’sTscoreisverymuchliketheZscoreadaptedtoasamplesizeunder30.Like thesamplestandarddeviation,itautomaticallyadjustsfortheexpectedvolatilityfoundin smallersamples.Moreover,liketheZscore,itcoversapercentageareaofacurveby multiplyingaspecificnumberbythestandarddeviation.Therefore,themeanplusor minusanumberofstandarddeviationscoversapercentageareaofthecurve.EachT scorealsocorrespondstothesizeofthesample,withasmallersamplesizeindicatinga higherTscore.Ourownsamplesizeisfive(forthenumberofyears),andthefollowingT scorescorrespondtothegivenpercentage: Table9-11 PERCENTAGEN=5,4DEG. FREEDOM TSCORE 99% 4.604 95% 2.776 90% 2.132 80% 1.533 50% 0.741 Forthesampledata,weneedtodeterminethemeanandsamplestandarddeviationwhich is1.294and0.53904respectively.Wedividethesamplestandarddeviationbythesquare 240 rootofthesamplesize,5,or0.53904/√5.Wemultiplythisnumberbythepercentage confidencelevelandcorrespondingTscore,andthenadditto,orsubtractitfromthe mean.At99%confidence,thisintervalwouldbe:1.294± ±± ±(4.604)(0.241)=1.294± ±± ±1.11or 0.184≤ ≤≤ ≤X≤ ≤≤ ≤2.404.Wecannotsaywithcertainty,thatifthenextoperatingmomentum indicatorisnotwithinthesenumbers,itwillrevertbacktothemean,butsinceitwill violatetheconstraints(0.184and2.404),wemaywanttodosomeextensiveexaminationif weareseriousaboutinvestinginthiscompany. Therelationshipbetweentheoperatingcomponents,salesandincome,determines thecharacterofthecompany.Largepermanentincreasesinthesecomponentsarerare,so thereisaheavyreversiontothemean.Anothermethodwecanapplyistodeterminethe individualgrowthratesforsalesandoperatingincome,dividethem,andthencreatean operatingmomentumoutofthequotient.Thismeanisnotatruemean,buta characteristicmeanthatcanbeusedtocompareindustriesorindividualfirms.The methodologyisasfollows: Table9-12 GlaxoSmithKline Year 1999 2000 2001 2002 2003 Operating Income 2702 5190 5508 8498 7365 Sales 8490 18079 20489 21312 21441 Todetermineanestimatedgrowthrate,weformaratiooftheneartermfigureinthe numeratorandthefartermfigureinthedenominator.Wethen"exponentiate"thisfigure withtheinverseofthenumberofperiodsbetweenyears(4).Theinverseof4is1/4or0.25. Thusforsales,thegrowthratewouldbe:(21441/8490)^0.25or1.2606.1.2606isthe growthrateandifwesubtract“1”,wedeterminethepercentagegrowthrateof26.06% peryear.Foroperatingincome,thefigureis(7365/2702)^0.25or1.2849fora28.49% growthrate.Dividingthetwo,wedeterminethecharacteristicoperatingmomentum: 241 (28.49/26.06)=1.0932.Noticethatwedidnotproclaimthepredictedgrowthratesof eithercomponent,butusedtheresultstoformamoremeaningfulnumber;theprobability thattheoperatingmomentumwillbearound“1”isfargreaterthantheprobabilityofsales increasingbyexactly26.06percentinthefollowingyear. OPERATINGBETA Thecapitalassetpricingmodel(CAPM)isveryadaptable.Whilenotprecisely accurate,themodelisflexibleenoughtocompriseawidevarietyoffinancialassetsandwill relatethemtointeractionsbetweentheriskfreerate(government),themarket(many buyersandsellers),andtheindividualasset(beta).Althoughsomewouldarguethatthe modelisinherentlyunstableandrevealsonlyafleetingglimpseoffinancialtruth,others woulddeclarethatthisvolatilityreflectstherealityofconstantfinancialchange. Acorporationcanbeviewedasaportfolioofassets,eachwithitsownresponseto economicandfinancialrisk.Thus,onedivisionofanoilcompany,explorationfor example,hasadifferent“beta”thananotherdivision.Thecollectivesumofbetasfrom eachdivision,weightedbyassetvalue,willmakeuptheoverallcorporatebeta.Canwe filteroutthefinancialriskandfindabetathatpertainstooperationsalone?This theoretical“unlevering”ofthecompanywaspursuedbyboththeteamof Miller/ModiglianiandHamadawellbefore1975.Theyfoundthattheycouldextractthe financialriskfrombetaiftheyfactoredinthedebttoequityratioaswellasthetaxrate. Themathematicaleaseofdoingsodependedonthelinearityofthefunction,andthe CAPMisconvenientlyastraightline.Nevertheless,evena“ballpark”riskmeasurement foroperationsopensupinnumerableotheroptionsbecausesalesandprofitcanbebetter relatedtothecostofequity.Suchameasurementcanbeusedformergersand acquisitions,performanceevaluation,andevenfortheevaluationoffirmsthatarenotin themarketandwhichneedanestimatedbetaforcomparison. THEUNLEVEREDBETAEQUATION Theconsensusequationissimpleandisjust: 242 UnleveredBeta=CurrentBeta/(1+[(1-T)(MVDebt/MVEquity)]).MVis“market value”,whileT=taxrate.Usingbookvaluesinplaceofmarketvaluesistheoretically improper.Inourexample,wehavemadeupascenariowherebookandmarketvaluesare thesamefortheeaseofcomputation.Inprofessionalriskmanagement,anydeviationfrom theidealwouldbeconsideredunsound. THEARDCOBARBELLCOMPANY:ANEXAMPLEOFUNLEVEREDBETA TheArdcoBarbellCompanyneedstoexpand.Thereislimiteddemandforold fashionedweightsalthoughthecompanykeepsitsleveragehighbyofferingnicheproducts. Thecurrentbetais1.4andthecompanyuseslittledebt,about10/40,or0.25debttoequity. Themarketvalueisthesameasbookequityinthiscase–40million.Withataxrateof35 %,whatwouldbeitsoperatingbeta? Table9-13 VARIABLE BETA 1.4 DEBT/EQUITY 0.25 TAXRATE .35OR35% UNLEVEREDBETA 1.4/(1+[(1-.35)(0.25)])=1.204 TheArdcoBarbellCompanytakesontwoacquisitions: • 1)Achainofgymswithabetaof1.6andadebttoequityof40/20or2.Themarket valueofthisenterpriseis20million. • 2)Atennisracquetcompanywithabetaof1.1,amarketvalueof10Million,anda debttoequityof5/10or0.5.Bothtaxratesare35%. Thelogicbehindbetasisthatiftheyareaddedlinearly,wecandeterminetheoperating betaofeachunit;allweneedtodoistoweighteachbetabyitsassociatedmarketvalueto deriveacombinedtotal. Step1:Findtheoperatingbetasofthenewacquisitions: GYMCHAIN:1.6/(1+[(1-.35)(2)])=0.696 TENNISRAQUETCOMPANY:1.1/(1+[(1-.35)(0.5)])=0.8301 243 Step2:Multiplyeachoperatingbetabyitsmarketvalueweight. Thecombinedvalueis40+20+10=70 (40/70)(1.204)+(20/70)(0.696)+(10/70)(0.8301)=1.0037.Thisistheoperatingbetaofthe newcompany. Step3:We“lever”thebetabackuptoreflectthenewdebtposition.Noticethatthe combinedcompanyhas10+40+5=55indebt.Forillustrativepurposes,weassumethese companieswerepurchasedwithcash,butArdcocouldhaveincurredmoredebttobuy thesecompaniesandwecouldhaveadjustedthatpropositionintoouranalysis.The equationis: (UnleveredBeta)(1+[(1-taxrate)(MVDebt/MVEquity)])=NewCombinedLeveraged Beta (1.0037)(1+[(1-.35)(55/70)=1.5163 1.5163isthenewlycombinedbeta.ItishigherbecauseArdcohadtoassumethegym chain’sdebt.ThebenefitofdiversificationcanbeobservedinArdco’smuchlower operatingbeta(1.0037vs.1.204),whichtheyhopewillcontributetopayingoffthelarger debtobligation. The“topdown“approachtobetawouldmultiplyeachseparateleveragedbetaby weightedmarketvaluetodetermineafinalbeta.Inthiscase,thatbetawouldbe (40/70)(1.4)+(20/70)(1.6)+(10/70)(1.1)=1.412.Whichapproachismorevalid?Whilethe “topdown”approachismoreconventional,itassumesthatseparateentitiesarecorrelated bymarketvalueandbetaresponse.Weknowfrompreviouschaptersthatthecoefficient ofdetermination,R 2 ,isusuallyavaluebetween0.2and0.4,andthatthe“alpha” componentsometimessupersedesthebetacomponentinimportance;weoftenassumethat thecorrelationfactorisstrongerthanitreallyis.Sinceweusetherequiredrateofreturn todetermineacostofequityandnotforpredictions,theeffectofalowcorrelationgets bufferedintheanalysis.However,the“topdown”approachassumesthata“singleindex” marketcorrelationisvalidenoughtocapturethediversificationeffectsofacombined 244 company.Ontheotherhand,theleveredbetatechniqueassumesthatbetaismadeupofa mathematicalsumofoperatingriskandfinancialriskwithoutadditionaloutliers.Thisis anoversimplification,butisvalidforthepurposeofanalysis.Infact,betaismadeupof somerandomfluctuationsintheeconomy,statistical“noise”,andotherundetermined factors.Sincebetaisunstable,aconsensusaveragebetween“topdown”andlevered approachesissufficient. “MOMANDPOP”STOREBETAS:COMPANIESWHOARENOTONTHEMARKET Mosttransactionsarenotenactedwithinacorporateenvironment.Anexample wouldbealocalbuildingcontractorwhowantstoaddaroofingcompanytohisportfolio. Thesebusinessdealsarenotglamorousbutareverytypical,andneedtobeevaluatedfor risk.Toexaminethem,analystsuseoneoftwoapproaches:1)Anaccountingbetacanbe determined.SalesorincomedatafromthebusinessisregressedagainsttheS&P500(or appropriateindex)aspercentagegainsandlosses.Aftertheaccountingbetaisdetermined, itisunleveredfordebt,andanoperatingbetaisderived;bookvaluesmustbeusedinstead ofmarketvaluesandweassumetheyarethesame.2)Theanalystusesthe“pureplay” technique.Theinvestortakesthebetasofthethreelargestmarket-tradedbusinessesinthe sameindustry,leversthemdownusingtheaveragedebttoequity,andderivesan unleveredbeta.Anexample: TheArdcoBarbellCompanynowhasabetaof1.52,amarketvalueof70,andadebtto equityof55/70.Itwantstobuyachainoftanningsalonswithnoactivelytradedmarket. Thesalonshaveadebttoequityof15/20andapriceof40Million.Thethreelargest activelytradedcompetitorsare: Table9-14 COMPETITOR SMITH BRONZE FRYE BETA 1.8 1.9 2.1 DEBTTO EQUITY 2 3 3.5 MARKETVALUE 60 60 40 245 Thecombined“topdown”betais(60/160)(1.8)+(60/160)(1.9)+(40/160)(2.1)=1.913 Theunleveredbetaforthishypotheticalcombinationis1.913/(1+[(1-.35)(2.8333)])= 0.673. Notethatthe2.8333figurewastheaveragedebttoequityforthethreefirms.Alsonote thatfinancialleveragecanhaveasignificanteffectonbeta.Withoutdebt,thethreefirms haveanoperatingbetaofonly0.673whichseemedso“safe”thattheypiledonthefinancial leveragetoimproveEPS. Thus,theoperatingbetaofthenewfirmisonly0.673.Thefirmisveryattractiveto Ardcobecausetheyhavesolittledebtinanindustrythatseemstothriveonit (hypothetically).Toformacompletelyleveredbeta,wemultiplytheunleveredbetabythe firm’sdebttoequitymultiple:Thatis(UnleveredBeta)(1+(1-.taxrate)(D/E)) Thecompletedbetais0.673x(1+(.65)(.75))=1.00111.Theacquisitionisprobablya “young”firmintheindustry,andhasnotestablisheditselfenoughtoincurthemassive amountofdebtofitspeers.ThisisanopportunityforArdco,andtheyimmediatelysnatch itup. OPERATIONSRESEARCHFORTHEINVESTOR Mostinvestorsdonothaveaccesstodetailedcorporatedatathatitemizesthetypes ofassets(prepaidinsuranceforexample)orthetypesofprojectsthatneedcapital budgetingattention.Thereisabondofmutualtrustbetweentopmanagementand shareholderswhenacquisitionsaremade;synergyhasbeenevaluatedandtherisksofa purchasehavebeenevoked.However,goodmanagementisfearless.Theywelcome questionsfromtheleastofemployeesandthesmallestofinvestorsbecausethereisan objectiveprinciplethatsubstantiatestheirposition. Partofriskmanagementistofindbalanceandcommongroundbetweentwosetsof data.Whenacquisitionsaremade,theremaybeahistoryofoperatingincomeandsales forbothcompanies,andtheinvestorcanusethisquarterlydatatomeasurethecovariance 246 betweenpercentchanges.Iftheacquiringcompanyishighrisk(operatingandfinancial leveragesarehigh),alowcovariancewillindicateastrategyofriskreduction. Analogously,alowriskcompanymayattempttostepupitsprofitmarginbytargetinga companywithahighcovariance.Otherstrategiesinclude:sectortiming.Ahighrisk companymaywanttotargetanotherhighriskcompanyifthatsectorisanticipatedto growassoonasassetsareefficientlyintegrated.Sincefewcombinationscanbethis strategicallyfacile,theinvestorneedsto“looktwice”,whenthisscenariooccurs.Similarly, whenalowriskcompanybuysacompanywithlowcovariance,itmaynotneedtheimplied diversificationasmuchasitneedstoseekouthighprofit,“riskier”,projects.Thenatural inclinationistotrustoperationstothosewhoknowitbest;theyoftenhavelargestakesin thesuccessofamerger.Nevertheless,askingquestionsistheprerogativeofthe shareholder,anddoingsowillhelpmanagementaswell. Thefollowingexamplecontinueswithatableofcashflowdata,comparingArdco Barbellwiththeproposedacquisitionofyetanothertennisracquetcompany,andalsoa vitamincompany.Themeans,standarddeviationsandcovariancesareatthebottomof thechart. Table9-15 QUARTER(Period) ARDCOBARBELL TENNISRAQUET VITAMINS 1 22 14 20 2 17 10 16 3 19 9 18 4 31 11 25 5 14 21 12 6 15 24 13 7 22 20 14 8 26 30 19 9 15 31 17 10 11 32 4 Mean 19.2 20.2 15.8 StandardDeviation 6.106 8.94 5.61 R(withArdco) -0.5882 0.77294 COV -32.108 26.477 247 *ALLDATAAREPERCENTCHANGES Inthisshortenedexample(atleast20quartersshouldbeused),itisobviousthatthe vitamincompanyismoreriskyfromadiversificationstandpoint.Ontheotherhand,the tennisracquetcompanysellstennisracquetsinthespringandsummer,whileArdcosells fitnessequipmentinthewintermonthsaspeople“getinshape”forsummer.Also,the chainofdistributorswouldbesimilarforArdcoandthetennisracquetcompany,butthere wouldbeaneedfornegotiationswithahealthfoodchainifthevitamincompanywere purchased.Additionally,anyfixedassetsfromthevitamincompanywouldbeuniqueto thatmanufacturingprocessandwouldneedtobeseparatedfromArdco.Similarly,trucks andwarehousesthatcarryfitnessequipmenthavedifferentprioritiesthatmaynotinclude themoredelicatestorageofvitamins.Therefore,Ardcomakesthedecisiontopurchasethe tennisracquetcompany-and-triestoputitslabelongenericvitaminswithanother companyinchargeofdistribution. TWOMASTERS:FISHERANDBUFFETT Twoofthegreatestanalystsofoperatingriskdisplayedanalmostintuitive quantitativesense.Theywerewellknownforstrategicinvesting,butshareda mathematician’sperceptionforminutedetail.TheyarePhilipFisherandWarrenBuffett. PhilipFisherwasafinancialanalystduringthedepthsofthedepression-almosta prerequisitefortenacity-andascendedfromthatmorasswithdedicationandvision.His book,CommonStocksandUncommonProfits,becamea“mustread”inthefinancial community.Hisbestknowntenetwastoevaluateacompany’slong-termprospectsby studyinggrowthandsalespotential,andthenfavoringthosecorporationswhowere dedicatedtoproducingatthelowestcost.Thisqualitativeapproachhadasitsfoundation, athoroughknowledgeofmanagementskillsandmanufacturingprocesseswhichFisher wouldexhaustivelyresearch.Fisherrejectedwhathetermed,“marginalcompanies”- thosethatshowedhighprofitsduringeconomicupswings,onlytolagbehindintherestof thebusinesscycle. 248 Fisheralsoexcelledindetectingoperatingsynergy-areasofgrowthinwhich productsorprocessescomplementedeachother.Forexample,whiletheaverageinvestor mightnoticeanagingpopulationandinvestinnursinghomes,Fisherwouldhaveinvested intheproductsandservicesneededtooperatethosehomesandthatwereusedin collaboration-perhapsdisposableneedlesand“sharps”containers.Twoofthebest examplesofthisacumenwereinvestmentsinDuPontandAlcoa.DuPontwasa gunpowdermanufacturerwhostandardizedprocessesandwentontoproduceunique syntheticmaterialslikecellophaneandLucite.Inaddition,theirmanagementteamwas responsiblefordevelopingthepreliminaryanalysisusedtoevaluatecapitalstructure,the aforementioned“DuPontequation”.Alcoawasanaluminumcompanywhocapitalizedon theairplanemanufacturingthatwasattheforefrontoftechnologyinthe1930sand40s. Bothcompaniesremainedflexibleandadaptedtheiroperationstonewproductdemand andtechnologies. WarrenBuffettisinaclassbyhimself.Fewinvestorshavematchedhisdedication togaugingriskandpositioningforlong-termgains.Curiously,hedoesnotfretatallabout economicfluctuationsorFederalReserveactions,andappearstohaveunbridledfaithin thenotionofa“goodidea”.However,hedoespossessaneconomist’sinsightaboutmost macrotopics,includinginflation.Hisnaturalinclinationistoavoidcompanieswithlarge amountsoffixedassets,becauseinitialrisesinassetturnoveraredepletedwhencapital expendituresfinallyneedtobemade.Inflationactslikeasmokescreenblindingthose firmswithahighpercentageoffixedassets,becausesaleswillinitiallyoutpacetheneedfor investment. AsecondmajorprincipleofBuffett’sistoinvestinfirmswithconsistentoperating history.Hereasonsthatthosecompanieswhoareintheprocessofamajorchangein operationsarenotgoodinvestmentprospectsbecauseofhigherrisksassociatedwithcost andrevenueinconsistencies. 249 Thirdly,andmostimportantly,Buffettseeksoutcompanieswithwhatheterms, “economicgoodwill”.Thesearecompaniesthatproduceaboveaveragelong-termprofits becausetheypossess“franchisevalue”-theabilitytoraisepriceswhenneeded.Theytend toproduceunregulatedproductsforwhichthereisnoclosesubstitute,andofteninducea “brandloyalty”amongtheircustomers.Appropriately,Buffettshunswhathecalls “commoditytype”businesses-lowprofitbusinessesthatchurnoutundifferentiated products.Hereasonsthatthesebusinessescanonlycompeteonthebasisofpriceandcan onlyprosperduringtherareeventofashortsupply. Whilemostanalyststreatthefutureasanextensionofthepast,bothBuffettand Fishercreatedadifferenttypeofmentalcalculus.Theyappearedtocomparetheratesof changeamongseveralvariableswhichcoalescedintoaninvestmentstrategythatproduced long-termgainsamongmanycommonlyknownbrands-Coke,TheWashingtonPost, ABC,Etc.Buffetsdesireforboth“pricingpower”andoperatingconsistency,seemsto seekacompanywithahighbutconsistentoperatingleveragethatwillnotdeflateatthe prospectofadownturn.Hightechnologyfirmswouldbeeliminatedbythatconstraint,but “niche”companieswhohavecreatedabrandconsciousnessamongconsumerswouldnot. Ahigh“mean”andalow“standarddeviation”amongbothoperatingmomentumand operatingmarginswouldmeetthatobjective.However,asmallcapitalintensityratio(one withoutalotoffixedassets)wouldalsofitBuffett’sobjective,exceptthattheseoften designatethevery“commoditytype”industriesthatheavoids.Naturally,aproductive balanceisimplementedwhentheseseveralconstraintsaresatisfied,butsuchastrategic combinationisdifficulttofind. ABRIEFOPERATINGANALYSISOFFED-EXANDSTAPLESFORTHEYEAR2000 Mostanalystsexaminecompaniesinthesameindustries,becauselargedisparities maybemoreindicativeoftheindustryratherthanthecompany;thereissamplebiaswhen differentcompaniesindifferentindustriesarecompared.Theneedforprecisionrequires focus,butwhenonecompanyissubjecttoforces(demand,legislation,tradebarriers)that 250 theotherisnot,theanalystcannotformvalid,actionableconclusions.Weclosethis chapterbysummarizingourindicatorsthroughanalysisoftwoseparatecompaniesintwo separateindustries.Whilesuchcomparisonisstatisticallyunconventional,itisrealistic fromaninvestor’sperspective. WeofferafiveyearoperatinghistoryofbothFed-Ex,thenextdaytransitspecialist, andStaples,theofficesupplyfixture.Salesandoperatinggrowthisdetermined,andthen changesinfixedassetsandcapitalintensityarederived.Finally,wedomean-variance statisticsontheoperatingmarginsofeachcompany.Thethemethatrunsthroughthe analysisisthechoicebetweengrowthandstability.Aretherisksofslowinggrowthgreater thantheeffectofstagnation,forexample?Doesaperiodoflonggrowthpreclude investment,i.e.,notgettinginsoonenough?Whenthestatisticsbecomecontradictory,we addsome“qualitative”analysisandtrytoperceivethesituationasFisherorBuffettwould. STAPLESANDFED-EXOPERATINGHISTORIES1994-1999 Table9-16 FED-EX Year 1994 1995 1996 1997 1998 1999 Sales 9392 10274 11520 15873 16773 18257 Op. Income 1244 1344 1477 2047 2198 2376 Current Assets 1728 2133 2880 3141 3285 Total Assets 6699 7625 9686 10648 11527 Table9-17 Fed-ExBeta BETA(2000) 0.98 EQUITY(2000) 8191 DEBT(2000) 1899 251 Table9-18 STAPLES Year 1994 1995 1996 1997 1998 1999 Sales 2000 3098 3968 5181 7123 8937 Op. Income 110 191 260 355 514 708 Current Assets 926 1151 1666 2064 2192 Total Assets 1403 1788 2455 3179 3814 Table9-19 StaplesBeta BETA(2000) 1.03 EQIUITY(2000) 1837 DEBT(2000) 2152 ANALYSISANDSTATISTICS Allbetasarenotcreatedequal.Withoutunleveringthebetas,wecanreadilyobservethat Staples’operatingbetaislowerbecausetheyhavemuchmoredebtintheircapital structure,andyetbothcompanies’betasarenearlythesame.SinceFed-Ex’soperations areorientedaroundtransportation,whichrequiresaheavyinvestmentinfixedassets, Fed-Exhasmoreeconomicrisk.Additionally,theyareexposedtotheriskofhigherprices intheoilmarketwhicharetotallyoutofthefirm’scontrol.Theseareinherentrisksthat comewiththenatureofthebusinessandcanonlybeminimallydiminishedby diversification.However,aswewillobserve,Fed-Excompensatesforthisdetrimentwitha highlevelofefficiencyandstability. 252 Table9-20 FED-EX Percent Change YEAR 1995 1996 1997 1998 1999 Operating Income% 8.04 9.9 38.59 7.38 8.1 Sales% 9.39 12.13 37.79 5.67 8.85 Table9-21 Staples Percent Change YEAR 1995 1996 1997 1998 1999 Operating Income% 73.64 36.13 36.54 44.79 37.74 Sales% 53.4 29.34 30.57 37.48 25.47 ThefollowingtabletabulatestheextraordinarygrowthofStaples’salesandoperating incomeandcomparesitwithFed-Ex’sratherordinarygrowth.However,whatthe investorshouldobserveissustainability.Nocompanygrowsattwentypercentforever. Althoughthegrowthoccurredinthecontextofalongbullmarket,fivetosixyearscanbe thelengthofanentirebusinesscycle.Staples’growthwasderivedfromastructural changeinthewayAmericadidbusiness:theInternetspawnednumeroushomebusinesses thatneededsmallamountsofofficesupplies.Suchatransitionwouldnottakenearlyas longasIBM’s,forexample,whichcomputerizedentireindustries.Competitionwould cropupeasilyinofficesupplies,butnotinindustrieswhichrequiredhugeinvestmentsin fixedassets.Thus,whileStaples’growthratelookedlikeatemptinginvestment,itwould havebeenfarriskiertoputcapitalintoitin2000thanin1996or1997. 253 Table9-22 FED-EX MEAN- VARIANCE STAPLES MEAN- VARIANCE SALES 1.697 SALES 24.22 OP.INCOME 0.898 OP.INCOME 29.797 *Wemeasuremean-varianceas(mean-standarddeviation)whichisatypicaladaptation. Table9-23 AverageOperating Momentum Lastyear(1999)operating Momentum FED-EX 0.99 0.915 STAPLES 1.265 1.4817 Staples’growthisbothhighandsteady,anditwouldbeeasyforaninvestortobedeceived bytheseoperatingstatistics.Incapitalstructuralism,ourdesireisalwaystoanticipate growthandneverto“chase”it.Capitalizingonanoverlylonggrowthcycleisreservedfor the“lucky”few;ahandfulofinvestorscanmakemoney“withoutknowinganybetter”, becausetheyhappentobeintherightplaceattherighttime.Whilethereismoneyinhigh riskmomentumtrading,itisnotrecommendedforeventhemostskilledinvestors,because itissuchagamble:theoddsarebetterplayingblackjackinLasVegas.However,there maycomeatimewhentheinvestorobservesthreetofouryearsofgrowth,andmayhave informationaboutmovementtowardanoptimalcapitalstructure.Atthispoint,both equityandthecostofequitymaybelowenoughtojustifyinvestment.Theinvestor appearstobecapitalizingonmomentum,butisnotinvestingonthecriteriaofprior growth:heorshehascorrectlyanticipatedaprofitablechangeincapitalstructurewhich shouldbeactualizedbyajumpineconomicprofit. THECONFIDENCEINTERVALTOOL 254 Whenwesightaneartermnumberthatisuncharacteristicofothersinthe sequence,itiscauseforconcern.Newdevelopments(changesincostsordemandfor example),mayaffectourinvestment.Therefore,werecommendusingthe99%confidence intervaltogaugeanyodditieswemayencounter.Withsmallsamplesizes,itissuspectasa measurement,butwemerelyuseitasachecktoseewhetherfurtherinvestigationneedsto bedone.Thefollowingcapitalintensityandnoncurrentassetratiosapplytothe companies: Table9-24 FED-EX YEAR 1995 1996 1997 1998 1999 Capital Intensity 0.652 0.662 0.61 0.6348 0.6314 Non Current AssetRatio 0.7421 0.7201 0.7027 0.705 0.715 Table9-25 STAPLES YEAR 1995 1996 1997 1998 1999 Capital Intensity 0.457 0.4506 0.4796 0.4463 0.4267 Non Current AssetRatio 0.3399 0.3562 0.3214 0.3492 0.4252 Alloftheseratioslookliketheyareinanorderlysequence,exceptforthe1999valueof noncurrentassetsforStaples.At0.4252,itislargerthantheothers,andwouldbethe mostlikelytoaffectourpotentialyear2000investment.Wedecidetosubmitittothe student’sTtestandderiveconfidencelimitsat99%. Table9-26 255 STAPLES NONCURRENTASSETRATIO Mean 0.358396 SampleStandardDeviation 0.03956 SampleSize 5 DegreesofFreedom 4 SquareRootofSampleSize 2.23606 Student'sT 4.604 ConfidenceInterval 0.358+/-(4.604)(0.03956)/2.236 Thelimitsoftheintervalare0.2765onthelowsideand0.4394onthehighside.Since 0.4252isbetweenthatinterval,wewillnotinvestigatethemeasurementanyfurther. However,theeccentricityisevenmoreevidentbecausethislastnumberispartofthe sampleitself.Also,thelargechangeinnon-currentassetsisnotconfirmedbyajumpin capitalintensity.Ifbothmeasurementsrose,wewouldinvestigatetheiracquisitions, distributionchanges,etc.Infact,thesimpleexpenditureofalargeamountofcashforan acquisitioncandepletecurrentassetsandinflatethenon-currentassetratio,whichisthe reasonforobtainingcorroboratingevidence;oneratiocanacttoconfirmtheother. OPERATINGMARGIN Thesizeandstabilityofoperatingmarginsisparamount.Sofarouranalysishas encompassedtheamountandstabilityofgrowth.Asgrowthisoftenstimulatedbya sector’spositioninthebusinesscycle,ittendstowaverandattractinvestorsintheshort- term.However,intheabsenceofotherinformationaboutfinancialleverageandcapital structure,decisionsmadeonthebasisofoperatingcharacteristicsshouldbelong-term. Rememberthatfinancialleveragewillrespondtothesizeandstabilityofsalesandincome andnotnecessarilyitsgrowth;theprobabilityofdefaultisloweredwheninterestcoverage ratiosrise.Thefollowingtablesexhibittheoperatingmarginhistoriesforbothcompanies: 256 Table9-27 OPERATING MARGINS (Decimal) YEAR 1995 1996 1997 1998 1999 FED-EX 0.1308 0.1282 0.129 0.131 0.1301 STAPLES 0.0623 0.0655 0.0685 0.0722 0.0792 Alargeandstableoperatingmargincanleadtothreepositivescenarios:1)Alargercash flow.2)Moredividendsflowingtoinvestorsand3)Retainedearnings,insteadofcredit, thatcanbeusedforpurchases.Thus,whilegrowingsalesleadstoalargeroperating income,thecoststructureofabusinesscanbeoptimizedbutnotchanged.Stapleshad growingoperatingmarginsthatwouldpotentially“topout”,whenallefficiencieswere realized.Ontheotherhand,Fed-Exhadanoperatingmarginthatwasnearlytwicethe sizeofStaples’.Theyweremakingastableprofitinabusinessthathadmuchmore inherenteconomicrisk.Amean-varianceanalysisrevealsthedifference: Table9-28 OPERATINGMARGIN % FED-EX STAPLES Mean 12.982% 6.954% SampleStandard Deviation 0.1197 0.6523 Mean-Standard Deviation 12.8623% 6.3017% ThehigheroperatingmargingivesFed-Exmuchmorefinancialflexibility.They arealowdebt,highequitycompany,andyettheirearningsarelargeandstableenoughto financewithmuchmoredebtiftheyneededto.Choosingtofinancewithretainedearnings lowerstheriskofinterestratechangesandbufferstheeffectofhavingahighleveloffixed assets.Inthisrespect,theirinherentbusinessriskmightbetoohighforaninvestorof 257 Buffett’scaliber,buttheydohave“pricingpower”asoneofthefewplayersinalimited field;overnightshipmentstoGreenlandorAfricaensurethatFed-Exwillbeknownasa “brand”.Such“franchisevalue”and“economicgoodwill”wouldberightupFisher’sand Buffett’salley Ratherthantheproverbial“cointoss”,ahigheroperatingmarginisanarbiterof investment.However,whatifFed-Ex’smarginsdeclineasStaples’rise?Forthatvery reason,wetrytocoalesceasmuchinformationaspossible,neitherdependingonearning forecasts,companyhype,orevenlongandshort-termmovingaverages.Enoughdatamust becoordinatedtoformacohesivepicture,andoperatingcharacteristicswillnottellthe entirestory.Andyet-ifweaddfinancialleverageandmarketinformationtothemix, muchofwhatweneedtoknowwillbeinthemargins. (BacktoTableofContents) 258 10 OPERATINGMOMENTUM Operatingmomentumhasneverbeenastandardmeasurementofrisk.Whileit mirrorsoperatingleveragewhencostsarestable,itsownvarianceissovolatilethatit defiesrationaluse.Inanygivenyear,itcanbealternately,small,large,orevencomposed ofnegativenumbers,leadingtotheconclusionthatanapplicationofthisratioisan exerciseinfutility-andnotutility! Moststatisticianswouldarguethatitisanirrelevantlyimprecisemeasurement.As anexample,considerthepremisethatwedefineoperatingmomentumasthepercentage changeinoperatingincomedividedbythepercentagechangeinsales.Thus,50/50,15/15, and-11/-11areallthesamenumber.Sinceitisameasureofvelocityratherthan magnitude,weneedtoobtainlong-termaverages(atleastfiveyears)tousethe measurementinameaningfulway.Anincreaseofthisratiorequiresthreeyearsofdata. Infact,itsveryimprecisionandrelativityforcesustoobtainasmuchinformationas possible.Theneedfordatacreatesalinkbetweenyearsthatallowsthisratiotobemore comprehensiveandforward-lookingthanitwouldfirstappear. Whileoperatingleverageusuallychangesinresponsetonewtechnologythat changestheproportionalrequirementoffixedcosts,anychangeincost,priceorquantity willaffectoperatingmomentum.Thisinherentvolatilitycausesittoriseandfalltenfold insomeperiods,especiallywhensalesandearningsareoutof“sync”.Atcertaintimes,a onepercentincreaseinsalescanleadtoafifteenpercentincreaseinoperatingincome, creatinganuncharacteristicratiooffifteen.Sincetherealoperatingleveragemaybeonly 1.5,andthelong-termaverageofmomentummaybeonly1.2,theseoutlyingmeasurements createvolatilityandtendtomakewildswingsincompaniesthatareundergoingmajor changes.Wesmooththisvolatilitybytakingseparategrowthmeasurementsforboth operatingincomeandsales,insteadofaveragingindividualratiosforeachyear. 259 REASONSFORSTUDY • Operatingmomentumisrelativelyunexploredterritoryamongacademics.Statistical methodswilloftenneglectitsimportancebecauseitdoesnotfittheparametersof “normality”.Althoughboththenumeratoranddenominatorcontain“essential” measurements(salesandoperatingincomerespectively),thebehaviorofthecombined ratiohasnotbeenexaminedthoroughly. • Analystsoftenmisunderstandoperatingmomentum,confusingitwith“operating leverage”,whichitmirrorsintheory.Whileoperatingleveragereflectsthestabilityof acoststructureinaproductionprocess,operatingmomentumreflectstherealcost over-runsandvolatilityofthatprocess.Litigation,restructuringcharges,and abandonedbusinesseswillallextractcashfromoperationseveniftheyareaccounted forseparately. • Operatingmomentumhascashvaluefortheinvestor.Oncewecanrelateoperating momentumtootherratiosandtheearningspotentialofacompanyinthenear-term,it isplacedinamoremeaningfulcontext. Considerasmallmanufacturingcompanywiththefollowingoperatingincomeandsales distributions: Table10-1 Variables YearOne YearTwo- Probability=1/2 YearTwo- Probability=1/2 Sales 100 50 150 OperatingIncome 10 5 15 TotalCost 90 45 135 Inthehypotheticalabsenceoffixedcosts,operatingmomentumandoperatingleveragefor bothscenarioswouldbeequalto“1”-althoughonescenarioinvolvesanincreaseoffifty percentincostsandsales,whiletheotherinvolvesadecreaseofthesamemagnitude. Neithermeasurementwouldfullyreflectrisk.Iffixedcostsbecametenpercentoftotal 260 costs,thetrueoperatingleveragewouldincreaseto1.9,butoperatingmomentumwould remainat1.Fromtheperspectiveofcommonsense,itwouldseemthatthedecreasing capacityscenariowouldbemoreriskymerelybecauseunusedcapacityismoredifficultto managethantheprospectoffullcapacity;fixedcostsstillneedtobepaidalthough revenueshavedeclined.Whilethosefixedcostsareanexplicitpartofoperatingleverage, theyremainimplicitinoperatingmomentum:wemeasuresensitivitytototalcostwiththe latterratioandinferthatfixedcostsareincreasingwhenoperatingincomeincreasesata fasterratethansales.Operatingleverage,ontheotherhand,isdependentontheindustry andchangesgraduallyovertimeasnewtechnologyisadded,whichultimatelyspreads fixedcostsoveragreaterquantity.Whenastateofover-productionisreached,priceswill decrease,causingfirmstoeitherattempttoreducefixedcostswithevennewertechnology, ortoleavethebusinessentirely.Ultimately,operatingmomentumrespondstothevagaries ofthemarket,reactingtoextraneousfactorslikelawsuits,discontinuedbusiness,and foreigncurrencytranslation.Whilesuchvariableswouldnotbereflectedinbasic operatingleverage,theyareaninherentpartofoperatingmomentum;eachofthese outlyingfactorsdrainscashfromcollectiveproductionprocesses,evenwhentheyare countedasonetimeseparatecharges.Amulti-billiondollarsettlementagainstatobacco company,foranexample,maynotbereportedaspartofnormaloperations,butwillaffect productionforyearstocome.Thus,thecloseroperatingmomentumistooperating leverage,themorestabletheproductionenvironment. Comparingindustry-averageoperatingleveragetocompany-specificoperating momentumwillrevealproductionrisksnotapparentineventhestandarddeviationsof salesandincome.Reasonsfordivergenceinclude:changingthebasicnatureofthefirm throughacquisitions,andanover-dependenceonasmallbaseofeithervendorsor customers.Whilediversificationcanlowerrisk,anoperatingmomentumthatisactually lowerthanindustryoperatingleveragecansignifythatthecompanywillunderperform theindustrywhenthatsectorisfavored.Sincetheobjectiveofdiversificationistolower 261 overallriskinthelongrun,companieswithloweroperatingriskareatatemporary disadvantagewhentherelativesectorsurgesaheadintheeconomy,becauseearningswill tendtoincreaseataslowandsteadypace.Those“sacrificial”profitsaregainedbackin long-termviability.Forexample,compareGallowinestoasmallwineryinFrance.The smallwinerymustdependonafavorablegrowingseason,isnotdiversified,andmakes largeprofitswhenconditionsare“right”.Galloisadiversifiedwinerymuchless dependentonafavorablemarket,andcansellwinefarintothefuture.Itsoperating momentummaybemeasuredtobelessthantheindustryaverageof“operatingleverage”, butittradesimmediatehighprofitsforlong-termviability.Therefore,itisimportantto combineinformationaboutthisdivergencewiththestandarddeviationofbothincomeand sales.Ifallthreemeasurementsarehigherthanaverage,theinvestmentrequiresextensive analysis,ormightwarrantshelvingaltogether. Inafewindustriesthereisthepowertochangepriceswithoutaffectingdemand. Theseelitecompaniesareusuallygoodinvestments,possessingwhatWarrenBuffettwould call“franchisevalue”.Toillustratetheeffectofchangesinpriceonbothoperating leverageandoperatingmomentum,considerthefollowingscenarios: Table10-2 HIGHOPERATINGLEVERAGE VariableCosts=700 OperatingLeverage= OperatingIncome=300 (1200-700)/300=1.66 Sales=1200 Price=4 OperatingMomentum= Quantity=300 1(GivenasaHypothetical) FixedCost=200 262 Table10-3 LOWOPERATINGLEVERAGE VariableCosts=800 OperatingLeverage= OperatingIncome=300 (1200-800)/300=1.33 Sales=1200 Price=4 OperatingMomentum= Quantity=300 1GivenasaHypothetical) FixedCosts=100 THEEFFECTSOFA12.5%PRICECUTFROM4TO3.5 Salesbecome1050from1200,whileoperatingincomedeclinesto150from300.Other variablesremainunchanged. Table10-4 HIGHOPERATINGLEVERAGESCENARIO (1050-700)/150=2.33 LOWOPERATINGLEVERAGESCENARIO (1050-800)/150=1.66 OPERATINGMOMENTUM -50/-12.5=4 Noticethatbothratiosdisplayedthegreaterriskofapricecutbyincreasing,butthat operatingmomentumreactedmoreviolently,increasingfourfold. THEEFFECTSOFA12.5%PRICEINCREASEFROM4TO4.5. Salesbecome1350from1200,operatingincomeincreasesto450from300.Othervariables remainunchanged. 263 Table10-5 HIGHOPERATINGLEVERAGESCENARIO (1350-700)/450=1.44 LOWOPERATINGLEVERAGESCENARIO (1350-800)/450=1.22 OPERATINGMOMENTUM 50/12.5=4 Theessenceofpricingpoweristhatitsignificantlyreducesrisk(operatingleverage)while raisingoperatingincome.Sincevariablecostsincreasewhenquantityincreases,thehigher coststhataccompanyquantitydrivenchangesinsalesareavoided.Whenpricecan increaseatafasterratethanquantity,thefirmmayhavecreatedanichethatcompeteson thebasisofquality-drivenproductdifferentiation. However,itshouldbenotedthatoperatingmomentumincreaseswhetherpricesare increasedordecreased.Thismovementhaspredictivevalue.Wheneveritmoves substantiallyawayfromthemean,whichinthiscaseweassumedtobe“1”,wecanexpect ittocompensatebyrevertingtowardtheaverageinthefollowingyear.Whilethe reversionisbynomeansa“hardandfast”rule,investorswillactuallydemanditbecauseit meansproductionisstabilizing.Thus,afteraquickriseto“4”,analystswouldbelooking fortheconditionsthatwouldloweroperatingmomentum-greaterquantitiesandmore variablecosts-hopefullywithincreasedearnings.Althoughoperatingmomentumdoes notdifferentiatebetweenspecificrisksaswellasoperatingleverage,anylargeshiftswillso upsetcorporateequilibrium,thatthefirmwillbeforcedtocontrolcost,quantity,orprice toredirectmomentumbacktoitsmean.Evenapriceshiftthathaspositiveacceptance fromthecustomermustbemetwithacountervailingforcebecausemoreriskhasbeen incurred;suchriskscouldentailincreasedsalarydemands,discontinuanceofsome operationsorevenrestructuringcharges. 264 Agoodexampleofoperatingmomentum-shiftsoccursinindustriesthatsell commodities.Whensupplyofacommodityisparticularlylow,theentireindustrycan raiseprices,whichwillincreaseoperatingmomentum,becausedemandisessentially inelastic-itwillnotdecreaseaspricesincrease.Theconsequentriseinoperatingincome relativetosalesoccursbecausesalesrisewithoutaquantity-drivenchangeincostsi.e., variablecostsdonotincrease.Thestockpricewillriseforabriefperiod,onlytobe followedbyaselloffbecausethesituationistemporaryandtheinherentrisksreturnto normal.Ifthecompanyfailstocounterthe“positive”riskofincreasedprices,itmust suffertheconsequencesofanothershiftinoperatingmomentum. Anotherexampleofcounteractingtheriskofhighergrowthoccurswithaprice announcement.Marketingstrategiesaregearedtowardtheriskofoperatingmomentum: Canthecompanycutquantityandstillmakeaprofitifcustomersfindalessexpensive substitute?Canthecompanycounterthismoveinthefollowingperiodwithaquantity drivenproductlinethatappealstoamorecost-consciouscustomer?Thesearethetypesof questionsthatfueldiversificationandarethereasonthatmorediversifiedcompanieshave aloweroperatingmomentumandinherentlylessriskthan“oneproduct”companies. Anothertopicthataffectsbothleverageandmomentumistheimpactofinflation. Thestudentwillobservethatintimesofinflation,companieswithmorefixedassetsmay actuallybenefit.Thedistinctlagtimebetweeninitialsalesandtheneedtoupgrade equipmentwilltemporarilyboostincomeabovethosecompaniesthatmustdealwithmore immediaterisesinvariablecosts.Ontheotherhand,risktothesecompaniesactually increasesbecausenotonlydotheyhavetomakethenecessarycapitalexpenditures,they mayhavetodosoallatonetimetokeepoperationsrunningsmoothly;sacrificinglong- termviabilityforshort-termgainsisthusariskystrategy. OPERATINGMOMENTUMSENSITIVITY Acomparisonbetweenoperatingmomentumandleveragerevealstheimplicit volatilityofoperatingmomentum.Thevariablesinthebreakevenequationare:price, 265 quantity,variablecostperunitandfixedcosts.Ifwechangeeachbreakevenvariablebya factorof5%andthen25%,andkeeptheotherbreakevenvariablesconstant,wecan observetheeffectsonbothleverageratios.Thechangesinvariableandfixedcostsassumes aonepercentchangeinsalesWeestablishahypotheticalbaseandthendisplaytheeffects ofbothfiveandtwenty-fivepercentincreasesanddecreases:. Table10-6 BREAKEVENPOINTVARIABLES-BASE Quantity=100 Price=4 Sales=400 VariableCosts=2perunit FixedCosts=100 OperatingIncome=100 266 Table10-7 OPERATING MOMENTUM VariableChange=5% Increase Decrease Price 4 4 Quantity 2 2 VariableCosts -6 14 FixedCosts -1 9 Table10-8 OPERATING MOMENTUM VariableChange=25% Increase Decrease Price 4 4 Quantity 2 2 VariableCosts -48.5 52.5 FixedCosts -23 29 Table10-9 OPERATING LEVERAGE BaseOp.Lev.=2 VariableChange=5% Increase Decrease Price 1.83 2.25 Quantity 1.91 2.11 VariableCosts 2.11 1.91 FixedCosts 2.105 1.9 267 Table10-10 OPERATING LEVERAGE BaseOp.Lev.=2 VariableChange=25% Increase Decrease Price 1.5 NA Quantity 1.67 3 VariableCosts 3 1.66 FixedCosts 2.66 1.6 Operatingmomentumisinsensitivetopriceandquantitychanges,buthyper-sensitiveto changesinvariablecosts.Naturally,isolatedchangesoccuronlyinthelaboratory.Reality dictatesthatthebreakevenvariablesinteract,andasvariablecostsrise,thecompanymay cutproduction,raiseprices,orboth.Whenacompanycanonlycompetebychurningout moreofanundifferentiatedproduct,thatproductbecomesacommodity.Asquantities increase,variablecostsbecomemoreprevalent,eventuallyeclipsinganyoperatingprofit. Themuchnarrowerrangeofvaluesmakesoperatingleverageabettergaugeofrisk thanmomentum.However,operatingmomentummaybethebetterforecastingtool, becauseofitsveryinstability.Whiletrueoperatingleveragemaychangeoverthelongrun throughchangesintheproductionprocess,momentumcapturesthevolatilityofyearto yearchanges;periodsofstabilitycanbereadilycontrastedwithdisturbances.Besides reversiontothemean,whichassumesstatistical“normality”,thereareothertechniques from“extremevaluestatistics”thatmaybettercharacterizethisfleetingmeasurement. Ifweassume“stable”productionprocesses(aconceptthatmanyoperations managerswouldscoffat),thenextexamplewillshowthemirrorrelationshipbetweenthe twomeasurements. Infouryears,TheBeeGoodHoneyCorporationhasthreesalesincreasesandsalesfor eachyearof1000,1500,1700and1800respectively.Eachyear,fixedcostsare300and variablecostsare60%ofsales,exceptinthelastyearwhenfixedcostsincreaseto400. 268 Table10-11 YEAR OP.LEV.=(S-VC)/(S- VC-FX) OPMOM.= %OP/%SALES Year1 (1000-600)/100=4 NA Year2 (1500-900)/300=2 200%/50%=4 Year3 (1700-1020)/380=1.78 26.666%/18.33%=2 Year4 (1800-1080)/320=2.25 -15.79%/5.9%=-2.6 Observehowoperatingmomentumwillperfectlymirroroperatingleveragewhencostsare stable.OncefixedcostschangeastheydoinYear4,thatmirrordiminishesandthe predictivepowerofoperatingleverageislost.Sincethereisequivocationaboutthe definitionof“fixed”and“variable”inrealaccountingcosts,operatingleverageisrelegated toahypothetical.Ontheotherhand,a“real”measurementlikeoperatingmomentumis lessreflectiveofriskandhardertointerpret. Thepredictiveabilityofoperatingmomentumispredicatedontheprobabilityof increasingoperatingmargin.Operatingmarginisaffectedby: • 1.Thetypeofbusinessandindustry • 2.Thelevelofcompetition,especiallythebarrierstoenteringthefield • 3.Theeconomicoutlookandbusinesscycle Withinthatframework,managementcanmaximizepotentialmargins,butmustcounterits limitationswithadditionalstrategies;highermarginswillnotcontinueindefinitely,andit isanastutemanagementthatwilllowermomentumwithalargesalesincrease,i.e.,the introductionofaquantity-drivenproductthatlowersoperatingrisk.Thefollowingtable illustratestheprobabilitiesofoperatingmarginandmomentumincreasesinadatasetof about180differentsamplepoints. 269 Table10-12 INCREASING OPERATINGMARGIN STATE NUMBER PERCENTAGE TotalOperatingMargin Increases 104/182 57.14 WithIncreasing OperatingMomentum 64/91 70.33 WithDecreasing OperatingMomentum 40/91 43.96 Therefore,theprobabilityofanoperatingmarginincreaseisgreaterwhenoperating momentumincreases.Bydefault,anyoperatingincomedecreasewillleadtoanegative momentumnumberifsalesincreaseatthesametime.Analogously,simultaneous decreasesofboththecomponentswillleadtoapositivenumberandapotentialincrease. Inessence,wearemeasuringhowcorrelatedincomeiswithsales,anddesignatingthe amountofriskbythecovariance,butwithoutrecoursetothechangesinfixedandvariable costs.Thereadershouldnoticethatthesampleof182datapointsdisplayedadistribution ofexactlyonehalf(91)momentumincreasesanddecreases–anindicationofrandom variation. REGRESSION Whenlinearregressionisappliedtooperatingmomentum,comparativeriskis vague.Withoutreferencetobothsizeandvariabilitytogether,novaliddecisionruleexists. Thepercentageincreasesofperiodicoperatingincomearethe“Y”variable,while percentagesalesincreasesaretheindependent“X”variable.Thefollowingtablesare comparisonsbetweenthreecompaniesthathavelittledebt,andwhosemajorriskis economici.e.,operatingrisk. 270 Table10-13 1996 1997 1998 1999 2000 MOLEX OP INCOME 7.69 16.37 5.63 -4.36 29.62 SALES 15.44 11.35 5.39 5.48 29.50 BIOMET OP INCOME 12.66 14.61 25.49 23.05 14.29 SALES 8.41 12.24 16.28 21.66 11.94 FAIR, ISAAC OP INCOME 37.69 38.27 11.92 15.16 8.15 SALES 30.70 33.56 23.62 12.60 7.58 Table10-14 COMPANY YINTERCEPT (ALPHA) R COEFFICIENT MOLEX -4.223 0.881 1.133 BIOMET 4.359 0.842 0.9684 FAIR,ISAAC -2.177 0.871 1.129 Theproblemwithregressionisininterpretation.Allnumbersinthedatatableare percentageincreases.Asthestudentwillobserve,thereislittlethatregressioncantellus, evenwhenoperatingriskisconsideredsogreatthatnolong-termdebtisincurredatall. Thebestindicatorsofriskremainthestandarddeviationsofsalesandincomerespectively, andtheirvariousadaptationslikethecoefficientofvariation,andthemean-variancerule. 271 Table10-15 COMPANY MEAN STANDARD DEVIATION MOLEX SALES 13.43% 9.92 OPERATINGINCOME 10.78% 12.66% FAIR,ISAAC SALES 21.68% 11.26 OPERATINGINCOME 22.23% 14.59 NoticethatMolexandFairIsaachavealmostidenticalregressionprofiles,butonelookat theirrespectivedistributionsmakesuschoosethelatter.MostoftheriskforFair,Isaacis ontheupside.Withouttheabilitytointer-relatesalesandincomethroughspecificcosts, ouronlyrecoursewouldbetoevaluateoperatingmomentumbycreatingaratioof coefficientsofvariationforbothoperatingincomeandsales.Theratioofstandard deviationsalonewouldmisstatetheriskforpercentagevaluesthatwereparticularlylarge; smallvalueswouldhaveasmallerstandarddeviationandappearlessrisky.ForMolex, thecoefficientofvariationforoperatingincomeis1.1619,whilethatforsalesis0.7385. Dividingthetwoyields1.5731.ForFair,Isaac,thecoefficientofvariationforoperating incomeis0.656,whilethatforsalesis0.5211.Dividingthetwoyields1.259.Thus,Molex hasariskierstreamofincomeandwewouldusethesefigurestoevaluateoperating momentum. THEGENERALELECTRICSOLUTION Eventothisauthor,thecoefficientofvariationseemslikeanincompletesolution. Whileitenumeratestheseparaterisksofbothsalesandincome,itseemstocreate independencebetweenthetwofundamentalsthatdoesnotexistinreality;weknowthat salesandincomeareinterdependentandtheriskthatweneedtomeasuremustflowfrom theirworkingtogether.GeneralElectric,acompanywithenoughpersonneltopopulatea 272 majorcity,hastheirowndefinitionofoperatingleverage.Intheir2006prospectus,they defineitasthepercentchangeinrevenueminusthepercentchangeintotalcost,or%∆ ∆∆ ∆ Revenue-%∆ ∆∆ ∆TotalCost,whichisautilitarianconcept.;operatingmargin,definedas operatingincomedividedbysales,mustriseanytimethepercentchangeinrevenueis greaterthanthepercentchangeintotalcost,andbothofthosevariablesarepositive. Consequently,whenthatconditionarises,operatingmomentummustbegreaterthanone. Whileoperatingmarginscanincreaseovertime,theyusuallymovecyclically,pushedby thechangesinmomentum.Thefinelinebetweencost,risk,andhighermarginstranslates toabetterreturnonequityandhigherstockprices. The%Revenue-%TotalCostsolutionhasgreatpracticalvalue.Itwillyielda numberthatishighlycorrelatedwithoperatingmargin,ROC/ROE,andstockprice.Like thepreviouscoefficientofvariation,anytypeofperformanceevaluationcanrevealthe returnandstabilityofbothmarginandmomentum-butwillnotexplicitlydivulgetherisk offixedassets.Again,thespecificityoftrueoperatingleveragecannotbesupersededby anyothermeasurement.Thefollowingtwotablesshowtherelationshipbetweenoperating marginandmomentum,andthe“GEversion”ofoperatingleverage.Performanceis measuredthroughthemean-variancemethod,andthe%Revenue-%TotalCostformula istermed“Distance”. Table10-16 “COMPANYA” Sales Total Cost OpInc. Op. Marg. Op. Mom. %Rev. %TC Distance 100 90 10 0.1 NA NA NA NA 120 102 18 0.15 4 20 13.33 6.67 132 110 22 0.167 2.22 10 7.84 2.16 140 120 20 0.143 -1.5 6.06 9.09 -3.03 165 130 35 0.212 4.199 17.86 8.83 9.53 150 110 40 0.266 -1.57 -9.09 -15.38 24.47 273 µ µµ µDistance=7.96,σ σσ σDistance=10.38,µ µµ µ-σ σσ σ=-2.42,Thus,themean-variancefor CompanyAis-2.42%.Thestudentshouldobservethatinthelastdatapoint,the operatingmomentumwaslessthan“1”andoperatingmarginstillincreased.Thesecond conditionoftherelationshipwasnotmet;theDistanceparameterrequiresthatthepercent changeinrevenueispositive. Table10-17 “CONPANYB” Sales Total Cost Op.Inc. Op. Mar. Op Mom %Rev. %TC Distance 80 70 10 0.125 NA NA NA NA 84 75 9 0.107 -2 5 7.14 -2.14 99 90 9 0.0909 0 17.86 20 -2.14 121 100 21 0.1725 5.99 22.22 11.11 11.11 140 130 10 0.0714 -3.34 15.7 30 -14.3 160 135 25 0.15625 10.5 14.28 3.85 10.43 µ µµ µDistance=0.592,σ σσ σDistance=10.54,µ µµ µ-σ σσ σ=-9.948.Thus,themean-variancefor CompanyBis-9.948%.TheoperatingmarginisgrowingmorevigorouslyforCompany A.Bycoordinatingoperatingmomentumand“Distance”,anycompanycanviewthe necessaryrequirements,i.e.,salesandincomechanges,forincreasingoperatingmargin. Althoughmarginsincreasewhenweimplementourmethod,thereisstillthechance ofmorevariabilitybecauseriskisnotexplicit;therearenoindicationsofchangesinfixed costswhichcanhavelong-termeffectsonproduction.Whenoperatingmomentumis growingovertime,ahigherlevelofcapitalexpendituresandfixedassetsmaybeneeded, butonlyoperatingleveragecanspecifytheamount.Thefollowingchartisafouryear comparisonofchangesinfixedcostsandvariousoperatingriskmeasurements: 274 Table10-18 Year Fixed/Total Fixed/Variable Op. Leverage Op.Mom. %Rev.- %TC Year1 0.33 0.5 4 NA NA Year2 0.25 0.33 2 4 16.67 Year3 0.2272 0.2941 1.78 2 3.33 Year4 0.2702 0.3703 2.25 -2.6 -6.24 Ofthethreeoperatingriskmeasurements,onlyoperatingleveragefullyreflectsthe increasesinfixedcosts.Thetwoothermeasurementsreacttotheincreasesthroughthe interfaceoftotalcosts,buttheirmeasuredresponseismorevolatile,atfirstdecliningand thengoingnegative. Ifcompaniesspecifiedfixedandvariablecostsonfinancialstatements,theinvestor wouldnotworryaboutmomentumbecausealloftheinformationtoincreasemarginsin thedomainofriskiscontainedintheoperatingleveragemeasurement.Sincesomeassets havedualuseandhavebothfixedandvariablecosts,derivingagenericmeasurementisa monumentaltask.Otherdilemmasoccurbetweeninvestinginthesizeandstabilityof income,andinvestingingrowthpotential.Liketheequivocationbetweenlargecapstocks andsmallcaps,theanswerismoredependentonthesituation,thanaformula.Picking growthoverstabilityisfineaslongasthegrowthisXeroxinthe1970sorGoogleinthe newmillennium.Sincesuchchoicesarerare,theinvestorwhostickswithstabilitycando well-aslongasthat“RockofGibraltar”typecompany“upstheante”bybalancingstable incomewithgreaterleverage.Themarketwillrewardrisk-seekingbehaviorwhen financialleveragehasa“counterweight”,i.e.,steadyinterestcoverageandalowdefault probability. Whileitisfutiletotimemarkets,itisnevertoodifficulttogaugetheinteraction betweeninterestrates,inflationandmarketactivity,anddecidewhetheroneisintheearly orlatestagesofarecovery;everyoneiswellawareofwhenthelast“downturn”occurred. 275 Itiswhentheanalystattemptstotimethepeaksandshiftsbetweencyclicalstagesthat predictionsgoawry.Mostinvestorsfindthatgrowthestimatesareasmuch“art”as “science”,andthatconcentratingongrowthfactorslikethe“PEG”ratioisarisky propositionatbest.Professionalanalystsgetdirect“guidance”fromcompaniesbecause accurateforecastsofrisingearningswillattractcapital.Formanyyears,estimatesmade fromfundamentalswereoffbyasmuchastwenty-fourpercent-andnotthroughthefault ofanalysts.Thereweresimplytoomanyuncontrollablevariablestomakeanaccurate forecastwiththeconsistencyneededforinvesting.Thedebatebetween“value”and “growth”investingthathasbeenongoingsincethedaysofPhilipFisherandBenjamin Grahamisstillragingtoday;exceptthatitisdirectedtowardtheflowofincome.While thegrowthofoperatingmarginsisconstrainedbytheproductioncharacteristicsofmost industries,themethodsbywhichreturnismaximizedarenot.Therearenumerousother variablesthatcanbeenhanced,evenasmarginstemporarilydecline.Suchclassic argumentslike“leasingversusowning”ormodernmethodslike“outsourcing”,are essentiallywaysofreducingfixedcostsandminimizingriskThus,whentheargumentis phrasedas“riskversusreturn”ratherthan“valueversusgrowth”,newavenuesof approachareencounteredbecauseinterdependenceisrecognized. CLASSICALMICROECONOMICSANDOPERATINGMOMENTUM Studentsoftentakeintroductorymicroeconomicsasaprerequisitetoupper-level courses.Frequently,onewillhearcomplaintsthatprofitmaximizationtheoryisnot “realistic”,that“marginalrevenueneverequalsmarginalcost”.Infact,suchascenariois valid,particularlyincommodityindustries.Asanexampleofthismarginalanalysis, consideranagri-businesswhoserevenueinoneyearis200with40inoperatingprofit,and 160intotalcosts.Thenextyearisveryfertileandrevenuesgoupbyfiftypercent.The companydecidesto“rampup”production,andseesariseintotalcostsoffiftypercent. Whathappenstooperatingincome,momentum,andtheGEinvention,“Distance”?The answercaneasilybedecipheredfromthenumbers: 276 Table10-19 VARIABLE YEARONE YEARTWO Sales 200 300 OperatingIncome 40 60 TotalCost 160 240 OperatingMomentum NA 1 Distance(%Revenue- %TC) NA 0 Inanalternativescenario,whencostsarerisingfasterthanrevenues,operatingmomentum willdecreaseand“distance”willdeclinepastzero. Table10-20 VARIABLE YEARONE YEARTWO Sales 200 300 OperatingIncome 40 40 TotalCost 160 260 OperatingMomentum NA 0 Distance(%Revenue- %TC) NA -12.5 Whiletheprovisionthatoperatingmomentumequalsoneand“distance“equalszero,does notimplythatprofitsarebeingmaximized,anegativechangeindistancecombinedwitha movementofoperatingmomentumtowardzeroindicatesthatquantityisbecoming counterproductive. 277 Figure10-1 Figure10-2 $ TOTALREVENUE (TR) TOTALCOST (TC) $ Quantity OperatingProfit Quantity 278 Thetopgraphshowstotalrevenueandtotalcost,whilethebottomgraphshowsoperating profit.Whenmarginalrevenueequalsmarginalcost,theslopesarethesame,andthe distancebetweentheconcretevaluesismaximized,asisoperatingprofit.When∆ ∆∆ ∆Total Revenue>∆ ∆∆ ∆TotalCost,thereispressuretoincreasethequantityproduced.When∆ ∆∆ ∆Total Revenue<∆ ∆∆ ∆TotalCost,thereispressuretoproduceless.Althoughtheequalityis consummatedwithabsolutesandnotpercentagechanges,thereissomeoptimumof percentagechangesinsalesandoperatingincomethatallowsthis“hypotheticalideal”to occur.Infact,thepercentagechangeinoperatingincomewillapproachzero,becausethe increaseintotalcostsexactlyoffsetstheincreaseinrevenue.Thus,whenoperating momentumapproacheszero,producingmoreofanitembecomescounterproductive–a nightmarishscenarioforanyfirm. (BacktoTableofContents) 279 11 STRATEGICCAPITALREQUIREMENTS Takenoutofcontext,anafter-thefact,“ex-post”analysiscanbebothhumblingand beguiling.Wepresentseveralsimplifiedmethodsofcomparinganidealizedcalculationof acompany’scapitalrequirementstoitsactualadditions.Attheoutset,theinvestorneeds torealizeseveralconceptsbeforeweproceed.Short-termmanipulationofcorporate fundamentalscandamagethelong-termviabilityofacompany.Whilethematerial presentedinthischapterentailsmethodsofraisingcapitaltomeettheobjectivesofthe investor,itdoesnotadvocatethem. Moreover,thecalculationsaremodelsthatareimperfectintheirsimplicity:they leaveoutmanyvariablesandcontainmanyassumptions.Theyfunctionmerelyas guidelinestoeducateinvestorsastowhyafirmbehavedinacertainmanner,andarenot meanttobefinalarbitersfordecision-making-fromeitherinvestors’ormanagements’ perspectives.Nevertheless,wecanextractacomparativelogicbehindsettingshort-term benchmarksforcapitalrequirements,ifweobservesomeoftherealitiesbehind“funding”. THEREALITIESOFFUNDING • 1)Thebestcompaniesoftenmoveawayfromtheirtargetcapitalstructuresforayear ortwo.Certainlargerisksmustbeincurredtoensurelargereturns,andthatentails “takingonestepbacktomovetwostepsforward”,i.e.,acumulativethreeyeargainof fiftypercentisbetterthanatwentypercentgainovertwoyears. • 2.Determiningcapitalrequirementscanbeacomplicatedmathematicalexercisegiven thecomplexityofmostlargefirms.Often,individualassetsareitemizedandthe requirementsofsmallerunitsareaggregated. • 3.Theidealmodelassumeslinearitywhilerealityimposesajaggedcurve.Economies ofscaleanddifferentutilizationratesoffixedassetscreateanexponentialrelationship betweensales,assets,andprofits. 280 • 4.Acquisitionsrequiremuchmorecapitalthanwouldbewarrantedbynormal increasesinsales.Sincemostpubliclytradedcompaniesgrowasmuchthrough acquisitionsasthroughinternal“organic”growth,theneedforfundsescalates. • 5.Onthesamenote,overcapitalizedcompaniesoftendobetterthanundercapitalized firms,becauseinvestorsexpectlargereturnsfromthesecapitalinfusions. • 6.UsingtheEVA/CapitalDynamictodeterminecapitalrequirementsispremisedon theaccuracyoftheCAPM-whichhasbeenshowntoberandomlyunstableandnot alwayscorrelated.Itisalsopossibletohaveanoverabundanceofeitherequityor debt,andstillimproveafirm’sEVA. • 7.Capitalstructureoptimizationintheshort-termmaynottakeintoaccountthe exigenciesofthecompetitiveenvironment,ortheoutlookfortheeconomy.For example,ifinterestratesarehistoricallylow,itmaymakesensetotakeonmoredebt thanisdeemedoptimal,andbuybacksharesofstock.Evenifnetincomeislowerin theimmediateyear,thecapitaldynamicmayrisetogreaterheightsinfutureyearswith fewersharesonthemarket. THEPROPERAMOUNTOFCAPITAL Thegreatestimpedimenttocapitalstructureoptimizationistheadditionoftoo muchcapital.AlthoughmostexecutiveswillconcentrateonminimizingtheWACC,once fundingoccurstherearepracticalconstraintsthatprohibitefficientcombinationsofdebt andequity.Forexample,arestrictivecovenantinanexistingbondcontractmaypreclude theaccumulationofdebtaboveaspecificlevel.Tomaintaincontrol,shareholdersmay wanttorestrictstockissueseveniftheyarewarranted.Short-terminterestratesmaybe sohighthatnormaloneyearloansarediscardedinfavoroflessexpensivelong-termdebt Thus,optimalcombinationsoffundingarenotalwayspossibleunless“ideal”internal conditionsexistforthefirm. Thenwhyistheamountofcapitalsoimportant?TheEVA/capitaldynamicisa universalfilterwithafewlimitedvariables.Itisamathematicalexercisethatcarrieslittle 281 restraint:thehigherthenetincome,andthelowertheamountandcostofequity,the higherthefunction.Whencomparedtopreviousyears’performances,itmustworkwithin specificlimitationsinordertoimprove.Anexcessofadditionalcapitalwillbetoocostlyin relationtotheincomeitisrequiredtoproduce,andwillbemirroredbyadecreased EVA/capitaldynamic.Ifprospectsaretrulyfavorable,thepotentialforefficientcapital utilizationincreases,andEVAwillincreaseaccordingly. IfthestudentexaminestheSpearmanrankcorrelationsinthesectionabout probability,heorshewillfindthatmorecapitaliscorrelatedwithstockpriceincreases, whichoccursbecausebothcapitalandnetincomeareheavilycorrelated.Evenlarge equityincreaseswhencoupledwithlargeearningsgainsusuallyleadtohigherstockprices. Ontheotherhand,largedebtincreasesarerarelyconsistentwithhighnetincome increasesbecausegreaterinterestpaymentscutintooperatingincome,inadditiontothe inclinationfor“troubled”companiestotakeonmoredebt.Thus,greaterequityincreases willaccompanyhighernetincomebecausethereisatendencytobothpayoffloans,and retainearnings;ahigherincomeattractscapitalintothestock.Conversely,mostofthe marketgainswillbespeculativewhenacompanyincursmoredebt:taxbenefitsare immediate,butthe“value”oftheinvestmentmaynotpayoffuntilearningsareactualized. AlsointhechaptersconcerningtheSpearmanrankcorrelations,weobservedthe effectsofcapitalrationing-adiminutionofmarketvaluebecausecorporations“over spent”intheprioryear.Whilethesituationwashypothetical,impliedbytransitivity,it representsthetruedangerofovercapitalizing,becauseacompanymaybeperceivedas moreriskyifincomehasyettocometofruition;otherprofitableprojectsmaybediscarded inordertoconcentrateonthecapitalinfusion.ThisisanotherdangerofusingEVA expectationstomakecapitalrequirementsdecisions:itmayimposeartificialrestraintson capitalthatnormallywouldberaisedtomeetcompetitiveneeds.Inessence,itcreatesa chasmbetweeninvestorswhowantimmediateresultsandfinancialexecutiveswhoare 282 concernedaboutlong-termviability.Ultimately,itmayleadtoa“worstcasescenario”: themanipulationofbalancesheetfundamentalstoappeasethelargestshareholders. THEDEBT/EQUITYTRADEOFFANDEVA Intheveryspeculativeyearof1998,companieslikePMCSierra,SunMicrosystems andLucentTechnologieswerethedarlingsoftheinvestmentset.Thesewerethe companiesthatweregoingtochangeAmericaandbringusboldlyintothetwenty-first century.Fromacapitalstructureperspective,theirunbridledearningspotentialseemedto warrantamassiveinfluxofequitycapital-eventhoughhistoricalearningsfluctuations andalackofoutstandingcreditmighthavedeemedotherwise.Technologyprovided investorswithanelixir-thevisionofatransformedeconomythatwouldalwaysout perform“thepast”.Therestofthestoryiswellknown.Whenthemarketfortechstocks finallycollapsed,millionsofinvestorswereleftwithalargeamountofsharesworthonlya fewdollarseach.Howcouldthishavebeenprevented? Whilenetincomewasclimbingrapidly,thelackoflong-termdebtwouldhavebeen avaguesignalthattheincomestreamwasrisky.Moreover,thelargebetasforthesestocks wouldhavebeenatip-off,becausetheyindicatedalevelofriskthatonlyafewinvestors couldhavesafelytaken.“Buyandhold”strategiesdidnotworkinaneraofmergerand acquisition.The“safe,littlecompany”thatGrandpaputintohisretirementaccountfive yearsago,wasnowahightechbehemoththatwasaboutto“bettheranch”.Evenasa higherEVAmayhaveindicatedmovementtowardtheoptimal(throughabnormallyhigh earnings),athoroughanalysisoftrendsinboththecostofequityandtherateofchangein equity,wouldhaverevealedotherwise.Hadmostofthetechstockscarriedbetasof0.5 insteadof1.5,theimminentdisastermayhavebeenprevented.Althoughthemarket wouldnothave“skyrocketed”,itmighthavestabilizedataslightlyaboveaveragelevel whichwouldhaveattractedcapitalfarintothenewmillennium.Thusthecapitaldynamic /EVAdoesnothavegreatpredictivepowerbecauseitfocusesonimmediateearnings 283 ratherthanpotentialcapitalutilization;itisaconcurrentindicator,andnotaleadingone, andwillnotforecasttrendsoverthecourseofthebusinesscycle. Sinceequityisproportionatelymoreexpensivethandebt,thefirmispresenteda tradeoffbetweentwoformsofrisktoshareholdervalue:theriskofdilutingmarketvalue withmoreshares,andtheriskofcompromisingdividendincomewithmoreinterest expense.Lessdebtwillreduceinterestpaymentscreatingaconsequentriseinnetincome. Iftheearningsthatareretainedfromthisincomerisefasterthanthecostofequity,EVA willincrease:theneteffectofincreasesinincomemorethanoffsetthecorresponding increasesinstockholders’equity.Thisistheleveragemechanismthatissocorrelatedwith movementtowardanoptimalcapitalstructure.Bydefinition,itoccurswhenthedifference betweennetincomeandtheabsolutecostofequityislarge. Onrareoccasions,acompanywillincreasedebtbutdecreaseequitybyaneven greateramount,causingashiftupwardsintheEVA/capitaldynamic.Sinceitisimperative toprotectEPSanddividends,theseareusuallyisolatedoccurrences,indicativeofa recapitalizationsuchasaleveragedbuyout.Manipulationscananddooccursoitis importanttoknowthecontextoftheseshifts.Infact,EVA/capitaldynamicanalysiswillbe impervioustotechniqueslikeshiftingexpensesfromincomestatementstocapitalaccounts simplyto“pumpup”immediateearnings.However,theinvestorshouldcertainly differentiatebetweenequityissuesandretainedearnings,becausetheformerisoftenused as“currency”foracquisitions,whilethelatterhasonlyacomparative,“opportunitycost” whichrarelydilutesmarketvalue. GivenaprojectedEPS,wecansetlimitsontheamountofequitycapitalthatcanbe raisedwithoutdiminishingmovementtowardtheoptimalstructure.Withincapital constraints,aknowledgeoflimitsonequitywillproduceacorrespondingcalculationof allowabledebt-giventhedefaultratingsofagencieslikeMoody’s.Together,wecanuse thisdatatosetarough“guesstimate”forcapitalrequirementsthatwouldbeidealfroman 284 investor’simmediateperspective-andmayormaynotbeconducivetoafirm’slong-term growth. Sincetheinvestors’immediategoalistoatleastpreservethedifferencebetweenthe costofequityandnetincome,itisasimplemattertotake:1)Ananalyst’sprojectednet income;2)AnestimatedrequiredrateofreturnfromtheCAPM;and3)Lastyear’s dataonnetincomeandstockholders’equity-tocalculateanabsolutelimitonthisyear’s equity.Thefollowingexamplewillsuffice: COMPANYXYZ Table11-1 YEAR ONE YEAR TWO NetIncome 100 NetIncome (estimate) 115 Equity 500 Equity ? CAPMPercentage 7% CAPMPercentage 8.5% CostofEquity 35 CostofEquity ? EVA/Capital dynamic 65 EVA/Capital dynamic Goal=65ormore Thecostofequityisderivedbysubtraction.115-X=65,X=50.Thatcostof50isthen dividedbytheestimatedCAPMpercentagetoobtainavalueforequity.50/.085=588.235. Thusgiventhecorrectestimates,themaximumthatstockholdersequitycouldincreaseand stillmaintainthesizeofEVAwouldbe88.235. TheestimationoftheCAPMpercentagecanbedifficultbecauseoneiscorrelating therelationshipbetweeninterestrates,themarket,andafirm’sbeta,butprofessional opinionisusuallyavailableforthecomponentpartsofthisrate.Inthescenarioabove,if XYZcomesinat600foranequityfigure,itdoesnotnecessarilymeanthatitistimetosell thestock,butitwillsignaltheneedtoexaminethecontextofthedeclineinEVA-sales, sectorandmarket-moreclosely.Relianceonestimates,whethertheybefromanalysts,or evenone’sownresearchistreadingonshakyground,andsothevalueofthisanalysis 285 comesafterthefact-expost.IftheEVA/capitaldynamicindicatorisprimarilya concurrentgaugeofstockperformance,whyisitimportanttoknowwhatitisafterthe fact? InordertoanalyzethepotentialforanincreaseintheEVAandbyassociation,the marketvalueofthestock,itisimportanttoknowhowmuchleewayisavailableforan equityincrease.TheXYZexampleabovesuggestedthata17.6%increaseinequitywas enoughtoabsorbthe15%increaseinnetincome.Butwhatifequitywerealreadyhigh? Whatifitcouldonlymoveupfiveorsixpercenttoabsorbtheincreaseinnetincome?In thatsituation,capitalfundingwouldhavetobeprovidedthroughlong-termdebtwhich increasesthevariabilityofEPS.Thus,highequity,whichusuallyactsasabufferagainst risk,canhavetheoppositeeffectifthelevelishighenoughtowarrantmoredebtfinancing. Whenthatscenariooccursatinopportunetimes,suchasinahighinterestrate environment,theentirecompanysuffersthroughhighercapitalcostsandadepressed stockprice.Thelowerthelevelofequityincomparisontonetincome,thegreaterthe potentialformovementtowardanoptimalcapitalstructure. EVA/CAPITALDYNAMICBASEDIMPROVEMENT Inpreviouschapters,wesuggestedthattheEVA/capitaldynamicimprovementis basedupontherelationshipbetweeninterestratesondebt,andtheriskpremiumofthe capitalassetpricingmodel.Moreover,wesuggestedinthischapterthatanoptimization basedonEVAincreasesmightsetthecompanyupforalaterfallbecausecapitalfunding wasnotcontingentonthenetpresentvalueofprojects.Adisconnectoccursanytimethat capitalbudgetingbecomesbasedontheshort-termgoalsoftheinvestor,ratherthanon soundfinancialprinciples-suchasacceptingthoseprojectswithapositivenetpresent value. Tousethismaterialproperly,wewouldneedtoknowthedefaultlimitssetondebt byratingsagencieslikeStandardandPoor’s-andestablishthethresholdlimitsofthe company’sratingclass(AAA,AA,BBB,etc.).Sinceweknowthelimitsonequity,itis 286 imperativetohavealegitimatestandardfordebt/equity-whichcanbeindustryaverages, companyfiveyearaverages,orevenastudyofcyclicalpeaksofthestockprice.Asinthe previoussection,themaximumamountofequityisderived,givenaspecificlevelofincome. Additionally,wedeterminewhatlevelofretainedearningsmustbeachievedtoimprove EVA.Wesetupasimplealgebraicequationandsolveforretainedearningsunderthe constraintthatnoadditionalsharesareissued;retainedearningsareaddedto stockholders’equityoncedividendsaresubtractedfromnetincome.Fromthegivenlevel oftotalcapital,wedeterminethechangeindebtbysubtractingthenewlevelofequity (previousequity+newretainedearnings).Whenwemultiplythisfigurebythecurrent interestrate,wewillobtainaninterestexpensethatcanbeusedtocalculatethenew financialleverageratio.Inessence,weareworkingbackwardstofindthelevelofdebtand equitythatwouldincreaseEVA,aswellasdetermininganylevelofadditionalequity (issues)thatwouldbewarrantedbythedifferencebetweentheoptimalandproposed amounts.Suchanequationisa“win-win”forinvestorsbecauseitessentiallydisplays whereabreakinthemarginalcostofcapitalwilloccur.Theamountofdividendsactsasa stresstestfornetincomeandadditionalsharescanbecomparedwiththenecessary dividendsthatwouldjustifythem. Threeotherassumptionsarerelevanttothisanalysis:1)inagivenyear,theinterest ratedoesnotriseincorrespondencetotheamountofdebt;2)betadoesnotrisewhenmore debtisincurred;and3)operatingincomeshiftstoaccommodateeithermoreorless interestexpense.Thoseareunrealisticassumptions,whichmaketheminvalidinputsfora workingcorporatemodel.The“ballpark”figureisahypotheticalidealmeanttogauge thepotentialforshiftingcapitalproportions.Indeed,anyuseofanEVAdetermined capitalmodelmustharborsomeassumptionsbasedonforecastsandwouldbeproneto error.Whilecorporatecapitalrequirementsarecontingentonsalesexpectations,EVA definedrequirementscoalesceseveralotherforecastsinvolvingvariablessuchastaxes, dividendsandnetincome–eachofwhichcanbefundamentallywrong. 287 INCREMENTALEQUITYIMPROVEMENT Thedefinitionoftheincrementalequityimprovementistheamountofequityto raise,anditsrelationtonetincome,thatwouldmaintainorimproveafirm’sEVA.Inthe example,Year1referstothecurrentyearwhileYear2isconsideredthenextyear. Table11-2 VARIABLES Year1NetIncome Year1PercentageCostofEquityasdeterminedbytheCAPM Year1Stockholders'Equity Year2NetIncome(projectedorestimated) Year2PercentageCostofEquity(estimatedfromchangesincomponentsofthe CAPM) Year2Stockholders'Equity?Determinedbyanalgebraicsolution Year1=(yr1),Year2=(yr2) Year2Stockholders’Equity=((Equity(yr1)x%CostofEquity(yr1))-(NetIncome (yr1))+(NetIncome(yr2)))/%CostofEquity(yr2) ThisfigureisthemaximumamountofequitythatcanberaisedinYear2foran improvementinEVA.AneasiermethodistosetupanEVAequationandsolveforequity: NetIncome-((newcost%)(X)=EVA.TheEVAshouldbeasleastasgreatastheprevious period’s.ThisistermedthethresholdEVA. THEIRRATIONALITYOFRATIONINGCAPITAL Withoutreferencetotheinternaldynamicsofcapitalbudgeting,allocatingcapitalonthe basisofimmediateeconomicprofit,i.e.,EVA,isdoomedtofail.TheEVA/capitaldynamic makesimprovementsbasedonchangesintheshort-termcostofcapital,butacompany’s long-termviabilityiscenteredondevelopingproductsthatmeetmarketdemandoverthe entirebusinesscycle.Thatprocessentailsfundingprojectsbasedonpositivenetpresent valuesandhavinganinternalrateofreturn(IRR)thatisgreaterthanthecostofcapital. Ifindeednetincomerisesandthecalculatedamountofcapitalprovesinadequate,thereis 288 adangerofprojectionsnotbeingmet.Ontheotherhand,thereturnonbothequityand capitalmaybegreaterintheshortrunsinceonlyprojectswiththehighestnetpresent valuewillbeaccepted.Thisrationingofcapitalmayfavorprojectsthatonlyreturnan amountabovethecurrentcostofcapital-whichmaybeacostthatisactuallybelowthe costthatwascalculatedwhensomeoftheseprojectsweredeveloped. INCREMENTALDEBT Themaximumamountofequitytoraisepresentsanextremecornersolutionto capitalfundingifnodebtwereraised.Raisingdebtwoulddilutenetincomewithinterest paymentsbutalsodecreasetherequiredamountofequityandthusitscost.Bysettingthe equationtothepreviousperiod’sEVA,thechangeinequityimpliesachangeindebt. Subtractinganidealizedequityfromagivenamountofcapitalwillyieldaresidualdebt figurethatwemultiplybytheinterestratetodetermineinterestexpense.Thereisno implicationthatthisfigurerepresentstheideallevelofdebt,becausetheamountofraised capitalmaybeincorrect.However,withinthelimitationssetbythemodel,itistheonly realisticchoiceifEVAisgoingtoimprove. DIVIDENDSANDRETAINEDEARNINGS Whenwesolvefornetincome,givenalevelofEVA,wealsoaddretainedearnings tothepreviousyear’sequity.Thedividendamountthatwesubtractfromnetincomemust projectthenextdividendbutusethelastperiod’ssharesoutstanding;wewanttodevelopa netincomewithoutreferencetoanequityissue.Forinvestors,thisassumptionis significantbecauseitcreatesaleveloffundingwherethemarginalcostofcapitalwillnot beraised.Forsensitivityanalysisincorporations,thatstandardcanberelaxed,anda companycanfindalevelofadditionalsharesthatwouldleastdilutestockvalue. CAPITALFUNDINGFROMEVA:TWOMETHODS METHOD1:SOLVINGFORTHEOPTIMALEQUITY Usingtheactualnetincomefromacompany,wehavealreadyidentifiedthe maximumamountofequity.Thisoptimumcanbesubtractedfromactualcapitalto 289 determinetheresidualamountofdebt.Ifweuseanalysts’forecastsforearnings,ouronly alternativefordeterminingcapitalistoassumethatthereturnoncapital(ROC)willbeat leastthesameaslastperiod’s.Whilethisassumptionispurespeculation,ithassome validitybecauseweareimprovingEVAaswell.Thederivedoptimalequityvalueisthen subtractedfromthisquotientofproposednetincomeandpreviousROC.Thereisnoneed tocalculatedividendsinthismethodbecauseretainedearningsareimplicitintheequation. Oncedebtisdetermined,itismultipliedbyanestimatedorcurrentinterestratetoderive interestexpense.Actualorproposednetincomeisthendividedby(1-taxrate)toproduce anewEBT.OnceinterestexpenseisaddedtoEBT,a“target”operatingincomeis determined. METHOD2:SOLVINGFORTHEOPTIMALNETINCOME WelookforaminimumleveloffundingthatwouldimproveEVA,andsowesetthe newequationtothepreviousEVAandsimultaneouslysolvefornetincome,andretained earnings.Dividendsaredeterminedbymultiplyinglastperiod’snumberofshares outstandingbythenextproposeddividend.Thestudent/investorshouldnotethatany amountofhypotheticalEVAcanbeset,aslongasactualshares,proposeddividends,and realisticequityamountsareused. 290 Table11-3 VARIABLES EXPLANATION NETINCOME The"X"Factor PROPOSEDNEXTDIVIDEND ProjectedorKnown OUTSTANDINGSHARES(Year1) LastPeriod'sNumberofShares Outstanding STOCKHOLDERS'EQUITY(Year1) LastPeriod'sStockholders'Equity CAPM%(RequiredRateofReturn) Eitherbasedonpastregressionsor estimated RETAINEDEARNINGS EqualsNetIncome-Dividends NEWEQUITY EqualsRetainedEarnings+Stk.Hlds' Eqty.Yr.1 ThefullequationisNetIncome-((Yr2%CostofEquity)x(NetIncome-(Yr2 Dividends))+(Yr1Equity))=Yr1EVAAlgebraically,wesolvefor“netincome”asall othervariablesareknown.TheYear2Dividendsvariableistheproductofthelast period’soutstandingshares,multipliedbytheprojecteddividend.Again,thereisroomfor adjustmentandsensitivityanalysisonafullspreadsheet. Table11-4 DETERMININGCAPITALPROPORTIONSANDREQUIREMENTS VARIABLES EXPLANATION NETINCOME DerivedfromPreviousCalculation TAXRATE CurrentEffectiveRate YEAR2EBIT Derivedfrom(EBT+InterestExpense) EBT EqualsNetIncome/(1-taxrate) IMPLIEDINTERESTEXPENSE (InterestRate)x(long-termDebt) CURRENTINTERESTRATE TheEffectiveRateEqualsActual InterestExpense/Actuallong-termDebt LONG-TERMDEBT EqualsCapital-Equity Oncenetincomeisderived,wesimultaneouslycalculatethenewamountofequity,because wehaveobtainedtheamountofretainedearningsthatsolvestheequation.Inorderto solveforthenewlevelofoperatingincome,,weneedthecurrenteffectivetaxrate.We 291 subtractthisratefrom“1”anddividethisdifferenceintoourderivednetincometoyielda newEBT(earningsbeforetaxes).Themethodis:NetIncome/(1-taxrate)=EBT.We thenaddtheimpliedinterestexpensetoEBTtodeterminetheamountofoperatingincome neededtoachievethisEVAimprovement PROJECTEDANALYSIS Whilethedeterminationofoptimalnetincomeisusedtoevaluatetheperformance ofcompaniesinanex-postmanner,itisasmallsteptoactuallyprojectEVAscenarios. Thethreemainvariablesneededwouldbe:1)aprojectionofthenextdividendpershare; 2)anaccurateforecastofthenextEBIT;3)somedeterminationofcapitalrequirements;as withtheoptimalequitymethod,wewouldmostlikelyformulateaminimallyacceptable levelofcapitalfundingbydeterminingthequotientoflastperiod’sROCandanalysts’ projectedearnings.Onewouldhavetoextrapolateboththecostofequityandinterest rates,whichisonlyachievableinastableeconomy. FINANCIALENGINEERING:SETTINGCAPITALREQUIREMENTSFROMEVA Ifweassumethatoperatingincomeisconstant,thenouroptimalnetincomewill implyaspecificlevelofinterestexpense,andbydefault,anamountoflong-termdebtthat correspondstoit.Inessence,weworkdeductivelybackwards.Westartwithanoptimal netincome,determineacorrespondingequity,andcalculatetheimpliedlong-termdebt. Whenweaddtheequityamounttothedebtamount,wedetermineaderivedamountof capital.Thismethodisparticularlyinsidiousbecausenoanalysisofcapitalbudgetingor economicviabilityneedstobedone.Wedeterminecapitaloutlaysfrompotentialprofits regardlessoftheirproductivesource. CONOCOPHILLIPS:2005-2006AREALWORLDEXAMPLE Afterfouryearsofrecordprofits,theoilcompanyConocoPhillipsbought BurlingtonResourceswithalargeamountofstock,andassumedthatcompany’s outstandingdebt.ConocoPhillipshadtoraiseenoughcapitaltopayforthepurchase, whichmoveditawayfromitsoptimaltargetstructure.Asapercentageofsales,average 292 earningbeforetaxesis6.4%(EBT,notEBIT),andthecompanywouldnotnormally descendbelowthislevelwithmoreinterestexpense.Thefollowingdataappliesto2005- 2006. Table11-5 CONOCOPHILLIPS Variables 2005 2006 NetIncome 13529 15550 CAPM% 6.82% 8.25% Stockholders'Equity 52731 82646 Long-termDebt 10758 23091 InterestRate(effective) 4.62% 4.71% TaxRate(effective) 42.1% 45.1% EVA/CapitalDynamic 9933 8732 InterestExpense 497 1087 Sales(fromoperations only) 179442 183650 OutstandingShares 1388.98(millions) 1581.25(millions) DividendperShare 1.18 1.44 DividendsPaid 1639 2277 FinancialLeverageRatio 1.0212 1.038 WeobservethatConocoPhillipsincreaseditsfinancialleverageratioonly1.6%to1.038 andthatalargerincreasewouldhavepossiblydeleteriouseffects,giventheprevailing higherinterestratesintheeconomyandshrinkingmarginsonoil.Thegoalistocreatea leveloffundingthatwouldpreservetheprevious9933inEVA.Atenpercentincreasein thefinancialleverageratioiswithinhistoricalbounds,whichwouldmakeitequalto approximately-1.14. METHOD1:CAPITALPROPORTIONSFROMOPTIMALEQUITY STEP1:Determinetheoptimalequity ((Yr1EquityxYr1%Cost)-(Yr1NetIncome)+(Yr2NetIncome))/(Yr2%Cost)= ((52731x.0682)-13529+15550)/.0825=68088.Asecondwayofachievingnearlythe sameresults:15550-(.0825)(X)=9933.(0.0825)(X)=5617,X=68084.85 293 STEP2:SubtractFromcapitaltodeterminedebt 105737-68088=37649 STEP3:DetermineInterestExpense 37649(.0471)=1773.27 STEP4:DetermineOperatingIncomeTarget EBT=15550/(1-taxrate)=15550/0.549=28324.23.Addinterestexpense28324.23+ 1773.27=30097.5ConocoPhillipswouldneed677.5(million)moreinoperatingincometo achievethisEVAideal. METHOD2:CAPITALPROPORTIONSFROMOPTIMALNETINCOME STEP1:Determinethethresholdlevelofnetincome. TheequationisX-((Yr2%CostofEquity)x(X-(Yr2 ** Dividends))+(Yr1Equity))=Yr 1EVA.InthisexampleitisX-((.0825)x((X-2000.13)+(52731)))=9933whichreducesto .X-(0.0825X+4185.3)=9933or0.9175X=14118.3.,X=15387.79.Yr2 ** dividendsare determinedbymultiplyingthelastperiod’ssharesoutstandingbythelatestdividendor 1388.98(1.44)=2000.13 STEP2ComputetheamountofEquity.(Xstkequalsnewequity) NetIncome-(CostofEquity%)(Xstk)=PreviousEVA=15387.79-(.0825)(Xstk)=9933, .0825(Xstk)=5454.79,Xstk=66118.67 STEP3:Determinetheamountofdebt. 105737-66118.67=39618.33 STEP4:DeterminetheamountofinterestExpense (39618.33)(0.0471)=1866.02 STEP5:Determinethetargetoperatingincome 15387.79/.549=EBT=28028.76,EBT+InterestExpense=OperatingIncome,28028.76+ 1866.02=29894.78 FINANCIALENGINEERINGMETHOD 294 DerivedOptimalNetIncome=15387.79.DerivedEquity=66118.67,ActualOperating Income=29420,ActualEffectiveTaxRate=45.1%,(1-taxrate)=0.549 EBT=15387.79/0.549=28028.76.OperatingIncome-EBT=InterestExpense InterestExpense=29420-28028.76=1391.24 ImpliedDebt=InterestExpense/InterestRate,1391.24/0.0471=29538 ImpliedCapital=DerivedEquity+ImpliedDebt,66118.67+29538=95656.67 ActualROE=18.82%,ProposedROE=23.27%,OldLTD/CAP=21.83%,New LTD/CAP=30.88%. ACOMPARISON Whilebothmethodsaddeddebt,theoptimalnetincomescenarioproducedalowercostof capitalandimpliedanoperatingincomethatwasonlyslightlymorethantheoriginal (29894.78Vs.29420).Thelowercostofcapitalwasproducedbyusingslightlylowercost debt. ThecapitalthatConocoPhillipsactuallyraisedwas42248whichisthedifferencebetween thetwoyears’sumsoflong-termdebtandequity.Thisanalysisraisesthesameamountof capitalwithfargreaterdebt.Onadefaultprobabilitybasis,theadditionaldebtmaybe excessive.Bankruptcyriskcancrippleastockandsomostincreasesinleveragearewithin theboundariesoftheindustry,andcorrespondtothehighestobtainablecreditrating. Withinthoseratings,operatingincomedeterminesinterestcoverage,whichfurther determinestheamountofallowabledebt.Thetwoboundariesforoperatingincomeare operatingmarginononesideandthefinancialleverageratioontheother.Toobtainthe optimallevelofdebt,operatingincomemustbehighenoughtoabsorbinterestpayments withoutexceedingindustryaveragesforfinancialleverage,andyetlowenoughsothatthe taxadvantagesofthosepaymentsexceedtheadvantagesofextraincome.Thisbalancing actcontinueswithoperatingmargin:Whilegreatermarginsleadtomoreretained earnings,lowermarginsimplythatlessequityfinancingwillbedone;thesemarginshave tobewithinindustryboundariesorinvestorswillpresumethatthefirmismorerisky. 295 Whenanalystsforecastyearlyearnings,theyextrapolatefromprojectedsales,takinginto consideration,margins,debtandinterestpayments.Therefore,debtbecomesimplicitin mostnetincomeprojections.Effectivetaxratesremainfairlystableformostcompanies, andsoafifteenpercentincreaseinoperatingincomewillimplyfifteenpercentincreasesin bothinterestexpenseandnetincomerespectively. Table11-6 COMPARISON 2005 2006 Financial Eng. Optimal NetInc. Optimal Equity Long-termDebt 10758 23091 29538 39018.33 37649 Equity 52731 82646 66118.67 66118.67 68088 Capital 63489 105737 95656.67 105737 105737 Interest Expense 497 1087 1391.24 1866.02 1773.27 NetIncome 13529 15550 15387.79 15387.79 15550 Financial LeverageRatio 1.0212 1.038 1.0496 1.0666 1.0626 LTD/CAP 16.95% 21.84% 30.88% 36.9% 35.6% Operating Income 24044 29420 29420 29894.78 30097.5 ROE 25.66% 18.82% 23.27% 23.27% 22.84% ROC 21.31% 14.71% 16.086% 14.55% 14.71% EVA/Capital Dynamic 9933 8732 9933 9933 9933 Therefore,withamajorincreaseindebt,EVAwouldhavebeenmaintained,andequity wouldnothavebeenissued.Capitalrequirementswouldhavemetthe105737standard, andthisconformancemayhavebeencrucialtofutureEVAs.Whilewedonotknow whetherConocoPhillips’projectswereproperlyfunded,wedoknowthatcapitalrationing canmakethings“lookgoodonpaper”,underminingtheexigenciesofcapitalbudgeting. WithoutathoroughexaminationofConocoPhillipsleverage,wecannotsecondguesstheir management.Theirproposedcapitalrequirementsweremostlikelyproper,considering 296 thatlargeacquisitionswerebeingmade,andthatfundingwentwellbeyondexisting operations. THEADDITIONALFUNDSNEEDEDEQUATION Thestandardmethodologyforassessingcapitalrequirementsisthepercentageof salesmethodanditscloserelative,theadditionalfundsneededequation,orAFN.These techniqueshavebeenmainstaysoffinancialtextbooksforoverfiftyyears,andseemalmost passégivenWallStreet’smoderndaypenchantformergersandacquisitions:theequation rarelymatchescapitaloutlaysbecauseassetaccretionthrough“organic”salesgrowthhas beensupersededbytheoutrightpurchaseofothercompanies. Withinthecomponentpartsoftheequationliesawonderfulfinanciallogicthatis elegantinitssimplicity.Itsbasicpremiseisthat“nothinghappenswithoutasale”-a principleoftenforgottenbycontemporaryhedgefundswhoaremoreenamoredbyrisk managementthanthemundane“upsanddowns”ofthebusinessworld.Infact,theAFNis almostacomprehensive,shorthandversionofthepercentageofsalesmethod,which itemizeseachbalancesheetitemasapercentageofexistingsales,andthenusesthesales forecasttoprojecttheestimatedsizeofeachitem.Onlythoseitemsthatincrease spontaneouslywithsalesareincluded;theseassetsarecomparedwiththeinternally generatedfundingavailablefromshort-termcreditandretainedearnings,andthedeficitis madeupfromthe“additionalfundsneeded”-externallygeneratedcapitallikebond issues. CorporateanalystsusetheAFNequationtoprojectneededfunding,butinvestors canuseitasascreentoidentifythosecompaniesandindustriesthathaveanabsolute advantageovertheirpeers.Thosecompanieswiththehighestaveragesalesincrease, coupledwiththeleastneedforadditionalfundsaremoredestinedtomovetowardan optimalcapitalstructure;theircapitalcostsarelow.Toillustrate: 297 THEVARIABLES Table11-7 VARIABLE EXPLANATION A/S(assets/sales) AssetsthatIncreasewith Sales ∆ ∆∆ ∆S ChangeinSalesfrom Year1toYear2 (AbsoluteandConcrete- notaPercentage) L/S Liabilitiesthat SpontaneouslyIncrease withSales(notNotes Payable,forexample) S1 AbsoluteandConcrete FigureforProjectedSales M ProfitMargin(net income/sales)Thisisa Projection (1-D) TheRetentionRatioonce DividendsarePaid D DividendPayoutRatio (Dividends/NetIncome Sincepreciseaccuracywiththisequationrequiresanaccountant’sknowledgeofthe company,familiaritywillbreedflexibility,i.e.,onecanderiveaballparkfigureby assumingthatL/Simpliescurrentliabilitiesandthatprofitmarginisafiveyearaverage. 298 Ifdividendsareconsistent,theycaneasilybeprojectedthroughthegeometricgrowth method(seestatisticschapter). THEADDITIONALFUNDSNEEDEDEQUATION (A/S(∆ ∆∆ ∆S))–(L/S(∆ ∆∆ ∆S))-(M(S1)(1-D)) Essentially,theequationiscomposedofassetsthatincreasedwithsalesminusthe internallygeneratedfundingavailablefromtradecredit,wagespayable,andother businessactivity,minustheinternallygeneratedfundingavailablefromincreasedprofits. Anegativeorlowfigureisusuallycorrelatedwithlowersalesasmeasuredonayearto yearbasis.But-onoccasiontherearecompanieswhoconsistentlygenerateinternalfunds abovetheirimmediateneeds,andrequirelittleoutsidefunding.Thesesamecompanies willovercapitalizeforthesakeofacquiringothercompanies-thetypeoffundingacquired byConocoPhillipsinourpreviousexample-andoffertheinvestoragreatopportunity. Overcapitalizationusuallyrequiresamovementawayfromtheoptimalcapitalstructure, andaconsequentloweringofthereturnoncapital,simplybecausethefundsarenotused fortheimmediategenerationofsales.Evaluationisbasedonthecompany’shistorical capitalturnoverrate,sales/capital;ifitishigh,thenovercapitalizationislessrisky.In essence,sincethefirmhasmorecapitalthanneeded,itscapitalcostswillbetoohigh,even iftheindividualcostofthecomponentsislow,andthatscenariocancauseadropinboth EVA/capitaldynamic,andthereturnoncapital. JustlikeEVAandthecapitalrequirementsmodel,theAFNlendsitselfto spreadsheetoptimizationthroughalinearprogrammingmoduleusinganequationsolver. Sensitivityanalysisisalsoeasilyperformedwithadatatableoraonevariablesolverlike “GoalSeek”.LikeEVAhowever,theAFNismechanisticandoblivioustocontext.Asan example,consideracompanywhoseemingly“overfunds”,alowprofitprojectmuchtothe consternationofvariouscorporate“quants”.Ifthatfundingoccursattheendofa businesscycle,itmaybejustthe“strokeofgenius”thatthefirmneedstobufferthemfrom adownturn;thedemandforthatproductmaybeconstant,andlowerfixedcosts.Thus, 299 therearemanyextraneousvariablesthatdeterminethelevelofcapitalfunding.TheAFN givesusacomparative“ballpark”figuretoaddtoourdecisionmakingandcanbea“red flag“tocheckmoreclosely. Mostindustriespossesseconomiesofscaleinwhichmoresalesareincreasingly generatedwithlessassets.Anylargedistributionchainknowsthebenefitsofconsolidating inventoryintoafewkeylocations,andnotduplicatingprocesses.Thesynergiesofmost mergersinvolvethecostsavingsofreducedinputs,asmoreunitsaredistributedoverthe sameamountoffixedcosts.Whilesalesdonotincreaseexponentially,theydoincreaseby amuchgreaterlinearfactor.Secondly,changesinthebusinesscyclecanoffsetadditional fundsbecauseexcessinventoriesbuildupduringunforeseendownturns.Notonlywill theseexcessesbeunaccountedforintheAFNequation,butsotoowilltheunusedcapacity offixedassets;machinerywillbeidlethatwilllowerexpectedsales. Thirdly,thecostandnatureoffixedassetsrequirethattheybeboughtand implementedinlarge“blocks”.Atfirst,theywillbeunderutilized,generatinglesssales andthengraduallyproducingatcapacity.Thisinconsistencycausesalargedropinthe assetstosalescurve,followedbyasteadyincrease. THEMODIFIEDADDITIONALFUNDSNEEDEDFUNCTION ThedeficienciesoftheAFNequationarereadilyapparent.Whileitisausefultool todeterminesomelevelsoffunding,holdingamoderncorporationtotheAFNstandardis myopicatbest.TherealvalueoftheAFNistouseitasanalgorithmtogaugecapital efficiency.Bydividingthelastfigure,retainedearnings,bysalesandthendividing,not subtracting,onefigureintoanother,wecanobtainameasureofcapitalselfsufficiency. Whenwecomparefirmswithinindustries,wefindittobeapreferredmeasureofrisk. Theequationreducesto:(A/S)÷ ÷÷ ÷(L/S)÷ ÷÷ ÷(RetainedEarnings/Sales). Thesignificanceofthesethreecomponentsisasfollows:TheA/Sfigureisacapital intensityratiothatgaugestheamountoffixedassetsinabusiness.Italsogivesa“rough estimate”totheamountofoperatingleverage.ThelargertheA/S,thelargertheamount 300 offixedcosts,withtheimplicationthatabusinessismore“capitalintensive”.Whenthis ratioisespeciallyhigh,earningsmaybemorevariable,andwewouldlookforlessfinancial leveragetobeemployed. TheL/Sratioisvital.Thegreatertheliabilitiesthatrisespontaneouslywithsales, themorefundsthatwillbeinternallygeneratedbybusinessactivity.Thereasonthatsome companiesquicklyturnlong-termdebtintoprofitisthattheycreateliabilitiesthat significantlylowertheircostofborrowing.Vendorrelationshipsthatdevelopsignificant tradecredit,alongwithvolumediscountsandtheabilitytoextendloans,createa“capital substructure”thatisneveradequatelymeasuredbyanalysts.Short-termcreditis normallylessexpensivethanlong-termdebt,andfinancialexecutiveshavelearnedto exploitdisparitiesintheyieldcurvethatofferarbitrageadvantagesincapitalfunding. Theretainedearningstosalesratiowillencompassthosefundsthataregenerated bymoreprofit.Theretentionratioisasignificantcomponentofgrowth;intheGordon model,wemultiplieditbyROEtoobtainagrowthfactorforthecostofequity.Wehave alsoobservedthattoogreatanamountofretainedearningscanraisethecostofcapital, especiallyattheendofabusinesscyclewhenthecostofequityishigh.Fewcompanies havetoworryabout“toomuch”retainedearnings;eveninthecaseofexcess,retained earningscanreduceshareissueswithoutincurringaccountingcosts.Whenthecompany’s returnoninvestmentislowerthanthecostofequity,theyshouldbedistributedas dividends,butforacompanywithbothahighcapitalintensityandvariableincome,they offertheleastexpensive,mostdependablesourceoffunding. ThegistofthemodifiedAFNfunctionistohavethelowestratiowithinaspecific industry.Onecanobservethemechanismifoneconsidersthecapitalintensityratio,A/S, asahypotheticalmeasureofoperatingrisk.Companieswithhighoperatingriskcan affordlessdebtandmustgeneratemoreinternalfunds.Ifwedividetheretainedearnings ratiointoahighA/S,andobtainalargenumber,itmaymeanthatretainedearningswere toosmalltoprovideadequatefinancing,especiallywhencomparedtosimilarcompanies. 301 ThemodifiedAFNfunctionissimilartoEVA/capitaldynamic,exceptonadifferentscale. Improvementisachievedwhentheratiodecreases,anditsabsolutesizedependsonthe typeofindustry;commodityindustrieshavelargemodifiedAFNs,whilehigherprofit businessesfallonthelowerendofthescale.Sinceeachentityiscapableofimprovement,a decreaseisrelativetotheprioryearandtheindustrythatcontainsit.LikeEVA,itruns concurrentwithstockprices,exceptthatitisnegatively,notpositivelycorrelated. Theelegantsimplicityofthemeasurementisthatitcanbecalculatedinthirty secondswithoutreferencetothecostofcapital,andassociatedregression,muchliketheDu PontequationforROE.Initssimplestform,theDuPontequationmultipliesprofit marginbyassetturnoverandtheequitymultiplier,or:(NetIncome/Sales)x(Sales/ Assets)x(Assets/Equity).Thisequationcontainsthesametypesofcomponentsbutgives themthecommonreferencepointofsalesandthendividesthem.Whilethemeasurement, ROE,doesnotdistinguishbetweenthetypesofindustry,themodifiedAFNseparatesthem bytheamountofoutsidefundingrequired,andshouldbeusedforintraindustry comparisons.Thefollowingexamplewillillustratethescopeofthenumbers.The student/investorisencouragedtolookfortrends. Table11-8 XEROX YEARS 1969 1974 1978 Assets 1516 4207 5578 SALES 1357 3505 5902 CurrentLiabilities 419 1050 1339 RetainedEarnings 113.07 249.53 315.85 A/S/L/S/R/S 42.98 56.25 76.39 302 Table11-9 MANPOWER YEARS 1994 1997 2000 Assets 1204 2047 3042 SALES 4296 7259 10843 CurrentLiabilities 668 1005 1522 RetainedEarnings 75.51 149.24 155.61 A/S/L/S/R/S 103.23 98.49 138.89 Table11-10 BARRA YEARS 1993 1996 2000 Assets 38.4 84.2 226 SALES 45.6 105 224 CurrentLiabilities 12.3 32.6 68 RetainedEarnings 3.6 13.5 45.3 A/S/L/S/R/S 39.67 20.06 16.47 Table11-11 WALMART YEARS 1992 1996 2001 Assets 20565 39604 83451 SALES 55484 104859 217799 CurrentLiabilities 6754 10957 27282 RetainedEarnings 1775.55 2567.04 5603.04 A/S/L/S/R/S 95.97 149.14 120.82 Table11-12 MICROSOFT YEARS 1993 1997 2001 Assets 3805 14387 52150 SALES 3753 11358 19747 CurrentLiabilities 563 3610 9755 RetainedEarnings 953 3439 7785 A/S/L/S/R/S 26.51 13.11 18.47 303 Noticethatcompanieswhoeschewoutsidefunding,likeBarraandMicrosoft,fall intoamuchnarrowerrangethanhighturnovercompanieslikeManpowerandWal-Mart. Higherassetturnover(smallercapitalintensity)allowsacompanytotakeonmoredebt withlessrisk,butcutsintoanalreadynarrowprofitmargin.Themarketwillseta premiumontheonequalitythataspecificcompanydoesnotpossess.ForWal-Mart,the premiumisonprofitmarginandmoreretainedearnings.ForMicrosoft,thechallengeis tolowercapitalintensity.Eachcompanyiscapableofimprovementwithintheirrespective degreesofAFN.Moreover,ahigherdegreeofAFNwillgiveacompanymoreflexibilityas tostrategiccapitalallocation;amorediversemixofcapitalfundingisavailablewhena companyusesexternalfinancing.Withnolong-termdebt,acompanylikeBarramustsell stockifretainedearningsareinsufficienttomeetanticipatedneeds.Theyaredependent onthevicissitudesofoperatingincometomeetcapitalgoals.Ontheotherhand,if Manpowerhasapooryear,butanticipatesbetterfutureprospects,theyhavetheoptionof turningtothecreditmarkets,fundingwithdebt,andthenpayingofftheloanwiththe proceedsofanequityissuewhenearningsincrease.Thus,thereisnoabsoluteadvantagein havingalowDegreeofAFN.Theadvantagecomeswhenoneeitherdecreasesthevalue,or possessesthelowestDegreeofAFNwithinasector.Intheformercase,adecreasewould mostlikelysignifythatacompanywasprofitableenoughtopayoffdebt-aslongas existingcapitalneedsweremet.Inthelattercase,thefirmmayhaveaprofitmargin slightlyaboveothersinthesamesector,whichwouldbelikelyifthefirmwereanindustry leader. DEGREEOFAFNLOGIC ThelogicbehindthedegreeofAFNisstraightforward.Mathematically,ifwe multiplybytheinversesoftheequation,wederiveaquotientbetweencostcomponentsand fundingcomponents:(Assets/Sales)x(Sales/CurrentLiabilities)x(Sales/Retained Earnings).Simplified,theexpressionis(AssetsxSales:/(CurrentLiabilitiesxRetained 304 Earnings).Sincetherespectivelevelsofretainedearningsandcurrentliabilitiesare dependentonsalesandassets,anabsoluteincreaseintheratioindicatesmorefinancingis derivedfromexternalsources-whichhasagoodprobabilityofraisingthecostofcapital. WhilewecanneverbecertainthatloweringtheAFNwillalsolowerthecostofcapital,a higherprofitmargin,moreretention,andincreasedbusinessactivityareallcorrelated withminimizingcapitalcosts. HowwellthedegreeofAFNstandsuptoconventionalindicatorslikeROEisthe subjectofcontinuedresearch.Onapurelyutilitarianlevel,bothofthesemeasurements areimperfect;theROEcanimprovedespitemassiveloadsofdebtbecauseleverage(Assets /Equity)isimplicitinthemagnitudeofthecalculation.Ifafirmbecomesobligatedto payinghighinterestratesfarintothefuture,suchanindicatorcanbemisleading.Onthe otherhand,thedegreeofAFNisdependentontherelationshipofsalestotheother variables;thisisabetterconcurrentindicatorofperformance.Whileitdoesnotmeasure majorchangesincapitalstructurelikeROE,itmaybeamoreaccurateindicatorofboth riskandreturn.Withouttheoptimizationofproportionaldebttoequity,ahigherROE canraisethecostofcapital.Toillustratehowclosethesemeasurementscanbe,consider thefollowinganalysisofManpowerin1999-2000.Thecostofequityisskewedupwards becauseitwascalculatedfromaoneyearrenditionoftheGordonModel;amoreaccurate fiveyearCAPMregressionwouldhavechangedthesizeofEVA,butnottheoverall relationshipbetweennetincomeandequity. 305 Table11-13 MANPOWER-1999–2000 VARIABLE 1999 2000 ROE 22.94% 23.12% ProfitMargin 1.53% 1.58% AssetTurnover 3.59 3.56 EquityMultiplier 4.18 4.11 Assets 2719 3042 Sales 9770 10843 Currentliabilities 1418 1522 RetainedEarnings 135 155.61 DegreeofAFN 138.98 138.74 Netincome 150 171 Equity 651 740 Costofequity 20.65% 21.04% EVA 15.56 15.3 NoticethatthedegreeofAFNmirroredtheROEasaperformanceindicator. However,EVAdecreasedeversoslightly;useoftheGordonmodelmakesEVA calculationsmoreexactingbutlessaccurate,becausetheydependoninternal,ratherthan marketdynamics.Mostinvestorswouldhaveavoidedthisstockbecausetherearenoclear indicationsofthedirectionofthecompany,andindeedStandardandPoor’sgavethis stockan“avoid”ratingin2001,becauseofconcernsabouttheemploymentmarketduring aconcurrentrecession. Amorethoroughanalysisindicatesthatthecompanyincreasedbothoperating momentumandfinancialleverage,aswellaslong-termdebttocapital.However,theasset /capitalratiodroppedto2.47in2000from2.7in1999.Whilesomecompaniesavoid short-termdebtwhentheyieldcurveisinvertedasitisbeforearecession,Manpower’s’ ratiowasprobablymoreindicativeoflowerbusinessactivity.Theslightdeclineinasset turnoverto3.56from3.59wouldnothavebeenasforwardlookingasadropincurrent 306 liabilitiesasanindicationofbusinessactivity.Additionally,thecompanywastakingon debtasinterestrateswerepeakingwhichmayhaveraisedthecostofcapital. THENEEDFORQUALITATIVEASSESMENT Althoughnumericalanalysiswasquiteneutral,itprovidedthelogicforqualitative assessment:aseriesoftradeoffsandstalematesleadingtouncertainty.Evenwithalow betaof0.8,Manpowerwas“inthewrongplaceatthewrongtime”.Theywereinthe employmentbusinessinthemiddleofarecession.Althoughtheywouldreceivesomeofthe overflowfrombusinessesthatdidnotwanttofullycommittheirresourcestohiring employees,theycouldnotcompensateforcutbacksintemporarylabor.Thus,capital structureanalysispossessessomepowerfulcomputationaltools,butitisnotasubstitutefor analyticalcommonsense.SellingChristmastreesinJulymaynotbesuchagoodidea, evenifthefundamentalstellyouthatitis. THEPROBLEMWITHOPTIMALITY Inanygiventimeperiod,analystscanconstructacollectiveprobabilitydistribution forsimilarcompaniesconsistingofinterestcoverageratiosandthefrequencyofdefault. Thisdataisthenturnedintoalogittypeprobabilityfunctionthatchoosesbetweendefault andsolvency.Theoretically,distressedeconomictimeswouldcreateahigherstandardand “tightencredit”,while“booms”wouldloosencreditandallowlowerTIE(timesinterest earned)ratios.However,unlessthesecreditstandardsareflexible,theywillbeunableto mirrorthechangesindemographicsandassetstructurethatcomprisethelargereconomy. Whendisparityoccurs,numerousfinancialinstitutionsundergoamassdislocation,with consequentfallouttosmallerbusinessesandtherequisiteFederalReserveinvolvement. Atthehighestlevels,mostmajorcompanieshavesomeoptimizingsoftwarethat attemptstoputconstraintsontheamountofcapitalspendingwithinthedomainofbetter performance.Bychangingthelevelsofconstraintstomeetneweconomicstandards,the companycreatesitsowndefaultprobabilitymodel.Forexample,ifIknowthatthe industrystandardis30%long-termdebttocapital,therewillbesomepenaltyinviolating 307 thatconstraintandthatsomecompensatingfactor-increasedmarketshare,more diversifiedoperations,etc.,-mustcounteractit.Eventually,thecompanyfacesasimilar dilemmaasmostfinancialinstitutions:thegreaterthecomplexityoftheeconomy,the morerandomvariationaffectsdecisionmaking.Probabilitymodelswillsometimesfailto worksimplybecausetheyareunabletoencompassnewinformation.Evenwhenvariables arecomputergenerated,thereisbothalagtimeandinadequacyintheirbreadthof coverage.Thus,wegetneuralnetworksthatprimarilygettheirinputfrompast performanceandareunableto“foresee”newdevelopments. Thecruxofproblemswithdefaultprobabilitymodelsliesnotwiththehuman frailtiesofanalystsorinmathematics,butinhowtheyareused.Basingconsumercredit onaprobabilitymodelthatisderivedfromonecompany,inonetimeperiod,hasthesame risksthatanynondiversifiedbusinessplanhas:itmayworkwellforaperiodoftime, whichencouragesoverdependence,andthencollapsescompletely,becauseitfailstoadapt tochangingcircumstances.Whenwebasecapitalstructureoptimizationonashifting parameterofminimizingcapitalcosts,giventheconstraintsofdefault,weequivocate betweenaddingdebtwhenitislessexpensive(usuallywheninterestratesarelow),and addingequitywhenitisattractive,andearningsarehigh.However,withouttheseself imposedconstraints,capitalspendingoptimizesatastructureofalldebt,simplybecause interestistaxdeductible.Thesolutionofcourse,istonotsomuchconcentrateonthecost ofcapitalasthecostofbankruptcy,whichisevenmoreundefined,unique,andarbitrary thantheformer.Sincebankruptcycostsimplytheuseofdefaultprobabilities,weagain entertheterritoryof“overdependence”-analgorithmthatcansendusinthewrong direction.Tocircumventthisdependence,wemustusetheprobabilityofdefaultina uniqueway:weneedtosolveforthevariablethatwouldmostimprovethealgorithm- earnings-andrelateittheamountofdebt.Secondlyweneedtojuxtaposeimmediatetax benefitstolong-termbenefits,realizingthehighernetpresentvalueoftheimmediate benefit.Lastly,weusethestockpriceasabarometerofvalue:anymarketvaluethatwas 308 createdwithoutlowerdefaultprobabilities,oratleastwithoutmoretaxbenefitsfromdebt, isviewedasatransitorynegative.Thus,theprobabilityofdefaultisasmuchapartofthe solution,asitisaconstraint.Anygenericprobabilityofdefaultcanofferinputonwhatthe bestlevelofearningswouldbe,althoughmathematicalprecisionwouldbelacking. Nevertheless,wecanatleastobtainaballparkfigurethatwouldenableustocomparean idealizedcapitalstructurewithourown.Suchanidealwilllackthemarketadaptability mentionedpreviously,butwillbelessdependentoncreatinganabsoluteconstrainton default;thelevelofdebtwilldependontheinteractionofprofitabilityandtaxadvantages, andonlythenonthelevelofdefault.Inessence,weuseourbiggestliabilities,taxesand debt,toouradvantage-asameasuringstickthatbalancesseveralothervariables- income,assets,stockpriceandtangibleassetsthatcanbeusedascollateral. MERGERMANIA Iftangiblebookvaluepershareisoneofthelinchpinsofouroptimization algorithm,itisbecauseithasbecomeprominentfortworeasons:1)hedgefundsand privateequityfirmshaveusedmergersandacquisitionstocreatetheillusionofgrowth;2) accountingpolicieshavechangedoverfromthe“pooling”methodtothe“purchase” methodofaccounting.Thedifferenceiscrucial:whencompaniespayapremiumoverfair assetvalue,theexcessistermed“goodwill”,andbecomesanintangibleasset.However,it isstillcountedinthe“bookvalue”ofassetsoverliabilities,whichmakesanymultipleof marketpricetobookpriceseemlessspeculative.Infact,theindicatorbecomesmore speculative,becauseitisbaseduponmanagement’sjudgmentofthemerger’spotential, ratherthaninvestors’choicestofundastock. Thelegitimatepurposeofmergersistoincreaseeconomiesofscaleandachieve synergisticearningsthatwouldnototherwisebepossibleinasmallerenterprise. Diversificationreducesriskandanyfirmthatdecreasesoperatingleveragecanultimately increasethevalueofthefirmbyaddingfinancialleverage.However,thescopeoftoday’s mergers,withthehooplaofCNBCcrowingabout“MergerMondays!”,isthatboth 309 investorsandmanagementgetconnedbytheasymmetryofinformation:manyofthese dealsaresimplymethodsofmakingalarge“commission”forafewparticipants.The investorsbelievethatmoreassetswillequalahigherstockprice,whilemanagementhasan unshakablefaithintheefficiencythatcomesfrompoolingresources.Infact,both investorsandmanagementcansuffertheconsequencesofamergerthatlookedgoodon paper,onlytorealizelaterthatgreatermarketsharedidnottranslateintogreaterprofit persharebecausethemarketwasdeclining. Defendersofmergershavethreevalidarguments.Iffixedassetscanbesharedby thetwocompanies,therewillbeeconomiesofscaleandsomesynergythathelpreduce costsandincreaseprofit.Secondly,evenifnosynergyoreconomiesexist,acashflowthat istimeddifferentlyfromtheacquiringcompany’swillhelpreducerisk.Thirdly,greater sizewillallowquantitypurchasesthatwillreducevariablecosts.Whatthesearguments haveincommonisthateachdependsonquantitativegrowth,notnecessarilytothe exclusionofthequalitative-butatleastemphasizingit.Whatinvestorslookforina companyisoftenan“emotionalideal”thatisdifferentiatedfromthecompetitioninsome discernibleway.ConsideralegalmergerbetweenFed-ExandUPS,orCokeandPepsi. Whatwouldbetheproducts?Howcouldinvestorschoosesides?Inessencegreater growthisnotalwaysacashcowforinvestors,especiallyifgreaterdemandisnot stimulated.However,itisalmostalwayslucrativeforthosewhobrokerthedeals.Infact, manymergersarebasedonwhatEugeneBrighamtermed,“theillusionofgrowth”.Here ishowitworks: CompanyXisnotgrowinginternallyanymore.Saleshavestagnatedandtheyneed toincreaseearningspershareorriskanalysts’downgradingthestock.TheirP/Eis20and theirEPSis$1/share.Foryearstheyhavebeengrowingat20%andinvestorsexpect thatratetocontinueinthefuture.Whattheydon’tknowisthataworldwideshortageof thecompany’smaininputwilleventuallyescalatecosts,cripplinganychanceof“organic” 310 growth.Ontheotherhand,companyYisdistressed,andataP/Eof10,lookslikea bargain.CompanyXmakesaplayforcompanyY. IfcompanyXhas20millionsharesoutstanding,eachearning$1pershare,their earnings,ofcourse,are20milliondollars.CompanyYhas10millionsharesoutstanding, eachalsoearning$1pershare,foratotalof10milliondollarsinearnings.After discussingwaystoconsummatethemerger,thecompaniesdecidethatthemostequitable methodwouldbetosetupanexchangeratioof10/20or0.5sharesofcompanyXstockfor eachshareofY.Themergerseemstogounquestioned;P/Eratiosareinlinewith earnings,price,andthenumberofshares.Thenewearningspershareofthecombined companyare:$20Million+$10Million/20millionshares+5millionshares=30/25= $1.20/share.Suddenly,companyXhasgrown20%inEPSsimplybycombiningtwo piecesofpaper! NowwhatifcompanyYisahighlyprofitablebioresearchfirmandcompanyXisa toolanddiemaker.CompanyYhaspatentsabouttoexpireandsothevalueofitsassetsis muchlessthanthe$100Million(5millionsharesx$20/share)instockthatcompanyX paidforit.Infact,supposecompanyYisonlyworth$50Million.CompanyXwould adjustforthedeficitlikethis: Table11-14 NETINCREASEINPLANTANDEQUIPMENT$50MILLION NETINCREASEINGOODWILL$50MILLION NETINCREASEINEQUITY$100MILLION IfcompanyXoriginallyhada4/3(1.33)markettobookvalue,itwoulddeclineto5/4 (1.25),andtangiblebookvaluewouldincreasebyonlyonehalftheassetvalue(50vs.100 Million).Ifwereversetheratioandgobybooktomarketvalue,bookvalueincreases relativetomarket,risingfrom3/4(0.75)to4/5(0.8).Butifwegobytangiblebookvalue pershare,theratiois3.5/5(0.7),whichislessthantheoriginalvalue. 311 Twosignificantpoints:1)BypurchasingacompanywithalowerPE,andpaying foritsmarketvalueinstock,wehaveincreasedEPSby20%withoutanycorresponding increaseinproductivity.2)Bypurchasingacompanyinexcessofitstruemarketvalue, andattributingthedeficittoanotherassetclass,“goodwill”,wehavetaken$50Millionin assets,andcreated$100MillioninmarketvalueInoursimplifiedmodelofbankruptcy costs,adecreasedtangiblebookvaluepersharewascoupledwithgreatermarketvalue, andwouldhaveincreasedthesedistresscostswithoutanyadditionalinterestpaymentsor ratechanges.Sincethedealwaspaidforinstockandnotdebt,therewerenoconcurrent taxadvantagesthatwouldhaveabsorbedthehigherdefaultcosts.Nevertheless,earnings wouldhaveincreasedby20%inthenear-term,leavingahandfulofshort-terminvestors happier,andthebrokersofthedealalotwealthier MERGERGROWTHILLUSIONANDEVA LikeearningsthereisnotelltaleoutlierthattellsusEVAisbeingmanipulated. Assumingthatthe“ruleofthumb”inverseofthePEgivesusanaccuratecostofequity,we canapplyEVAtobothcompanies.CompanyXhasa1/20or5%costofequity,while companyYhasa1/10or10%costofequity.CompanyXhada(20-(.05)(300))or$5 Millioneconomicprofit.CompanyYhada(10-(.1)(100))orzeroeconomicprofit. Togetherthecombinationproduced(30-(((.75)(.05)+(.25)(.!))(400))or30-25=$5 MillionEVAUltimately,thefinalmergerengenderedahighercostofequityforcompany Xwithnorealgainineconomicprofit.Tocontinuegrowingwithoutproductivesynergy, companyXwouldhavetotakeonincreasinglylargermergersorsuffertheconsequences oflessgrowthandalargercostofequity. Themainconcernofanoptimaltargetstrategyistoensurethatthetaxbenefitsof anyleverageusedtobuyanothercompanyareequaltotheadditionaldistresscosts.Ifthe gapbetweenmarketvalueandtangiblebookvalueincreases,thosemarginalbankruptcy costscanoffsetanygainintaxbenefits.Theaverageshareholdercannotbeawareofall thepurchasesmadebyafirm,butlargeinstitutionalinvestorshavetobe.Ifthereisno 312 realgainineconomicprofit,butearningsaretoutedas“growing”,itmaybetimetolook foranotherinvestment. (BacktoTableofContents) 313 314 315 SECTIONII:BUILDINGCAPITALSTRUCTUREMODELS 12 THEECONOMICPROFITLABORATORY:COMPUTER APPLICATIONS Thischapterwillrequiretheusertocreateaneconomicprofitmodelina spreadsheetprogramlikeMicrosoft’sExcel.Withoutsuchaccess,thestudent/investorcan stillfollowtheconclusionsoftheexperiments,andmaywanttoreviewtheentirechapter beforeproceeding.Thereisnospecialskillrequired,eitherincomputerscienceor mathematics-exceptforremedialknowledge.However,theuserwillfindthataworking modelisinstructivewhenheorshetriestofollowthelogicofthematerial. SETUP Theprogramissetupforbothsensitivityanalysisandoptimization.Itwillrequire twosectionswhicharenearmirrorduplicatesofeachother.Thefirstsectionisforinput ofactualvaluesfromfinancialstatementsforsensitivityanalysis.Thesecondsectionwill beasetupforoptimizationoftheactualfigures,andwillallowentryofonlyonevariable. SECTIONONE Sectiononehas14columnsand4rows.Thecolumnsarelabeled“A”to“N”,but columnAisjustatitlecolumn.Theusershouldinputtheword“ACTUAL”inboldprint inA1.Theothercolumnsandrowsaresetupasapatternofonerowoftitlesandanother rowofcalculationsbelowit.Row1isalltitles,androw2isallcalculations.Row3isall titlesagain,androw4iscalculations. 316 SECTIONONE Table12-1 Column/Row Title Calculation Column/Row Title Calculation B1-B2 EBIT ENTER B3-B4 Interest Expense D4xE4 C1-C2 NetIncome (B2-B4)-(C4 x(B2-B4)) C3-C4 TaxRate ENTER D1-D2 Numberof Shares ENTER D3-D4 Interest Rate ENTER E1-E2 EPS C2/D2 E3-E4 Long-term debt H4-J2 F1-F2 Book/sh. J2/D2 F3-F4 Costofdebt D4(1-C4) G1-G2 Risk-free rate ENTER G3-G4 Debt/Equity E4/J2 H1-H2 BETA ENTER H3-H4 Total Capital ENTER I1-I2 Market Rate ENTER I3-I4 WACC ((F4x E4)/H4)+(( K2x J2)/H4) J1-J2 Equity ENTER J3-J4 Costof Capital I4xH4 K1-K2 Costof Equity G2+(H2x (I2-G2)) K3-K4 TotalCost ofEquity J2xK2 L1-L2 EVA C2-(k2x J2) L3-L4 Stock E2/K2 M1-M2 LTD/CAP E4/H4 M3-M4 ROE C2/J2 N1-N2 ROC C2/H4 N3-N4 EVAStock Value L2/I4 SECTIONTWO Sectiontwoisforoptimization.Itstitlesformaperfectmirrorimageofthosein sectionone-exceptfortwocells:inN6thetitleis“EVADifference”andnot“ROC”,with acalculationof(=L7-L2)placedbelowitinN7.Thecalculationsintheoptimization sectionaredeterminedbytheinputsinsectionone,exceptfortheentryof“Equity”inJ7. Technically,sectiontwochangesBETAeachtimeanewproportionofdebttoequityis 317 inputted.Bychangingjustonecell(“Equity”inJ7),anewnetincomeandcostofequityis calculatedwhichchangesEVA. ThebetacalculationsaremerelyrestatementsoftheHamadaresearchthatwas coveredinthechapteronthecostofequity.Byholdingoperatingincome,capital,andthe interestrateconstant,allchangesinEVAaredrivenbychangesinequity.Properly speaking,themodelisbasedonthecapitaldynamicandnotEVA,butweusethese conceptsinterchangeablywheninterestpaidondebtmirrorstheinterestrate. Table12-2 Column/Row Title Calculation B6-B7 EBIT B2 C6-C7 NetIncome (B7-B9)-(C9x(B7-B9) D6-D7 NumberofShares D2 E6-E7 EPS C7/D7 F6-F7 Book/sh. J7/D7 G6-G7 Risk-freeRate G2 H6-H7 BETA H2/[(1+(1-C4)xG4)x (1+(1-C9)xG9)] I6-I7 MarketRate I2 J6-J7 Equity ENTER K6-K7 CostofEquity G7+(H7x(I7-G7)) L6-L7 EVA C7-(K7xJ7) M6-M7 LTD/CAP E9/H9 N6-N7 EVADifference L7–L2 318 Table12-3 Column/Row Title Calculation B8-B9 InterestExpense D9-E9 C8-C9 TaxRate C4 D8-D9 InterestRate D4 E8-E9 Long-termDebt H9-J7 F8-F9 CostofDebt B9x(1-C9) G8-G9 Debt/Equity E9/J7 H8-H9 TotalCapital H4 I8-I9 WACC (F9x(E9/H9))+(K7x (J7/H9)) J8-J9 CostofCapital I9xH9 K9-K0 TotalCostofEquity K7xJ7 L8-L9 Stock E7/K7 M8-M9 ROE C7/J7 N8-N9 EVAStockValue L7/I9 IncellA6,inputtheword“OPTIMIZATION”inboldletters.CellsA7,A8andA9willbe leftblank.Anotheroptionthattheusermayfindhelpfulistocopytheentire“ACTUAL” section,sectionone,andpasteitafewrowsbelowtheoptimizationsection.Inthismanner, thestudent/investorcanmakecomparisonsbetweenthesamefirmordifferentfirmssince themoduleis“standalone”andnotdependentonanoutsidesourceforcalculations. SAMPLEDATA Toemploybothoptimizationandsensitivityanalysis,wewillusesomesampledata toenterintothe“ACTUAL”section.Afterinputtingthevariables,anumberof“derived” variableswillbecalculated.Theyarelistedheretomakesurethatyouare“ontrack”. 319 Table12-4 INPUTED VARIABLES AMOUNT DERIVED VARIABLES AMOUNT Equity 500 Long-termdebt 500 TaxRate 30% EPS 4.375 InterestRate 5% NetIncome 87.5 TotalCapital 1000 BookValue/sh. 25 Shares 20 EVA 37.5 OperatingIncome 150 InterestExpense 25 Risk-freeRate 4% ROE 17.5% BETA 1 ROC 8.75% MarketRate 10% StockValue 43.75 EVAStockValue 35.13 Long-termDebt/ Cap. 50% WACC 6.75% Usethederivedfigurestocheckyourcalculations.Itisquiteeasytoplacean inappropriatesign.Theusercanconfiguretheprogramasdecimalsorpercentsbut shouldprobablysticktodecimalsinthebeginningtomakesureitworkscorrectly. SENSITIVITYVERSUSOPTIMIZATION Sensitivityanalysisismorerealisticthanoptimizationbecauseeachvariablecanbe changedtomeettheactualdemandsoffinancialstatements.Thus,arealisticrenditionof- say,GE’sEVAispossibleaswellasanindicationofthenecessaryinputstoimproveit.On theotherhand,optimizationismoretheoreticalbecauseweneedtokeepbothinterestrates andthenumberofsharesconstant;thesevariablesarederivedfromsourcesoutsidethe model.Thenumberofsharestoissueissometimesapoliticaljudgmentwithalittle mathematicsthrownin,whileinterestrateschangeonaperiodicbasis.Thereisno predictivemodelthatcanencompasseitherofthesevariableswhenmassivechangesin capitalstructurearemade.Therefore,optimizationassumesthatafirmcanraiseasmuch debtasrequiredatthegiveninterestrateandthattheamountofshareswillnotchange. 320 However,giventhoseassumptions,itcanpointthefirmintheproperdirection;without makingchangesinthenumberofsharesissuedorthecurrentinterestratethatispaid,the firmhastheoptionofmovingslightlyinthesuggesteddirection,andsomeimprovementin actualEVAisviable.Athirdnominalconstraintistheamountofcapital;thismodel assumesthattheproperamountofcapitalisbeingraisedandworkswithinthose limitations.Sincecapitalallocationisthemostimportantdecisionthatmanagementcan make,underoroverfundingprospectiveprojectscreatesnumerous“shocks”tothe system.Whiletheseprogramscanexhibithowthoseproblemscanbecircumvented,they cannotcalculatetheproperamountofcapital.“Fixing”theoverageorunderagewitha shiftincapitalproportionwillonlydelaythepainofinefficiencyuntilacompensating capitalinflowactuallyneutralizesit. PROVINGTHECAPITALDYNAMIC/EVAHYPOTHESIS Economicprofitisalinearconstruct.Byminimizingtheamountofdebt,wecan emphasizethe“netincome”sideofthefunctionbecauseinterestwillnotbededucted,and theearningsfigurewillgrowtoitsmaximum.However,ifthecostofdebtisinexpensive enough,thentheinterestdeductionsmaybebothsmallandtaxdeductiblesuchthatequity shouldbeminimized.Minimizingequitywoulddetractfromthe“Stockholders’Equity” sideofthefunctionandcreatealargerEVAbydefault.IfEVAdoesindeedmaximizeat extremesofthecapitalcomponents,thentheremustbesomecombinationofcostelements thatcreatesanindifferencebetweendebtandequity. SETTINGTHECONSTRAINTS Toexaminethishypothesis,wewillusethesampledatafromabove.Atthistime,it isnecessaryforthestudent/investortobefamiliarwithalinearprogrammingmodulelike “Solver”,whichislocatedinExcel’stoolsection.Solverisan“AddIn”forsolvinglinear problemswithconstraints.SinceEVAisalinearfunction,itisamenabletochangingits parametersandmaximization.Ifunfamiliarwiththeprocedure,gototheinformation sectionandreadtheinstructionsonSolver;otherspreadsheetprogramshaveverysimilar 321 modules,anditbehoovesthestudent/investortoatleasthaveaworkingfamiliaritywith these.Inessence,wewillbeabletochangeanyblankinputcell(withoutacalculationinit) thatispartoftheeconomicprofitfunctioninordertooptimizeit.Theprocedureisas follows: STEP1.Gototheoptimizationsection(section2)andinputincellJ7,anynumber between1andtheamountoftotalcapital.Say,500inthisexample. Step2.EngageSolverinthetoolssection.ClickonthecellwiththeEVAcalculationinit, L7.Thatisthefunctiontobeoptimized. STEP3.Wheretheoptionsgive“minimize”,“maximize”,or“setequalto”,click “maximize”. STEP4.Intheboxthatstates“bychanging”,wetypeinJ7 STEP5Wesettwoconstraints.WeaddthatcellJ7isgreaterthanorequalto(>=)the number,“1”.Wedothisbecausethenatureofbetaistodividedebtbyequityandweneed tohaveanonzerofunction.Inthesecondconstraint,wesetlong-termdebt,cellE9toany numbergreatertoorequaltozero(>=)becausewedonotwanttooptimizewithnegative numberswhichwouldbeimpossibletoobtain. STEP6.Click“Solve” Unlessthecombinationofcapitalcostcomponentsisveryunique,thefunctionwill maximizewhenequityiseither“1”oratthecapitallimit(inthiscase1000).Thisextreme cornersolutionmaynotbepracticalorobtainablebutitshowsinwhichdirectioncapital componentsmustmoveinordertoimproveeconomicprofit-withexistingparameters. TRIGGERPOINTS Inalinearfunctionlikethecapitaldynamic,theremaybesomecombinationofcost componentsthattotallyneutralizestheeffectofchangingcapitalproportions.For example,iftheinterestrateishighenough,EVAwilloptimizeatanallequityvalue,butat somelowerrate,optimizationwillbeintheotherdirection-acapitalstructurecomposed ofalldebt.Itishelpfulforafirmtoknowwhatthese“triggerpoints”aresothattheycan 322 setpriorities;attemptingtochangeEVAbychangingcapitalproportionsmaybe productiveatsometimesandnotatothers.Wheninterestratesareespeciallylow,for example,wealthiercompaniescanincurdebtandimproveEVAthroughleveragealone.It ishelpfultoknowwhatthisrateis.Ineffect,thestudent/investorwillfindthatthe relationshipbetweenthe“riskpremium”inthecapitalassetpricingmodel,andtheinterest ratethatafirmpaysonitslong-termdebt,isperhapsthemostimportantdeterminantof EVA.Moreover,thetaxratebecomessignificantnotjustforthedeductibilityofinterest expense,butbecauseitaffectsbetacalculationsandthecostofequity.Withoutariskof default,EVAmaximizationbecomesanexerciseincombiningcostcomponents,anherein liesitsgreatestliability;amarketthatlacksequilibriumcantemporarilyskewthesecost variablesandmisleadafirmintomakingbadjudgments. SETTINGTHETRIGGERPOINTMODULE Intheoptimizationsection,weinputeithera“1”orthecapitallimit(inthiscase 1000)inthe“Equity”cell,J7.WethenengageSolvertosettheoptimizeddifferenceto zerobychangingoneofthecostofcapitalcomponents.Thefollowingbriefprocedure takesusthroughadetermination: STEP1Asstated,entera“1”orthecapitallimitofH4incellJ7 STEP2.EngageSolverinthetoolssectionofExcel. STEP3.ClickoncellN7,“EVADifference”.ThiscellsubtractstheactualEVAfromthe optimizedversion. STEP4.Clickthe“setequalto”button,andenterazero. STEP5.Inthe“bychangingcells”part,pickanycellforwhichyouwishtofindatrigger point.Thecellneedstobeablankentry(nocalculations)andshouldbeacostofcapital variable.Inorder,thoseelementsare:TaxRate,Risk-freeRate,MarketRateBETAor InterestRate STEP6.Iftheconstraintsarenotsetastheywereintheoptimizationmodule,change themtothatconfigurationnow.Otherwiseleavethemastheyare:J7>=1,andE9>=0. 323 STEP7.Click“SOLVE” Asanexample,usecellD4,“InterestRate”asyourtriggerpoint.Enterthatcellinthe“By changing”section.Withthesuggesteddatastillinplace,thatcellwillchangeto7.23%. Keepingallothercostofcapitalcomponentsconstant,ifwechangetheinterestrateto 7.23%,EVAwilloptimizeatonehundredpercentequity.Ifwechangetheinterestrateto 7.22%,EVAwilloptimizeatonehundredpercentdebt.Thus,ifafirmisnearthis combinationofcostelements,theemphasisshouldbeonincomemanagementandnot capitalproportions-atleastinthedomainofEVAimprovement.Atthisjuncture,an indifferencecurveformsandthecostofmakingchangesbasedoncapitalproportionwould belessthanthebenefits. THEEARNINGSSOLUTION WhilemanagingcapitalproportionscanhaveaprofoundeffectonEVA,itisnot nearlyascorrelatedtostockmarketincreasesasearnings.Infact,throughthe managementofretainedearningsalone,afirmcangainacompetitiveadvantage; operatingincomefuelstheprofitabilityofequityfinancing.Althoughthecostofretained earningscanescalateanddiminishEVA,fewfirmswillbemateriallyhurtbyfundingin thatmanner.Theculpritinascenarioofcostlyretainedearningsisalwaysthepercentage increaseinnetincomewhichneedstobehighenoughtojustifyretention.Thus,what mightseemtobeahealthyincreaseinearningsmaynotbeadequatetoeclipsethepotential increaseinequity,andeconomicprofitdeclines.However,afirmwhoisworryingabout “toomuchretainedearnings”,isusuallywealthyenoughtohaveacompensatingdividend policyorasharebuybackprogram. Morenetincomecanalsogiveimpetustoawell-managedshareissuewhichwill usuallymoveacompanyawayfromitstargetstructure.Wheneverafirmexpands throughshareissuesratherthanretainedearnings,flotationcostsareincurredandtherisk ofowningthestockgoesup;theweightedaveragecostofcapitalrises.Moreearningscan bufferthedilutioneffects(bothinEPSandshareprice)ofastockissuewhileattracting 324 buyersandcreatingdemandatahighmarketprice.Anytimeretainedearningsare inadequatetocovercapitalrequirements,afirmhasachoicebetweenvarietiesofother sources:thechoicetofinancewithnewstockseemscosteffectiveifthepriceishigh enoughbecausemorecapitalisraisedwithlessshares.Duringtheissue,EVAcanstill increasedespitethenon-optimalityofcapitalproportions.Inthiscase,morecapitalwould beraisedthanwarrantedwhichwouldincreaseitscost.Thattypeofdecision-toraisea largeamountofcapital-representsatemporaryriskthatshiftsafirmawayfromitstarget capitalstructureinordertogainmorepotentialreturninlaterperiods. OPTIMIZATIONANDCORRELATION Inouroptimizationmodel,therewasaperfectsubstitutionofdebtforequity;aswe loweredonecomponent,weraisedtheother.Ineffect,wecouldincreaseEVAduringa debtissuebecausethereductioninstockholders’equitywasgreaterthanthereductionin netincome.Thisscenariooccurredatlowerinterestrates,andoncea“triggerpoint’was engaged,netincomewasreducedfasterthanequitywoulddecrease. Whileitiscertainlymathematicallypossibletoincreaseboththeproportionofdebt andEVAsimultaneously,itisimprobablethatsuchanincreasewouldberelativelylarge. Fromthenetincomeside,interestpaymentsmaybegreaterandmostcompanieswanta nettaxadvantageovertaxespaid,issuingdebtwhenearningsarediminished.; additionally,thereislessdemandforanequityissuewhenearningsaredepleted. However,thechoiceisrarelyonesourceofcapitaloveranother,butthequestionof,“How muchofboth?”.Firmsfundinadiversifiedmannertolowertheriskofdependence,just astheymayengageseveralvendorsforthesamepart.Equitywillbeincreasedduring mostdebtissues,andsincemoredebtraisesthecostofequity,itwillbemoreexpensiveas well.Thus,thestockholders’equitysideofEVAmayriseevenasnetincomeincreasesare minimalduringadebtissue.TheneteffectwillbeasmallerEVAandnota“managed substitute”ofdebtforequity. RAISINGCAPITALEFFECTIVELY 325 Insensitivityanalysis,wecanlowertheinterestrateenoughtoraisebothEVAand increasedebttoequityproportionally.Thisisthescenarioinwhichmanyfirmsfind themselvesearlyoninarecovery:“theFed”lowersratestothepointwherefirmscannot affordtoeschewdebt.Morecapitalcanberaisedatalowercostwhichincreasesthe optimalityforthosewhofundwithdebt.Thepreviousexampleoftoomuchcapital causingashiftawayfromtheoptimaltarget,becomesmuchlessofaproblem;thereare simplygreatertaxadvantagesatalowerprobabilityofdefault.Infact,theentirecostof bankruptcydiminishesbecauseassetpricesgetdepletedduringadownturn:whenthe prospectsofcoveringinterestpaymentsimprove,theadditionofdebtmaymovea companytowardamoreoptimalstructuresimplybyrestoringgrowthtoassets.Duringa markettop,raisingtoomuchcapitalimpliedtheuseofanon-optimalsourcelike convertiblebondsorastockissue.Inanenvironmentoflowcapitalcosts,however,raising morecapitalisencouragedbecauseitcanbedoneinexpensivelyandwillsetthefirmupfor acompetitiveexpansionwithotherfirms.Thus,itbehoovesanyfirmtokeeptheWACCto aminimumbecausemorecapitalcanberaisedwithsuchacombination;thehallmarkwill behighmarketvaluesaccompaniedbyalowprobabilityofdefault. Forthosewhocanrememberthe“TechBubble”ofthelate1990s,muchofthe damageoccurredbyraisingtoomuchcapitalatthewrongtime-attheendofabusiness cycle.Thepromiseofa“NewAmerica”throughtechnologyalmostmimickedthe“Better LivingthroughChemistry”axiomofthe1960swhenplasticswererevolutionizingthe productindustry.Theoutlookforstartupsandventurecapitalseemedpositivewithno endinsight.However,notonlywerethecostsofbothdebtandequityaccelerating upwardsbylate1998,butfirmswhohadlittlestableearningshistorywereissuingmany sharesofequitytofundwhateverInternet-relatedprojecttheyhaddreamedup.Over productioninfiberopticsrequiredhugeamountsoffixedassets,andyettheworldwas subtlyshiftingtowirelessaccess.Somethinghadtogive. 326 Whenthemarketcollapsedintheearlyyearsofthenewmillennium,thereasonwas notalackofinnovationorvision.Itwasnotbecauseafewspeculatorswerepumpingup themarketwithexcessiveoptimismandaquick”pull-out”.Itwasnotbecausewewere underproducingcomparedtotherestoftheworld.Thereasonwasasimpleneglectofthe principlesofthecostofcapital.Investorswerepumpingmoneyintocompanieswitha diminishedEVA,expectingexcessiverisktoproduceanexcessivereturnwherenonewas warranted. THECONNECTIONBEWTEEN,CAPITAL,STOCKPRICE,ANDEVA Inthecomputerprogram,thereisbothacellfor“StockPrice”and“EVAStock Price”,whichareartificialconstructsthatdivideanearningscomponentbysomeelement ofthecostofcapital.Inthecaseof“StockPrice”itisearningspershare(EPS)dividedby thecostofequity,while“EVAStockPrice”isEVAdividedbytheweightedaveragecostof capital(WACC).Thesevalueshavesometheoreticalvalidityina“no-growth”situation whereallearningsarepaidoutasdividends.However,hypotheticalunderpinnings notwithstanding,theyeachshowearningsaccelerationcomparedtoaccelerationinthecost ofcapital;inamicrocosm,thesearethemaindeterminantsofascendingstockprices. Ifthestudent/investorcanobservethehighcorrelationbetweenEVAandstock price,thenthepercentageformulaforEVAevincestheconnectionbetweenearningsand capital: (NOPAT/Capital-WACC)xCapital=EVA.Whenearnings(NOPAT)declinesand capitalstaysthesame,thenEVAwilldecrease.WhileNOPAT/Capitalisnotquitethe sameas“returnoncapital”or“ROC”whichisNetIncome/Capital,thesimilaritiesare greatenoughtoequatetheentities.Therefore,ifcapitalremainsthesame,andthecostof capitaldeclines,anyincreaseinearningswillhaveapositiveeffectonbothEVAandstock price.Inessence,bothEPS/CostofEquityandEVA/WACCwillalsobegreater. 327 SENSITIVITYANALYSIS:THEEFFECTOFCHANGESINOPERATINGINCOME ANDCAPITAL Increasedoperatingincomeisnotmathematicallylinkedtoequityissuesor increasesinequity.Netincomehasaspecificrelationshiptoequitybecauseinterestis deductedwhendebtisincurred,butthereisnodirectlinkbetweenoperatingincome increasesandariseinequity.However,astrongcorrelationexistsbetweenthetwo becausetheconditionsthatareconducivetotheirgrowthinteractconcurrently:when operatingincomeisup,debttendstogetpaidoff,andearningsgetretained.Usingthe sampledata,wecanreadilyobservetheeffectofonemoredollarofoperatingincomeon EPS,EVA,ROEandhypotheticalstockprice.Achangeinoperatingincomechangesthose variablesmorethananyothers. Table12-5 INPUTED VARIABLES AMOUNT DERIVED VARIABLES AMOUNT Equity 500 Long-termdebt 500 TaxRate 30% EPS 4.375 InterestRate 5% NetIncome 87.5 TotalCapital 1000 BookValue/sh. 25 shares 20 EVA 37.5 OperatingIncome 150 InterestExpense 25 Risk-freeRate 4% ROE 17.5% BETA 1 ROC 8.75% MarketRate 10% StockValue 43.75 EVAStockValue 35.13 Long-termDebt/ Cap. 50% WACC 6.75% Forthisexperiment,operatingincomeisat150,andwearegoingtobothaddandthen subtractoneunittoobservehowtheothervariablesreact. 328 Table12-6 CHANGEOPINC.TO 151 Variable DirectionofChange Amount EPS UP 4.4083 EVA UP 38.065 ROE UP 17.63% HypotheticalStockPrice UP 44.083 Table12-7 CHANGEOPINC.TO 149 Variable DirectionofChange Amount EPS DOWN 4.34 EVA DOWN 36.8 ROE DOWN 17.36% HypotheticalStockPrice DOWN 43.4 Whiletheamountofchangeincapitalisalsopositivelycorrelatedwithstockpricechanges intherealworld,intheeconomicprofitlaboratoryitisnot.Thereason?Thereareno forward-lookingmeasuresofprospectivesuccessandonlycurrentperformanceisgauged. Ifacapitalinflowdoesnotimmediatelytranslatetomoreprofit,theeconomicprofit laboratoryseesitas“deadweight”,neitherproducingmoreincomenorminimizingthe costofcapital.Inthefollowingtables,theproportionofdebttoequityispreservedby addingandthensubtractingtwounitsofcapital.Thusintheadditionalcapitalexample, bothdebtandequityare501,andinthedecrease,bothare499. 329 Table12-8 INCREASECAPITAL BY2 Variable DirectionofChange Amount EPS DOWN 4.3733 EVA DOWN 37.365 ROE DOWN 17.46% HypotheticalStockprice DOWN 43.733 Table12-9 DECREASECAPITAL BY2 Variable DirectionofChange Amount EPS UP 4.3768 EVA UP 37.635 ROE UP 17.54% HypotheticalStockPrice UP 43.768 Duringtheincreaseincapital,amovementtowardnon-optimalitywasenacted.Although betaremainsthesame,earningspershare(EPS)decreasedwithoutanychangeineither thenumberofsharesoroperatingincome.Theadditionalunitofdebtcreatedmore interestexpenseandlessnetincome.Moreover,theadditionalunitofequityeclipsedthe taxbenefitsofdeductibleinterestforcingEVAdownward. Creatinggreaterreturnsthroughtheefficientuseofcapitalisnotoptimizingcapital structure.Whilesuchefficiencyevokestheprocessofoptimization,andiscorrelatedwith ahigherstockprice,theparametersthatgovernEVAareverywide;earningsanomalies, lackofadefaultprobability,andanabsenceofforesightcanskewresults.However,itis nottooradicaltostatethateconomicprofitandstockpricegohandandhand,andthat gainsinEVAareheavilycorrelatedwithmovementtowardanoptimaltarget;false readingsofEVAaretheexceptionratherthantherule. 330 Forthosewhocravecertainty,usingadefaultprobabilitylikeAltman’sZScorein combinationwithEVAanalysiswouldhelpcorroborateit.But-theinvestorneedsto alwaysbewarybecausebothofthesemeasurementsarecoincidentindicators;wecannot makepredictionsfornextyearoffofthisyear’sEVA.However,wecantellwhentheEVA componentsareatapointwhereimprovementwouldbequiteeasytoachieve,given earningsoutlooksandnormalincreasesinboththeriskpremiumandcorporateequity. Alessdefinitivecombinationwouldbetousethecapitaldynamicformofeconomic profitwiththeWACC.SincetheWACCisimplicitinthecapitaldynamic,anerrorinone willcauseanerrorintheother,andsocorroborationinthisregardismoredangerous. Whentheeconomyisnotinequilibrium(thereareinefficiencies)riskmaynotbepriced correctly,andtheWACCwillrisewhenallfinanciallogiccallsforittofall.Additionally, theprecisionofthemeasurementdependsonhowwelltheusercalculatesthecostofequity whichispartiallysubjective.Thus,smallchangesintheWACCcanlargelybeignored. Thecorrelationvalueofeconomicprofitwithanoptimalcapitalstructureis primarilyderivedfromitsabilitytocounterpoiseearningswithcapital.Byforminga constraintbetweenvariablesthatwouldnormallybecorrelated(netincomeand stockholders’equity),economicprofithasalltheingredientsofanoptimizingfunction. Forexample,thebetacomponentinthecostofequitywilldeclinewithanincreasing proportionofequitytodebt,buttheriskpremiumwillrisetocounterthataction:when themarketasawholeimproves,morefirmsuseequityfinancingdespiteitshighercost. Analogously,bothnetincomeandthecostofequityriseconcurrentlybecausemore earningsleadtoabettermarketrateandapossibleincreaseininterestrates.Whenwe mathematicallyopposehighlycorrelatedvariablesweresolvetheconflictbygivingmore weighttoonethantheother.Ineffect,riskmanagementwilladjusttheratesofchangein eachvariablebydampeningitscorrelationvaluetomakenetincomerisefasterthaneither stockholders’equityorthecostofequity. 331 ESTABLISHINGGUIDELINES Unlessoneisinthethroesofabullmarket,investinginex-postEVAincreasesisa strategythatisdoomedtofail.Whilesomesuccessmaybegeneratedfrommomentum, ultimatelytherewillbeasmanytransitionsdownwardasupward,andtheinvestorwillnot beatthemarketandmayevenunderperformit. Abullmarketwillproduceaboutatwentypercentthresholdoffirmsthatwillout performtheirpeersoveraonetothreeyeartimeframe;atanygiventime,onlyasmall percentageofcompanieswillproducelargereturnsThe“miracle”forwhichinvestorsare lookingisthefirmwithsustainableearningsthatacceleratefasterthanthecostofcapital. Ifthecharacteristicsofthesectorandfirmarefavorable,firmsthattravelthelongest distancetoreachanoptimaltargetandyetaremovingrapidlyinthatdirection,tendto havethehighestrisksandreturns.Firmsthatareclosertothetargetbutmakefavorable useofcapitalaregenerallymorestableoverthelongrun.Thereforethefirstguidelineis tohavesomeestimateofwhattheoptimaltargetstructureshouldbe. • 1)Attempttodetermineatleasta“ballpark”figureofwhattheoptimalcapital structuretargetshouldbe.Thismayseemeasierthanpresumediftheanalystiswilling todosomeresearch.Byobservingtheaveragesoverfiveyearsforthethreeleading companiesinasector,orbycoordinatingstock-pricepeakswithcapitalstructurefor anindividualfirm,theinvestorcanachieveaworkable“guesstimate”.The mathematicalmodelinthistextmayconfirmrealworlddata,butthestudent/investor shouldrealizethatitoptimizesdebtandequityinthedomainofnetincome;itisnot forwardlookinganddoesnotconsidernewopportunities. • 2)Knowtheleverageposition.Theeconomicprofitmodelworksbestifequityislow andnetincomeisabouttoincreasesubstantially.Inthatframework,EVAcanincrease fortwotothreeyearswithoutthe“hazard”ofabuildupinstockholders’equity.When afirmisincreasingitsproportionofequity,betawilldecreaseputtingdownward pressureonthecostofequity.However,theinvestorneedstoanticipatethisfavorable 332 positionbyobservingtheleverageratiosandinvestingaccordingly.Seethe “Fundamentals”chapterformoreinformation. • 3)Beawareofthebusinesscycle.Anentirechapterisdevotedtothissubject.Simply observingthepatternofinterestratehikesordecreasesandthemarket’sreactionto themwillgivesomeindicationofwheretheeconomyisheaded.Firmswithlower operatingrisktendtodowellinadownturnandearlyrecovery,whileriskierfirmsdo wellatamarkettop • 4)Useanalyst’sforecaststoyouradvantage.Ifthereisaconsensusaboutthefuture prospectsforasector,byallmeansdosomeresearchonafewofthefirms’leverage positions.Oftentimes,afirmwhohasbuiltupdebtisreadytostartpayingoffits investmentandattractequityinterest.Moreover,earningsestimatesareprobably availableonaprospectiveinvestmenttarget.Comparingthoseestimateswithboth existingequity,andhistoricequitygrowth,willgivetheinvestoraperspectiveaboutthe capacityforEVAimprovement.Ontheotherhand,avoidjudgmentsaboutindividual firms.Theremaybesomanydowngradesandupgradesonafirmthattheevaluations becomemeaningless. • 5)Useexecutivetradesasasignal.Oncethestudent/investorcanspotaleverage positionthatmightyieldafavorableEVA,checktoseeifcompanyinsidersare investinginit.Maketheefforttoresearchtheleveragefirst,andusethelistedinsiders’ tradesasconfirmation.Donot,however,mistakeblindoptimismforaninvestment opportunity.Manyinsidersarerequiredtotakeastakeinthecompany. (BacktoTableofContents) 333 APPENDIX EVAOPTIMIZATIONWITHSAMPLEDATA A B C D E F 1ACTUAL EBIT NetIncome #Shares EPS Book/Share 2 150 87.5 20 4.375 25 3 InterestExpense TaxRate InterestRate LTD CostofDebt 4 25 30.00% 5.00% 500 3.50% 5 6OPTIMIZE EBIT NetIncome #Shares EPS Book/Share 7 150 70.035 20 3.501 25 8 InterestExpense TaxRate InterestRate LTD CostofDebt 9 49.95 30.00% 5.00% 999 3.50% 10 11COMPARE EBIT NetIncome #Shares EPS Book/Share 12 0 0 Div0! 25 13 InterestExpense TaxRate InterestRate LTD CostofDebt 14 0 0 0.00% G H I J K L M RF Beta K(m) Equity CostofEquity(Rate) EVA LtD/Cap 4.00% 1 10.00% 500 0.1 37.5 0.5 D/E TotalCap WACC CostofCap TotalCostofEquity Stock ROE 1 1000 0.0675 67.5 50 43.75 0.175 RF Beta K(m) Equity CostofEquity(Rate) EVA LtD/Cap 4.00% 411.94118 10.00% 1 24.75647059 45.279 0.999 D/E TotalCap WACC CostofCap TotalCostofEquity Stock ROE 999 1000 0.0597 59.7214706 24.75647059 70.724 70.035 RF Beta K(m) Equity CostofEquity(Rate) EVA LtD/Cap 0 0 #DIV/0! D/E TotalCap WACC CostofCap TotalCostofEquity Stock ROE #DIV/0 ! #DIV/0! #DIV/0! 0 #DIV/0 ! #DIV/0! N ROC 0.0875 EVAStock 35.12880562 EVADifference 7.778529412 EVAStock 42.72682084 ROC #DIV/0! EVAStock #DIV/0! 334 FINDINGATRIGGERPOINTFORTHEINTERESTRATEONDEBT A B C D E F 1ACTUAL EBIT NetIncome #Shares EPS Book/Share 2 150 79.706 20 3.9853 25 3 InterestExpense TaxRate InterestRate LTD CostofDebt 4 36.134 30.00% 7.23% 500 5.06% 5 6OPTIMIZE EBIT NetIncome #Shares EPS Book/Share 7 150 105 20 5.25 25 8 InterestExpense TaxRate InterestRate LTD CostofDebt 9 0 30.00% 7.23% 0 5.06% 10 11COMPARE EBIT NetIncome #Shares EPS Book/Share 12 0 0 #DIV/0 ! 25 13 InterestExpense TaxRate InterestRate LTD CostofDebt 14 0 0 0.00% G H I J K L M RF Beta K(m) Equity CostofEquity(Rate) EVA LtD/Cap 4.00% 1 10.00% 500 0.1 29.706 0.5 D/E TotalCap WACC CostofCap EquityCost Stock ROE 1 1000 0.0753 75.294 50 39.853 0.1594 RF Beta K(m) Equity CostofEquity(Rate) EVA LtD/Cap 4.00% 0.5882 10.00% 1000 0.075294 29.706 0 D/E TotalCap WACC CostofCap EquityCost Stock ROE 0 1000 0.0753 75.294 75.294 34.863 0.105 RF Beta K(m) Equity CostofEquity(Rate) EVA LtD/Cap 0 0 #DIV/0! D/E TotalCap WACC CostofCap EquityCost Stock ROE #DIV/0! #DIV/0! #DIV/0! 0 #DIV/0! #DIV/0! N ROC 0.079 EVAStock 27.626 EVADifference 0 EVAStock 27.626 ROC #DIV/0! EVAStock #DIV/0! 335 13 THEMARGINALBENEFITSEQUATION:ANEXPERIMENTAL MODEL Oneofthegreatestconflictsinourtimecomesfromthetensionbetweenobfuscation andtransparency.Manywhoholdpowerdesiretocreateasystemofasymmetrical informationwhereinspecializedknowledgebecomesabridgetoaccomplishingatask. Whentheinformationishoardedandisolated,itisusedforpoliticalpurposes,creatinga certainpositionorstatuswithinagroup.Thus,patents,copyrights,“secretformulas”, esotericclubmemberships,andcompromisingphotosallharborthecorrelationbetween specializedinformationandpoliticalpower.However,inanagethatdisseminates informationveryquickly,powertendstoriseandwanejustasfast,becauseofthe transformationalqualityoftheinformationitself Defaultalgorithmsarepowerfultools.Whenthestatisticiansuddenlybecame infusedwiththepowertodecidewhogetscreditandwhoisrefused,thestakesinthat occupationroseexponentially.Themassivepopularizationof“creditscores”andthe influencetheyentail,isjustoneexampleofan“invisiblehand”regulatingmostofour publicaffairs.Intheframeworkofcapitalstructure,theprobabilityofdefaultisthe “missinglink”betweentheresearchoftheMiller/Modiglianiteamandthepractical applicationofmovementtowardatargetoptimalstructure.Infact,alloftheconclusions fromtheirresearchwerearrivedatgiventheabsenceofbankruptcy,anditwasasimple “tweak”oftheirfamousequation,V(L)=V(U)+TBthatprovidedtheimpetusfor optimization:wesubtractthecostofbankruptcy. Thereinlaystheproblemofoptimization.Thecostofbankruptcyisnotageneric cost,norisitalegallyuniformterm;theremaybemanyunique“costsofbankruptcy”that applytoeachfirm’sspecificsituation.Onecompanymaysellallitsassetsandanother may“reorganize”;athirdmaymergewithacompetitorandbecomeacompletelydifferent 336 company.Incapitalstructureanalysis,wearetaskedwithcreatingacommoncostof bankruptcythatattemptstoreconciletwoentities:thebookvalueoftheproportionof debttoequity,andthemarketvalueofafirm’sstock.Consequently,wediscoveran interfacethatunitesthosevaluesintheprobabilityofdefault-theriskthatthefirmcannot maketimelyinterestpayments.Implicitinthatinterfaceistheutilizationofassets:the rateandamountofgeneratedincome. THEMODELEDCONCEPT Underthepremisethatthereexistssomeoptimalproportionofdebttoequitythat maximizesthevalueofthefirm,wecreateafunctionalmodelbasedonthe Miller/Modiglianiconcept.Taxadvantagesofdebtarebalancedwiththeproductofthe amountoflossandtheprobabilityofdefault,whichwehavetermed,“thecostof bankruptcy”.Therefore,anyadditionalvaluecreatedbydebtwillbeapparentwhentax advantagesexceedbankruptcycosts.Whenthechangeinbankruptcycostsbeginsto exceedthechangeintaxadvantages,anoptimumisfound,andthisisthelimitingfactoron additionaldebt.Ineffect,thechangeintaxadvantageswillequalthemonetarychangein thecostofbankruptcy,andthefirstderivativeoftheentirefunctionwillequalzero, becauseamaximumhasbeenobtained Ideally,theequationwewanttomaximizeis:[(TaxRate)(Long-termdebt)]-[(% ProbabilityofDefault)(AmountofLoss)].Furthermore,theamountoflossisconstructed ofthereciprocaloftheratiobetweentangiblebookvalueandmarketvaluemultipliedby marketvalue.Thatis-theamountoflossisequalto:[(1-(TangibleBookValue/Market Value))(NumberofSharesOutstandingxPriceperShare)].Fortheratiooftangiblebook valuetomarketvalue,wecaninterchangeablyusepersharevalues,andmarketvaluewill bedeterminedbytheproductinthelastparentheses.Ineffect,thisequationistellingus thattheoptimalamountofdebtisbasedonabalancebetweentaxadvantages,default performance,assets,andstockprice. THEMARGINALBENEFITSEQUATION 337 Onceweacceptthepropositionthattheincrementalvalueofafirmisequaltothe differencebetweenbankruptcycostsandtheproductoftheamountofbondsandthetax rate,weneedtoaccountforthosefirmswhocreatevaluewithoutdebt.Manycompanies lacktheassetstructuretosupportdebttothepointwhereityieldsataxadvantage;sales maynotbestableorassetsmaynotoffergoodcollateralvalue.Thus,weassumethatthe unleveredfirmhasthesameincomegeneratingcapacityastheleveredfirm,andthatthe onlydifferencebetweenthemishowtheyarefunded;aplantassethasthesameearnings potentialwhetherfinancedbydebtorequity.Withtheadditionofbankruptcycosts, however,thepropositionchangesbecausebankruptcyisnotalinearfunction;ithassome thresholdamountandrisesatanon-constantrate.Inessence,thetypeofassetsdetermines boththebankruptcyfunctionandtheamountoffinancialleverageavailable.Thosefirms whoproduceahighoutputatalowerfixedcostwillhavealowercostofbankruptcyand havethecapabilityofincurringmoredebt. Hypothetically,firmsthathavenodebtwillhavealargeramountofpotentiallossto shareholdersthatwillprohibitthemfromtakingonleverage.Byminimizingthenumber ofcommonsharesoutstandingandyetgeneratingahighlevelofsales,thesefirmscankeep bankruptcycoststoaminimum.Toanunleveredfirm,bankruptcycostsrepresenta thresholdamountthatmustbeexceededbypossessingahigherearningscapabilitythana leveredfirm:whattheyloseintaxadvantages,theymustmakeupinthepotentialfor generatingincome.Withoutsuchcapacity,therewouldbenoreasontobeinanybusiness thatdidnotrequireleverage,becauseallunleveredfirmswouldbevaluedlessthantheir leveragedcounterparts. Theincrementalvalueofafirmwhohasnodebtislessthanzero,becausetherewill alwaysbeaminimumlevelofbankruptcycoststhatwouldbederivedfromoperational incapacity.Thus,theunleveredfirmmustgeneratealevelofincomethatovercomesthis functionaldisadvantage.Whenwesubtractbankruptcycosts,weassumethattheyrise withthelevelofdebt,andareuniquetothatfirm.Thecostofbankruptcy,however,is 338 moremultidimensional;notonlywillbankruptcycostsrisewiththelevelofdebt,theywill adjustforthetypeofasset,andtheincomegeneratingpotentialthereof.Inthismuchmore realistic,non-linearcapacity,twicethedebtwillnothavetwicetherisk,butwilltakeon anynumberthatthebankruptcyalgorithmgivesit.Inthiscontext,“operatingrisk”isa confluencebetweentangibleassetsandtheprobabilityofdefaultandoperatingleverageis implicitlydefined. Althoughamodelcangiveanestimateofanoptimaltarget,itworks probabilistically,definedbythecomponentsofthealgorithm.Therelevancyofanycredit algorithmchangesperiodicallyastherelationshipbetweenincomeanddebtshiftsinthe greatereconomy.Therefore,anyworkingcapitalstructuremodelwillalwaysremain transientandexperimental.Notallcompanieswillmeetitsconstraints.Somemayevenbe inoperable.Banksandinsurancecompanies,forexample,haveassetstructuresthatare muchdifferentfromamanufacturingcompany.Likewise,autilitycompanymaytakeon moredebtatahigherdefaultratesimplybecausethelocalgovernmentinfusesitwithcash. Theseareanomaliesthatareoutsidethepurviewofthemodelandnegatethepossibilityof inventinga“onesizefitsall”algorithm.However,therearecommonaltiesamongall bankruptciesthatcanbeaddressed,especiallyconcerningthefateofthecommon shareholder,andthesearetheinputsintothemodel DEFAULTPROBABILITYANDBANKRUPTCY Amongthegreatnumberofbothgenericandcommercialdefaultalgorithms available,ahandfulwouldfittheneedsofcapitalstructuretheory.Thealgorithmneedsto bebothprohibitiveofexcessivedebt,butflexibleenoughtoinputandsolveforvariables. Italsoneedstobecheckedagainstsomeotherstandard,likesalesornetincome,for viability.AnadaptationoftheZmijewskialgorithm,hasthatcapacity. 339 TheZmijewskialgorithm * willallowforthesolutionofvariablesthathelpminimize it.Withtheadditionofthetaxbenefitsofimmediateinterestpayments,itoptimizesinthe domainofnetincome,allowingfortheoptimizationofbothnetincomeanddebtwhenthe nettaxbenefitsofinterestaredividedbythenettaxbenefitsofdebt.Byitself,withoutthe additionofinterestbenefits,thealgorithmisnotrobustenoughthroughoutitsentirerange tooptimizedebt.Likemostdefaultalgorithms,itisconstructedfromaveragesthatwill allowacompanytobecapitalizedentirelybydebtbeforeitregistersa“onehundred percent”chanceofdefault.Thus,itwouldnothelpmeetthetestofastandalonemodel thatmaximizeswhenthefirstderivativeisequaltozero.Nevertheless,whenadaptedfor interestbenefits,andusedintwocounterpoisedequations,itwillyieldanamountofdebt thatcreatesparitybetweentaxadvantagesandthecostofbankruptcy.Givenaspecific levelofcapital,theequationthenfindsthelevelofdebtthatkeepsdefaultvalueslowand incomelevelshigh.Theoptimizationisnthedomainofnetincome,avariableinthe defaultalgorithm,allowingittomeetacomparativestandardotherthanitsownincreasing value. THEINTERESTBENEFITSMECHANISM TousetheZmijewskialgorithmincapitalstructureoptimization,wemustmodify themarginalbenefitsequation.Ineffect,weformaratiobetweenthemarginalbenefitsof interestexpenseandthemarginalbenefitsoflong-termdebt.Inthenumerator,weinput thefullmarginalbenefitsequation,exceptthatwereplacelong-termdebtwithinterest expenseinthefirstpartoftheexpression.Wethendividebytheentiremarginalbenefits equation,andattempttomaximizethefunction.Thefinalfunctionlookslikethis:[(Tax Rate)(InterestExpense)]-[(%ProbabilityofDefault)(AmountofLoss)]/[(Tax Rate)(Long-termdebt)]-[(%ProbabilityofDefault)(AmountofLoss)].Anyincreasein theratioisinterpretedasmovementtowardtheoptimal.Alternatively,whentheratio * Zmijewski’soriginalanalysisof840bankruptandsolventcompaniesusedprobitanalysistoformanalgorithm. Thisisalogiisticversionofthatresearch. 340 decreaseswhilemarginalbenefitsareincreasing,theoptimumhasbeenpassed:thenature ofthealgorithmallowsmarginalbenefitstoincreasepasttheoptimumwhichisthereason itneedstobemodified.Inayeartoyearcomparison,themarginalbenefitsequationwill beliketheweightedaveragecostofcapital(WACC):movementisatentativeindicator becauseothervariablesmayinteractwithitandskewtheresults.However,whenusedin combinationwithinterestbenefits,theprobabilityofdefaultwillbeloweredinthedomain oftaxadvantages,anda“guesstimated”targetcanbeobtained CHECKINGRESULTSAGAINSTAVIABLESTANDARD Thegistofoptimizationoccursintherealmofbalancingthereturnonequity(ROE) withthereturnoncapital(ROC).Whendebtreplacesequity,greaterinterestpayments depletenetincome.However,equitydecreasesatamorerapidrateandthereturnon equityrises.Inthemeantime,ROCisdecreasedbecausetheeffectofmoredebtisto decreasenetincome.Althoughthecapitalbaseremainsstable,ROCdeclineswhenROEis atamaximum.Ontheotherhand,whenequityreplacesdebt,theoppositephenomenon occurs:ROCismaximized.Withlessinterestpaymentsfordebt,netincomeincreases untilitismaximizedinacapitalstructureofallequity.Toachieveanoptimumfordebt, theprogramwilltradebasispointsbetweenthetwomeasurementsbecauseitattemptsto createthegreatesttaxadvantagesinthedomainofalowerdefaultprobability.Thetax advantagesaregreaterwhenROEismaximized,andthedefaultrateislesswhenROCis maximized.Together,thetwomeasurementsareoptimizedbyinputtingthecorrect amountoflong-termdebt. DEFAULTMECHANICS TheZmijewskimodelhasaminimumnumberofvariables(four),buttheinherent flexibilityofthemodelmakesiteffective.Theproductofparametersandfundamental ratiosformsthelogarithmofaprobabilityofdefault.Wealgebraicallyeliminatethe logarithmandsolvefortheprobability.Thus,Ln[P1/(1-P1)]=X1BwhereP1isthe probabilityofdefault,X1arethefundamentalratios,andBarethecoefficientsofthe 341 algorithm.Toobtainaprobability,weturntheequationaroundandinputP1=1/1+ EXP[-XB],wherethefundamentalratiosaregivennegativesigns.Theinterceptisalso givenanegativesignbecauseitwasoriginallymultipliedby“1”,andnowitisbeing multipliedbynegativeone(-1)toformapositivenumber.Thefollowingtablecontainsa definitionofthefundamentalratiosandthecoefficientsofthealgorithm. Table13-1 ZmijewskiDefault NAME FUNCTION COEFFICIENT TL/TA TotalLiabilities/Total Assets 6.384 CA/CL CurrentAssets/Current Liabilities 0.069 NI/TA NetIncome/TotalAssets -1.06 Intercept NONE -9.479 Asanexample,wecaninputtypicalratiostoshowhowthebasicfunctionworks:TL/TA =0.5,CA/CL=1.5,NI/TA=0.07. P1=1/1+[EXP((6.384)(-0.5))+((0.069)(-1.5))+((-1.06)(-0.07))+((-9.479)(-1))]= 1/1+EXP(6.2577)=.00191or0.19% Whilewecanquestiontheaccuracyofthisdetermination,astudent/investorwho researchesgenericalgorithmswillfindthattheZmijewskiprobabilitysharessome commonaltieswithothers.Ohlson,ShumwayandMertoneachexplicitlyinputtedan assetsvariable,anincomevariableandanexistingdebtvariableintotheirdistributions. Infact,eachoftheseisdesignedtodetectmajorfinancialcatastrophes,butfallshortof predictingbankruptcyforafirmthroughoutitsentirerangeofdebt/assetcombinations. However,theyareexplicitriskindicatorsandespeciallyworkwellintermsofmeasuring themagnitudeofchangesinoverallfinancialposition. STRATEGICIMPLICATIONS:FINANCIALLEVERAGE 342 Whendebtisaddedtothecapitalstructure,achainofeventsensuesthataffects eachcomponentofthemarginalbenefitsequation.First,thetaxbenefitsrise,increasing thetotalincrementalvalueofleverage.Secondly,thequalityofpurchasedassetsbecomes paramountbecausetheinfusionwillinitiallycauseanincreaseintheprobabilityofdefault; incomemaynotbeimmediatelygeneratedfromapurchaseandthelagcausesthe probabilityofdefaulttorise.Thirdly,ifmorecommonsharesarenotissued,thebalance betweenmarketvalueandtangibleassetsmaychange,affectingtheamountofloss. Therefore,anycompanythattakesondebtmustmonitorthedefaultratetoseehow thesubtractionofinterestpaymentsaffectsnetincome,andhowtheotherratioscanbe bufferedtomaintainthecurrentlevelofsolvency.Ultimately,themorerapidlyafirmcan turnapurchaseintoanincomegeneratingasset,thelesseffectondefault,andthebetter wasthedecisiontousedebt.Intangibleassetsarenotnecessarilylessworthythantangible assetsbutwillbemoredifficulttocollateralizeandamortize.Moreover,anincreasein tangibleassetswillallowmarketvaluetogrowandstillmaintainmoremarginalbenefits. Inthismodel,ifmarketvaluegrowswithouta“legitimatereason”-thatis-withoutgreater taxbenefitsorlowerdefaultprobability,itisobservedtobeaspeculativerun-upandis expectedtofallprecipitously. Fromtheequityside,thereislessconcernaboutgeneratingincomefromassets becausenetincomeisimplicitinequitygrowthfromretainedearnings.Naturally,any stockissuewouldberegisteredasmoresharesoutstandingwhichinisolation,wouldraise theamountofloss.However,equityderivedfromretainedearningsshoulddecreasethe probabilityofdefault,andthatistheprimarysignaltoobservewhentheproportionof equityisincreased.Asecondsignalwouldbedebtneutralityorevenslightincreasesin long-termdebtthatwouldincreasetaxadvantages.Ineffect,payingoffdebtandreplacing itwithequityisnotconducivetoassetgrowth,andmayevenincreasethecostofcapital- but-increasingtheamountofbothwilloftenunitethetwinobjectivesofgrowthand optimalproportion. 343 STRATEGICIMPLICATIONS:OPERATINGRISK Weassumethattheparametersofthismodelarecorrect,buttheyareexperimental. Theremaybebetter,“standalone”defaultalgorithmsthatneednomodification.Amore precisebankruptcycostcanbedeveloped.Andespecially-amodulethatallowsforthe interactionofoperatingleveragecanbeinputted.Infact,withouttheinclusionof operatingleverage,acapitalstructuremodelisincomplete.Onlybyobservingtheeffectof increasedfixedcostsonafirm’sbreakevenpoint,cantheproperamountofleveragebe determined.Whatwehavedoneissubstitutedassetclass,anddefaultprobability,forthe effectofoperatingleverage,makingitanimplicitfactor.However,theeffectofoperating leverageisparamounttoincomegenerationbecauseitwilldeterminethe“base”from whichfinancialleveragecanoperate.Forexample,alargeoperatingleveragewillproduce alargerincreaseinoperatingincome,whichnaturallydiminishestherateofdefault.A lowerdefaultratewillimplytheuseoflessdebtandmoreretainedearnings. Consequently,afirmwhoachievesthisfinancialstatewithfewersharesoutstandingwill seetheirstockpricegrow. Whenoperatingleverageisbothlargeandunstable,therecanbenotaxadvantages fromdebt,andmarketvaluewillbeafunctionofmoreshareissuesandlessprice appreciation-thesamedebaclethathitmanyNASDAQstocksinthelate1990s.The problemwithhighoperatingleverageisthatitwaversandmaycreateanunstabledefault probability:nocreditorswanttotakeachanceonacompanywhosedefaultrateisless thanonepercentinoneyearandtenpercentinanother.Thelarge“intercept”inthe defaultalgorithmaccountsforthevarianceofinput,butalsoaccountsforalarge percentageofdefaultprobability. Stableoperatingmarginsleadtotheproperuseofleverage.Althoughfirmswith high,unstableoperatingleveragecanprosperatcertainpointsinthebusinesscycle,their stockwillbespeculativeatbest.Someofthebiggertechnames,Cisco,Google,Inteland Microsofthave“softened“theirapproachbydiversifyingintorelatedfields,attemptingto 344 maintainahighearningscapacitythatismorestable.Thosefirmswhocannottake advantageoftaxbreaksmustearnenoughtocoverathresholdamountofbankruptcy costsorbeatadisadvantagetofinanciallyleveragedcompanies.Whilethismodeldoesnot accountforthe“benefitsoffundingwithretainedearnings”inlieuoftaxadvantages,one oftheprovisionswouldbethegenerationofincomeaboveandbeyondthetaxbenefitsthat arelost-includingthesubtractionofinterestpaymentsandtheadvantagesofalower defaultrate. Figure13-1 OperatingLeverage Equity Debt DefaultProbability Amount of Loss Tax Benefit Thestudent/investormustkeepinmindthatthemodelisarenditionoftheincremental valueofdebt-andnotthevalueofthefirmitself.Ifafirmchoosesnottofundwithdebt, wecancheckthroughthemodelwhetherthedecisionisvalid,butwecannotgaugethe periodtoperiodperformanceofmarginalbenefitsbecausetheywillbelessthanzero.We can,however,usethecostofbankruptcysidetomeasureperformancebybalancingthe rateofdefaultwiththeamountofloss.Ifweseetherateofdefaultgodown,weshould lookforabalancingincreaseintheamountoflosssinceitrepresentsmostlymarketvalue. SPREADSHEETCONSTANTS 345 Themarginalbenefitsmodelispartlystaticandpartlydynamic.Whileallinputs canbechanged,somevariableswillremainconstantsimplybecausethereisnorealistic methodofmakingthemreact.Forexample,thevariable,“NumberofShares Outstanding”canbechangedbutwillnotreactastheothervariableschangebecauseitis deemedtobeoutsidetheconfinesofthemodel.Similarly,the“MarketPrice”ofthestock willnotreacttochangesinthemodelbecausenoknownformulacandeterminetheprice changeofacommonstock.Analogously,operatingincomerespondstosales,whichimplies interactionwiththegreatereconomy.Sincetherearenoproductionvariablesinthe model,EBITisagiven.Thus,variablesthataredeterminedtoberealistically uncontrollablearemadeintoconstants. Secondly,theamountofcapitalisandassumedtobethecorrectamounttoraise. Therelationshipbetweenassetsandcapitalisfixed,carryingovertotangibleassetsand marketcapitalaswell.Noadditionalamountofdebtorequityshouldaffectthequality andtypesofassetstopurchase.However,iftheamountthatafirmraisesiswellbeyondits requirements,thenweareoptimizingdebtattoolargealevelofcapitalanditwillbe incorrect.Byusinglong-termaverages,wediminishyeartoyearanomalies,i.e.,making calculationsdependonanunusualamountofcapitalraisedinanyoneyear.Alternatively, theamountsthatweinputascapitalarealsoactualamountsandsoitis“realistic”to attempttooptimizeatthatlevel;theterm,“optimal”isonlyoperativeinrelationtothe constraintsofthesituation. Thirdly,itisassumedthattheinterestratewillnotchangeasnewlevelsofdebtare engaged.Thatisarealisticassumptionifafirmisalreadyneartheoptimaltarget,butnot realisticifafirmmustgofromaleveloftentothirtypercentdebttoequity.Mostfirms woulddesiretogetmoredebtatthesamerate,buttheWACCispredicatedonincreasing ratesforincreasingrisk.Additionally,theinputtedinterestrateisthe“effectiverate”as appliedtolong-termdebt.Thisratio,(InterestExpense/Long-termdebt)iseffectivefor theperiodanddeniesinconsistenciesbetweeninterestexpenseandtheactualrate. 346 Fourthly,thetaxrateisaconstant.Whiletheaverageeffectiverateisused,itdoes notchangeinresponsetochangesinfundamentalsthatverywellmightdictateanewrate. Theeffectiverateforabanneryear(whenthemodelincreasestheamountofnetincome)is thesameforadismalyearofnearnegativeincome. Finally,wemakecurrentliabilitiesandcurrentassetsconstantsbecausetheyreact tomyriadvariables,muchlikethestockprice.Therecanbenodeterministicfunctionfor eithervariablebecausetheydependonsuchoutliersasinterestraterelationships,vendor credit,typeofindustry,etc.Theinteractionsaretoovariedtoinputamodelfunctionthat wouldberealistic.Thelistofconstantsis: • 1.Capital-Capitalisinputtedasagivenamount.Theamountofcapitalactsasa baseforthetradeoffbetweendebtandequity.Italsoimpliesthattheamountofassets andtherelationshipbetweentangibleandintangibleassetsisconstant. • 2.InterestRate-Theinterestratebecomesafunctionoftherelationshipbetween interestexpenseandlong-termdebtbutremainsunchangedintermsofthemovement ofothervariables • 3.TheNumberofSharesOutstanding-Inputtedastheactualamountattheendofthe measuredperiod. • 4.TheMarketPriceoftheStock-Themid-rangepriceduringtheperiodisinputted. Themid-rangeismerelythehighpriceofthestockduringtheperiod,plusthelow price,dividedbytwo. • 5.CurrentLiabilitiesandCurrentAssets-Thesevariablesdependoncomplex interactionsoutsideofthemodel. • 6.TaxRate-Thetaxrateremainsconstantdespitevicissitudesinnetincome. • 7.OperatingIncome(EBIT)Thisvalueformsthebasefornetincomebefore deductionsofinterestandtaxes. SPREADSHEETLOGIC 347 Whenlong-termdebtisincreased,interestisdeductedfromoperatingincomeatan appropriaterate.Thechangeindebtcausesacorrespondingchangeintotalliabilitiesand netincomewhichaffectstheprobabilityofdefault.Whilemoredebtwillcreatemoretax advantages,thebenefitsmustbeweighedagainstthepotentialforincreasingbankruptcy costs.Theappropriateamountofdebtwillbetheamountthatmaximizestheratioof interestbenefitstomarginalbenefitswhenbothofthosefiguresaregreaterthanzero. Forthosefirmswithlargeintrinsicvalueswellpastthevalueoftangibleassets,the programshouldregisterzeroornegativeinterestbenefitsandindicatethatacapital structureofallequityismosteffective.Infact,oneofthesavinggracesoftheZmijewski algorithmisthatitseemstobe“forgiving”enoughtoallowinterestbenefitsinthefirst place;thedefaultrateisnotsostringentthatinterestbenefitsaredenied.Moreover,we haveaddedmanyofthefunctionsfromEVAoptimizationthatallowsustoviewthe programinanothercontext.Forexample,weseeoptimizeddebtinthecontextofmore EVAand/orbalancingROEandROC;inotherwords,wecanviewtherecommendations inthedomainofrepercussions.Iftheprogramrecommendsanallequitystructure,wecan observehowfollowingthatrecommendationaffectsEVAandROC. DYNAMICVARIABLES Onceweinputagreaterorlesseramountoflong-termdebt,achainofeventsensues thataffectstheothervariablesassociatedwithmarginalbenefits.Thevariablesthat changewhenlong-termdebtischangedarecalled“dynamic”variables,andcanbe realisticallydeterminedfromeitheraccountingfunctionsorspreadsheetlogic.Mostof thesevariablesareexplicitlysetforthinthespreadsheet.Inthecaseoftheonethatisnot (taxespaid),theusercancalculatethefigureorcreateanothercellthatdoesthatjob.The followingare“dynamic”variables: • 1)Equity-Itispossibletoincludepreferredstockinthisfigure,becausetheprogram simplysubtractslong-termdebtfromcapital.However,weincludeWACCandbeta 348 calculationstosupplementthemarginalbenefitsinformation;preferredstockwillbe anoutlieranditisbetternottoincludeit. • 2)InterestExpense-Theprogrammultipliestheinterestratebylong-termdebtand thensubtractsthisamountfromoperatingincome. • 3)TotalLiabilities-Theoptimizationsubstitutesdebtforequityandaddsto(subtracts from)totalliabilities.Thebasefigureiscurrentliabilitieswhichremainsconstant. • 4)NetIncome-Wheninterestissubtractedfromoperatingincome,taxesarededucted andtheexpressionbecomes“netincome”. • 5)TaxesPaid-Asmentionedabove,theprovisionforincometaxesisimplicitbecause theaccountingtotalisnotneededformarginalbenefits.Anextracellcanbecreated forthecalculation. • 6)TaxBenefitsonDebt-Thisexpressionformsamajorpartofthemarginalbenefits equationandistaxratemultipliedbylong-termdebt. • 7)TaxBenefitsonInterest-Thisexpressionisinterestexpensemultipliedbythetax rate,andformspartoftheinterestbenefitsmechanism. • 8)TheProbabilityofDefault-Thisisaninteractionbetweensomeoftheother dynamicvariables,whichareconfiguredbytheZmijewskialgorithm. MODELSETUP Themodelissetupasa“standalone”optimizerforthepurposesofillustration.However, theEVAsensitivitymoduleinchaptertwelvecangivecorroborationofthe recommendationsfrommarginalbenefits.Together,thetwomodelscanbeintegratedif onemakestheEVAvariablesdependentontheZmijewskivariables,allowingthestudent/ investortoobservetheeffectonthecostofcapital.Theonlyadditionalentrieswouldbe thethreevariablesthatcomprisetheCAPM,whichare:therisk-freerate,themarketrate andthebetaofthecompany.Suchinterdependenceallowstheusertotesttheoptimum againstEVA,EPS,ROEetc.Theprogramisalsosetupforcomparisonssothatthe modulecanbecopiedandconfiguredfortheinputofacomparisonfirm.Thelistof 349 formulasintheprogram,andtheentireset-upofthespreadsheetareavailableinthe appendix. THEPROCESS:ENTRYVARIABLES Thereareelevenentryvariablesthatareenteredintothegraycellsofthe spreadsheetOncetheentryvariablesarecompleted,thereisone(redhighlighted) decisionvariableforlong-termdebtthatisenteredintheoptimizationsection.Itisthis cellwhichwillyieldtheoptimizedvaluefordebtandchangetheenteredvalues.Thus, elevenvariablesareenteredinsequenceinthe“ZmijewskiVariables”section;these variableswillcarryovertothe“OPTIMIZATION”moduleandbetransformedas differentnumbersareenteredinthered“LTD”cell. Itisbesttouseahistoricalfiveyearaverageinalloftheentryvariables.However, theprogramcanstillgiveafair“guesstimate”ifoperatingincomewastypicalfora particularyear,thusallowingperiodtoperiodinput.Theentryvariablesanddecision variablearelistedasfollows: • A)Elevenvariableslocatedinverticalsequenceinthe“ZmijewskiVariable”section.1. Assets2.CurrentAssets3.CurrentLiabilities4.LTD(actual)5.EffectiveInterest Rate(InterestExpense/LTD)6.EBIT(earningsbeforeinterestandtaxes)7.Market Priceforthestock,whichisanaverageoftherange.8.TheTangibleBookvalueper share9.Theeffectivetaxrate.10.Theamountofcapital(withpreferredstock deductedforbetteraccuracy).11.Thenumberofsharesoutstanding • B)Thedecisionvariable.Highlightedinred,thereisonedecisionvariablelocatedin theblueoptimizationsection,belowthetopsection. 350 THEPROCESS:OPTIMIZINGWITHSOLVER Optimizationoccurswithaminimumofconstraints.Thekeycellsare:B28which containsthedecisionvariable“LTD”inred;L16whichcontainstheoptimizedamountof marginalinterestbenefits:J16whichcontainstheoptimizedamountofmarginaldebt benefits;andB35,“Equity”,whichmustbesettoavaluegreaterthanorequalto“1”,so thatthechangeinbetacanregisterproperly(whenusedinconjunctionwithanEVA module).Thefullratiotobemaximized,marginalinterestbenefits/marginaldebt benefits,islocatedinO16.Wesettheparametersasfollows:1.MaximizecellO16;2.By changingcellB28;3.Subjectto:L16>=0,B28>=0,B35>=1.IfSolverhasnofeasible solution,itisbecausethereisnopositivemarginalinterestbenefitthroughouttheentire range.Therefore,theoptimalamountofdebtforthatcompanyiszero.Manyhighbeta stocksfallintothatcategory. THERESULTS:THREEEXAMPLES Whileitmaybenexttoimpossibletoproveandverifytherecommendationsas “optimal”,thestudent/investorcandetectchangesinEPSorbetathatreflectefficiency.In fact,theprogramdoesnotuniformlytrytolowertheproportionofdebttoequityinorder toraisenetincome.Itproceedsbyevaluatingtheprobabilityofdefaultwiththeexisting amountofincomeandthenmakesadeterminationabouttaxbenefits.Therefore,three separatecasesarise:caseoneiswherethereisarecommendationforlessdebt,which naturallyraisesthereturnoncapitalandEPSwhileloweringbeta;casetwoiswherethere isarecommendationformoredebtwhichlowersEPSandROCbutraisesbetaandROE; casethreeistherecommendationtousenodebtinthecapitalstructurebecauseno marginalinterestbenefitsareobtained.Inthislastcase,“nofeasiblesolution”isregistered inSolver,andeithertheprobabilityofdefaultistoogreat,ortheamountoflossisgreat, orboth. 351 CASEONE:LESSDEBTRECOMMENDED *Note*ineachcase,theamountofcapitaliscomposedoflong-termdebtandcommon equity. Table13-2 LOWE'S2000 (LOW) ENTRY VARIABLE AMOUNT (Mil) Key Measurements Actual Optimized Assets 11376 Long-term Debt 2698 1774 CurrentAssets 4175 LTD/CAP 32.93% 21.65% Current Liabilities 2929 Financial Leverage 1.1136 1.071 LTD 2698 EVA 153.87 117.73 Effective InterestRate 0.0541 ROC 9.88% 10.27% EBIT 1431 ROE 14.74% 13.1% MarketPrice 25.37 EPS $1.06 $1.10 TangibleBook Val./sh. 7.18 ROC-WACC % 0.75% 0.69% EffectiveTax Rate 0.37 ROE-Equity % 2.8% 1.83% Capital 8192 BETA 0.99 0.89 Numberof Shares 764.15 a.Risk-free Rate 5.5% b.Beta 0.99 c.Market Rate 12% Inamodelofthistype,decreasingdebtwillautomaticallyincreaseEPSandthereturnon capital(ROC),whiledecreasingthereturnonequity(ROE).However,the“distance ratios”thatmeasurethedifferencebetweenROCandthecostofcapitalforexample,did notimprove.Ontheotherhand,thelimitedreturnoncapitalof10.27percent(optimized) ismuchmoresustainableatthislevelofdebt.Theprogramregisteredafairlyhigh 352 probabilityofdefaultatthethirty-twopercentdebtlevelandpareditdownbylimitingthe amountofdebttoapproximatelytwenty-onepercent.Ahigheroperatingincomewould allowahigherdebtlevelbecausetheprobabilityofdefaultwouldbelower. CASETWO:MOREDEBTRECOMMENDED Table13-3 MERCK2001 (MRK) ENTRY VARIABLE AMOUNT (Mil) Key Measurements Actual Optimized Assets 44007 Long-term Debt 4799 6823 CurrentAssets 12962 LTD/CAP 23.02% 32.72% Current Liabilities 11544 Financial Leverage 1.04 1.0656 LTD 4799 EVA 6217.24 6174.95 Effective InterestRate 0.0968 ROC 34.93% 34.26% EBIT 10721.33 ROE 45.37% 50.93% MarketPrice 76.025 EPS $3.14 $3.08 TangibleBook Val./sh. 3.77 ROC-WACC % 28.23% 27.37% EffectiveTax Rate 0.29 ROE-Equity % 38.73% 44.02% Capital 20849 BETA 0.42 0.47 Numberof Shares 2319.1 a.Risk-free Rate 4.2% b.Beta 0.42 c.Market Rate 10% Incasetwo,theprogramsawthatMerckhadphenomenalearningpowerandalow probabilityofdefault.ItraisedbetaandROEattheexpenseofthereturnoncapital (ROC)andearningspershare(EPS).Theemphasiswasobviouslyonincreasingtax 353 benefits.However,analystsknowthatdrugcompaniescanfallprecipitouslybecauseof lawsuitsandexpiredpatents.Weassumethatthistypeofvolatilityisimplicitinthe relationshipbetweenearnings,assetsandshareprice,butotherintangiblemeasurements maycapturethisriskbetter. CASETHREE:THERECOMMENDATIONFORNODEBT Table13-4 CITRIX1999 (CTXS) ENTRY VARIABLE AMOUNT (Mil) Key Measurements Actual Optimized Assets 1038 Long-term Debt 314 0 CurrentAssets 570 LTD/CAP 37.1% 0% Current Liabilities 137 Financial Leverage 1.0699 1 LTD 314 EVA -2.42 -24.8 Effective InterestRate 0.0407 ROC 13.38% 14.32% EBIT 196.6 ROE 21.27% 14.32% MarketPrice 59.125 EPS $0.58 $0.62 TangibleBook Val./sh. 2.92 ROC-WACC % -1.22% -2.9% EffectiveTax Rate 0.38 ROE-Equity % -0.4% -2.9% Capital 847 BETA 1.52 1.11 Numberof Shares 195 a.Risk-free Rate 5% b.Beta 1.52 c.Market Rate 16% CitrixSystemsisaveryrepresentativecase.Throughoutthenineties,thiscompany escheweddebtbutdecidedtotakeonthreehundredandfourteenmillioninlong-termdebt 354 in1999.Likemany“tech”stocksitsoperatingleverageisfundamentallyhigh,whichis reflectedbyahighbeta.Theadditionofdebtintoitscapitalstructureaddedtomore equityriskwhichmadeitexcessivelydifficulttoachieveapositiveEVAeveninabanner yearlike1999.Bythistime,anoverheatedmarkethadpusheduptheriskofallequity. ThisprogramdoesnotfindafeasiblesolutionwithSolverbecausebankruptcycostsare greaterthanthemarginaltaxbenefitsofinterest,makingthemnegative.The recommendationofzerodebtbringsbetadowntoafarmorereasonable1.11. EVADISCREPANCIES Thestudent/investorwillnoticethatEVAplungedineachcaseof“optimization”. WeknowfrompreviouschaptersthatEVAoptimizesatextremesofequityordebt dependingontherelationshipbetweentheirrespectivecosts.Reconcilingalong-term optimizationwithvolatilechangesinthecostofcapitalistenuous;weassumethatthe weightedaveragecostofcapitalwillreflectriskoverthelong-run.Inoptimization, however,long-termchangesarecombinedwithcurrentEVAinformationthatwillbe“out ofsync”withtheperiodicimprovementsthatarenecessary.Forexample,ifoptimization tellstheusertoretainmoreearningsandshifttoalowerdebttoequityratio(Lowe’s)then itmustbeassumedthatsuchchangescannotoccur“overnight”,andthatacertain consistencyinoperatingincomemustbearrivedat.Analogously,norecommendationfor moredebtinthecapitalstructurecanbeundertakenduringa“creditcrunch”when interestratesareespeciallyhigh. Itisfallacioustoarguethatthismodelwillsetstandardsofoptimizationforevery company.However,ifthereaderwillobservetheaverageproportionsofdebttoequityin anindustry,orthosesameproportionsduringthetimingofastockpeakduringabusiness cycle,themodeloftenyieldsarealisticrendition.Ultimately,eveninthecasesthatitdoes notoptimize,itbalanceskeyvariablesandoffersmanagementseveralancillarytargets. (BacktoTableofContents) 355 APPENDIX:SPREADSHEETFORMULASANDZMIJEWSKIOPTIMIZATION A B D E G H 1Zmijewskivar. variable negative variable amount 2assets coefficient 3currentassets intercept 9.479 LTD 0 4currentliab. TL/TA -6.384 taxbenefit 0 5LTD CA/CL -0.069 6interestrate NI/TA 1.06 7interestexpense 0 8EBIT Actual 9EBT 0 ROE LTD/CAP LEV. ROC 10netincome 0 #DIV/0! #DIV/0! #DIV/0! #DIV/0! 11totalliabilities 0 12equity 0 13mktprice 14bookvalue #DIV/0! OPTO. 15marketvalue 0 ROE LTD/CAP LEV ROC 16tangibleprice #DIV/0! #DIV/0! #DIV/0! #DIV/0! 17taxrate 18capital 19#shares J L M O 1amountof loss default cost bank 2 3DIV/0! intermediary #DIV/0! #DIV/0! 4 exp #DIV/0! 5 final #DIV/0! 6 7 8 9Marginal Benefits Interest Benefit INT/ Marg. 10#DIV/0! #DIV/0! #DIV/0! 15Marginal Benefits Interest Benefit MAX 16#DIV/0! #DIV/0! #DIV/0! 356 A B D E F G H 23OPTIMIZE 24Zmijewski var. variable negative variable amount 25assets 0 coefficient 26current assets 0 intercept 9.479 LTD 0 27currentliab. 0 TL/TA -6.384 taxbenefit 0 28LTD CA/CL -0.069 29interestrate 0 NI/TA 1.06 30interest expense 0 31EBIT 0 32EBT 0 33netincome 0 34total liabilities 0 35equity 0 36mktprice 0 37bookvalue #DIV/0! 38market value 0 39tangible price 0 40taxrate 0 41capital 0 Marginal Tax Benefit Bank Cost EQUALIZE 42#shares 0 0 #DIV/0! #DIV/0! J L M O 24amountof loss default cost bank 25 26#DIV/0! intermediary #DIV/0! #DIV/0! 27 exp #DIV/0! 28 final #DIV/0! 29 357 FORMULASFROMRANGEA1:O42 H3.=B5 J3.=(1-(B16/B13))*B15 M3. =E3+(E4*(B11/B2))+(E5*(B3/B4))+(E6*(B10/B2)) O3.=M5*J3 H4.=B17*B5 M4.=EXP(M3) M5.=1/(1+M4) B7.=B5*B6 B9.=B8-B7 B10.=B9-(B17*B9) D10.=B10/B12 E10.=B5/B18 F10.=B8/(B8-B7) G10.=B10/B18 J10.=H4-O3 L10.=(B7*B17)-O3 O10.=L10/J10 B11.=B4+B5 B12.=B18-B5 B14.=B12/B19 B15.=B13*B19 D16.=B33/B35 E16.=B28/B41 F16.=B31/(B31-B30) G16.=B33/B41 J16.=H27-O26 L16.=(B30*B40)-O26 O16.=L16/J16 B25.=B2 B26.=B3 H26.=B28 J26.=(1-(B39/B36))*B38 M26. =E26+(E27*(B34/B25))+(E28*(B26/B27))+(E29*(B33/B25)) O26.=M28*J26 B27.=B4 H27.=B40*B28 M27.=EXP(M26) M28.=1/(1+M27) B29.=B6 B30.=B28*B29 B31.=B8 358 B32.=B31-B30 B33.=B32-(B40*B32) B34.=B27+B28 B35.=B41-B28 B36.=B13 B37.=B35/B42 B38.=B13*B42 B39.=B16 B40.=B17 B41.=B18 B42.=B19 E42.=H27-H4 F42.=O26-O3 G42.=E42-F42 (BacktoTableof Contents) 359 14 ANINTRODUCTIONTORESIDUALECONOMICPROFIT THEORY:USINGA CONSTANTDIVIDENDDISCOUNTMODEL ANINTRODUCTIONTORESIDUALECONOMICPROFITTHEORY Theopportunitycostofanactionisthebenefitlostbynotchoosingthebest alternative.Thatstandardconceptfromeconomicsbecamethefoundationforcapital evaluationtechniqueslikeEVAandthecapitaldynamic.Onemerelycomparedthe incomethatafirmwasactuallygeneratingtowhatsimilarlyriskyfirmsweregenerating; theopportunitycostwasthelinkbetweenthecomparisons.However,neithertechniqueis sufficientfor“handson”determinationofdividendpolicy,newshareissuesortheamount ofearningstoretain.Outsideoftheamountofleverage,fewofthevariableswere controllablebythefirm,andindeed,the“costofequity”wasverymarketdependent. Themorepragmatic“residual”economicprofit,hasthesamefoundationand frameworkasEVA,butusesthe“expectedrateofreturn”fromavaluationmodelasits costofequity.Itallowstheusertocontrolalevelofcurrentdividend,nextdividend,new stockissues,andtheamountofretainedearningsintherealmofnetincomeincreases. Thus,givenalevelofnetincome,thefirmcanoptimizeitsdividendpolicyandattemptto improveitsresidualeconomicprofitfortheyear.Oneofitsstrategicadvantagesistofind alevelofnetincomewhereastockissuewouldnotunderminecapitalstructure.Secondly, itcandecidewhetherthegrowthratefordividendsisoptimalorneedstobechanged. Thirdly,itcancreateanoptimalamountofretentionthatwillmaximizeeconomicprofit. Yearsago,analystswouldequatethecostofequitywiththereturnonequity(ROE) andusetheminterchangeably.But-itwasdecidedthatROEwasnotatrueopportunity 360 costbecause,NetIncome-[(ROE)x(Stockholders’Equity)]=0.Tobeatrueopportunity cost,thecostofequityhadtobesomelevelbelowthereturnonequity.Whenafirm maximizedthisdistance,itwasfunctioningabovethelevelofitspeers.Valuationmodels, especiallytheGordonmodel,yieldedanexpectedrateofreturnthatcouldbecompared withROEtogaugeafirm’sperformance.Sincethis“expectedrateofreturn’wasdriven byinternalfundamentalsanddidnothaveasmanymarketvariablesastheCAPM- oriented“requiredrateofreturn”had,itwasmoreamenabletocorporatecontrol;the onlymarketdrivenvariablewasthepriceofthestock.Althoughitwouldlackthe performanceaccuracyofEVAorthecapitaldynamic,itcouldbeusedtofinetunecapital structure.Theamountofretainedearnings,dividends,debt,andevendividendgrowth wouldbederivedfromthemodel. Theassumptionsinthemodelarenumerous; • 1.TheusermustacceptthatamodelliketheGordonmodel,whichisonlytheoretically applicabletofirmswhopayaconstantlygrowingdividend,canbeusedasanear-term proxyforthecostofequity. • 2.TheusermustacceptthatthegrowthfactorintheGordonmodeliscomposedofthe productofROEandtheretentionratio. • 3.Theusermustacceptthatthestockpriceisdeterminedoutsideofthemodel,bythe market,andthatitwillnotchangeinresponsetoanychangesmadetothe fundamentals.Itisaconstantinthemodel,althoughinreality,itmaychangedaily. • 4.Nondividendpayingstockscannotbeinputted.Firm’swhoretainonehundred percentoftheirearningswouldhavebothagrowthrateandacostofequityequaltoits ROE.Aswehaveseen,theeconomicprofitinthatsituationwouldbezero. Thethreepillarsofresidualeconomicprofittheoryare;opportunitycost,dividend discountvaluation,anddividendtheory OPPORTUNITYCOST 361 Theopportunitycostthatisconveyedbytheresidualeconomicprofitmodelismuch differentthantherequiredrateofreturn.Whiletherequiredrateofreturnused companiesofsimilarrisktocomparegeneratedincome,theresidualeconomicprofitcites the“opportunity”ofeitherreinvestingnetincome,orpayingitoutintheformof dividends.Ineffect,itisasmuchareinvestmentrateasitisanopportunitycost.Ithasa smallmarketinterfacewhenitusestheexpecteddividendyieldbecausethecurrentstock priceisthedenominatorofthatcomponent.However,itlacksthefar-reachingmarket orientationthattheCAPMgeneratedratepossesses.Theexpectedrateofreturnwill approachthetruecostofequityonlywhenthemarketisinequilibrium;companiesof similarriskwouldhavesimilarreturns.Atthatpoint,therewouldbenodifference between“required”and“expected“ratesofreturn. Fundamentally,residualeconomicprofittheorydictatesthatthechoicesmadein reinvestmentordistributionarederivedfrombalancingtheneedsofthecompanywiththe exigenciesofthemarket.Forexample,iftheeconomyisnearrecession,afirmmight decidetodistributeearningsasdividendsratherthantoretainthem,simplybecausethe outlookfornewprojectsissopoor.Thus,theoptimalamountoffundingisimplicitinthe model. VALUATIONMODELS Valuationmodelsofferanopportunitytogaugemarketresponsetoearnings fundamentalsbybalancingfuturegrowthwiththecostofcapital.Theyaddatime elementthatplacesacurrent-dollarfigureonanticipatedearnings.Theresultisafigure thatisreferredtoas“fair”valueor“intrinsic”value.Whencomparedtothemarketvalue ofasecurity,thisvaluewillhelptheanalystdeterminewhetherthestockisunderorover priced. Likemostanalyticaltools,valuationmodelscanbeveryaccurate,butontheother hand,theirprojectionsmaybesooffbasethatthemethodlosescredibility.Whilesome firmshaveverypredictabledividends,theanalystmustmatchthesewithasuitablecostof 362 capitaloveralonginterim.Sincepredictingsuchratesforevenoneyearisagargantuan task,forecastinginthismannerisopentonumerouserrors,andsomeanalystssimply reverttocapitalizingoperatingincometodetermineafairvalue.Thevalidityofvaluation models,however,isfoundedontheoreticalabsoluteswithoutrecoursetoperiodic disparities;itmeasuresgrowthoverlongtimespans. Thetheorybehindvaluationissound.Earningsbelongtotheshareholders.Unless theyarereinvestedandanticipatedtoreturnahigherratethanshareholderscanreceive oninvestmentsofsimilarrisk,theyshouldbepaidoutasdividends.Thus,itistheamount andgrowthofthesedividendsthatdetermineshowmuchaninvestoriswillingtopayfora shareofstock.Iftheinvestorpaysmorethantheanticipatedflowofdividends,thestockis overvalued.Ifheorshe(hopefully)paysless,thestockisundervalued,anditcanbesold ataprofitatafuturedate.Moreover,twoadditionalfactorsneedtobeconsidered:since astockisownedinperpetuityuntilitissold,itsonlysourceofreturnduringthatinterimis thedividend.Secondly,theword“anticipate”connotesthediscountingprocessof evaluatingfutureincomeintermsofpresentdollarswhichinvolves“exponentiating”the quotientofcash-flow(dividends)tothecostofcapital-foreachyearthestockremains active.Thus,almostlikeeconomicprofit,astockisrightfullyvaluedbytheproductionof incomecomparedtothecosttoproduceit. DIVIDENDTHEORY Dividendsaresacrosanct.Notonlyarestocksevaluatedbytheirpotentialstreams ofdividends,butthegrowthrateofdividendswilloftenmirrorthegrowthrateofthe company.Forthisreason,theaccelerationindividendgrowthwillbelessthanthe accelerationinearnings;mostcompanieswanttoensurepayment,andanydecreasesends anextremelynegativesignaltoinvestors.Infact,cuttingthedividendisperhapsthesingle mostextremesymbolofmanagerialdefeatoutsideofdeclaringchapter11.Evenwhen warrantedtosaveacompany,shareholderswillseetheactionasasellsignal.Ontheother 363 hand,dividendgrowthiscitedasasymbolofmanagerialconfidence,andthebest companieshaveasolidtrackrecordofsteadilyincreasingdividends Mostcompaniespegdividendgrowthatarateveryclosetoearningsbecausethe commonstockholderneedstobecompensatedfromtheprofitsofthefirm.However, alternativetheorieshaveoftencreptintotheacademicliterature,advocatingdifferent paymentrates,fixedpercentagesoftheprofits,andevenaconcepttailoredtostockprice optimization-theresidualtheoryofdividends. Theresidualtheoryofdividendsstatesfourprinciplesbehindcreatingacoherent dividendpolicy: • 1)Determinetheoptimalcapitalbudget • 2)Determinetheamountofequityneededtofinancethisbudget. • 3)Financethisamountofequitywithretainedearningstothegreatestextentpossible. • 4)Paydividendsonlywhenretainedearningsarenotfullyexhausted Forexample,ifCompanyXneeds100millioninfunding(thecapitalbudgetis100 million)andtheirtargetcapitalstructurecallsfor65percentequityand35percentdebt, then65millionofthe100millionwillbeinequity.If80millioninearningsisavailable (netincomeis80million)tomeetthatbudget,then(65/80)or81.25%willberetained earnings,whichleaves15millionfordividenddistribution.Fromamathematical perspective,suchapolicydoesindeedminimizethecostofcapital.Aslongastheamount ofequityisoptimal,fundingwithretainedearningswillnotincurflotationcostsnorwillit potentiallydilutethemarketvaluewithnewissues;anynewstockissuerepresentsa futureobligationtopaydividendsaswell.However,ifthecapitalbudgetweretogrow suddenly(perhapsbecauseofgreateropportunities),thefirmcouldrenegeondividend growth,whichmightbedevastating-dependingoncontractualdemandsandthehistoric patternofpayments.Ineffect,afirmpaysoutdividendsevenwhenthereinvestmentrate maybehigherontheissueofnewstock.Whilethatconceptopposestherationalityofan opportunitycost,itisarealisticaffirmationofthepsychologicalmilieu;thefirmcannot 364 assumethatshareholdersdesiretoincurtheriskofpostponingimmediateconsumptionfor futurepotentialreturns. Inreality,netincomewillnotalwaysbesufficienttoprovideretainedearnings,or willinterestratesandcreditavailabilitybefavorableenoughtoincurdebt.Shareholders willcontinuetodemandhigherdividendsdespiteanuncharacteristicallybadyear.Infact, dividendpolicyisanunsolvedconundruminthebusinessworld;no“onesizefitsall” policyexistssinceeachentityisstructureddifferently.Older,establishedcompaniescater topensionfundsandretireesandrarelycuttheirdividends,whilenewer,“startups”may establishapolicythatistailoredtotheirearningsflowthatmightincludeonly“special dividends”whentheyareperiodicallydeclared. Theresidualeconomicprofitmodelallowsflexibilityinthisarea;theusercan declareanydividendforthisyearornextyear,butconsistencyisimplicitinthemodel. ThevalueofusingtheGordonmodel,evenwhenitisnotapplicabletotheactualdividend distribution,isthatitprovidesashort-termanswer;thedividenddistributionimplication mayevenberealisticgiventhevolatilityinsomemarkets.Sincetheeconomicprofitinthe currenttermisbeingcomparedwithadjacentyears,thelossinaccuracywillnotbeas profoundastheutilityindeterminingproperproportions. RESIDUALECONOMICPROFIT Theresidualeconomicprofitmodelappliestheexpectedrateofreturnto stockholders’equityandthensubtractsitfromnetincome.Inthismanner,ithasthesame frameworkastheothereconomicprofitmodels.However,itisfarmorevolatilethanthe “requiredrateofreturn”;adirectrenditionofcurrentgrowth(ROEmultipliedby retention),withoutanysmoothingfromafiveyearmovingaverage,willgivetheexpected rateofreturnacurrentbias.Withfiveyearsmoothing,amorereliablefigureemerges, albeitonethatismuchmoredependentonthefundamentalsofthecompanythanmarket averages.Theanalystcanusethisdependenceoninternalfundamentalstogaugethe 365 effectsofdividendpolicyandsharebuybacks.Indeed,theonlydifferencebetweenROE andtheexpectedratewillbehowacompanyretainsearningsanddistributesdividends. THEDIVIDENDTRAP Toensureastable,growingdividend,afirmwillattempttoaccelerateearningsata fasterpacethandividends;suchapolicyalsoraisescapitalintheformofretained earnings.Afterseveralsuccessfulyearsofthispolicy,afirmwillfindthatstockholders’ equityhasbuiltuptoanuntenablelevel:retainedearningskeepgrowinginperpetuity, whilenetincomeistabulatedonayearlyschedule.Almost“outofnowhere”,itappears thatanylevelofnetincomewillcauseenoughearningstoberetainedsuchthateconomic profitdeclines.Thecompanyhasfallenintothe“dividendtrap”. Fromamodelperspective,thereareonlytwowaysoutof“thetrap”.Eitherthe companycanpayaspecialdividendthatdepletesequitybyanadequateamount,oritcan buybacksharesofstock.Mostcompanieschoosetodothelatter,andbuybackshave becomestandardpractice.WhenMicrosoftpaidaspecialdividend,analystshadnoidea thatthedecisiontopayanydividendatall,originallyimplementedataboutthesametime thetaxlawchanged,putthefirmintoanentirelydifferentcategory.Withoutahugely activemarketfornewissues,circa1997,Microsoftwasretainingearningsandbuildingup equityattoorapidapace.Thechoicebetweenspecialdividendandbuybackissignificant. Asmostfinancialobserverswillnote,themarketvalueofthecompanyisimmediately depletedbytheamountofthedividendpayout.Ontheotherhand,abuybackwillraise marketvaluebytakingsharesoffthemarketbutcreatesaglutoftreasurystockand makesitdifficulttoissuenewshareswhenthetimeisadvantageous.Bothtechniqueswill requiremassiveamountsofcash. Mostcompaniesendup“takingitonthechin”andhaveamediocreyear,great earningsnotwithstanding.Ineffect,thetotalcostofequitybecomestooexpensive-not onlybecauseearningshaveacceleratedthepercentagecostofequity,butbecause stockholders’equityistoohigh.Somecompanies,liketheoneinthefollowingexamples, 366 willusethisopportunitytoloadupondebtandmoveawayfromtheoptimalcapital structure:alargeacquisitionthatpaysoffrapidlywillmovethecompanybacktothe target,usuallywithahigherstockpriceintow.Lesswealthyfirmswillundergoa downturnthatexpandsliabilitiesbutcontractsassets,ascenariothatwilldiminishequity andlaterallowittorespondtoimprovedearnings.Thus,the“dividendtrap”isa fundamentalparadox.Afirmisinbusinesstogenerateincome,butiftoomuchisretained andreinvested,theabsolutefactorsofgrowth,thefirmwillsuffertheconsequencesofa lowereconomicprofit. MODELOPTIMIZATION Mathematically,thismodelwilloptimizewhenthepayouttotheshareholderisone hundredpercentandnoearningsareretained.Forthatreasonitismorerealisticfor sensitivityanalysisthanoptimization.Nevertheless,itispredicatedongeneratingincome withoutregardstogrowth;ittreatsgrowthasacostandnotasourceoffutureincome. Thestringentparametersstrictlyconstraincorporateperformancetoasetofactions, whichifviolated,willdiminisheconomicprofit.Therefore,aseriesofimprovements,with thisdefinitionof“residualeconomicprofit”,willleadtostockpriceappreciation. Besidestheinadequaciesofusingadividendgrowthmodelfordeterminingthecost ofequity,thedefinitionofgrowthfactors,ROEandretention,aresubjecttodebate.While themarketpriceofthestockisaverysmallpartofthemodel,ittooisaconstant,and subjecttodebate.Notwithstandingthosediscrepancies,itishardtoimaginean improvementinthisversionofeconomicprofitandnotobserveparallelimprovementsin othermethodslikeEVAandthecapitaldynamic.Theparametersaresostrict,thata smalldeclineinresidualeconomicprofitmayevenleadtoanincreasewhencalculatedas EVA.Forthatreason,EVAisamuchbetterinvestmenttooleventhoughitlacksthe operationalcapacityofresidualeconomicprofit;EVAisintunewithwhatthemarket demands,whileresidualeconomicprofitdisplaystheinternaldynamicsofthecompany. MODELBACKGROUND 367 Exceptforallowingchangesinthenumberofsharesandamountofdividends,the standardGordonmodelisthebasisforthe“opportunitycost”.Wederivean“expected rateofreturn”byshiftingthevariablesintheequation,P=D1/(K-G).Theexpected rate,“K”becomes:K=(D1/P)+G,wherePisthecurrentpriceofthestock,D1isthe nextexpecteddividend,andGisthegrowthrate.Wefurtherdecomposethegrowthrate, “G”intoaproductofreturnonequityandretention.Mathematically,theretentionratio isdeterminedbysubtractingthecalculation,[[(numberofsharesoutstanding)x(Dividend pershare)]/Netincome]fromthenumber,“1”.Ifwemultiplythisnumberbynet income,weobtaintheamountofretainedearnings.Thedividendpershareisthecurrent dividendandnotD1,whichrepresentsthenextexpecteddividend. Thestandardmethodologyofeconomicprofitapplies:wemerelymultiplyour derivedexpectedrateofreturnbystockholder’sequity,andsubtractthisproductfromnet income.Theideaistoimprovethisfigureyearoveryearanddeterminethereasonsit mightnot.ItispossibletousethemoreexactingcalculationsusedinEVAmodels,i.e., operatingincome,interestdeductionsetc.,buttheconceptisbestexemplifiedinthe investor-friendlyformatof:thecapitaldynamic,NetIncome-[(opportunitycost) (stockholders’equity)] Theconceptofthemodelistomakechangesinequitybychangingexistingshares, newshares,dividendsandnetincome.Theseshiftswillcauseconcurrentchangesin residualeconomicprofitandallowustoobservepotentialyearoveryearimprovement. Anychangeindividendscausesashiftinretainedearningsasdoeschangesinnetincome. AsROEandretentionchange,theresidualeconomicprofitascendsordescends. 368 MODELSETUP Table14-1 ROW COLUMNA COLUMNB 1 INPUTS 2 3 NetIncome (ENTER) 4 CurrentDividend (ENTER) 5 OldSharesOutstanding (ENTER) 6 NewSharesOutstanding (ENTER) 7 OldBookValue/Sh. (ENTER) 8 MarketPrice/Sh. (ENTER) 9 GrowthRate (ENTER) 10 NewIssuePrice/Sh. (ENTER) 11 12 DERIVED 13 14 Old(Last)Equity EqualsB5*B7 15 NewIssuedEquityBook Val. EqualsB6*B10 16 TotalShares EqualsB5+B6 17 TotalCurrentDividends EqualsB4*B16 18 RetainedEarnings EqualsB3-B17 19 PayoutRatio EqualsB17/B3 20 RetentionRatio EqualsB18/B3 21 ROE EqualsB3/B24 22 BookValue/Sh. EqualsB24/B16 23 ExpectedDividend EqualsB9*B4 24 TotalStockholders' Equity EqualsB14+B15+B18 25 26 ECONOMICPROFIT 27 28 Growth EqualsB20*B21 29 ExpectedDividendYield EqualsB23/B8 30 ExpectedRateofReturn EqualsB28+B29 31 TotalCostofEquity EqualsB24*B30 32 33 ResidualEconomicProfit EqualsB3-B31 COLUMND LABEL COLUMNE D1 Capital E1(ENTER) D2 Long-termDebt E2EqualsE1-B24 D3 LTD/CAPITAL E3EqualsE2/E1 369 MODELADAPTATIONS Thismodelcanbeusedasa“standalone”forsensitivityanalysisoradaptedfor comparisonsbetweenyears.Whenadapted,threeseparatecalculationsaremadeoutside ofthemodelforentryintotheinputsection: • 1)Determinethegrowthrateofdividends.Thisisafactorsetupas(1+Decimal percentage).Thefactorcanbeanestimateforthenextyear,oritcanbecalculatedasa trend.Forexample,inthenextsection,wecalculateageometricmeanoverfiveyears: Dividendfiveyearsago=0.74.Currentdividend=1.44.Numberofyearsbetween interim=4.Inverseofthisnumber=1/4or0.25.Thecalculationisasfollows:(1.44 /0.74) 0.25 =1.1811.Theproperfiguretoinputis1.1811 • 2)Inputabookvaluepershareforlastyear’s(period’s)equity.Simplydividelast year’sstockholders’equitybythenumberofsharesoutstandingduringthatperiod. Example:Lastperiod’sequity=52731.Lastperiod’sshares=1377.8.Calculation: 52731/1377.8=$38.27/sh. • 3)Inputanissuepricefornewsharesifany.Takethisyear’sendingequityand subtractbothretainedearningsforthisyearandthetotalvalueoflastyear’sequity. Dividethatfigurebythenumberofnewsharesoutstanding,whichisthedifference betweenthisyearstotalsharesandlastyearstotalshares.Example:Lastperiod’s equity=52731,lastperiodsshares=1377.8.Thisperiod’sequity=82646,thisperiod’s shares=1646.1,thisperiod’snewretainedearnings=13179.62.Differenceinthe numberofshares=(1646.1-1377.8)=268.3Calculation:82646-52731-13179.62 =16735.38Dividebythechangeinshares:16735.38/268.3=$62.376/share Thesethreecalculationsstandasinputsintothenextsection. THECASEOF:CANYOUTOPTHIS? 370 ConocoPhillipswaspartoftherunawayoilindustryinthenewmillennium.By 2005,thestockpricehadtopped$110pershareandwasstillgrowing.Withrecordprofits andnoendinsight,ConocoPhillipsdecidedtosplitthestock.By2006,however,theyhad hitabitofawall:inthepastfouryears,earningshadgrownatabouta40%pacewhile dividendshadgrownatabout20%.Thedisparitycausedabuildupofretainedearnings sothatitwasnexttoimpossibletoimproveontheeconomicprofitfigurefor2005.No amountofincomecouldhavebeatenthediscrepancyunlessretainedearningswere severelyrestricted.Ontheotherhand,ConocoPhillipsusedthis“wall’asanopportunity toexpand.Ratherthanattemptingtobeatthephenomenal2005figure,theydecidedto raiseevenmoreequityandbuyBurlingtonResources.Atthisjuncture,thedecisioncould havebeentobuyupalotofstockorgiveaspecialdividend,butConocoPhillipsoptedto moveawayfromtheiroptimalcapitalstructureandraisebillionsindebt.Infacttheyalso issuednewsharesofstockwhichtheyautomaticallybeganbuyingbackin2007.The decisionwassound.Ifmoreincomewasnotthesolutiontoimprovingeconomicprofit,it wastimetomoveawayfromitsoptimalcapitalstructureandformanewoptimum. ConocoPhillipswasafavoredcompanyinamostfavoredindustry,andthisriskwasnot insurmountable. COMPARINGSPREADSHEETS ThefollowingfigureswereapplicabletoConocoPhillipsresidualeconomicprofitin 2005: Table14-2 2005 QUANTITY NetIncome 13529 Stockholders'Equity 52731 TotalDividends 1639 MarketPrice(Endofyear) 116.36 2006ExpectedDividend 1.44 371 Table14-3 2005 QUANTITY RetentionRatio (13529-1639)/13529=0.8789 ROE 13529/52731=0.2566 ExpectedDividendYield 1.44/116.36=0.0124 ExpectedRateofReturn (0.2566)(0.8788)+0.0124=0.2379 ResidualEconomicProfit 13529-[(0.2379)(52731)=984.29 Thus,theapproximateresidualeconomicprofitof984.29canbeinputtedasacomparison figure.Spreadsheet#1showstheinputforConocoPhillipsin2006withoutanychanges; The416.82inresidualeconomicprofitisnotnearlyaslargeasthe2005figure.However, thecompanyraisedasubstantialamountofadditionalcapitalandonlyraisedlong-term debttocapitalbyasmallamount.Raisingalotofcapitalwithoutgreatriskwillbodewell forthefuture Inthisspreadsheet,wecanchangethethreemainparameters,dividends,net incomeortheamountofshares.WemerelygotothetoolssectioninExcelandclickon GoalSeek.Wethensetcell#33to984.29bychangingoneofthethreecellsthathavethe keyvariables.Spreadsheet#2showsthesolutiontoConocoPhillipsPhillipsproblem;we changedthenumberincellB4,thecurrentdividend.Accordingtothecalculation,if ConocoPhillipshadpaidaspecialdividendof$2.97pershareforatotaldifferenceof 4896.141-2370.384or2,525.757(Million),economicprofitwouldhaverisensubstantially. Thatdecisionwouldhavetakenconsiderablecashthatwastiedupinthepurchaseof BurlingtonResources.Alternativedecisionssuchasraisingnetincomeorbuyingback shareswouldnothaveworkedasexemplifiedinSpreadsheet#3.Inthatcase,weraised theissuepricetothemarketpriceofapproximately$71.95,(sharesarebeingtakenoffthe market)andwehadGoalSeekchangecell#B6,newlyissuedshares.Wewouldexpect thatwewouldobtainanegativenumberinthatcellthatwoulddiminishstockholders’ 372 equity.Sinceweobtainedanonsensicalnumber,thedecisionwasnotvalid.Hadwetried tochangenetincome,thesameimpossibleresultwouldhavebeenobtained. Toconcludethissection,wedonotrecommendthatacompanyadheretothe residualeconomicprofittheoryunlessitisinthefirm’sbestinterestsandcoalesceswithset goals.The“bar”maybesetsohigh,thatitmaybenexttoimpossibleforafirmtobeata priorfigure.ConocoPhillipsdisplayedamasterfulendaroundbychangingtheir capitalizationentirely.Nevertheless,economicprofittheoryexhibitsanarrayofoptions, andmaysolveproblemsonceitisdevelopedintoacohesivestrategy. (BacktoTableofContents) 373 SPREADSHEET#1 A B D E 1)INPUTS Capital 105737 2) Long-term debt 23093.4972 3)NetIncome 15550 LTD/CAP 0.218405073 4)CurrentDividend 1.44 5)OldShares 1377.8 6)NewShares 268.3 7)OldBookValue/sh 38.27 8)Price/sh 71.95 9)GrowthRate 1.1811 10)NewIssuePrice/sh 62.376 11) 12)DERIVED 13) 14)Old(Last)Equity 52728.41 15)NewlyIssued Equity 16735.48 16)TotalShares 1646.1 17)TotalCurrent Dividends 2370.384 18)RetainedEarnings 13179.62 19)PayoutRatio 0.152436 20)RetentionRatio 0.847564 21)ROE 0.188158 22)BookValue/sh 50.20564 23)ExpectedDividend 1.700784 24)TotalStockholders' Equity 82643.5 25) 26)RESIDUAL ECONOMIC PROFIT 27) 28)Growth 0.159476 29)ExpectedDividend Yield 0.023638 30)ExpectedRateof Return 0.183114 31)TotalCostof Equity 15133.18 32) 33)ResidualEconomic Profit 416.8225 374 SPREADSHEET#2 INPUTS Capital 105737 NetIncome 15550 Long-term debt 25619.25382 CurrentDividend 2.974388 LTD/CAP 0.242292233 OldShares 1377.8 NewShares 268.3 OldBookValue/sh 38.27 Price/sh 71.95 GrowthRate 1.1811 NewIssuePrice/sh 62.376 DERIVED Old(Last)Equity 52728.41 NewlyIssuedEquity 16735.48 TotalShares 1646.1 TotalCurrent Dividends 4896.141 RetainedEarnings 10653.86 PayoutRatio 0.314864 RetentionRatio 0.685136 ROE 0.194089 BookValue/sh 48.67125 ExpectedDividend 3.51305 TotalStockholders' Equity 80117.75 RESIDUAL ECONOMIC PROFIT Growth 0.132978 ExpectedDividend Yield 0.048826 ExpectedRateof Return 0.181804 TotalCostofEquity 14565.71 ResidualEconomic Profit 984.29 375 SPREADSHEET#3 INPUTS Capital 105737 NetIncome 15550 Long-term debt 215870.0799 CurrentDividend 1.44 LTD/CAP 2.041575607 OldShares 1377.8 NewShares -2502.16 OldBookValue/sh 38.27 Price/sh 71.95 GrowthRate 1.1811 NewIssuePrice/sh 71.95 DERIVED Old(Last)Equity 52728.41 NewlyIssuedEquity -180031 TotalShares -1124.36 TotalCurrent Dividends -1619.08 RetainedEarnings 17169.08 PayoutRatio -0.10412 RetentionRatio 1.104121 ROE -0.14119 BookValue/sh 97.95161 ExpectedDividend 1.700784 TotalStockholders' Equity -110133 RESIDUAL ECONOMIC PROFIT Growth -0.15589 ExpectedDividend Yield 0.023638 ExpectedRateof Return -0.13226 TotalCostofEquity 14565.71 ResidualEconomic Profit 984.29 376 SECTIONIII:REALWORLDCASES 15 ANALYTICALTOOLS:PRACTICALAPPLICATION Capitalstructureisdependentonthecostofequity.Forthosewhoareeducatedin thesciences,itmightbehardtoconceiveofatheoreticalnumberthatcanshiftfromdayto dayandthenuseitasadecision-makingtool.Ofcourse,wecanusemanyconcrete measurementstoconfirmourdecision,butthecruxofanalysisrevolvesaround determiningthechangeinariskbasedopportunitycost-thecostofequity.Inessence,we arepricingtheriskofonefirm’sequityandthencomparingittoanother,usingreturnasa criterion.Ifthatconceptseemsabstract,thencompareittoahumidityindexinweather forecasting;theindexrangesbetweentwonumbersdependingonlocation,butonceit reachestheupperlimitsofitsrangeforaconsiderableamountoftime,raincanbe expected.However,thelimitsforSeattlewillbemuchdifferentfromthoseinPhoenix. Thatdifferenceinboundariesisanalogoustohowthecostofequityworksincorporations; onefirmdoesverywellwithinarangeofeleventothirteenpercent,butforanother companysuchnumberswillspelldisaster. THETOLERATIONOFIMPRECISION Sincethecostofequitychangesfromdaytoday,theanalystmustbeabletotolerate imprecision;heorshewillmostprobablynotdetermineanexactfigureforthecostof equity.Thepremiumisplacedondeterminingacostthatwillreflectthereactionofthe firmtotheearningsofothercompanieswithsimilarrisk.Thus,itmaybebettertobe imprecisewithaconcretevalue,but“deadly”accuratewhenitcomestodetermining whetherthefigureisgrowingordeclining.Forexample,anaccuratemeasurementof10 %withamisjudgmentofa1%decrease,mightbelesshelpfulthananimprecise guesstimateof13%butwithaveryprecisegrowthestimateof2%.Thecapitalstructure analystislessconcernedwithstaticvaluesthanwithmeasurementsofdynamicchange. 377 Forthisreason,periodtoperiodchangesinEVAaremorevaluablethantheabsolutesize ofEVA,whichwouldbenefitfromaprecisemeasurementofthecostofequity.Theanalyst candetectmovementtowardtheoptimalcapitalstructureaslongasthemethodologyof determiningthecostofequityreflectsanaccurategaugeofchangesinthecostofcapital. Weillustratethisconceptbyusingradicallydifferentmethodologiestoarriveatthesame decision. Inthechapteronthecostofequity,wementionedthatthedifferenceinthevarious methodsofdeterminingthatcostistheresultofthemarketandthecompanynotbeingin equilibrium-thestockmaybeoverpricedorthecompanymaybeoutperformingthe marketintermsofearnings.Forexample,ifacompanyhasareturnonequity(ROE)that ismuchgreaterthanthemarketforanumberofyears,anditscostofequityisdetermined throughadividenddiscountmodel,thefigurewillbesignificantlyhigherthanone determinedthroughthecapitalassetpricingmodel(CAPM).Infact,savvyanalystscan exploitthisdifferencebyobservingthatthe“expected”return,determinedfromthe discountmodel,isgreaterthanthe‘required”returnasdeterminedthroughtheCAPM. Theirconclusionwouldbetobuythestockbecauseitisoutperformingcompaniesof similarrisk-orinotherwords-itsrisk/returnprofileismuchbetterthancomparable firms. Forcapitalbudgeting,mostexpertssuggestusingaconsensusmethodthatwill balanceanearnings-derivedcost(dividenddiscountmodel)withamarketderivedone (CAPM).Sincethesameforcesaffecteachmethodsimilarly,usingadividenddiscount modelshouldnotchangeadecisionevenifitdeterminesacostwellaboveorbelowthe marketdetermination.Interestratechangesandtheproportionofdebttoequitywillbe theprimefactorsthatchangeeachrespectivecost.Foranon-growthstock,evenusingthe simplisticproxy,earningspershare/pricepershare,theinverseofaP/Eratio,canleadto acorrectinvestmentdecision-iftheforcesthatmoveafirmtowardtheoptimalcapital structurearestrongenough. 378 ERRINGONTHESIDEOFCONSERVATISM Whenusingdifferentmethodstodeterminethecostofequity,webuilda conservativeapproach: • 1.Wecanusethehighestdeterminationofthecostofequityoverlowercost calculations.Aslongastheperiodtoperiodchangeissimilar,wecancompareany absolutefiguresforEVAwiththehighestcostpossible.Theworstcasescenariois usingalowcostofequityforonecompanyandcomparingittoanotherusingahigher estimate. • 2.IntheCAPM,theriskpremiumisdeterminedbytherisk-freerate(tenyearbond) subtractedfromthemarketrate.Whatifthatmarketratehasbeenabnormallylow foranumberofyears?Thatwouldgiveabiastothedownside,evenifthemarketwas surging.Tocombatsuchaskew,researchershavedeterminedthatinvestorsnormally needafivepercentagepointgainofequitiesovertherisk-freerateinordertoinvest. Thus,fivepercentshouldbetheminimumriskpremium,althoughlargernumberscan beusedasappropriateandwillrealisticallyescalatethiscost. • 3.Inashort-termdividenddiscountmodel,weneedtodetermineagrowthratebefore determiningthecostofequity.Justasweusefiveyearsofdatatodeterminethe CAPM,weuseamovingaverageoffiveyearsofreturnonequityandretentionratio data.Suchanaveragewillsmooththedataandpreventaparticularlyinordinateyear fromdeterminingthecostofequity. • 4.ThepropermethodologytodeterminetheCAPMinvolvesobtainingthefive-year averagesforboththemarketandtherisk-freerates.However,wemaysometimesuse thecurrentrisk-freerateinourcalculations,inordertogiveita“FederalReserve bias”.Thisistechnicallyimproper;riskistobeassessedovertheentireperiodofthe regression.However,thetenyeartreasuryrateisnowherenearlyasvolatileasthe marketrate,andusingacurrentrateallowsustospotpotentialtrends,especiallyshifts inFederalReservepolicy. 379 BRIEFMETHOLOLOGIESFORDETERMININGTHECOSTOFEQUITY 1.THECAPM:Thepreferredmethodisthecapitalassetpricingmodelbecauseitbrings somuchmarketinformationtothetable. • a)Wedeterminetheaveragerisk-freeratebydownloadingrelevanttenyeartreasury datafromawebsitelikewww.federalreserve.gov.Expertsadviseusingtheaverage rateduringtheperiodofregression. • b)WedeterminethemarketratefortheperiodbydownloadingmonthlydatafortheS &Poverthelastfiveyears.[(lastfigure-firstfigure)-1)]willyieldarate.Wedivide thisoverallpercentagechange,bythenumberofyearsintheregression(five)toobtain anaverage. • c)Wedoaregressionbetweenafirm’sfive-yearmonthlystockpriceandtheS&P (percentageincreases,notabsolutefigures). • d)Weassembleourcostofequity:Risk-freerate(stepa)+[(betafromstepc)(Market ratefromstepb-Risk-freerate)] Oncethestudent/investorgoesthroughtheprocedureafewtimes,itwillbeautomaticand effortlessandoneneednotbeatrainedstatistician.TheregressionwillproduceaY interceptandabeta,andeachofthesewillbeexaminedforchangesbecausetheyare separatetypesofriskindicators 2.THEGORDONMODEL:WhiletheGordonmodelwasdevelopedspecificallyforuse withconstantdividendstocks,itseaseofimplementationmakesitidealfortheinvestor.It maybeconceptuallyincorrecttousethismodelwithalltypesofdividendpayingstocks, butthenwehavetoviewanyestimationofnextyear’sdividendwithsuspicion;toassume constantgrowthintheshort-termwillnotproducealargedisparityandmayevenbe actualized. • a)Weestimatenextyear’sdividendfromafive-yeargrowthtrend:(LastDividend/ DividendFivePeriodsago)^0.25-1.Thisequationwillyieldapercentagevaluefor 380 yearlygrowth.The0.25exponentistheinverse(1/4)ofthenumberofperiodsbetween years(4). • b)Wedeterminetheaveragemidrangepriceforthestockovertheappropriateperiod byaddingthehighandthelowanddividingbytwo. • c)Wetakeafiveyearhistoryofthereturnonequityandtheretentionratio,[1- (dividendspaid/netincome)],andmultiplythetwoforeachyear.Thisproductisour growthrateforeachyear.Wethentakethefiveyearaveragetodeterminethegrowth rate.Forthenextdata,wedetermineanotheryearlygrowthrateanddropthefirst datapointtomakeitafiveyearmovingaverage. • d)Weassembleourcostofequity:(NextDividendEstimate/CurrentMidrangePrice) +Growthrate.Thegrowthrateisafiveyearmovingaverage. Normally,thisfigurewillbewellbeloworabovethefigureobtainedthroughtheCAPM becausethetwowillnotbeinequilibrium.However,thesameforcesworkoneach method:ifretainedearningsareinsufficient,afirmmustgotothecreditandequity marketstoraiseadequatecapital.ThegrowthratethatisexplicitintheGordonmodel, becomesimplicitinthebehaviorofbetaintheCAPM.AlthoughtheCAPMisthe preferredmethod,dataisnotalwaysavailableandtheGordonmodelisagoodsubstitute; itmirrorsmanyoftheinteractionsfoundintherelationshipsbetweenCAPMvariables. 3.THEE/PEARNINGS/PRICEMETHOD:Inapinch,theanalystcanresorttousing theP/Einverse,earnings/priceasasubstitutemeasurement,butshouldnotexpect premiumresults.WeaddthismethodbecauseitrationalizestheuseofP/Eanalysisandit workswellifacompanyisnotagrowthstockandexhibitsstrongmovementtowardan optimalcapitalstructure.Forexample,ifearningsrisetwentypercent,butE/Prisesforty percent,wecaninterpretthecostofcapitaltobewellbehindtheaccelerationinearnings andthefirmwouldbeagoodinvestment-atleastuntilthemarketfactoredintheincrease. UsingthisratiousuallyunderestimatesthecostofequityandthusoverestimatesEVA,but aperiodicdifferencewillreflectsomebasicchangeinthecostofcapital.Sinceitisso 381 quicklycalculated,itcanbeusedtoidentifyfavorableinvestmentsituationsthatwarrant furtherexamination,butitshouldneverbeusedinisolation. WhilemanyanalystsstillusetheP/Eapproach,measuringtheperformanceofthe inversionshouldnotseemunusualandmayclearupsomeoftheidiosyncrasiesobserved overtheyears.IndividualP/Esareoftencomparedtothemarkettocheckifafirmisover soldoroverbought.Inmanymarkets,theP/Eisaproxyforthegrowthratethatisneeded tomaintainprice.Forexample,the“Peg”ratioofP/Einthenumeratortoactualearnings growthinthedenominatorwouldbea“buy”indicatorifitwaswellunderoneanda“sell” ifitwasoverone.However,growthstocksinparticularseemimmunetosuchanalysis becausecapitalflowsintothemsoerratically,pumpinguppriceswhenasectorisfavored, andallowingthemtofallprecipitouslywheninvestorsmoveon.Often,arisingP/Ewill indicateapotentialinvestmentifearningsaregrowingevenfaster,andsotheinvestor mustweightheaccelerationfactorsofthetwojustaswiththecostofequityandnet income.Infact,EVAisamirrormeasurementofaproper“Peg”applicationwhenone considersthatitisoptimizedwhilenetincomeisrisingandthecostofequity(readE/P)is declining.Again,itisnotpropertoconsidereitherE/PorP/Easastaticmeasurement; onlywhenoneexaminesthedynamicrelationshipbetweenearningsandP/Ewillsuchan indicatorbeuseful. PuttingP/Einthedomainofthecostofequity,wecanclearupsomeoftheeccentric behaviorthatP/Eshaveexhibitedovertheyears:RaisingdebtlowersaP/Ebecauseit raisesthecostofequity(E/P).HigherinflationandinterestrateswillleadtoalowerP/E andahighercostofequity-E/P.Agreatervariabilityofearningsistheproductofmore financialleverageandindicatedbyalowP/Eorahighercostofequity.Unfortunately,a blindfaithinP/Eanalysiscancreatehavocbecauseinsomemarkets,P/Escanbedistorted andbecomemeaningless.Withinaspecificrange,theymaypredictcompanybehavior,but assoonasthatboundaryisviolated,theindicatorbecomessuperfluousandinvestor emotionstartscontrollingthemarket.Thisdiscrepancycanbestbeobservedby 382 comparingtheGordonmodeltoE/Pforagrowthstock.Asthereaderwillnoticethe judgmenttousetheCAPMcostofequityasmuchaspossibleisderivedfromabasicneed foruniformity.Inthisexample,weuseXerox,adarlingofthe1960sandnowastruggling company.Atthetime,theywerethepremiergrowthstock,growingatarateofoverfifty percent.In1969,earningswere$2.08pershareandthecurrentpricewas$58.00.The expecteddividendwas0.58andsotheGordonmodellookedsomethinglikethis: (0.58/$58.00)+.53(growth)=54%.TheE/Pruleofthumbwouldcalculatethesamecost as: $2.08/58=3.59%.Thus,whengrowthisabovefifteenpercentitisalmostderigortouse theCAPM,fornotonlydoesitcreatemoreuniformity,butitisapplicabletonon-dividend payingfirmsaswell. THEHURDLERATE Thedangerofstaticinvestingistoblindlyinfusecapitalwhenafirmisataspecific, measuredlevel-whetheritissales,EPSorassetvalue.Thedangerisevenmoreemphatic tothosewho“value”investinlowP/Es.Withoutknowingthecontextofthepriceand earningsbehavior,investorsmayinvestinastocksimplybecauseithasthelowestP/Ein theindustry.Amorethoroughcheckmayrevealpricingproblems,toomuchdebtoreven acorruptbutthoroughlyentrenchedmanagement.Thereis,however,oneconcretefigure thatcanbeusedasabenchmarkwhenevaluatingthecostofequity.Justasafirm’sIRR establishesthe“hurdle”rateforcapitalbudgeting,thereturnonequity(ROE)establishes ahurdlerateforthecostofequity.TheratioNetIncome/Stockholders’Equityisthe maximumrateatwhichEVAwillbezero.Anycostofequitypercentagebelowthisrate willproduceapositiveEVA.Consequently,movementtowardamoreoptimalcapital structureismeasuredbythechange(again,notthestaticversion)inthedifferencebetween ROEandthecostofequitypercentage,(ROE-costofequity%),overatleasttwoperiods. Industrieswithcharacteristicallyhigherdifferenceswillnotbefavoredoverthosewith smallerdifferencesunlesstheirdynamicsarebetter;therateofchangeinthedifference 383 betweenROEandthecostofequitymustbeincreasing.JustaswithEPS,ifWallStreet expectsacompanytohavetenpercentagepointsdifference,itwillnotrewardafirmthat meetsthecriteria.However,afirmwhousuallyhasasevenpointdifferencewillbehighly rewardedifitshiftstoanelevenpointdifference.Capitalflowsintounexpectedgrowth. Thusthepremiumisforanalystswhocanforeseechangesinearnings,stockholders’ equity,orthecostofequity.Whilemostanalysisconcentratesonearnings,correctly forecastingastockissuecanbejustasprofitable.Infact,salesandsectorvolatilitymake earningssodifficulttoforecastthatanalystsoftenresortto“guidance”-gleaning informationfromcompanyofficialsabouttaxes,plansforlong-termdebt,andoperating efficiencies.Ontheotherhand,equitychangesareimplementedmoregraduallybecause companyofficialsexerciseatleastsomecontrol-overbothnewissuesandtherateof earningsretention.And-althoughthecostofequitycanbevolatile,ittoohassome predictabilitysinceitlagsbehindearningsandfollowsbothinterestratebehaviorandthe proportionofdebttoequity. THEEVA/CAPITALDYNAMIC Werefertoourinvestor’sversionofeconomicprofitasthecapitaldynamicbutwemay useitinterchangeablywithEVAbecauseityieldsthesameresult.Inourapproach,wedo notfiguretheweightedaveragecostofcapital(WACC)nordowefigurewhatistermed “NOPAT”ornetoperatingprofitaftertaxes,althoughitmaybeessentialtohavethese numbersonhandforcomparativepurposes.Wemerelysubtracttheproductofthe percentagecostofequityandstockholders’equityfromnetincome.Wethencomparethis derivedfigurewiththenextperiodanddeterminethepercentageincreaseordeficiency. Thefinalequationis:NetIncome-[(PercentageCostofEquity)(Stockholders’Equity)]. THEWEIGHTEDAVERAGECOSTOFCAPITAL(WACC) Itisextremelydifficultfortheaverageinvestortodetermineanaverageweighted costofcapitalbecauseaprecisefigurerequirestheenumerationofeachpercentageof capitalateachspecificrate-includingpreferredstockandshort-termdebt.However,the 384 professionalanalystshouldordinarilyhaveknowledgeofthisbreakdownfroma10Kor prospectusandcaneasilydetermineanychangesonaspreadsheet.Weknowfrom previouschaptersthattheWACCisthesummationoftheproductsofthevariouscostsof capitalcomponentsandeachpercentageofthecomponenttypeinthecapitalstructure. Forourpurposes: • RetainedearningsandcommonstockarepricedatthederivedCAPM“required”rate. • Preferredstockisaperpetuity-figuredbydividingthepreferreddividendbythe issuingprice-thatisnetofanyflotationcost.TheequationisCostofPreferred= PreferredDividend/NetIssuingPrice • Thecostofbothshorttermdebtandlongtermdebtis(interestrate)(1-T)withT representingtheeffectivetaxrate.Theanalystiscautionedtousethecurrentinterest ratesthatarepaidonlong-termandshort-termdebtrespectively.Dividinginterest expenseintodebtisnotanaccuratesubstitute. Abriefexample: Table15-1 TYPEOFCAPITAL AMOUNT PERCENTAGE SHORT-TERMDEBT 5 5 LONG-TERMDEBT 35 35 PREFERREDSTOCK 10 10 COMMONSTOCK 50 50 TOTAL 100 100 385 Table15-2 MARKETCOMPONENT RATE PREFERREDDIVIDEND $8.00 PREFERREDISSUINGPRICE $100.00 MARKETRATE 0.11 BETA 0.95 10YEARTREASURY(RISK-FREE RATE) 0.0425 ONEYEARBANKNOTE(SHORT- TERMDEBT) 0.06 15YEARAARATEDBONDS(LTD) 0.09 EFFECTIVETAXRATE 0.4 Table15-3 TYPEOFCAPITAL FORMULA COST SHORT-TERMDEBT .06(1-.4)= 0.036 LONG-TERMDEBT .09(1-.4)= 0.054 PREFERREDSTOCK $8.00/$100.00 0.08 COMMONSTOCK .0425+[(.95)(.11-.0425)]= 0.1066 Summingtheproductsoftherelativepercentagesandthecosts:[(.05)(.036)]+[(.35)(.054)] +[(.1)(.08)]+[(.5)(.1066)]=.082or8.2%. OneusestheWACCwarily.Evenwiththemostprecisecalculation,interpretation ofitsmovementisambiguous.Inamarketwithrisinginterestrates(anormalyield curve),amovementdownwardoftheWACCcoupledwithanupwardmovementinnet incomewillmostlikelyindicateamovetowardamoreoptimalcapitalstructureanda higherstockprice.However,anyupwardmovementdoesnotcertainlymeanthata companyismovingawayfromthatstructure.Minimizationofthecostofcapitalcanoccur bothdownwardandupward,dependingonthecorrectproportionofdebttoequityin coordinationwithinterestratechanges.Onereasonthatweconcentratemoreonthecost ofequityisthatitgivesusmorefocusedinformation,butliketheWACC,upward 386 movementcannotbeinterpretedasessentiallynegativebecauseitmustbeexaminedinthe contextofearningsacceleration;atthetopofthemarket,thecostofequitywillhaverisen significantly,buttherateofchangeinearningsshouldbeevengreater. COMPARINGRISK:JUSTIFICATIONFORTWOCOSTSOFEQUITY Ratherthanbreedingconfusion,theusualdiscrepancyinvaluesbetweenthe GordonModel(oranydividenddiscountmodel)andtheCAPMoffersanopportunity. Sincetheserespectivecostsofequitywillbesimilaronlyduringbriefperiodsof equilibrium,thedisparityreflectstheinternaldynamicsofacompanyeitherout performingorunderperformingthecollectivemarket.Forexample,consideracompany whoisclosetoequilibrium.IftheGordonModeldictatesatwelvepercentexpectedreturn versusanelevenpercentCAPM-derived“required”return,andthecompanyfollowsup withjusttwopercentgrowth(ROEtimesretentionrate),thesavvyinvestorrushestosell thestockbecausethemarketwillcertainlyrepriceitsrisk.Likewise,ifthedividend discountmodeliswellabovetherequiredreturninthemarket,theremaybeupward pressuretobuythestocksuchthatbothmarketanddividenddiscountmodelarein equilibrium.Thisisnottosaythatastockcannotlanguishdespitealargedisparity,butif thecorrelationintheoriginalregressionislarge(marketindexandstockprice),therewill bestrongpressuretounitethetwocosts. CHANGESINTHECAPM WhentheFederalReservecutsinterestrates,stocksimmediatelypickupvalueand certainspecialistswillprofitfromthismove.However,thereisfargreaterprofitincapital structureforesight:theabilitytocoordinateseveralvariablesthatindicatecapitalflows andthegaugingoftheriskthatweencounter.TheCAPMoffersareal-timeindicatorthat isfarmorecomprehensivethanmosttechnicalanalysisbecauseitbringsinexogenous variableslikeinterestrateandmarketrisk,andisnotjustdependentonpriceandvolume. Moreover,theCAPMgivesusfreedomfromtheevaluationoffinancialstatementswhich arenotpublisheduntiltwomonthsafterthefact,Suchalagcreatesa“lookaheadbias”; 387 anyearningsinformationwegleanfromastatementisnotuptodateenoughtoactupon. Themarketwillfactorinearningsinformationalmostinstantly-and-sometimesbefore accountantsevenfinishthetallyifrumormillsarestrong.Thefollowingadaptationsmay workfortheanalyst: • Thepropermethodologyistousesixtydatapointsfrommonthlyindexdataandthen regressafirm’sstockpricechangeagainstthemarket’schangeoverafive-yearspan. Everyquartertheanalystcanupdatetheregressionwiththreenewmonthlypoints, droppingthethreeoldestonestocreateamovingaverage.Thisupdateallowsthe investortoexaminechangesinalpha,betaandthemarket.However,makingthe marketindexandrisk-freeratesyearlyaveragesratherthanfive-yearaverages,while conventionallyimproper,willgivethecostofequitya“currentbias”.Ifwewereto averagethemovertheperiodoftheregression,moreweightwouldbegiventoperiods ofsustainedvolatility.Forexample,ifthemarketisexceptionallyhighduringsix monthsofoneyear,theCAPMwouldbeskewedinthatdirection. • Asacomparisonfigure,thestudent/investorisencouragedtodoanotherregression withdailydataoveraoneyearperiodandtousecurrentrisk-freeandmarketrates. Theoutcomewillbeamuchmorevolatilefigurethatismostlydeterminedby conditionsinthecurrentstockmarket.Whilethismethodisinsufficienttopricerisk,it willallowtheindividualtoobservepressureonthecostofequity.Anyradicalchange inmarket,interestrateorbetawillcreatemorevariationinthecostofequity,butmay alsoindicatethedirectionittakes. • Changestowatchinclude:1.Achangeinalpha.2.Achangeinbeta.3.Achangein Rsquaredand(1-Rsquared).4.Achangeintheinterestrate.5.Achangeinthe market.Toelaborateon“alpha”:althoughalphaisnotpartoftheCAPM,observing changesallowstheinvestortogaugenon-systematicrisk.Ifanalphaisquitelargeand growing,acompanyisnotonlylessdependentonthemarket,itsstockmayimprove evenwhenthemarketisdown.Suchfavorablestatusisusuallytemporarybutmay 388 happeninscenarioswherethefirmreceivesspecialtreatment(tariffs),orhasspecial pricingpower(goldminingwhenfinancialconfidenceisshaken).Ontheotherhand,if (1-Rsquared)isgrowingandalphaisnot,thestockmaybeunstable,reactingviolently inacalmmarketornotatallduringperiodsofmarketappreciation. THECOMPARATIVECAPITALDYNAMIC Thestudent/investormayfaceadilemmabetweeninvestingintwogoodprospectsin separateindustries,oneasmallmanufacturerandtheotheralargemultinational.The investorhasdeterminedthateachisabouttoincreasetheirrespectiveEVAs,butisunsure aboutthesignificanceofactualsize.Inthiscase,onecouldusethequotientformofthe capitaldynamic,whichthisauthorterms,“thecomparativecapitaldynamic”.Remember thatthecapitaldynamicisaninvestorfriendlyversionofEVA,whichisadifference operation.Inthecomparativecapitaldynamic,wedividethetotalcostofequity,thatisthe productofstockholders’equityandthepercentagecost,intonetincomeandderivearatio. TheratioNetIncome/TotalCostofEquitycanbeusedasacomparativeindicator.The highertheratio,thegreateristheeconomicprofitgiventheinternaldynamicsofcapital structure.Whenonecompanygoesfromasmallernumberlike1.5,toamuchlarger numberlike3.5,upwardpressureonthestockmaybefollowedbysubsequentdownward pressure;thefirmfindsitdifficulttofollowsuchastellarperformance.Infact,ifthe student/investorobservesthisnumberovertime,heorshewillfindthatmanyappreciating stockswillhaveasimilarcomparativecapitaldynamic,andsoeachmarketseemstoputa premiumonreachingaspecificnumber-asitrewardsthecompanywithahigherstock price. THEMARGINALBENEFITSEQUATION Manyfinancialinstitutionswillguardtheirprobabilitymodelslikestatesecrets whichhaveledsomeacademicstoexaminethemin“deconstructionistmode”,attempting todebunktheirefficacy.Continuingwiththethemeofadaptation,thegeneric,publicly availablealgorithmscansuitourneedforestimation;theymaynotperfectlyinformusifa 389 firmisabouttobecomeinsolvent,butourmethodologyisfardifferentfromthedecisionto grantcredit.Inessence,weuseadefaultprobabilitymodelastheprimevariableinthe determinationofthecostofbankruptcy.Whenthetaxbenefitsofdebtexceedthiscost, marketvalueiscreatedandthestockappreciates. Withanalgorithmthatmatchesthespecificprobabilityofdefaultforthecompany andindustry,optimizationofdebttoequityoccurswhenthefunctionismaximizedandthe firstderivativeisequaltozero.Bothstockholders’equityandlong-termdebttocapital shouldbecomponentsofthatalgorithm.However,fortheevaluationofmovementtoward theoptimal,weonlyneedtoexaminethebehaviorofthecomponentsandtoobservethat thefunctionisincreasing.Ofparticularsignificanceisthebehaviorofthedefault probability.Ifitincreasesatall,itmustbeaccompaniedbyalargedebtissueandonlya smallstockincreasebecausetherewillbecountervailingforcesinthenextyeartobalance theequation.Inmostcases,ifdebtincreases,andthedefaultprobabilitydeclines,itis becauseearningspressureispositiveandthecompanyhasaddedequitytobufferthedebt toequityratio.Inarisingmarket,thisscenarioisidealbecauseitrepresentsbothasset growthandearningsgrowth,withadeclineinthelong-termdebttocapitalratio-more returninthedomainoflessrisk.Thekeytotherelationshipistheinteractionbetween defaultprobabilityanddebt.Ifmoredebtleadstoarapidincreaseincash-flow,the defaultprobabilitydecreases,andthestockwillappreciate.However,ifadebtissue languishesandrequiresmoreinvestmentwithlittlereturn,theprobabilityofdefaultwill rise,andthestockwillunderperform.Tocreateaworkingmarginalbenefitsequation: • 1.LiketheEVA/CapitalDynamic,weseektoincreasethefigureyearoveryear.This authoranalyzesperformancebasedonthefiscalyearofthecompanysimplybecause managementstrategyisexhibitedinthebehaviorofdefaultprobabilitycomponents overthatperiod. • 2.Theleftsideoftheequationismerelytheoutstandinglong-termdebtmultipliedby theaverageeffectivetaxratefortheindustry.(Long-termdebt)x(EffectiveTaxRate) 390 • 3.Todeterminetherightsideoftheequation,theinvestormustcalculatethe“tangible bookvaluepershare”.Inordertodecipherhowmuchofmarketvalueisintrinsic valueandhowmuchismadeupofassetsthatareunclaimedbycreditors,wesubtract intangibleassetsandunamortizeddebtfromtotalassets.Wedividethisfigurebythe totalnumberofsharesoutstandinganddeterminethe“tangiblebookvaluepershare”. Thefullfunctionis:(Totalassets-Intangibleassets-Unamortizeddebt)/Numberof sharesoutstanding. • 4.Wedeterminethedifferencebetweencurrentmarketvalueandtangiblevalueto determinetheamountofpotentiallossincaseofbankruptcy.[1-(Tangiblebookvalue pershare/Marketvaluepershare)]x[(Numberofsharesoutstanding)x(Marketvalue pershare)] • 5.Weuseaprobabilitydefaultmodeltodeterminea“ballpark”figurefordefault. Sincewearenotdoingcreditanalysis,agenericalgorithmcandisplaytheincreasein riskfromyeartoyear.Altman,Shumway,Merton,OhlsonandZmijewskihaveall publishedusablealgorithms,butthisauthorprefersZmijewski’sbecauseitiseconomic andcapturesdefaultriskinafewexpressions.Inthefollowingmethodology,wewill showthestudent/investorhowtotransformtheparametersintoalogitexpressionand thenaprobability. Table15-4 COMPONENT(ZMIJEWSKI'S) VALUECOEFFICIENT Intercept -9.479 TotalLiabilities/TotalAssets 6.384 CurrentAssets/CurrentLiabilities 0.069 NetIncome/TotalAssets -1.06 Logitprobabilitiesareexpressedinlogarithms.Wecanalgebraicallyeliminatethe logarithmandreplaceitwiththeEXPfunctionwhichisacommandtomultiplyEuler’s number(2.7182818...)byanexponent.WhenwesubstituteEuler’snumber,wealsomust 391 makeallthecomponentsinZmijewski’salgorithmnegative.Thus,ifcurrentassets/ currentliabilitiesis“2”,itbecomes-2. Themechanicsofthisoperationareasfollows:Ln[P 1 /(1-P)]=X i ß P 1 =1/[1+EXP(-X i ß)] Theinterceptisalsonegativebecauseitisinferredthatithasacoefficientof“1”.Withso muchoftheexpressiondependentontherelativelylargeinterceptof-9.479,itshouldbe quiteobvioustothemathematicianthatthepredictionofbankruptcyisnotanexact science!.However,“throwingthebabyoutwiththebathwater”getsusnowhere,and usingthealgorithmtoobservechangesinriskrelativetodebtandstockpricemakesita valuabletool. • 6.Wedeterminethecostofbankruptcybymultiplyingthedefaultprobabilitybythe potentiallossor:(ProbabilityofDefault)x[1-(Tangiblebookvaluepershare/Market valuepershare)]x(NumberofsharesoutstandingxMarketvaluepershare) • 7.Thefinalexpressionisthedifferencebetweenleftandrightsideswhichweexpectto seeincrease.(Long-termdebtxEffectivetaxrate)-[(ProbabilityofDefault)x(1- (Tangiblebookvaluepershare/Marketvaluepershare))x(Numberofshares outstandingxMarketvaluepershare)] Ifthestockpricefalls(perhapsasareactiontotheeconomy)onecanreada“false increase”intothisequationsoitisimportanttounderstandtheinteractionbetweenthe componentparts.Thebasicpremiseisthatthevalueofaleveredcompanyisgreaterthan anunleveredonebecauseinterestistaxdeductible.Ifafirmisstilladdingdebtevenasit increasesearningsandequity,theprobabilityofdefaultwilldecreasewhichwillput upwardpressureonthestock.Thissamefunctioncanbeusedtofindatargetproportion ofdebttoequityaslongastheproperdefaultalgorithmisused;thefunctionmaximizes whenthefirstderivativeequalszero.Inmanyfinancialtextbooks,theauthorwillreferto thissituationas“marginaltaxbenefitsequalmarginalbankruptcycosts”. LEVERAGESTATEANALYSIS 392 Withoutquestion,theanticipationofaprofitableleveragestatewillbethemost formidableweaponinyourarsenal.Whileotheranalyticaltoolsareprimarilyconcurrent indicatorsofstockprice,aleveragestatethattransitionsfromrequiringmorecapitalto payingoffinvestorswithhigherprofitsispredictive.Thepremiumistosearchforastate thatwillnotonlygeneratemoreearnings,butdoessowithaminimumofriskthatwill sustaintheaccelerationofearningsforaslongaspossible.Minimizingthecostofcapital, however,oftenrequiresadifferentleveragestateforeachphaseofthebusinesscycle;the costofcapitalcanactuallyrisewhileafirmisminimizingit.Thisvolatiledichotomy betweenthecostofcapital,theamountofcapital,andhowtheyinteractwitheachotherin thebusinesscyclecanbeperplexing. Thelogicchainbehindleveragestatesisbasic.Astheeconomyimproves,the FederalReserveraisesinterestratesmakinglong-termdebtmoreexpensive.Therefore,a premiumiscreatedforequityfunding.Equityisbuiltbyeitherattractinginvestorstothe stockthroughhigherearnings,orretainingthosesameearningsanddecreasingthe proportionofdebttoequity.Whentheeconomicoutlookdeclines,theFederalReserve lowersrates,rewardingthewealthiestcompanieswhocanaffordthemostdebtwitha lowercostofcapital.Companiesthatproducethemostoutputduringarecovery, compensateshareholderswithnetincomethatisgeneratedfromlowerinterestratesand increaseddemandforitsproducts.Theirgreaterfinancialleverageallowsthemtofinance withlessequity,andthesharepriceescalates. Betweenthesetwoextremesareintermediatepositionsthatarebasedonphase, sectorandtransition.Operatingleverage,forexample,isextremelyimportanttofirms whodonotfundwithdebt.Ahigherrelativeoperatingleverageforthesefirms,whentheir respectivesectorsarefavored,willguaranteehigherprofitsbecausedemandisbothstable andhighduringthatperiod.However,thesamefirmmaywantaloweroperatingleverage duringasectordownturn. Thetwomostbasicindicatorsare: 393 • 1.Thefinancialleverageratio.Thisratioisformedbysubtractinginterestexpense fromearningsbeforeinterestandtaxes(EBIT)anddividingitbackintoEBIT.The fullfunctionis:EBIT/EBIT-I.Analystsmightrecognizethisexpressionasthe inverseoftheDuPontequationsimileEBT/EBITwhichisacomponentpartofthe returnonequity(ROE).Theimportantpointtorealizeisthatasearningsincrease,this ratiobeginstoshrink-evenifbyaminusculeamount.Theprobabilityofdefault decreasesasmoreoperatingearningscoverinterestexpense.Ashifttoasmallerratio oftensignalsamovementtowardanoptimalcapitalstructurewithlessinherentrisk. • 2.Thelong-termdebttocapitalratio.Thisratiowillmeasuretheproportionofdebtto equityinameaningfulway.Ifearningsarehigh,theywillberetainedoncedividends arepaidandtheLTD/CAPratiowilldecrease.Ifinterestratesarelowenough,an increaseinthisratiowillactuallyentaillowercapitalcostswithatypicalearnings increaseinsubsequentyears.Thus,likethefinancialleverageratio,theeconomic contextofachangeintheratioisparamount.Forexample,asimultaneousshift upwardinbothratios(moreinterestandmoredebt)alongwithafavorableearnings forecastwillbeaprecursorforatransitiontoproportionallylessdebtandmore earnings. Twosecondaryindicatorsalsoexist.Theyare: • 3.OperatingMomentum.%∆ ∆∆ ∆EBIT/%∆ ∆∆ ∆Sales-Whilethisratioisoftenused interchangeablywithoperatingleverage,itisonlythesamefunctionwhenafirmisin equilibrium-arareoccurrenceconsideringmoderncorporatevolatility.Nevertheless, itofferstheinvestorameasurablefunctionthatindicatesashort-termtrendin earnings.Itissignificantbecauseitaffectsfinancialrisk;whenacompanytakeson greaterdebt,anoperatingmomentumthatsuddenlydecreaseswillincreasethe company’sdefaultrisk-itsabilitytopayoffinterestexpenses.Likewise,ifmoredebtis incurred,arisingoperatingmomentumwillbufferthatsameexpense.Althoughthe absolutevalueintotalleveragemaybesimilarforeachsituation,thechangeinratiosis 394 indicativeofoperatingmargins.Nofirmwantstoincurdebtfromapositionof decreasingmargins.Firmswhousedebtstrategicallywanttoincurdebtfroma positionofstrength. • 4.Theassettocapitalratio.Whilemuchismadeoftheassettoequityratio,thisratio maybeindicativeofa“pick-up”inbusiness.Itistruethatmorecurrentliabilitieswill decreaseworkingcapitalandleadtoahigherprobabilityofdefault,butitisequally validthatmorelabor,morevendorcontractsandmoreshort-termcreditsignalsa potentialincreaseinbusiness.Short-termdebtisalegitimatesourceoffinancingthat maylowerthecostofcapital.Short-termdebtisalsoaprecursortomoresales.When suchsalesareactuated,theybecome“accountsreceivable”,andcurrentassetsagain balancecurrentliabilities.However,whenthisoccurs,itwillbetoolateforthe shareholdertoinvestbecauseearningswouldhavebeenincreasedandshareprice wouldhavemovedconcurrently. Sincetheinvestorcannotdependonaleveragestatetoprovidemomentum (investinginastatewhereearningsarehigh)heorshemustanticipatethetransitionfrom a“debt”statetoamoreprofitablestate.Gaugingriskbyobservingthebehaviorofthe financialleverageratioandlong-termdebttocapitalratiooveranumberofquartersis standard,butthereisnoformula;theproportionoflong-termdebttocapitalismoretrend worthybecausefirmsmustaddorsubtractcapitalinlargeincrementstobecosteffective. ThatisnotaninvitationtoinvestassoonastheanalystobservesalowerLTD/CAPforone quarter.Infact,ifcreditstandardsarelowerintheoveralleconomy,thecompanymay actuallybemovingtowardahighercostofcapitalinthatscenario.However,thisisa situationinwhichtheinvestorcandependonanalyst’sforecasts.Ifthecompanyhastaken onleverageandanalystspredictcomparativelyhigherearnings,thepayoffisexpectedto berapid,andtheinvestorcanlookforwardtoahighershareprice. Secondly,aninvestorcanobservethecurrentmostprofitablesectorintheeconomy. Whatwastheleveragestatebeforeitbecameprofitable?Whatisitsleveragestatenow? 395 Notallcompaniesinasectorwillfallintothesamepattern,butifthestudent/investoris tenaciousenough,heorshewillderiveasolidindicationofhowtheseleveragestate componentsshouldbecoordinated.Therationaleforacoherentstrategyisthateach economypricesriskandreturninaspecificway;thepropercombinationofleverage factorswillleadtothehighestpointonthe“efficientfrontier”-thehighestreturnperunit ofrisk.Inessence,thereisonestatethatwillencouragetheminimizationofthecostof capitalperunitofincomeaboveallothers.Therewillalsobeastatethatactuatesthe movementtowardanoptimalcapitalstructure.Theinvestor’simperativeistobeinthe gameearlyenoughtocapturetheaccelerationofearnings. Lastly,observing“inside”activityisamust.Whenaleveragestateissolid, companyexecutiveswillbegintoaccumulateshares.Althoughmassivesellingactivityis oftenfortaxpurposesandisnota“sell”indicator,thepurchaseofmoresharesisa“buy” indicator-iftheanalystexaminestheleveragestatefirst.Theseinvestmentsarefarmore lucrativeatthebeginningandmidstagesofarecovery/expansionthaninthelaterstages, atthetopofthemarket.Theinvestorshouldavoidbuyingsharesifearningacceleration hasalreadyoccurred,butmightconsiderinvestingwhenthecompanyhasbothinsiderand debtactivityoccurringsimultaneously.Investingafterearningshavealreadybeen actuatedcanoccurwhenexecutivesmake“goodfaith”investments,butunbridled optimismisnotthepurviewofanobjectiveanalyst. THELOOKAHEADBIAS Thedifferencebetweeninvestinginleveragestatesandotherfundamentalslike earningsorsalesisthattheleveragestatelackswhatistermed,a‘lookaheadbias”. Frequently,investorsmakedecisionsafterreceivingreportsonearningsoranother fundamentalandfailtorealizethatthefirmisinaconstantstateofflux;thereportrefers toinformationthatwasreceivedlongagoandmaynolongerbeactionable.Moreover,any screenorstrategythatisconstructedonthatbasismaybeobsoletebecausethemarketwill nolongerreacttosuchinformationinthesameway.Fundamentalsareconcurrent 396 indicators;theyperformconcurrentlywiththestockprice.Ontheotherhand,leverage statesarepredictiveindicatorsbecausetheyaredevelopedwellbeforeapriceisexpectedto rise.Infact,thereisatenmonth“windowofopportunity”ineachfiscalyearbecausea10 Kwillcomeoutabouttwomonthsaftertheyearisover.Socalled‘smartmoney”investors maybeaddingsharesatthistime,buttheinvestingpublicwillbeunawareuntilearnings begintoescalate.Thepredictivenatureofleveragestatesversusconcurrentindicatorslike EVAisthataleveragestatewillforeshadowthedirectionofthecostofequityinrelationto earnings.Whileitisdifficulttopredictpreciselywhenprofitswillriseorevenifasector willbestrongatallduringabusinesscycle,weknowthatbetawillreacttoachangeinthe proportionofdebttoequity.WealsoknowwhentheFederalReserveisloweringratesor raisingthemandthattheymoveinadiscernibletrend.Whenthecostofequitybecomes lowenoughcomparedtothechangeinearnings,theequitymarketbeginstorise. Therefore,aleveragestateissimplyanextensionofthesamecapitalstructurethemeof earningschangesinrelationtothecostofequity;thatstatewhichencouragesearningsto risefastestinternsofthecostofequity,willbethestatethatoffersthemostreturnforthe leastamountofrisk.Atthesametime,itwillbethestatethatmovesthefirmtowardan optimalcapitalstructurewhichmaximizesthepriceofthestock. MICROANALYSIS:QUARTERLYOBSERVATION Fewtechniquesarefraughtwithmoreriskthanquarterlyextrapolationof performance.Althoughwecandetectchangesinearningsandsalesfromperiodtoperiod, short-termpredictionsareverysusceptibletorandomvolatility,simplybecausefactors outsideofthemodelmayhaveasmuchaffectonthestockpriceastheinternaldynamicsof thecompanyitself.Considerthe2007-2008creditcrunch;firmsthathadexcellent earningsandcreditweredamagedbyspeculationinthehousingmarket.Adecreaseinthe costofequityovertwoquartersmayhavebeenacomparativeadvantage,butitcouldnot preventastockfromfalling.Moreover,thereisnofundamentalthatissostablethatit offersareliablepredictionvariable.Evenchangesinthefinancialleverageratio,which 397 canbeminute,mustbeaccumulatedoveraseriesofquarterstobeconfirming.Thus, tryingtobe“inthegame”early,byanticipatingafavorableleveragestate,canbecostly. Forexample,consideracompanywhogoesthroughthreequarterswithlowerLTD/CAP andlowerfinancialleverage.Theinvestormakesamoveandaccumulatesstock,onlyto findoutthatintheforthquarterthecompanytakesonanunsoundacquisitionandpays foritwithanequityissue. ThehallmarkofquarterlyanalysisistoanticipatealargejumpinEVAbeforeit occurs.Professionalanalystsconstantlyforecastearningsandsoanear-termoutlookis almostalwaysavailable.However,predictingchangesinthecostofcapitaland stockholdersequitymaybevexing.Whilemanagementhasincrementalcontrolover equityanddesiressmoothchanges,thereisnoguaranteethatthegrowthratewillnot suddenlybecomeeccentric.Financialprofessionalsknowthatthestockpriceisamultiple ofbookvalueandmaywantittoconformtoindustryaverages,ortheymaywantto exerciseoptionsinanticipationoffurthergrowth.Ontheotherhand,thecostofequity willmirrorthechangeininterestratesandstockprices,aswellastheproportionofdebtto equitywithinthefirm.However,itlagstheperformanceofearningsandisdifficultto measureonashort-termbasis.Thecostisnotalwaysreflectiveofimmediateequityrisk becauseitiscalculatedoverasixtymonthperiod,butobservationoftheriskpremium,the differencebetweenthemarketindexandthetenyearbond,canindicatethedirectionof change.Again,thepremiumisplacedonearningsacceleratingfasterthanthetotalcostof equity,andthenextsectionswillfocusontheattempttopredictsuchaphenomenon. NAIVEEXTRAPOLATION Innaiveextrapolation,wecomparethetotalcostofequityfivequartersagotowhat itisnow,andderiveagrowthrate.Wethencompareanalyst’sestimatesofearningsasa growthratetothegrowthrateintotalcostofequity.Weplacethechangeinearningsin thenumeratorandthechangeinthetotalcostofequityinthedenominatorandcreatea percentagechangeversionofthecomparativedynamic.Ifitisover“1”,weanticipatethe 398 EVA/capitaldynamictoriseandthestockpricetodolikewise.Infact,wecanevenmake concretepredictionsofEVAifwemultiplethegrowthratesbylastperiod’sbaseofnet incomeandtotalcostofequity.Howeversuchpredictionsdonotanticipatetransitional changesinequityandearningsthatmaybetotallyskewed.Analystscanoftenchange forecastsonce“guidance”isreceivedfromthecompany,buttheaverageinvestordoesnot havesuchrecourse.Anaiveextrapolationisverymuchlikeproclaimingthattheweather tomorrowwillbejustlikeitistoday.Fouroutoffivetimes,thatpredictionwillbecorrect buttwentypercentofthetimeitwillfailbecauseitdoesnotanticipatethetransitionsfrom sunnytostormyorviceversa. Weapplytheexponentialgrowthratetechnique(AKAgeometricmean)overaspan offivequartersalthoughfourperiodsisacceptableifdataisunavailable;inthiscase,we arelookingforaconfirmation,notadecisionindicator.Growthrateextrapolationis accurateforastabledividend,butisnotappropriateforvolatilemeasurementsunlessitis usedasacomparisontoadifferentbutrelatedratio.Forexample,extrapolatingaten percentgrowthrateforsalesandassumingthatsalesareexpectedtorisebytenpercentis amisuseandwillbeinaccurate.However,ifweextrapolatethetenpercentsalesrate,and thenextrapolateathirteenpercentincreaseforvariablecosts,therelationbetweenthetwo issignificant. Tousethegrowthratetechnique,weneedboththepercentagecostofequityfrom fivequartersagoandthebalancesheetitem,stockholders’equity.Wealsoneedthe currentdataforthoseitems.Wethenmultiplyeachpercentagecostofequitybyeach entryforstockholders’equity.Wethenmakearatiobetweenthelatestdatainthe numeratorandthehistoricaldatainthedenominator,andmultiplybyanexponentthatis theinverseofthenumberofperiodsbetweenthedataentries.Inthecaseoffiveperiods, thenumberbetweenperiodsisfourandtheexponentis“1/4”or0.25.Iftheperiodlength weretwelveyears,thenumberbetweenperiodswouldbeelevenandtheexponentwouldbe “1/11”or0.0909.Toillustratethetechnique,examinethefollowingdata: 399 Table15-5 VARIABLE CURRENT 5QUARTERSAGO %COSTOFEQUITY 10.6% 9.4% STOCKHOLDERS' EQUITY 100 84 Wedeterminetherespectiveproductsandusetheresultsasthecomponentsoftheratio. (.106)(100)=10.6and(.094)(84)=7.896.(10.6/7.896) 0.25 =1.0764.Asthisisthegrowth factor,wesubtract“1”toobtainadecimalpercentage=7.64%perquarter.Whenwe examineanalysts’earningsestimates,theywillusuallybeanEPSfigureforayear,andso weneedtodeterminethatgrowthrateanddividebyfourtoputitonaquarterlybasis. Forexample,lastyearsEPSwas$1.00pershareandtheforecastisfor$1.19.(1.19/1.00) equalsanineteenpercentyeartoyeargain.Whenwedividebyfour,weobtaina4.75 percentquarterlygain.Thefinalratioisthen4.75%/7.64%.Theconclusionisthatthe totalcostofequityisgrowingfasterthanearningsandsothiswouldconfirmanegative evaluationonthestock.However,ifotherindicatorspointtoabanneryear,thegrowth ratestudyshouldbedisregarded.Thepremiumisonjudgment.Thepercentagecostof equitynaturallyrisesuntilthenextdownturn,butthecompanycanloweritsbetainthe interimandbuffersomeofthatincrease.Ifthefirmisnowintheprocessofbuyingback stock,theprospectswouldbebetterthanourgrowthstudyconcluded. EARNINGSPRESSURE Theartandscienceofforecastingearningsisdifficult.Perhapsthebest methodologybeginswithademandforecastfortheindustrythatconsidersthedirectionof theeconomy.Nextthemarketshareoftheindividualfirmmustbeexamined.Finally,the internaldynamicsofthefirmitselfareconsidered.Thecomplexityofthisfeatcreates variabilityinforecastsbecauseitisessentiallyamathematicalversionofMurphy’sLaw: morevariablesinthemodelcanleadtomoreinaccuracyespeciallyifthosevariablesare 400 independentofeachother.However,analystsdoformconsensusopinionsanditismuch bettertousetheircollectivejudgmentthandependonnaiveextrapolation.Ifearningsare a“hitormiss”proposition,anyattempttoachieveprecisionwithoutathoroughlyvetted modelwillsurelybea“miss”.Whenweusenaiveextrapolationofearnings,itbecomesa benchmark;itissimplyanaveragegrowthrateoverasetperiod.Wecanaddthis extrapolationtoourownfundamentallyderivedestimate,andthencomparebothofthese estimatestoanalysts’expectations.Again,wealwaysuseourestimatesasaconfirming indicator.Anyforwardlookinginputsforearningsshouldbefromtheanalystsandnot fromextrapolation Thefirstprocedureistoconstructanestimatefromfundamentals.Thiswillyielda percentageincreasethatwouldoccuriftherelationshipbetweentheinputsremained unchanged.Oneadaptation-weusethebookvaluesofdebtandequityratherthanthe marketvalueasaninput.Wearelookingfora“ballpark”benchmarkandnotaforecast. Toconstructthisestimate,weneedonlytoplugalistoffundamentalsintoastraight expression: Table15-6 VARIABLE FORMULA A)PAYOUTRATIO Dividends/NetIncome B)RETURNONASSETS(ROA) NetIncome+((Interestexpense)(1-tax rate))/TotalAssets C)TOTALDEBT/STOCKHOLDERS' EQUITY Bookvalueofdebt/Bookvalueofequity D)INTERESTRATEONTOTAL DEBT InterestExpense/InterestBearingDebt E)TAXRATE Decimaleffectivetaxrate Intheexpression,itismucheasiertoworkfromrighttoleft: (1-A)x(B+(Cx(B-(Dx(1-E))))) 401 Whenweparsethisfunction,thefactorsthatimproveanearningsoutlookarequite obvious:lessinterestandmoredebt,butagreaterreturnonassetsaswell.Thenatureof thefunctionistocreatetensionbetweenitscomponents,becauseelementswithnegative correlationareaddedtoormultipliedagainsteachother.Likethecostofequityas determinedbytheGordonmodel,muchofthegrowthcomesfromimprovingtheretention ratio-(1-A). Oncewehaveafigurefromthefundamentals,wecandoafive-yearexponential growthrateonearningsanddeterminehowearningswouldgrowiftheyfollowedatrend line.Again,thisisthesamefunctionthatweusedinquarterlyextrapolation,exceptthat weuseyearsinsteadofquartersasaperiod.Wenowhavetwoofourown“earnings estimates”tocomparewithprofessionalanalysts’estimates. OnWallStreet,whenacompanyperformsaboveanalysts’expectations,theyare rewardedwithanincreasedshareprice.Theobjectiveinthisexerciseistocompare analysts’forecastswithourmathematicaldeterminations.Whenanalysts’estimatesare abovebothourfigures,weinterpretthatdifferenceasapotentialbuyindicatorbecause theremaybeupwardspressureonearnings.Whenanalysts’estimatesarebelowbothour figures,itisapotentialsellindicator,becausetheremaybedownwardpressuretounder performearningstrends.Ifanalysts’estimatesareinthemiddle,weinterpretthesituation asneutralandgleanmoreinformation.Ourestimatesarederivedfromtrendsand fundamentalsandlacktheextensiveinformationanalysts’areprivyto.However,our calculationsalsostandasabenchmarkbecausetheyareobjectiveandmechanical,and formthefoundationforcomparison. (BacktoTableofContents) 402 APPENDIX:DIVIDENDDISCOUNTMODELS Sincemanyanalysts“gobythebook”andusethevariousdividenddiscountmodels todetermineastock’s“fairvalue”,weincludeanexplanationofabasicmodel.However, themodelsrequireextrapolationongrowthrates,sometimesyearsintothefuture.The readerisreferredtoBurtonMalkiel’sARandomWalkDownWallStreettoobservesome oftheodditiesofthisprocess.Infact,mostinvestmentbanksstillusesomeformof discountmodeltoevaluatestocksbecausethemodelsdoindeedhavelegitimacywhenthe marketisfairlyvalued(aboutthemiddleofabusinesscycle).Thetheoreticalbackground ofthemodelsissound.Astockisvaluedatthepresentvalueofdividendsthatare anticipatedtobepaidinthefuture;thatvalueisbasedonthegrowthrateofthecompany andthecosttoborrowmoney-theweightedaveragecostofcapital(WACC).The“chink” inthearmorcomesfromanticipatinggrowthratesandthecostofcapitalwhichareso volatilethattheyoftendefyprediction.Moreover,theeconomymaybeinaphasewhen stocksareoverorundervalued,makingempiricalapplicationlessthanviable. Thereaderisreferredtoafundamentalfinancetexttounderstandtheconceptof presentvalue:thebasicconceptrevolvesaroundtheknowledgethatadollarreceivedin thepresentisworthmorethanadollarreceivedinthefuture,becausethepresentdollar canearninterest.Anyamountinthefutureisdiscountedbyafactorof(1/(1+borrowing rate)^numberofperiodsinthefuture).Inthiscase,the“borrowingcost”isequatedwith thecostofequity.Thus,ourgrowingdividendsarediscountedbyoneplustheborrowing ratemultipliedtothepowerofthenumberoffutureperiods.Wesumallofthepresent valuesovertheperiodofonegrowthrateandthenaddthisfiguretothepresentvalue derivedfromthenextgrowthrate. PRICE=[ΣD(1+G1) T-1 /(1+K) T ]+[(D(1+G1) N-1 /(1+K) N )x((1+G2)/(K-G2))] Thefollowingtablewillitemizethevariables.Ofparticularinterestmaybethefunctionof “T”.Tisacommandtodothecalculationinnumericalsequenceuntilthenumber specifiedby“N”isreached.Forexample,ifN=3,thentherespectivedenominatorsinthat 403 calculationwillbeinsequence:(1+K) 1 ,(1+K) 2 ,(1+K) 3 .Thereadershouldalsonoticethat (N–1)representstheyearsofgrowthandshouldnotconfusethatfigurewith“N”,whichis oneperiodafterthenumberofyearsofgrowthisoveror(YearsofGrowth+1). Table15-7 SYMBOL EXPLANATION Σ Acommandtosumthesequenceof calculations D Thepresentdividend K Thecostofequity N-1 Thenumberofyearsthatgrowthwill occurataspecificrate(G1) T StartingatT=1,acommandtodothe calculationinintegersequenceuntil"N" isreached G1 Thegrowthratethatoccursduring"N- 1"years N NumberofYearsofGrowth+“1” G2 Thegrowthratethatoccursafter"N- 1"years Dividenddiscountmodelscanbequiteinvolvedandcomplex,enumeratingthreeorfour differentgrowthratesatatime.Approachthesewithcaution. (BacktoTableofContents) 404 16 KIMBERLYCLARK-“TOOMUCHOFAGOODTHING” EconomicProfitandMarginalBenefitsAnalysis Therecessionof91-92wasfollowedbyoneofthelongestbullmarketsinhistory. Thespeculativeexcessesthathelpedfueltheeconomyencouragedtheissueofmassive amountsofequity.Althoughprofitswererising,optimismwasrisingfaster,andthelarge issuesdidnothingtodiluteanalreadyoverheatedmarket.Oncethebottomfelloutin 2001,shareholders(especiallythosewhohadinvestedheavilyinNASDAQstocks)wereleft holdingthebag.Techstocksthathadbeentradingforoveronehundreddollarsashare couldoftenbeboughtforlessthantendollars.Infact,asofthiswriting(2008),NASDAQ tradesataboutone-halfitsvaluein1999. UNDERPINNING1:POSITIONINTHEBUSINESSCYCLE Tothosewhosawthisdebaclecoming,goodinvestingsenseledthemintodefensive sectorstocks,mostlyconsumerstaples,healthcareandhouseholdproducts,whosedemand wouldoutpacealaggingeconomy.Kimberly-Clark(KMB)wasalarge,well-knownpaper productscompanywithmorethanthirteenbilliondollarsinrevenue.Itsbetaof0.44made itaperfectcandidateforportfoliorebalancinginplaceofhigherbetatechandtelecom equipmentstocks.Asalowbetahouseholdproductscompany,itwaspositionedwellinthe businesscycle-whichisthefirstoffourbasicunderpinningsthatdeterminesour investmentanalysis.Itsprofitswerestableenoughtotakeadvantageoffinancialleverage, butitsequityriskwaslowenoughsothatdebtwasnotexcessive;Kimberly-Clarkcould takeonmoreleverage,especiallyinalowinterestrateenvironment. Thefirm’spositiongaveitafavorablecombinationofcostanddemandfactorsthat wouldyieldastrategicadvantage;positioninthebusinesscycleissoimportantbecauseof itsintegrativeeffectonallotherfactors.Whenafirmhastherightcombinationof 405 operatingandfinancialleveragetheinterfacebetweensalesandcapitalisstrengthened. Hence,inadownturn,thosefirmswithsmallbutsteadygrowtharelessrisky. UNDERPINNINGS2,3,AND4:OPPORTUNITIESFORANALYSIS Mostfirmsareintegratedwiththegreatereconomythroughoperatingandfinancial risk.Duringeachphaseofthecycle,somefirmwillpossessacombinationofextraordinary factorsthathelpsitdominateothers,andwhentheeconomychanges,thosefactorsrecede. However,outsideofsectorrotation,thereislittletheanalystcandotopredictafirm’s reactiontoachangingmarketWithoutknowledgeoftheforcesthatmoveanindividual company,ourdecisionswillbelimitedtopriceandvolumemovements.Although macroeconomicconditionshavethegreatestimpactoncorporateactivities,weneedatleast threeothertenetstogovernouranalysis-amethodologythatcorroboratesafirm’s possessionof“favorablefactors”ifyouwill.Theotherthreeunderpinningsare:shiftsto anoptimalcapitalstructure;potentialincreasesincomparativeeconomicprofit;and increasesinsalesandearningspotential. Researchingcorporatehistoryforaveragesincapitalproportionscanbe painstaking.Thereisnomandatethatacompanywillperformatthataverageorthatthe averageisevenrelevantgiventhecontingenciesinthecurrenteconomiccycle;interestrate changescanshifttheoptimalproportion.However,certainpatternsofleveragechanges arecorrelatedwithperformance,andanychangeintheprobabilityofdefaultisusually accompaniedbyaconsequentchangeinstockprice.Sincedefaultisafunctionof probability,itisquitedifficulttoconfiguretheproperlevelofdebttoequity deterministically;notwodefaultfunctionsarealike.Forthatreason,werelyonseveral measurements,realizingthattheemphasisshouldbeondetectingmovementtowardan optimalproportionratherthanoncalculatingadecisiveratio. Oneofthecorroboratingtechniquesistoanalyzeeconomicprofitforthepotential toincrease-whichmaynotbeso“esoteric”asitseems.Whenweinvestonthebasisof 406 earnings,webaseourjudgmentononevariable-income-whichistheprojectedoutcome ofcoordinatingcosts,salesandtypeofindustry.Witheconomicprofitanalysis,wehavean interactionofatleastthreevariableswhichalsorepresentsthecollectiveanalysisof numerouscomponents-beta,retainedearnings,newissuesandtaxes.Ourjudgmental riskbecomesdiversified.Wecanbewrongaboutnetincome,forexample,butprecisely accurateaboutariseinthecostofequityandstillmaintainsomepredictivecapability. Statistically,weknowthatalargemoveawayfromthemean,willrevertintheopposite directionbecauseeachindustrysetsapatternfordividends,retentionandtheamountof equityafirmcansafelyissue.Balancesheetsneedto“balance”,andweusethat conformancetoexaminechangesineconomicprofitthatshowshiftsinafirm’slevelof risk. Becauseoftheirinherentvolatility,changesinsalesandearningsarethemostrisky predictionsthatanalystswillmake;mostprofessionalanalystswillreceive“guidance” fromboththeirindustryandtargetcorporation.Incapitalstructureanalysis,wetrynot tomakeearningsforecastsourselves,butwillrelyonconsensusopiniontoprovideinputs forsensitivityanalysisineconomicprofitmodels.Wearemuchmoreconcernedwiththe changingpatternofoperatingriskthatoccurswhensalesandearningschange.Andwe attempttointegratesalesandearningschangesintomarginalbenefitsanalysisbytreating themasriskfactors.Inanydefaultmodel,earningsbecomesthelinchpinindetermining howmuchdebtispermissible;werecognizethatmaximizingearningsmayincreaseriskto intolerablelevels,creatingareboundeffectthatunderminesgrowthinfutureperiods. Thus,weattempttoobservesustainablelevels,amountsthatcanoptimizecapital structureandyetcreateshareholderwealth. THELEVERAGESTATE Thestudent/investormayestimateKimberly-Clark’soptimaltargetstructurefrom thechapteronmarginalbenefitsmodeling.Fromthatmodel,we“guesstimate”Kimberly- Clark’stargettobeapproximately26%long-termdebttocapital,butrecognizethatthe 407 constraintsareartificial,especiallyintherealmofinterestratesandcapital.Nevertheless, weattempttoestablishthatKimberly-Clark’slowriskstemsfromitsproximitytothe optimalcapitalstructureandthatonlyahigh-riskmoveawayfromthattargetandthen backagain,wouldgreatlyappreciatethestock.Sincethisinvestmentisanon-speculative portfoliorebalancingduringadownturn,theparametersfitourneedsperfectly. Inanormalmarket,wemightlookforafirmwhoisabouttoenteraproportionate equitybuildingcycle,acompanythatisgoingtoloweritsdebtratiowhileprofitsare increasing.In2000,however,manyfirmswerecomingoffoflongstretcheswherefunding wasdoneprimarilythroughretainedearningsandequityissuesthatfloodedthemarket withnewstock.IftheFederalReservebeginstolowerrates,thosefirmswhocanmost affordnewdebtwillhaveabigadvantage;thecostofcapitalwillbelessexpensive.Thus, Kimberly-Clark,whoseprofitshavebeenexceptionalandwhohoversnearanoptimal capitalproportion,wouldbeaperfectchoiceforaninvestment.However,theimportance ofobjectiveanalysiscannotbeunderestimated;weneedabreakdownoftheircurrent leveragestate,andlater,wemustmakeananalysisofbothEVAandmarginalbenefits. 1.OperatingRisk Toestablishameasureofoperatingrisk,weusethefirm’soperatingmomentum: (%∆ ∆∆ ∆OperatingIncome/%∆ ∆∆ ∆Sales)whichgivesanapproximationofoperatingleverage. Wealsoestablishthefiveyeargeometricaveragegrowthwhichis5.2percentforoperating income,and6.33percentforsales.Thisratioof5.2/6.33equals0.82whichis comparativelylow.Wethencomparethisfiguretothemostrecentmomentumfiguresfor 1999and2000. 408 Table16-1 Kimberly-Clark Year 1998 1999 2000 OperatingIncome 2320 2815 3203 Sales 12298 13007 13982 Operating momentum 21.34/5.77=3.7 13.78/7.5=1.84 Change1999/2000 -50.34% GeometricAverage 5.2/6.33=0.82 Kimberly-Clark’soperatingmomentumhasbeenbuttressedbyahighoperatingincome whichisintheprocessofrevertingtoitsmean.Theexpectationofaneconomicdownturn wouldfurtherdimprospectsforahigherincome,andsowewouldexpectalower operatingmomentumandasmallerincreaseinEBIT.Thetaxeffectsoftakingondebt, wouldbejustifiedbythelowertaxespaidonincomeandthepotentiallylowerinterestrate thatwoulddecreasecapitalcosts.However,itisalwaysa“crapshoot”anytimedebtis raisedproportionately;incomemustbeincreasedenoughtolowertheprobabilityof defaultorthefirmcanmeander-withmoredebtmakinguptheshortfall. Table16-2 2.FinancialLeverageRatio KMBFinancialLeverage YEAR 1999 2000 InterestExpense 213 222 EBIT 2654 2844 FinancialLeverageRatio 1.087 1.085 Change1999/2000 -0.18% Althoughthechangeinthefinancialleverageratiowasslight,itstillreflectsa smallerprobabilityofdefaultwhenallotherfactorsareheldconstant.Despitethehigh- 409 interestrateenvironmentof2000,Kimberly-Clarkmovedtoreducethisratio;thehigher EBITallowedthefirmtobothincreaseitstotalamountoflong-termdebt(andtax advantages)andreduceitsfinancialleverageratiosimultaneously.Itis,however, necessarytoobservethefinancialleverageratiointhecontextofproportionalchangesin long-termdebt.Whenthetwomoveinoppositedirections,moreanalysisisrequired becausethereisequivocationinboththemovementofthecostofcapitalandinearnings; suchanomaliesrepresentashiftinbalanceandaredifficulttointerpret.Forexample, whenafirmtakesonzerocouponbonds,financialleverageratiosmaydecreaseatthesame timethatdebtbecomesahigherproportionofcapital;interestexpensemaybekept artificiallylowuntila“dayofreckoning”.Inthatcase,theleveragestatedoeslittleto reflecttheinherentrisk,andtheanalystwouldproperlydependmoreoneconomicprofit analysistomakeadetermination. 3.ProportionofDebttoEquity Table16-3 KMBDebt/ Equity YEAR %LTD/CAP Long-term Debt Stockholders' Equity Capital 1999 27.44 1926.6 5093.1 7019.7 2000 25.75 2000.6 5767.3 7767.9 1999/2000 Change -6.15% AnineyearaverageshowedthatKimberly-Clark’saveragelong-termdebtto capitalwas23.35%.Theapproximately26%optimumthatwedeterminedfrom marginalbenefitsanalysisisa“ballpark”figurebasedonaveragesfromthepastfive years.Thus,themovementto25.75%appearstobeinthe“right”direction,butthe marginoftoleranceissosmallthatweneedtodependonothermeasurementstoconfirm 410 it.Fromabusinesscycleperspective,takingonlessdebtwheninterestratesarehigh encouragesoptimalitybecausemoreincomeisfreedforshareholdersintheformof dividendsorstockbuybacks,andlessispaidoutinmoreexpensiveinterest.Thehigher amountofretainedearningsbecomesthesourceforfunding,butmustbeaccompaniedby alowcostofequitytobeeffective.WithoutknowledgeofEVAandthecostofequity,any analysiswouldbeincomplete.Nevertheless,wecanbegintoclassifytheinvestmentitself- verylowriskwithahighprobabilityofexceedingthereturnonatenyeartreasurynote. 4.TheAssets/CapitalRatio Theimportanceofthisindicatorcannotbedismissed.Raisingshort-termdebtcan beatemporarysubstituteforlong-termfundinginperiodsofuncertainty;firmswho wouldotherwiseraiselong-termdebtdelaypurchasesandnewprojectsuntiltrendsbegin tounfold.Theassettocapitalratiohelpsinatleastfourotherways: • a)Itsignalsapotentiallygreaterreturnoncapital • b)Itmaysignalanincreaseinfree-cashflow • c)Itmayhelplimittheamountofexternalfinancingbycovering“shortfalls’inthe capitalbudget. • d)Itmaysignalgreatervendoractivityandpotentialsalesincreases. SinceKimberly-Clarkraisedthisratiofromapositionofstrength(closetoanoptimal capitalstructure),itcanonlybeviewedasa“plus”andnottheinsolvencymeasurement thatmightoccurinleanertimes.Kimberly-Clarkraiseditsassetsby12.98percentand capitalby10.66percent;theincreaseinassetstocapitalwasminimalandprobablywasnot greatenoughtocoveranycapitalshortfall.Withsuchproximitytoitsoptimalcapital target,therewassimplynoneedtousethisratioasanadjunct. 411 Table16-4 Assets/Capital YEAR Assets Capital Assets/Capital 1999 12816 7019.7 1.83 2000 14480 7767.9 1.86 CHANGESINECONOMICPROFIT Despiteamovetolowerriskwithgreaterequity,therequirementstoimproveEVA arestringent;afirmneedsthecorrectamountofequityinthedomainoftwoother variables-thecostofequity,andnetincome.Observethefollowingcomparisonfor percentageincreasestobettergaugethesituation: Table16-5 Percentage Changes YEAR NetIncome Percentage Change Stockholders' Equity Percentage Change 1999 1668 5093.1 2000 1801 7.97% 5767.3 13.23% Inthelastsection,wedescribedcapitalasa10.66percentchange,andnowwecancontrast itwiththechangeinequitywhichwas13.23percent.Already,wecanseethatpotentialfor anEVAincreaseisgreatlydiminished,becausewewouldmostlikelyneedadecreaseinthe costofequitytoaccomplishit.However,thecostofequitywillalwaysberaisedtowardthe endofabusinesscycle,notonlybecausethemarketis“overheated”,butbecausethe FederalReserveraisesratestocombatinflation;evenanearrisk-lessstockwillundergoa percentagecostofequityincreaseThus,wehaveaset-upforthe“perfectstorm”:the companydoeseverythingright,butthesituationisuntenablebecauseretainedearnings havebuiltuptoahighlevelrightatapointwheretheyaremostexpensive(comparatively). THEEXTREMECONSENSUSMETHOD 412 Toobtainacostofequityfigure,weutilizewhatthisauthorterms“theextreme consensusmethod”.Thechainoflogicforthiscombinationisasfollows:boththe“ruleof thumb”,E/PmethodandtheGordonmodelcanbederivedfromthesameequationofP= D1/(K–G).Whenusedalone,thecostofequityfortheGordonmodelwouldreduceto (D1/P)+G.Ifwemaketheassumptionsthat“G”,orgrowth,istheproductofreturnon equity(ROE)andretention,andweequatethecostofequitywithROE,thenROEwould reducetoE/Pwhenboththebookpriceandmarketpriceareequal.However,inthe Gordonmodel,pricemakesasmalldifferenceinthecostofequity,whileintheE/P methoditisadeterminingfactor.Inreality,thetwomethodsproduceextreme comparativecostswhicharehypotheticallylinkedbythedifferencebetweenmarketand bookvalues.Ifweaveragethetwopercentages,wecangetafairlycomprehensiveideaof whatthecostofequityactuallyis,andmoreimportantly,agaugeofrisk.Therelationship betweenthecomponentpartsofeachmethodisstableandwilldisplaychangesinequity riskwhenthemeasureascends.TheGordonmodelismorefundamentallydrivenbecause ofitsdependenceonthebookvalues,netincomeandstockholders’equity.Ontheother hand,E/Pisquitelikeitsinverse,“P/E”:itisvolatileanddifficulttointerpretandyet dependsonthemarket.Together,thetwomethodsformauniqueconsensus:bothmarket drivenanddependentoninternaldynamics,butoriginallyformedfromthesamefunction (withassumptions). 1.GordonModelDeterminations TomakedeterminationswiththeGordonmodel,weneedthreemeasurements:1. AnaverageofgrowthoverthepastfiveyearstobedeterminedbytheproductofROEand retention.2.Prospectivedividendgrowthfor2000(alreadyknown)and2001.3.An averagepricetocomparewiththatdividendfortheyears,1999and2000. 413 Table16-6 1999Growth Year RetentionRatio ROE Growth 1995 1 0.001 0.001 1996 0.63 0.345 0.2174 1997 0.39 0.205 0.0799 1998 0.54 0.244 0.132 1999 0.67 0.371 0.249 Average 0.1349 Table16-7 2000Growth Year RetentionRatio ROE Growth 1996 0.63 0.345 0.2174 1997 0.39 0.205 0.0799 1998 0.54 0.244 0.132 1999 0.67 0.371 0.248 2000 0.68 0.322 0.2257 Average 0.1806 Next,anaveragerangepriceisdeterminedforthepurposeofderivinganE/Pfigureand forinputintotheexpecteddividendyield,“D1/P”.Theaveragepriceintermsofrange maybebetterthanthemeanasanindicatorofthepotentialdistributionofastockbecause volatilityisexpressedthroughouttherange.Fewpeoplebuylowandsellhighand investorstendtobuyspasmodically,increasingvolumeasastockclimbshigher.However, earningsareaccumulatedthroughoutaperiod,anditmaybemorepropertomatchthe endofperiodpricewiththeendofperiodearningsfigure.Astrictfundamentalistwould certainlychoosethelattermethod,whileananalystwhowasfixatedontheprocessof changewouldpicktheaveragerangemethod. 414 Table16-8 AverageRange Price YEAR LOW HIGH AVERAGE 1999 44.81 69.56 57.185 2000 42 73.25 57.625 Sinceourhypotheticalanalysisbeginsin2000,wehavealreadyobtaineda prospectivedividendfigurefor1999-theactualdividendof$1.08,paidin2000.Toobtain anexpecteddividendfor2001thatwillbeappliedtotheyear2000(thenextexpected dividend),weneedtoextrapolateagrowthrateandmultiplyitbytheactualcurrent dividend.Weneedthreepiecesofdatatoobtainthisrate:thecurrentdividend,the dividendfiveyearsagoandthenumberofyearsbetweenfiveyears.Inrespectiveorder, theyare:$1.08,0.92,and4.Wethenmakearatiooutofthedividendsandusetheinverse offour(1/4)asanexponent:(1.08/0.92) 1/4 =1.0408whichbecomesthegrowthrate multiplier.Growthextrapolationtendstoworkwellwithdividendsbecausefirmspride themselvesonsteadiness.Theexpecteddividendfor2001is$1.08(1.0408)=$1.12.We nowhavetwonextexpecteddividends:$1.08whichisthenextexpecteddividendfor1999, and$1.12whichisthenextexpecteddividendfor2000.TheD1/Pfigurefor1999isthus, 1.08/57.185=1.89%andfor2000,itis1.12/57.625=1.94%.Thesefiguresarethen addedtotherespectiveROEfigurestocompletetheGordonmodel: 415 Table16-9 Gordon Model Year NextYield(D1/ P) Growth(ROEx Ret.) EQUATION Expected Rate 1999 0.0189 0.1359 0.0189+ 0.1359 15.49% 2000 0.0194 0.1806 0.0194+ 0.1806 19.96% 2.“RuleofThumb”,EarningstoPriceorE/PDeterminations OncetheGordonmodelcalculationsarecompleted,therestofthemethodis remedial.Wemerelymatchearningspersharefortheyearwiththemid-rangeprice. Table16-10 E/PAnalysis Year EPS Mid-rangePrice E/P 1999 3.09 57.185 3.09/57.185=5.4 % 2000 3.34 57.625 3.34/57.625=5.8 % Inanormalanalysis,theCAPMwouldbeusedtoderiveacostofequity.However,with non-growthstocksthatarerelativelystable(mostDOWcomponents)thestudent/investor shouldobservethatextremeconsensusisaviableoption.Theobjectivemethodologyis thatwhichbestcapturesboththechangesinthemarket,andtheinternaldynamicsofthe firmsimultaneously.ThecombinationofthemarketderivedE/Pandthebookvalue- drivenGordonmodelcancreateaworkingcostofequitybecausetheyarefunctionally related. 416 3.CreateaCostofEquity Wenowaveragethetwofiguresforeachyearandcompare: Table16-11 CostofEquity Year E/P GordonModel EQUATION CostofEquity 1999 0.054 0.1549 (0.054+0.1549)/2 10.45% 2000 0.058 0.1996 (0.058+0.1996)/2 12.88% Higherinterestrates,greatervolatility,andanover-heatedmarketpushedthecostof equityupin2000.Comparatively,Kimberly-Clark’slowbeta,lowprobabilityofdefault andadherencetoatargetcapitalstructurewereasavinggrace.However,asweshall observeinthenextsection,eventhebestcompaniesmayhavelimitedoptionswhen constrainedbythesequentialeconomicdeclineofcompaniesaroundthem.Atthesetimes, itmaybebettertomove“sideways”thaneitherupordown. ECONOMICPROFIT Atthisjuncture,thecalculationofthecapitaldynamicisremedial;wemerelyplug inthenumbers:NetIncome-[(%CostofEquity)(Stockholders’Equity)]. Table16-12 Economic Profit Kimberly- Clark Year NetIncome CostofEquity Stockholders' Equity Capital Dynamic 1999 1668 0.1045 5093.1 1135.77 2000 1801 0.1288 5767.3 1058.17 Change 1999/2000 -6.8% 417 Inthechapteroneconomicprofitoptimization,westipulatedthatearningsarethe liberatingforceforequity.Hindsightcangiveusanideaofwhatthemaximumequity couldhavebeengiventhesameincreaseinnetincome.OLDEVA=NewNetIncome- [(New%CostofEquity)(X)].Wesolvefor“X.”1135.77=1801-[(0.1288)(X)].Then“X” =5164.82 Thus,Kimberly-Clarkwouldhaveincreasedeconomicprofitwithamaximumequity increaseofonly71.72versusthe674.2thatactuallyoccurred:(5164.82-5093.1=71.72), (5767.3-5093.1=674.2) Wasthe71.72feasible?Absolutelynot.Tofundexistingsales,Kimberly-Clark mustraiseaminimumamountofnewcapital.Iftheadditional748.2thatwasraisedis goingtowardprojectswithapositivenetpresentvalue,therewasnowaythatthefirm shouldrationcapitaltomakea“paperprofit”.Fromthenetincomeside,wecan determinetheshortfallbyusingthesamesensitivityanalysisandmakingnetincomethe “X”variable:1135.77=X-[(0.1288)(5767.3)],X=1878.59.Thus,a78milliondollar increaseinnetincomewouldhaveprovidedtheimpetusforagreaterincreaseineconomic profit. Giventhehighercostofretainedearningsandthepossibilityofaneconomic downturnloomingonthehorizon,itisprobablethatKimberly-Clarkmayhaveraised muchmorecapitalthanrequired.Asuperficialexaminationindicatesthatnearlyallofits expandedsalescouldhavebeenfundedwithlessretainedearnings.However,raisingthe payoutratiowouldhavecommittedthecompanytofuturedividendpaymentsthatmay havebeenunwieldy,andwouldhaveimpededfinancialflexibility.Declaringaspecial dividendwouldhavediminishedretention,butthebestdefensewastheroutethat Kimberly-Clarkactuallytook:theybeganaseriesofsharebuybacksthatnotonly diminishedthebookvalueofequity,buttooksharesoffthemarketaswell TOOMUCHOFAGOODTHING 418 Andhereinliestheproblemofboomandbustcycles.Withrecordsalesandprofits, Kimberly-Clarkwas“victimized”bytheeconomicrealityofanover-heatedmarket.The costofequityskyrocketedbecauseinvestorswereattractedtohigherprofits,buyingup sharesandincreasingthedemandforequity.The“internal”financingthatKimberly- Clarkwasimplementingthroughretainingearningswassubjecttothesamemarket turmoilasothersourcesoffunding;itwasonlyascosteffectiveascomparisonswould allow.SinceKimberly-Clarkwasgrowingearningsata7.97percentpaceandnotthe 12.64percentthatwasneeded,themarketdidnotrewardthefirmwithastockprice increase.However,thefirmhadsuchasolidfoundation,thatitsstockdidnotdecrease either-andrepresentedalowriskcomponentinanyportfolio. Thecapitaldynamicprovedtobeavalidmeasurementofcorporaterisk.Whileit didlittletoindicatemovementtowardanoptimalcapitalstructure(thefirmwasalmost there)itreflectedthehighpriceofequityandtheinterfacewiththegreatermarket.Italso reflectedtheneedtocompensateshareholderswithsomecompendiumofbenefitsrather thanretainearningsatahighprice.SinceEVAwassohighin1999,therewaslittlethat Kimberly-Clarkcoulddoin2000toincreaseit.Thereturnonequitywasover37percent in1999versusthestillveryhigh33percentin2000.Whilesomecompaniesdocommit financialmanagementtorisk-less,stable,increasesinEVA,manyfactors(interestrates, typeofindustry,market)aresouncontrollablethatanyimprovementinthemeasurement becomesthegoal,ratherthansomespecificamount.Kimberly-Clarkmanagedthe situationasitarose,avoidingsomeofthespeculativeexcessesofitspeers.Companiescan anddoget“painted”intocorners.Historicallocationsofcapital,timelinessinthebusiness cycle,andhonoringcontractualdemandscanallstymiethebestintentionsofmanagement. MARGINALBENEFITSANALYSIS Nodefaultmodelisperfect.Whilemostmodelscovermajorvariableslikeassets andincome,eachisdevelopedinadifferentperiodwhichdeterminestheoveralleffectof eachvariable.Thus,amodeldevelopedinthe1970’smightemphasizeassetvaluesand 419 inflation,whileamodelcreatedin2008wouldemphasizesubprimeexposure.Thebetter genericmodels(Ohlson,Shumway,Altman,MertonKMV)willagreewithinafew percentagepointsastotheprobabilityofdefaultbutwillnotevincethesameamountof accuracythroughouttheentirerange.Theyareaccurateenoughtodetectlargemovesin theprobabilityofdefaultwhichistheirprimarypurpose.Aswiththecostofequity,the usermusthave:1)sometoleranceforimprecisionasthederivedfigureisanestimateand notadecisiontool.2)corroboratingmethodsandanalysisthatconfirmafindingand3) theabilitytoobservechangeinthemeasurement. Asintheeconomicprofitanalysis,theyears1999and2000werecomparedby enteringthenecessaryfundamentalsintoafunction.Wemeasuredthecostofbankruptcy andtheinherenttaxadvantagesinthedecisiontousedebtbyformingamarginalbenefits equation.Wethenobservedanyimprovementwhenthesubsequentperiodmeasurement waslargerthantheinitial. Theslightchangesinmarginalbenefits,economicprofit,andthemarketpriceof thestockconfirmthecorrelationvalueofthemeasurements.Ineffect,thedifference betweenthemeasurementswasslightenoughtobeinconclusivebutmatchedthe performanceofthecompany-tentativebutstable.Thestateofthecompanyseemedtobe ina“holdingpattern”thatwouldneitherconfirmnordenyitasaninvestmentvehicle. However,atthejunctureof2000-2001,thesoundestjudgmentwastoseekastockthat wouldminimizerisk,andKimberly-Clarkfitthatbill BASICMETHODOLOGY Fordetailedinformation,seethechapteronmarginalbenefitsanalysis.The student/investorsetsupafunctionthatequatestheproductoflong-termdebtandthetax rate,withtheproductofaprojectedamountoflossandtheprobabilityofdefault.The functionmaximizeswhentheincrementalchangeonbothsidesisequaltozero.Thisisthe pointwherethefirstderivativeofthefunctionisequaltozeroandistheprime determinantofanoptimalcapitalstructure.However,wecangaugeyeartoyear 420 improvementsimplybyobservingwhethernextperiod’smarginalbenefitsarelargerthan theinitialperiod’s.Moreover,wecanalsotestthemarginalbenefitsofinterestinthesame functionandusetheresultasthenumeratorinanimmediatebenefitsratio.Whenwe dividethisnumberbytheoriginalmarginalbenefitsfigure,wecanlookforimprovement; anestimatedoptimumoccurswhentheratioismaximized. Inessence,wehavetwofunctionstolookat: 1.[(Long-termdebt)(TaxRate)]-[(ProbabilityofDefault%)(AmountofLoss)]=X1 2.[(InterestExpense)(TaxRate)]-[(ProbabilityofDefault%)(AmountofLoss)]=X2 WelookforimprovementsinX1,butmoresignificantly,welookforanimprovementinX2 /X1.Incertaincases,thetaxadvantagesofinterestexpensewillbelessthanzerowhich willsignifythatthecompanyshouldnothavedebt. TAXBENEFITSFORKIMBERLY-CLARK ThetaxbenefitcalculationsforKimberly-Clarkarethemostremedial,requiring multiplicationofthetaxrateandtheamountoflong-termdebt.Thecalculationis replicatedwithinterestexpense. Table16-13 Tax Benefits Year TaxRate Interest Expense Long-term debt (Tax)(Interest) (Tax)(Debt) 1999 0.3 213 1927 63.9 578.1 2000 0.29 222 2000.6 64.38 580.174 AMOUNTOFLOSSFORKIMBERLY-CLARK Creatingagenericamountoflossforinputintoabankruptcycostisapurely experimentalendeavor.Sincetheword“loss”issubjective,thereisroomfor interpretation,buteachdefaultisalegalconstructwithuniquerequirements.Our 421 constructionofa“loss”isbasedonthelosstoshareholdersaboveatangibleassetvalue whichisassumedtobethepropertyofcreditors. Theconstructionofthisfunctionis(1-(TangibleBookValue/MarketValue)) (NumberofSharesOutstandingxMarketPricepershare).Tangiblebookvalueis constructedbysubtractingallintangibleassets,goodwillandunamortizeddebtfrom assets,andthendividingbythenumberofsharesoutstanding.Themarketvaluepershare isdeterminedbythesamemethodasinprevioussections;itisanaveragebetweenthelow stockpriceandhighstockpricefortheyear. Table16-14 Amountof Loss Year Tangible Book/sh. Market Price/sh. Numberof Shares Amountof Loss 1999 7.12 57.185 539.8 27025.09 2000 7.04 57.625 539.22 27276.443 THEPROBABILITYOFDEFAULT Mostdefaultprobabilitiesareverygoodindicatorsofrisk,butfewwillbeboth accurateandflexibleenoughforcapitalstructuremodeling.Theuniquerequirementsof capitalstructurearevariablesthatemphasizetheconsequencesofincreasingdebtby allowingsolutionforlong-termdebtorequity.Ineffect,weneedanalgorithmwhichwill curveupwards,displayingincreasingriskatanincreasingrate,andyetoptimizeinthe domainofearnings. TheZmijewskimodelissimpletoapply,andalthoughtherearemoreaccurate defaultalgorithms,fewpossesstheinherentflexibilityofthismodel.Inanexperimental mode,itworkswell.Thebasiclogicbehindthefunctionisthattheproductofparameters andfundamentalratiosformsthelogarithmofaprobabilityofdefault.Wealgebraically eliminatethelogarithmandsolvefortheprobability.Thus,Ln[P1/(1-P1)]=X1Bwhere 422 P1istheprobabilityofdefault,X1arethefundamentalratios,andBarethecoefficientsof thealgorithm.Toobtainaprobability,weturntheequationaroundandinputP1=1/1+ EXP[-XB],wherewegivenegativevaluestothefundamentalratios.Thefollowingtable containsadefinitionofthefundamentalratiosandthecoefficientsofthealgorithm. Table16-15 ZmijewskiDefault NAME FUNCTION COEFFICIENT TL/TA TotalLiabilities/Total Assets 6.384 CA/CL CurrentAssets/Current Liabilities 0.069 NI/TA NetIncome/TotalAssets -1.06 Intercept NONE -9.479 Table16-16 Kimberly- Clark Year TL/TA CA/CL NI/TA Default Probability 1999 0.45045 0.92616 0.13015 0.00126= 0.126% 2000 0.45405 0.8286 0.12438 0.00129= 0.129% THECOSTOFBANKRUPTCY Tocalculateourexperimentalcostofbankruptcy,wemultiplytheamountoflossby theprobabilityofdefault.Theresultwouldbeafigurethattheshareholderswouldlosein theeventofliquidationofassets. 423 Table16-17 Kimberly-Clark Year Probabilityof Default AmountofLoss Costof Bankruptcy 1999 0.00126 27025.09 34.052 2000 0.00129 27276.443 35.187 MARGINALBENEFITS Theprimaryobjectiveincomparativeanalysisistotestwhetherthetaxadvantages havegrownrelativetobankruptcycosts;eachbankruptcycostissubtractedfromeachtax advantageandtheperiodicfiguresarecompared.WiththeZmijewskialgorithm,wecan alsoobservewhethertheimmediatetaxeffectsofinterestaregrowingincomparisontothe overalltaxadvantagesofdebt.Somealgorithmswillnotpermitthiscalculation,butthe Zmijewskifundamentalsoptimizeinthatdomain-wherethisratioismaximized. Table16-18 Kimberly- Clark Year Tax Benefitsof Debt Bankruptcy Costs Marginal Benefits Interest Benefits Interest/ Debt Benefits 1999 578.1 34.052 544.048 29.848 0.05486 2000 580.174 35.187 544.987 29.193 0.05357 Theinterestbenefitswerederivedbysubtractingbankruptcycostsfromthetax advantagesofinterest(63.9and64.38for1999and2000respectively),whilearatiowas formedbydividingthisfigureintothemarginalbenefitsofdebt. Aswouldtypifyacompanythatishoveringaroundanoptimaltarget,thereislittle changeindefault,taxbenefitsortheamountofloss.Allindicationsfrombotheconomic 424 profitandmarginalbenefitsanalysiswouldpointtothenecessityofmovingawayfroman optimaltargettoseekmorerisk-ifindeedKimberly-Clarkwantedtoappreciateitsstock. However,stayingatthetargetwouldprovideminimalriskwiththepossibilityofspecial dividendsandbuybacksaswellasthenaturalappreciationoftheregulardividend.Such strategicdecisionscannotbetakenlightlybecausefutureprospectsneedtobeweighed againstthepotentialtemporarydiminishmentofshareholdervalue. CONFIRMATION Ascorroborationofourfindings,perhapsnodefaultmeasurementhasstoodup betterthanAltman’sZscore.Itslinearitymakesituntenableforuseincapitalstructure models,butasasimplemeasureofchangingrisk,itisunsurpassed.A“Z”Scoreisthe summedproductsofratiosandcoefficientswherealargerscoreisbetterbecauseit indicatesalowerprobabilityofdefault.AnyincreaseinAltman’sZScorecanbematched withcorrespondingdefaultprobabilitiesfromothermethodstoseeiftheyconfirmone another. ALTMAN’SZSCORE:BOOKVALUEVERSION ThestandardZscoreissetformarketvaluesandiswidelyavailable.However, marketvaluesgetinflated,andsothebookvalueversionismoreconducivetoobserving theeffectofcapitalstructurevariables.Wemerelypluginfundamentalratiosandthen addupthescore.Thebasicfunctionis:0.71(X1)+0.847(X2)+3.10(X3)+0.420(X4)+ 0.998(X5).Thefollowingtabledescribesthecomponentratios: 425 Table16-19 ALTMAN'SZSCORECOMPONENTS(BOOKVALUEVERSION) X1=WorkingCapital/Assets***Workingcapitaliscurrentassetsminuscurrent liabilities. X2=RetainedEarnings/Assets X3=EBIT/Assets X4=BookValueofEquity/BookValueofLiabilities X5=Sales/Assets Evenifeachelementinthenumeratorstaysstrongandstable,increasingdebtwill increaseeachdenominatorleadingtoalowerZscore.Intheyearofadebtissue,tax benefitswillbalancetheincreaseindefaultprobability,butgreaterriskoccursinthenext period;ifearningsdonotimproveenoughtoincreasetheirrespectivenumerators,alarge amountofassetswillremainunderperformingWithoutanewinfusionofdebt,therewill benotaxbenefitstocounteractthechangeindefaultprobabilityandtheresultwillbea stockthatfalters.Althoughafirmcanissuemoredebtinsubsequentperiods,thefirm mustbeginpayingbackandloweringtheproportionofdebttoequity.Atthispoint,the risk/returnratiofortheinvestorrisesbecausethefirmcaneitherwallowinlow profitability-ifthewrongprojectshavebeenimplemented-orbeginpayingoffrapidly andincreaseitsstockprice.Thus,theZscoreisbothmathematicallysoundandintegrated withcapitalstructuretheory. 426 KIMBERLY-CLARK’SZSCORE Table16-20 1999FUNDAMENTALS AMOUNT WorkingCapital -284 Retainedearnings 6764.6 EBIT 2654 BookValueofEquity 5093.1 BookValueofLiabilities 5772.6 Sales 13007 Assets 12816 Table16-21 2000FUNDAMETALS AMOUNT WorkingCapital -784 RetainedEarnings 7982 EBIT 2844 BookValueofEquity 5767.3 BookValueofLiabilities 6574.6 Sales 13982 Assets 14480 427 Table16-22 Kimberly- ClarkZScore YEAR(X1) (X2) (X3) (X4) (X5) ZScore 1999 -0.022 0.528 0.207 0.882 1.015 2.4567 2000 -0.054 0.551 0.196 0.877 0.996 2.364 Fortheyear1999,thenumber2.4567canbecomparedwith2000’snumber,which declinedto2.364,indicatingahigherprobabilityofdefault.Thestudent/investorwill noticethatAltman’sfigureagreeswithZmijewski’s,evidencingslightlymorerisk. INVESTMENTCONCLUSION Kimberly-Clarksfinancialacumenisadmirable.Theyknowhowtogaugerisk. However,inorderforthestocktoappreciate,theyneedtomakeamoveawayfromthe optimaltargetandthenbackagain.Ifitseemsunusualnottotaketheriskneededto appreciatethestock,considerthetimingofsuchamove-atthebeginningofadownturn. Thefirmiscorrecttoplayitconservativelywhichmakesitanattractiveadditiontoany portfolio.The“planoftheday”inearly2001wastodiminishrisk,andfewcompanies wouldaccomplishthatbetter. (BacktoTableofContents) 428 APPENDIX:EXTRAPOLATEDRISK-When“Normal”istoorisky Investingonthebasisofforecastsistenuous.However,iftheinvestorcanadapta contingencyplantoeachforecast,andrecognizeitsinherentfragility,heorshewillhave moreoptionsandgreatersuccess.Businessesthatadoptacontingencyplanareflexible enoughto“shiftgears”whenconditionschangeandthereinliesthesecretofcorporate longevity;contingencyplanningwillhelpdiversifyafirm’sactionsandlowersitsoverall risk Whenweusethegeometricmeantoextrapolatethethreecomponentsofthecapital dynamic,wearenotforecastingperse.Wearefollowingthetrendlineformedovera numberofperiodsanddeterminingtherisktoeconomicprofit.Whenwecomparethat risktoothercompanies,wearemerelystating,“Thisistherisktoeconomicprofitifthings keepgoingastheyare.”Thus,itispurelyamathematicalexercisedevoidofnumerous outlierssuchassectorrotation,theeconomy,orthecostofcapital.However,somefirms createanenvironmentwhereeconomicprofitcannotgrow,anditisinthesecasesthat extrapolatedriskisvaluable;toomuchequityissuedattoohighapricewillshowupasa trendthatcancripplethecapitaldynamic.Atthispoint,thefirmmustoverhaulitscapital structurethroughacquisitions,stockbuybacksorevendivestiture Justasweextrapolatednextyear’sdividendintheGordonmodel,wecanfinda trendlinefornetincome,capital,andstockholders’equity.Sincethecostofequityisless affectedbycompanytrends,weleaveitasis.However,thecorrelationbetweencapital,net incomeandequityisveryhigh;theinteractionbetweenretainedearnings,additional equity,andcapitalformationisquitestrong.Welookateachelementoverafiveyear interim.Ifwespotabadearningsyearaseitherthefirstorlastyearintheperiod,wedrop thatfigureanduseamoretypicalyear;again,theemphasisisnotonprecisionbutto produceagrowthfigurethatencompassesthetrendinthatelement. Theconstructionofthegeometricmeanissimple.Wemakearatiooutofthelast yearinthetrend(current)anddivideitbythefirstyear(fiveyearsago).Wethenusethe 429 inverseofthenumberbetweenperiodsasanexponent.Thisfigureisagrowthfactor.If wesubtract“1”weobtainanaverageyearlypercentagegrowth.Inthecaseofafiveyear period,thenumberbetweenperiodsisfourandtheinversewouldbe1/4or0.25.Ifthe periodweretenyears,thenumberbetweenwouldbeninewithaninverseof1/9or0.1111. Thestudent/investorneedsacalculatortousethefractionalexponent.Thegrowth functionis:(CurrentNumber/NumberXperiodsago) 1/(X-1) . Atthisjuncture,wepluginacompany’sfiguresintothegeometricmeanfunctionandthen doacapitaldynamiccalculationontheextrapolatedfigures,leavingthecostofequityasis. Toobtainariskcalculationforcomparison,wedividetheextrapolatedcostofequityinto extrapolatednetincome;thisfigureisthecomparativecapitaldynamicforthatcompany. Table16-23 Kimberly- Clark Year NetIncome Stockholders' Equity Capital CostofEquity 1996 1404 4483 6222 N/A 2000 1801 5767 7767.6 0.1288 Table16-24 FUNDAMENTAL FUNCTION GROWTHFACTOR NetIncome (1801/1404)^0.25 1.0642 Stockholders'Equity (5767/4483)^0.25 1.0649 Capital (7767.6/6222)^0.25 1.057 Theextrapolationofthecapitaldynamic:1801(1.0642)-[(0.1288)(5767)(1.0649)]= 1916.62-790.99=1125.63.Theriskmeasurementis1916.62/790.99=2.423 430 Thecurrentcapitaldynamicis1801-[(0.1288)(5767)]=1058.21,withariskfactor of1801/742.79=2.424;avirtualtiethatisdescribedbytherateofgrowthofnetincome andequity.Ifthecostofequitygoesdownin2001,theeconomicprofitmayrise. Asecondimplicationoftheanalysisistheriseinlong-termdebt.Wemultiplythe growthfactorsbythefundamentalsforcapitalandequityandthensubtracttoobtainan extrapolatedfigurefordebt.Ifcapitalisgrowingfasterthanequity,thefirmistakingon debt,andanextrapolatedfigureprovidesinformationonwhata“normal”increasewould be. Table16-25 EXTRAPOLATEDLONG-TERMDEBT ExtrapolatedEquity=5767(1.0649)=6141.28 ExtrapolatedCapital=7767.6(1.057)=8210.35 Capital-Equity=Long-termdebt=2069.07 Withoutreferencetomanyotherfigures,debtisnotanabsolutemeasureofrisk,butwe canatleastobservewhata“normal”increasewouldbe.Andthatisthecruxof extrapolation-toobservewhatisnormal,anddecidewhether“normal”maybetoorisky. Thetechniqueismoreapplicabletolowrisk,lowbetafirmswhohavesteadyincomes(like Kimberly-Clark).However,mostfirmsdesireastableequitybecausethemarketpriceof thestockwillbesomemultipleofthebookvalueandstabilityhelpsmaintainthe difference. (BacktoTableofContents) 431 17 FULLSTEAMAHEAD:ANANALYSISOFCONOCOPHILLIPS, 2002-2006 Whenafirmhasearningsthatareacceleratingmuchfasterthanthemarketrate, themeasurementsareforthright;performanceiscapturedinbothfinancialstatementsand stockcharts.However,signsofincipientandsustainablegrowtharemoreelusive.The risksofgrowthafterPhillipsPetroleummergedwithConocoin2002wereapparent. Investmentsuccesswasnotguaranteedinaneconomicenvironmentplaguedbyboth recessionandthepreparationforwarintheMiddleEast.Whilethepricefluctuationsof anynaturalresourcecancripplecommodityrelatedbusinesses,volatilityintheoilmarket causescomparativelymorerisk.Notonlydoshortageshavefarreachingmacroeconomic consequences,butanoilglutcreateslesscompetitionwithintheindustryitself-and ultimately,lessprofit.Consequently,hedgingoilrelatedcommoditypricesisanimportant partofthebusiness.Asoilpricesrise,industryplayerswantto“lockin”alowerpriceand mayprofitfromthesedecisions.Anythreattosupply,whetheritisitfromOPEC,orthe warinIraq,willpropelfuturespricesupward.However,nothingpreparedinvestorsfor thesteepriseinoilpricesthatwereacombinationofbothpoliticalriskandglobaldemand afterthewarstarted.ThatConocoPhillipswouldestablishitselfasapremierplayerinan industrythatwoulddominateanentirebusinesscyclewassimplyunforeseeable. THECONTEXT Rarelydowehaveachancetoassociateaspecificcompanywiththeperformanceof anentirebusinesscycle.Byobservingcapitalstructuredecisionsfromtroughtopeak,the student/investorcancorrelateindividualcorporatebehaviorwiththeforcesthatworkon thegreatereconomy.Mostcyclesaredistinguishedbytheinterspersingofseveralsectors, eachofwhichdominatesoneortwophases.Inevitably,thefallfromgraceoccurswhen capitalcostsbegintoeclipseearnings,andthenextsectortohaveearningsaccelerationwill 432 bethe“risingstar”.InthecaseofthemergerbetweenPhillipsPetroleumandConocoin 2002,themeteoricriseinearningswaspunctuatedbyfiveyearsofsolidriskmanagement thatexploitedagrowingeconomy.Severalriskrelated“hurdles”weredeftlyhandledby ConocoPhillips: • 1.ConocoPhillipswasformedintheaftermathofthe911attacksandincreased tensionsintheMiddleEast.Forlackofabetterword,thestockmarket“crashed”in 2001and2002,leavingarecessiontrailingbehind.Equitybasedcompanieswere especiallyhardhit. • 2.TheFederalReservebegantolowerthefederalfundsratetolevelsnotseeninfifty years.Ontheotherhand,PhillipsPetroleumdidnothaveexcessivedebtonitsbalance sheet,norhadtheyissuedalotofequityassomecompaniesdidinthelate90s. • 3.OilpricesbegantorisewellbeyondthepointwarrantedbytheriskintheMiddle East.Infact,IndiaandChinaweretobecomemajoreconomicplayersontheworld stageandemergingmarketsbegantogrowatafuriouspace.Thedemandforoilwas nolongeranisolateddomesticconcern. • 4.Oilbecameapoliticalissue.Tosaythattheoilcompanieshad“friends”in Washingtonmaybeanunfaircriticismoranunderstatementdependingontowhom onespeaks.However,publicoutcryabout“BigOil”becamelargeenoughtowarrant post-KatrinahearingsinCongress.Infact,theformerassistantSecretaryofStatesat ontheBoardofDirectorsatConocoPhillips. Oilcompaniesmanageriskaswellasthebestfinancialinstitutionssimplybecausethe natureofthebusinessissospeculative.Exploringforoilisperhapsthearchetypalrisky businessandcallstomindsuchimagesasTexaswildcattersand“instant”wealth.For ConocoPhillips,theriskswerefarmorecomplicatedthanmerespeculation:theyhadto foreseetheeconomicrecoverythatwasforeshadowedbyasurgingstockmarketinMarch of2003.Secondly,theyneededseveralplanstonegotiatethroughtheriskofMiddleEast turmoil:aplan,forexampleifoilweretobeshutoff-theworstcasescenario.Any 433 extremesinpricewouldentailmoreriskbecausechangesindemandwouldhavean exaggeratedeffectonvariablecostsontheupsideoralternatively,onunusedcapacityifthe recoverywereshort-lived. Theeconomicrecoverythatbeganin2003wasclassicinmanyways;theFederal Reserveloweredinterestrates,andthefirmsthatcouldmostafforddebttookadvantageof lowercapitalcosts.Inthe“normal”sequence,lowbetastocksbegantorecoverfollowed byriskierconsumerdiscretionarystocks.However,thisrecoverywasmoreunusualthan othersinseveralrespects: • 1.Employerswereslowertohire,andtherewaslittlewage/pricepressureatthe anticipatorylevel.Workersdidnotcausehigherinflationaryexpectationsand employersdidnot“bidup”thepriceofwages.Thecruxofthisanomalyliesinthe alternativesprofferedbyaglobaleconomy.Notonlywereillegalimmigrants competingforlowwagejobs,butinformationservicejobsmovedoutofthecountry;a computerglitchcanbefixedfromIndiaaswellasfromSanFrancisco.Moreover, manufacturingjobsmovedtoChinawhereinexpensivelaborcouldmakeawidget,and thensenditbacktotheUnitedStates. • 2.TheFederalReserveloweredinterestratestohistoriclowsandthenkeptthem there.Thiswasarealestatespeculator’sdream.Interestsensitivestockstookoffwith housingleadingtheway.Anybodywithamortgagewasrefinancingatalowerrate. Whenhousingpricessoared,thebubblebegantoburstbecausecorporateearnings couldnotprovidetheequilibriumthatwouldmatchhigherhomepriceswithworker income.Achainreactionoccurredthatbecame“thecreditcrunch”of2007-2008with acombinationofhighratesofforeclosures,defaultsonloans,andhousingderivatives (collateralizeddebtobligationsorCDOs)thatwereworthless. • 3.Thegovernmentwentonaspendingspree.InordertopayforthewarinIraq,the UnitedStatesgovernmentfloatedlargebondissuestoChinaandranuphugedeficits simultaneously.Inthemeantime,thedollarsanktocomparativelows.When 434 multinational“profits”weretranslatedfromanascendingcurrencytoonethatwas depreciating,earningswereinflatedabovetheirtrueproductivevalue.Additionally, theU.S.exportmarketpickedup,notonthequalitativebasisofbetterproductsand services,butbecausethedollarwasatanalltimelow,andAmericangoodswerea relativebargain. Thestagewassetforarecoveryandexpansionthatwerehighlyskewed.Allsectorsinthe economydidnotparticipatenordidtheentireworkingpopulation.Someareas,like Michigan,remaineddepressedthroughoutthe“recovery”.Ontheotherhand,areasof heavyconstructionlikeLasVegasprofitedenormouslyfromthehousingboom.Infact,the economyfavoredtwosectorsthroughoutthebusinesscycle-oilandhousing-whileothers, likenewspaperpublishing.,begantofallbythewayside.Whilesuchcyclicaldominance doesnotbodewellfortherespectivefuturesoftheseindustriesinthenextcycle,their collectiveresponsetoachangingeconomywasintegratedintospecificcapitalstructure decisionsthatpositionedthemsowell.ConocoPhillipsmayhavebeeninthe“rightplaceat therighttime”,butitwastheforesighttomanageriskthatallowedthemtocontinueto profit. THEDECISION ThedecisiontomergePhillipsPetroleumwithConocowasamasterpieceofboth timingandoperatingsynergy.Thegreaterriskintheoilindustrywasthwartedby counterbalancingitwithlessriskasamergedcompany.Ifthestudent/investorwill observejustthreefundamentalsfromthisperiod(2001-2006),thecasecanbemade: Table17-1 YEAR 2001 2002 2003 2004 2005 2006 OP. Income 4937 4953 12638 18713 28297 36704 Sales 25030 57201 105097 136916 183364 183650 Assets 35217 76836 82455 92861 106999 164781 435 The2001figuresareforPhillipsPetroleumalone.Besidesthetremendoussurgeingrowth aftermergingwithConoco,noticethatPhillips’operatingmarginin2001was19.72%(Op income/Sales).Ontheotherhand,in2005,atthepeakofgrowth,itwasonly15.43%. Assetturnover(sales/assets),however,wasjust0.7107in2001,whileitwas1.714in2005, almosttwoandonehalftimesaslarge.Whathappened?Thecharacterofthemerged companywasradicallydifferentthantheoldPhilipsPetroleum.Withasmallsacrificein marginalprofit,themergerdecreasedthecapitalintensityratio(Assets/Sales)requiring lessfixedassetsperdollarofincomegenerated.Thisincreasedproductivitycreateda moreevencash-flow,allowinggreaterflexibilityinfundingbutwithoutchangingthe leveragecharacteristicstowardgreateruseofdebt.Ineffect,theConocoPhillipsmerger increasedreturn,butreducedriskwhichisarareandformidablecombination. Onebenchmarkofanyinvestmentisthetimeittakestoreturnit.Whenany companypurchasesalargeamountofassets,thereisusuallyalagbetweentheintegration ofthoseassetsandprofitability.Itisatthisstage,thatstockpricesdeclineorbecome stagnatebecausecapitalistiedupinaprojectwithlittlecurrentyield.Inthecaseofthe ConocoPhillipsmerger,thereturnswerealmostimmediate.Bythefourthquarterof2002, operatingincomewassurging,settingthenewlyformedcompanyupformoreprofitability tofollow.Infact,ConocoPhillips“recovery”almostperfectlycoincidedwiththatofthe greatereconomy.Hadthelagbetweenassetpurchaseandprofitabilitybeenayearortwo, thecompanywouldhavealmost“missedtheboat”,failingtoparticipatesuccessfullyinthe firstphaseoftheexpansioncycle.Thatscenarioalonecouldhavedevastatedshareholders becausetheremainderoftheeconomywasbeginningto“heatup”.Fromatechnical standpoint,anyrisingdefaultprobabilityquicklybegantodissipatebythefourthquarter of2002assalesandprofitspickedup.Theslightlyincreasedleveragefromthe combinationofbalancesheetsreducedthepotentialforshareissuesaswellasincreasing taxbenefits. PRICEPERFORMANCE 436 Inthemajorityofbusinesscyclesahigherstockpricereflectssectordominance whichmayspananintervalfromsixmonthstotwoyears.However,assumptionsabout stockpricescanblindustotherealityofdefiedprobabilities.Weassumethatafirmlike ConocoPhillipscanmovetowardanoptimalcapitalstructureforonlyashorttime,butin thiscase,thetraderswhospeculatedonthebasisofmomentumwonthe“jackpot”:the naiveinvestorwhokeeps“playingthesamenumber”wasrewardedbyanoilmarketthat soared. Table17-2 YEAR START END LOW HIGH % CHANGE 2002 60.12 48.39 44.66 63.73 -19.51 2003 49.33 65.57 45.31 65.57 32.92 2004 65.48 86.83 64.78 90.99 32.61 2005 (SPLIT) 84.11 116.36 84.11 116.36 38.34 2006 60.61 71.95 56.03 73.07 19.71 ANTICIPATINGPERFORMANCE:LEVERAGESTATES ConocoPhillipsdisplaystheclassicprogressionfromadebtorientedstatein2002to asequenceofprofitgeneratedequitystatesin2003to2005backtoadebtorientedstatein 2006.Whileinterestrateswereincreasing,debtwasneverprohibitivelyexpensive;equity statesweregeneratedbymoreretainedearningsandanattempttomaximizethe efficienciesofthemerger,whicheffectivelyreducedrisk.Thetaskfortheinvestoristo forgethetransitionbetween“debtstate”and“profitorequitystate”byrecognizingthe shiftincapitalstructure. 437 Table17-3 YEAR 2001 2002 2003 2004 2005 2006 Financial Lev. 1.0734 1.129 1.0766 1.03 1.0179 1.0305 LTD/CAP % 37.52 39.05 32.22 25.17 16.94 21.84 OpMom. -0.5973 0.00519 1.853 1.59 1.51 190.48 Assets/Cap 1.5345 1.5864 1.6261 1.6265 1.6853 1.5504 Themostprofitableyears(2003-2005),werecharacterizedbysteadyoperatingmomentum withinanarrowrange.However,whenacquisitionsweremade,firstwithConocoin2002, andlaterwithBurlingtonResourcesin2006,thestablerelationshipbetweensalesand operatingprofitwasupsetbecausethesenewassetsmustbeintegratedintotheexisting structure,requiringrestructuringcharges,layoffsetc.ToConocoPhillips’credit,the “debt”yearswerecharacterizedbyincreasingoperatingmomentumwhichsignifiesthat additionalinterestexpensecanbepaid,andwillnot“ratchetup”theprobabilityofdefault. Additionally,operatingmomentumwasraisedduringtheinitialyearoftheexpansion (2003)evenasfinancialleveragewaslowered.Inessence,ConocoPhillipswasabletoraise betaduringaperiodofmarketexpansion,butreducetheotherrisksassociatedwith financialleverage.This“miracle”ofmarkettimingproducedanetgaininbetaattheonly timeinthecyclewhensuchagainwouldbeadvantageous-whentheentiremarketis risingandsectorperformanceisnotyetcritical AlsonoteworthyishowthemergerleftPhillipsPetroleum’s’debtstructure relativelyuntouched.Themergerwasconsummatedwithmoredebt,butitwassoon bufferedbyprofitgeneratedequitysothatlong-termdebttocapitaldecreased.This decreaseallowedgreaterflexibilityintheuseofdebtwhenConocoPhillipsneededtomake anacquisitionastheydidin2006withBurlingtonResources.Thatdealrequiredmore equitythandebt,althoughConocoPhillipsmorethandoubledexistinglong-termdebtto 438 23091.TheconsequentLTD/CAPratiowentuptoonly21.84%fromadiminutive16.94 %. Incaseswherebothdebtandoperatingmomentumareincreased,greaterassetsto capitalactsasanadditionalplusfortheinvestorratherthananegativeindicator.Such increaseswillescalatethereturnoncapital(ROC)andcanofferalessriskysourceof financing.However,whenitisdecreased,asitwasin2006,ConocoPhillipsnolonger neededtohedgetheriskinlong-termdebtastheydidin2002;profitshadbeensurging andthelong-termdebttocapitalratiowasrelativelylow.Thediminishedratiowouldhave beenmoreofa“judgmentalnegative”hadthecompanyorsectorbeensufferingandthe stocklanguishing. QUARTERLYLEVERAGERESULTS Therationalconnectionbetweenspeculativegamblingandprobabilisticinvestingis greatertodaythanatanytimeinfinancialhistory.Optionstrading,currencyswapsand exoticderivativeshavegainedlegitimacybecausetheirproperuseoffersriskreductionand protectionthataninvestormightnothaveotherwise.Itdoesthestudentadisserviceto denythatspeculationissometimeslucrativeandthat“highrollers”existonWallStreet whomakeitaprofessionalcareer.However,themathematicsofshort-termdecision makingfavorsthevariationcausedbyuncontrollablerandomfactors.Ifoneinvestsina stockthatistrendingforthreeweeks,itisacrapshoottoassumethatitwillbehave similarlyinweeksfourandfive. Whileleveragestateinvestingoffersbetterchoices,anticipatingastate(andthe transitionbetweendebtandprofit)isfraughtwiththesamepitfallsasanyshort-term investment:thecompanywillhavehigherprofitsfortwoquarters,onlytomakeabad debtladenacquisitioninthethirdquarter.Notallcompaniesworktooptimizetheir capitalstructures.Withthatcaveat,thefollowingquarterlydataindicatesthatspeculation onConocoPhillips’leveragestatewouldprovetobethecorrectandobviouschoice.: 439 Table17-4 QUARTER 3rdQTR2002 4thQTR2002 1stQTR2003 2ndQTR2003 Operating Income 1058 2065 3651 2756 Sales 15678 20688 27077 25595 Financial Leverage 1.145 1.1186 1.0607 1.0715 LTD/CAP% 37.31 39.057 32.757 31.32 Operating Mom. N/A 2.978 2.55 4.48 Assets/ Capital 1.6196 1.5864 1.742 1.7308 Anyfundamentalsthatmimicthistypeofrelationshiparenottobeinterpretedasworthy ofspeculation.However,weobserveearningspressurecoincidentwithadeclineinlong- termdebttocapitalandthefinancialleverageratioovermorethantwoquarters.The firmisreducingitsprobabilityofdefaultaswellasloweringthefinancialleverage componentofbeta.Aftertakingondebtin2002toconsummatethemerger,itwouldbe improbablethatConocoPhillipswouldbeginincurringgreaterlong-termdebttocapital afterreducingitinquartersoneandtwo.Suchanassumptionisproblematic;ifprofits begantofalterinthethirdandfourthquartersof2003,theinvestorwouldbeleft“holding thebag’withlesslong-termdebtbutagreaterfinancialleverageratiobecausetheability topayinterestexpensewouldbeundermined.Suchatransitionstateof“halfandhalf”is muchmoredifficulttoreadandisonereasonwhyspeculationshouldbeavoided. INTERPRETINGREGRESSION Withoutafullestablishmentofthesecuritiesmarketline(CAPM),wecanalso interpretthequarterlyregressionin2003.Theseregressionsaretheprecursorstothe determinationofacostofequity.Theyaresimplyfiveyearregressionsbetweenthe percentagechangeinthestockandthemarketpercentage,withthreedatapointsdropped andreplacedasanewquarterisadded.WearesearchingforchangesinRsquaredor 440 alphathatwouldindicatelessormoredependenceonthemarket,andchangesinbeta whichwouldbeattributabletogreaterleverage. Table17-5 PERIOD LINE CORRELATION ACTUALMKT% 2002 Y=0.3057+0.3814(X) 0.288006 -9.995% 1stQTR2003 Y=0.3331+0.3408(X) 0.237738 4.2% 2ndQTR2003 Y=0.4405+0.3554(X) 0.246158 5.11% 3rdQTR2003 Y=0.5256+0.5226(X) 0.35043 9.03% 4thQTR2003 Y=0.9668+0.6477(X) 0.408869 7.65% Inthisdisplay,thenumberthatprecedes“X”isbeta,whiletheothernumberisalpha.The correlationis“R”andnotRsquared.Theregressionlineitselfusesfiveyearsofmarket datatoestablishabeta,buttheactualmarketchangesforthatperiodareposted.Ineffect, ConocoPhillipsincreasedalpha,beta,andthecorrelationcoefficientsimultaneously,while themarketwasincreasing.Sincethefirmwasreducingitsdebttoequityratioatthistime, theincreaseinbetaisderivedfromincreasingoperatingincomerelativetothemarket.As thisinitialrecovery“surge”inthemarketbegantosputter,ConocoPhillipswouldcontinue toreducefinancialleverageaswellasitsdependenceonthemarket-thecorrelation coefficient,“R”.However,non-systematicrisk-asmeasuredbyalpha-wasincreased.A “cartel-like”auraformedaroundoilwhichseemedtoimmunizeitfromcyclical downturns:Thestockswouldriseevenifthemarketdidnot.Thatanomalywasthe outgrowthofpriceanddemandeffectsthatwereintheoilcompanies’favoratthetime, althoughpublicfearsofcollusionwereforeverimminent.Thehigheralphaindicatesthat thecompanyisgoingtoproduceareturnevenifthemarketreturnsnothingandisa functionoftheintegrationofcompanywithindustry.Anyindustrythatproducesan especiallyvaluablecommoditythatisinconstantdemand,andcommandsahighprice,will haveahigheralpha.ConocoPhillipsbeganreducingrisktothepointwhereprofitswere nolongersystemic,butdependedontheirpositionintheoilindustry“peckingorder”. 441 ESTABLISHINGACOMPARATIVECOSTOFEQUITY Ratherthandependingononemethodofdeterminingthecostofequity,the interactionbetweenearnings,priceandmarketcanbestbeshownwiththree.Weshow thatasurgingcompanylikeConocoPhillipsisresilientenoughtoharborradicallydifferent costsofequityandstillestablishanincreaseinEVA.However,forinvestmentpurposes, webelievethattheCAPMoffersthemostequitablemethodformostfirmsbecauseofthe tendencytodependonthemarket:Nineoutoftenstockswillgodownduringabear market,andthreeoutoffourstockswillgoupduringabullmarket.Whenusingthe GordonmodelortheE/Pmethod,idiosyncrasiescreepinthatareidentifiedwithmarket skeworthespecificsituationofthecompany.Theinvestorwantsacomparativerisk figure,anopportunitycostthatisbestderivedfromasamplepoolthatcomprisesallstocks -theCAPM. • THECAPM.Belowarethetalliedcostsofequityfortheyears2002-2006.Amonthly fiveyearregressionwasperformed,startingin1998forthe2002figureandcomprising 2002to2006forthelatterfigure.Inthisperiod,themarketriskpremiumnever exceededfivepercent,andsothatfigurewasestablishedasafloor.Theaveragerisk- freeratesaretabulatedaswell. Table17-6 YEAR BETA RISK PREMIUM RISK-FREE % CostofEquity 2002 0.331477 0.05 0.0461 0.0652 2003 0.647655 0.05 0.0401 0.0725 2004 0.823431 0.05 0.0427 0.0839 2005 0.505491 0.05 0.0429 0.0682 2006 0.690769 0.05 0.048 0.08254 442 • THEGORDONMODEL:TheGordonmodelseemstopricethecostofequitytoo high,butestablishesameasureofperformanceabovethemarketifthefirmisdoing comparativelywell.Thereasonthatthemeasurementisplacedwellabovethatderived bytheCAPMisthatitestablishesaninternalbenchmarkforthecostofequity- dependenceonboththereturnonequityandtheretentionratioforearnings-which maynotreflectmarketaverages.LiketheCAPM,moreearningswillincreasethecost ofequity,butunliketheCAPM,underperformancewilldecreaseit.Forthatreason alone,itcanbeusedasacomparativefigurewiththeCAPM,butisnotareliable measureofcomparativeriskwiththemarket.ThefollowingtablesgiveROEsand retentionratiosfortheappropriateyears.Thestudent/investorshouldbeawarethat theproductofthesetwofiguresisaddedtothe“expected”dividendyield-thatis-the dividendexpectedinthenextfiscalperiod. Table17-7 YEAR 1997 1998 1999 2000 2001 NetIncome 959 237 609 1862 1661 Equity 4814 4219 4599 6093 14340 ROE 19.92% 5.617 13.24 30.56 11.58 Retention Ratio 68.19% 0 43.51 81.41 75.74 Growth 12.59% 0 5.76 24.88 8.77 Table17-8 YEAR 2002 2003 2004 2005 2006 NetIncome 698 4735 8129 13529 15550 Equity 29517 34366 42723 52731 82646 ROE 2.364% 13.66 19.03 25.66 18.82 Retention Ratio 0% 76.62 84.84 87.89 85.36 Growth 0% 10.47 16.15 22.55 16.06 443 Table17-9 YEAR 2002 2003 2004 2005 2006 Average Price $54.20 $55.44 $77.89 112.965* $64.55 Next Dividend $0.82 $0.90 $1.18 $1.44 1.70* *Price:IndicatesSplit *NextDividend-IndicatesAuthor’sProjection Wethenassemblethegrowthinformationinafiveyearmovingaverageandaddeach figuretothemodifieddividendyieldforeachyear. Table17-10 YEAR 2002 2003 2004 2005 2006 Gordon Model 9.39% 11.6 13.59 12.85 15.64 • EARNINGS/PRICEOR“E/P”:Inanyirregularmarketwherestocksareeither undervaluedorovervalued(justaboutallmarkets),theE/P“ruleofthumb”failsto work.However,itisthecapitalstructureequivalentofaP/Eratio,andtherelationship betweenearningsandpricecarriesalotofinformationaboutthecostofequity.When investorsvalueonestock’searningsaboveanother,theyarewillingtobiduptheprice ofthestockandthefirmcaneasilyattractequity.Equityissuesbecomeprofitable becausethegreatestamountofcapitalisraisedfortheleastamountofsharesissued. Theuncertaintyofthismeasurementisderivedfromthevolatilityofthemarketwhich 444 makesthepricingofequityunreliableasanopportunitycost.Althoughtheprice changesonadailybasis,theriskofequityismorestable.Nevertheless,inasurging company,eventhisunreliablemeasurementwillcreateaballparkfigurefor percentageincreasesinEVAbecauseittendstoriseandfallsimilarlytotheCAPMand theGordonmodel.Thebasiclogicbehindthiscorrelationisthatbothpriceand earningsriseintandemthroughoutatypicalbusinesscycle,reachapeak,andthenfall whentheriskoflowerearningseclipsesprice.Thatsub-cycleisanalogoustohowa firmattractsmorecapitalwithbetterearnings.Unfortunately,therelationshipisnot mathematicallyprecise,andusingthismodelwithcertaingrowthstocksorintimesof excessivespeculation,willbadlyskewanycomparisontocompaniesofsimilarrisk. Table17-11 YEAR 2002 2003 2004 2005 2006 EPS 0.72 3.45 5.8 9.55 9.66 PRICE 54.2 55.44 77.89 112.905 64.55 E/P 1.32% 6.22 7.45 8.45 7.49 DIFFERENCINGROEANDTHECOSTOFEQUITY Table17-12 YEAR 2002 2003 2004 2005 2006 ROE 0.02364 0.1366 0.1903 0.2566 0.1882 Costof Equity 0.0652 0.0725 0.0839 0.0682 0.0825 ROE-Cost ofEquity -4.16% 6.41% 10.64% 18.84% 10.57% Theperformanceofthismeasurementmirrorsstockperformanceasmuchasthe percentagegaininEVA.Theyareverysimilarmeasurements;adecreaseinthisindicator (orEVA)doesnotnecessarilysignifyadecreasingstockprice,althoughthecorrelationis 445 strong.Ifthemarketissatisfiedthatevenadecrease(asin2006)willoutperformthe market,equitycapitalmaystillflowintothecompanyprovidingthatthenear-term outlookisoptimisticenough. CONTRASTINGTHEREQUIREDRETURNWITHTHEEXPECTEDRETURN TotakeadvantageofthedifferencebetweentheCAPMandtheGordonmodel,we recognizethattheCAPMexpressesamarketcomparative“requiredreturn”,anddividend discountmodelsliketheGordonmodelwillexpressan“expectedreturn”thatismore dependentontheinternaldynamicsofthecompany.Wheninternaldynamicsoutpacethe market,thereisoftenupwardpressureonthestock.Whenastockunderperformsthe market,theGordonmodelcalculationwilloftenbebelowtheCAPMderivedcostofequity. Thislackofequilibriumcanbeaninvestor’sbestfriend;ifwesubtracttheCAPMcostof equityfromthedividenddiscountmodel(inthiscase,theGordonmodel)wecanusethe differencetocomparethetworisks-marketandinternal. Table17-13 YEAR 2002 2003 2004 2005 2006 GORDON 9.39% 11.6 13.59 12.85 15.64 CAPM 6.52% 7.25 8.39 6.82 8.254 DIFFERENCE 2.87% 4.35 5.2 6.03 7.386 Whilethedifferencedoesnotrepresentaperfectrenditionofpressureonthestock,this indicatorissimilartoEVAandtheROEdifferencingthatwedisplayedpreviously.The growthfactorintheGordonmodelisderivedfromearningsandretention,whilethe CAPMismoredependentonforcesinthemarket.Effectively,wearemeasuringthe accelerationofearningsincomparisontothecostofequity.Noticethough,thatthefigure in2006wasabovethatin2004and2005.Sinceitisaconcurrentindicatorofstock performance,similartoEVA,theinvestordoesnotknowifthefigurehaspeakedandifthe 446 stockwillsubsequentlydecline.Inthisregard,theinvestorneedstofollowthemarket imperativeandrecognizethesignsofamarket“top”(seethechapteronthebusinesscycle) aswellasexaminingleveragepositions. THEEVA/CAPITALDYNAMIC Institutionalinvestorssuchaspensionfundsmustdiversifyriskandsearchfor companieswhohavestablecash-flows.Ex-postevaluationofseveralyearsofEVA increasesisaprovenindicatorofviability.Notonlydoincreasescorrelatewithpositive changesinstockprice,butthesizeandstabilityofEVAcanbeaforcethatguaranteesa growingdividend.However,incapitalstructure,wearemoreconcernedwiththe movementofEVAthatitsabsolutesize.ApositiveincreaseinEVAdoesnotalways translatetomovementtowardanoptimalstructure,becauseitisdependentonthecostof capitalratherthanthecostofbankruptcy.Whilebankruptcycostsarealmostalways minimizedwhenthecostofcapitalisminimized,theremaybetimeswhenthecapital structurecannotconformtothemarket.Forexample,ifinterestratesareespeciallylow,a companymayneedtoshedsomedebtinsteadoftakingonnewdebtatalowerrate.Inthis case,theprobabilityofdefaultwouldbetoohighandwouldeclipsetheneedformoretax benefits.Moreover,EVA/capitaldynamicwouldpenalizethecompanyfornottakingon thedebt,becausethetaxadvantagewouldoutweighthedeclineinnetincome.ThusEVA woulddeclineevenasthefirmattemptstominimizethecostofbankruptcy. ForasurgingcompanylikeConocoPhillips,thethreemethodsofcalculatingthe costofequityrevealthesameearnings/riskpressure;iftherequireddecisionisinvestment grade,ratherthananacademicillustration,theCAPMshouldbechosentocalculatethe costofequity.But-ajudicioususeofE/PinfirmswithaP/Erightaroundthemarket averagecanidentifycompaniesthatmightwarrantfurtherexamination.Inessence,both E/PandtheGordonmodelarevaluablecomparativetoolstoquicklylocateaprospective investment.Atthattime,afullregressionisdoneandapreciseassessmentismade. 447 InthecaseoftheGordonmodel,thestandardEVAcalculationwillrevealthatthe companymustretainthecorrectamountofearningsandpayoutalargeenoughdividend, incomparisontonetincomeandstockholders’equity.Thishighbenchmarkwillplacethe calculationveryclosetoROEandproduceasmallEVA.However,periodiccomparisons ofthismethodindicatethecompetencyoffinancialmanagementbecausethereisalow thresholdforerror.Forexample,ifdividendsgrowatanexorbitantpaceincomparisonto priceandearnings,thedeficiencywillberevealedinasmallerEVAcalculation.Similarly, ifstockholders’equityballoonstonewheights,thismethodwillexaggeratetheincrease. ThefollowingtablesitemizenetincomeandequityforConocoPhillipsandthenapplythe threedifferentmethodstodeterminetheEVA/capitaldynamic.Thereaderwillremember thatthecalculationis:Netincome-[(%costofequity)(stockholders’equity)] Table17-14 YEAR 2002 2003 2004 2005 2006 NetIncome 698 4735 8129 13529 15550 Stockholders' Equity 29517 34366 42723 52731 82646 Table17-15 1.CAPM YEAR 2002 2003 2004 2005 2006 CAPM 0.0652 0.0725 0.0839 0.0682 0.0825 EVA -1227 2243 4545 9933 8732 448 Table17-16 2.GORDONMODEL YEAR 2002 2003 2004 2005 2006 GORDON 0.0939 0.116 0.1359 0.1285 0.1564 EVA -2074 749 2323 6753 2624 Table17-17 3.EARNINGS/PRICE(E/P) YEAR 2002 2003 2004 2005 2006 E/P 0.0132 0.0622 0.0745 0.0845 0.0749 EVA 308 2597 4946 9073 9360 Thestudent/investorshouldnoticehowtheCAPMandGordonmodelsboth“punished” ConocoPhillipsforissuingtoomuchequityin2006.Themarginalbenefitsfunctionwill laterrevealthatindeedthetaxbenefitsoftheacquisitionofBurlingtonResourcesfarout weighedanyincreaseinbankruptcycosts;theprobabilityofdefaultwasvirtually unaffected.Therefore,beforetotallydismissingtheE/Pindicationsofmovementtoward anoptimalcapitalstructure(anincreaseinEVA),moreindepthanalysisneedstobedone. NAIVEEXTRAPOLATION Ifthecapitalstructuralistwereastute,heorshewouldhavepickedupon ConocoPhillipsinearly2003.Themarketsoaredandthefirmwaspositionedtotake advantageofthesurge.Theleveragepositionwassolidinanexpandingeconomy-the proportionofequitywasbuildingupthroughretainedearnings.However,bytheendof 2004,thecompanywouldhavehadtwoyearsofoverthirtypercentgains.Atthispoint, “smartmoney”investorswouldsurelytakeprofits.Hadinvestorshesitated,theywould havemissedoutonanotherthirtypercentgain:theyear2005turnedouttobe ConocoPhillips’biggestyetwithastocksplit,andrecordsalesandprofits.Howwouldan 449 investorplaythisdilemma?Fromaleverageperspective,theprobabilitywasagainst anotherbanneryear. Asinlate2002,therewasaconfluenceofriskfactorsinthefirm’sfavor.Thewar intheMiddleEastdidnotproduceuncontrollablesupplyanddemandissues-justasteady riseinoilprices.TheFederalReservewasstillraisingratestohedgeinflation;the economywasheatingup.IndiaandChinawereinthethroesofcompetitionformoreoil; theirrespectiveeconomieswerebooming.Althoughonecannotanticipateapositive corporatemovementafteryearsofstockgainswithleveragetheoryalone,thecollective judgmentofprofessionalanalystsofferssomesolace.Theincreaseinearningswasa66.42 %gainin2005-whichcouldnotgounnoticedevenifthepredictionwereoffbyhalf. Thus,analystswouldhaveperceivedtheinordinatedemandforoilandextrapolateditinto ausableforecast.Ifweaddedthatinformationtoourownnaiveforecast,investorswho watchedConocoPhillipswouldhavebeeninfor“theride”in2005.Investorswhojudge primarilyonthebasisofprobability(thisauthor)wouldhavetakenprofits. Thenaiveextrapolationprocedurerequiresustofindthegeometricmeanofgrowth forbothearningsandthetotalcostofequityinthenearterm.Anytimethatwecompare quarterlyperformanceoneyearapart,theeffectisthesameasseparatingthe fundamentalsbyfiveperiodswithfourintervalsbetweenperiods.Therefore,the calculationisthesamethatweuseforyearlyperiods:[(ThisPeriod’sX/5Period’sAgo X)^0.25]-1=PercentageGrowthperPeriod.Inthiscase,wearetestingtoobserve whetherornotearningsaregrowingfasterthanthetotalcostofequity.Whilenoindicator canguaranteethatthefirmwillnotbeginacceleratingequitythroughstockissuesassoon asaninvestmentismade,theprobabilityofanEVAincreaseisgreaterwhenearningsare leadingequity. 450 Table17-18 QUARTER 4thQUARTER2003 4thQUARTER2004 NETINCOME (Cumulative) 4735 8129 EQUITY(Cumulative) 34366 42723 CAPM% 0.0725 0.0839 TOTALCOSTOF EQUITY 2492 3584 Thethreecalculationsareasfollows: 1.NetIncome:[(8129/4735)^0.25]-1=14.47% 2.CostofEquity:[(3584/2492)^0.25]-1=9.51% 3.ComparativeRatio:14.47/9.51=1.52 Thereaderwillnotethatnetincomeaccumulatesthroughouttheyeartodeterminean annualfigure,butstockholders’equityisacumulativetotal.Sinceaquarterlynetincome figuremightnotbepartofanactivetrend,weusethecumulativeyearlytotalsandtreat themasagrowthprogressionmadeoverfourperiods.Forexample,ifthefourthquarter in2003hadarestructuringcharge,earningsmightbenegativeforthatquarteronly. Ratherthanresorttousingthethirdquarterfigure,simplyusingtheyearlyfiguresfor 2003and2004willsmoothoutanyerrantnoise.Unfortunately,naiveextrapolation mirrorsthesizeoftheincreaseinEVAandassumesthatnextyear’sEVAwillbelikethis year’s.Themajoradvantageofnaiveextrapolationisthatwecangaugetherelativesizeof componentchangesandthencomparethemtoanalyst’searningsforecasts.Exceptfor duringaneconomicdownturn,weneverassumethatthecostofequitywilldecelerate,and soanalysts’consensusexpectationsbecomeabenchmarkforfutureEVAforecasts.While earningscanbevolatile,thecostofequityismorestableinthenearterm;financial managementdoesnotwantequitytoradicallyshiftandpotentiallydilutethepriceper 451 marketshare.However,periodicchangesaremadewhenitiscosteffectivetoissuenew stock,andthosechangesareunpredictable.In2006,forexample,ConocoPhillipsissued stockasamethodforpurchasingBurlingtonResources.WithConocoPhillipslowlong- termdebttocapitalratioofapproximatelytwenty-onepercent,onlythosewhospoketo companyofficialswouldhavepredictedthelargeequityissue.Theinstitutionalimperative wouldhavebeentoissuemoredebt. THEMARGINALBENEFITSFUNCTION Themarketcantemporarilymispriceriskcausingadivergencebetweenthecostof capitalandtheprobabilityofdefault.Inessence,ifloweringthecostofcapitalincreases thecostofbankruptcy,itshouldbeavoided.Overthelongterm,suchdisparitieswillwork themselvesout;moredebtwillincurahighercostandincreasebankruptcycosts simultaneously.However,arisingmarginalbenefitsfunctioninthedomainofadeclining EVAmayhaveupsidepotential.EVAalwaysoptimizesonthebasisofthelowestcostof capital,butremainsaconcurrentindicator;anddoesnotportendachangeinthe income/capitalrelationship.Ontheotherhand,themarginalbenefitsfunctionanticipates newdynamicsbecauseitmaximizeswhenthechangeoneachsideofitisequal.Theneed forbalancedistinguishesthefunction.Iflong-termdebtdeclines,wewouldlookfora reductionineitherstockpriceordefaultprobabilityontheotherside. ThisanalyticaldichotomybetweenEVAincreasesandthemarginalbenefits functionwasillustratedinfiscalyear2006.ConocoPhillipsboughtBurlingtonResources, increasingbothlong-termdebtandequity.FromanEVAstandpoint,ConocoPhillips issuedtoomuchequityandhadthecapacitytoincreasedebt.Althoughthepotentialto increasetheEVA/capitaldynamicwasapparent,thatfiguredeclined.However,fromthe perspectiveofmarginalbenefits,ConocoPhillipsissuedthepreciseamountofdebtto forestallanincreaseintheprobabilityofdefaultandincreasetaxbenefitsatthesametime. Infact,moredebtwouldhavewouldhaveunderminedtheincreasesinassetsandnet income,andConocoPhillipseffectivelyoptimizedlong-termdebtinthedomainsoftax 452 benefitsanddefaultprobability.Toexhibittheseconcepts,wewillgothroughthestepby stepcalculationsfor2005and2006.ThereaderisreferredtothechaptersonAnalytical ToolsandCapitalStructure,respectively,forhelpwiththemechanics.Toreiteratethe equation:MarginalBenefits=(TaxBenefits)–[(DefaultProbability)X(AmountofLoss)]. Table17-19 2005-2006DATA ZMIJEWSKIFUNCTIONS ZMIJEWSKIPARAMETERS Intercept -9.479 TotalLiabilities/TotalAssets 6.384 CurrentAssets/CurrentLiabilities 0.069 NetIncome/TotalAssets -1.06 Table17-20 YEAR 2005 2006 TotalLiabilities/Total Assets 0.4962 0.4912 CurrentAssets/Current Liab. 0.9182 0.9484 NetIncome/TotalAssets 0.1264 0.0944 MidrangeStockPrice 98.215 129.1(withoutsplit) NumberofShares (millions) 1455 1706 Long-termDebt 10758 23091 TaxRate(decimal) 0.421 0.451 UnamortizedDebt 53093 80940 IntangibleAssets 16439 32439 1.ESTABLISHTHETAXBENEFITS 2005:(10758)x(0.421)=4529 2006:(23091)x(0.451)=10414 453 2.ESTABLISHTHETANGIBLEBOOKVALUEPERSHARE 2005:(106999-16439-53093)/1455=25.75 2006:(164781-32439–80940)/1706=30.13 3.ESTABLISHTHEAMOUNTOFPOTENTIALLOSS 2005:MarketValue=(1455)x(98.215)=142,903 2005PotentialLoss:[1-(25.75/98.215)]x(142903)=105437 2006MarketValue=(1706)x(129.1)=220,244.6 2006:PotentialLoss[1-(30.13/129.1)]x(220244.6)=168842.82 4.ESTABLISHTHEPROBABILITYOFDEFAULT 2005:1/[1+(EXP(9.479)-((0.4962)x(6.384))-((0.9182)x(0.069))+((0.1264)x(1.060)))]= 1/[1+(EXP6.3818874)]=1/592.042=0.001689069=0.169% 2006:1/[1+(EXP(9.479)-((0.4912)x(6.384))-(0.9484)x(0.069))+(0.0944)x(1.060)))]= 1/[1+(EXP6.3778)]=1/588.6334=0.00169885=0.169% 5.COMBINETHEEXPRESSION 2005:4529-((105437)x(0.001689069))=4350.91 2006:10414-((168843)x(0.00169885))=10127.16 Bymaintainingadequateearningsandissuingtherightamountofdebt, ConocoPhillipscreatedamergerthatdidnotincreasetheriskofdefault.Iflessdebtis incurredinthenextyear,thepremiumwillbeonincreasingstockpricewhiledecreasing theprobabilityofdefault.Suchataskisdifficulttoaccomplishsosoonafteramajor acquisitionbecauseitentailstherapidassimilationoftheotherfirm’sassetswhile increasingprofitability.Often,costoverrunscoupledwithslowintegrationwillrequire 454 morecapitalandretainedearningswillnotbeadequatetocoverthedeficit.Theresult?- Ahigherproportionoflong-termdebttocapital.Thisisonemorereasonthatdebtladen leveragestatestendtobesequential.Thereadershouldalsonotethatthecomparison between2005and2006entailedanadjustmentforthe2005stocksplitbyputtingthe2006 priceonthesamelevel. WecanconcludethatnoneoftheaddedassetvaluewasreflectedinEVA. ConocoPhillips’improvedpotentialwasmoreapparentinthemarginalbenefitsfunction, butthatcalculationcannotpredictfutureprofitabilityeither.Toreconciletherisksofthe mergerwouldentailexaminingBurlingtonResourcespastprofitswhileitemizingfixedand variablecostsandlookingforsynergy.Anevenmoredetailedexaminationwouldentail producingatimelineforimplementation.Inmostcases,theonlyrecoursefortheanalyst istodoacomparativehistoryofthecapitalturnoverratio%∆SALES/%∆CAPITAL andensurethatthecompanyiscompetitive.Justaseffective,butpronetoexaggeration,is toweightheestimatesfromseveralcompanysources-aninformaltypeof“guidance”. THECOMPARATIVECAPITALDYNAMIC:GAUGINGUPSIDEPOTENTIAL Therelationshipbetweenearningsandthecostofequitycannotimprove indefinitely.AlthoughEVAincreasesareconcurrentwithstockpriceappreciation,thereis littleabilitytodetectadownsideshiftinearningsandalmostnoabilitytopredictanequity issue.Whileweattempttocompensatefortheinformationdeficitwithgrowthrate comparisonsandleveragestateanalysis,abettermethodmightbederivedfromusingthe comparativecapitaldynamic. Throughoutthistextwehaveemphasizedthepointthatstockpricemaximization occurswhenthetargetcapitalstructureisreached.Theactualizationoftheoptimum remainsafunctionofearningsacceleratingfasterthanthetotalcostofequity,or%∆Net Income/%∆TotalCostofEquity>1.Theratiooftheabsolutes,NetIncome/TotalCost ofEquityisanadaptationofEVAthatoffersachancetocomparemagnitude.Whena firmtakesondebt,itscomparativecapitaldynamic(CCD)tendstobesmallerthanwhenit 455 isgrowingthroughretainedearnings;bydefault,earningsmustbegrowingtoproduce moreretainedearnings.Laterinthebusinesscycle,moreearningswillenlargetheCCD untilitcomestoapeak,wheremorefinancingiscosteffectivefromanequityissue.Inthe caseofConocoPhillips,theCCDpeakwasreachedin2005whenthestocksplit,twofor one.In2006,thefirmacquiredBurlingtonResourceswithstockanddebtthatdiminished theCCD.Althoughthestockstillappreciated,muchofthegainwasfromtheparticular circumstancesintheoilmarketwithrisingpricesandpentupglobaldemand. Ineffect,eachindustryhasadifferentstandardofCCDbasedontherelationships betweenthreespecifictypesofreturns:returnonassets(ROA),returnonequity(ROE), andreturnoncapital(ROC).Whilesomeintersectorcomparisonscanbemade,the analystmustbecarefulaboutcomparingdifferentfinancialstructureslikeinsuranceor investmentbanks,bothofwhichhaveveryhighCCDs,withahighturnoverCCDsuchasa restaurantchain.Nevertheless,theprimefunctionoftheCCDistodointraindustry comparisons,andcomparisonsbetweenintervalsforaspecificcompany.Asthereader willobservefromthedata,theCCDhitsapeakwhereearningsbecomemorerisky.Also significantwouldbetheprevailinglevelofinterestratesintheeconomy.Iftheyare generallylow,thentheCCDwilltendtobelarger. Table17-21 YEAR 2002 2003 2004 2005 2006 NetIncome 698 4735 8129 13529 15550 TotalCost ofEquity 1925 2492 3584 3596 6822 CCD 0.363 1.9 2.27 3.76 2.28 %Stock Price -19.51% 32.92 32.61 39.34 18.71 456 Sincetherearenoindustrystandards,theresearcherislefttointerpretwhat“toohigh”a numberwouldbe.LikeP/Es,thesefiguresriseandfallbuthavemorestabilityand rationalitybehindthem.Forexample,ifConocoPhillipsrosebacktoabovea“4”,the investormightquestionwhetherthenextyearwouldbe“bearish”. EARNINGSPRESSURE Ifwecomparenaivelyextrapolatedearningsgrowthrateswithanalysts’forecasts, weobtaininformationthatcontrastsatimeprogression(naiveextrapolation)withan integrateddemandforecast(theanalysts’).Wedonothavetosellastockifanalysts’ forecastsarelessthanthetimeprogression,butweneedtorecognizethatchangesoccur afterearningsaccelerationpeaksandweneedtoassessrisksmoreclosely.Moreover,we alwaysassessearningspressureinthedomainofanalyst’sforecastsbecauseconsensus opinionismoreinformedthananytimeseriescanbe.However,anothertechniquecan produceafundamentallyderivedforecastthatcomparesfavorablywithnaive extrapolation,and-iftheanalysthasanideaaboutthedirectionofearningscomponents- withforecasts.Thefollowingequationisa“pluginthenumbers”functionthatworks easiestifdonefromrighttoleft: (1-A)x(B+(Cx(B-(Dx(1-E))))) Table17-22 FUNCTION EXPANATION A)PayoutRatio DividendsPaid/NetIncome B)ROA NetIncome+((int.expense)x(1-tax rate))/Assets CInterestBearingDebt/Equity D)InterestrateonTotalDebt InterestExpense/InterestBearingDebt E)TaxRate EffectiveTaxrateexpressedasa decimal 457 Theequationwillproduceagrowthrate,whichwillbecurrentwhenguidanceaboutthe directionofthecomponentsisgiven.Forexample,ifaccountantstelltheanalystthatthe effectivetaxrateisestimatedtobe39%thisyear,thatisonepieceofinformation.If interestratesaregoingupbytwenty-fivebasispointsaquarter,thatisanotherpieceof information.Inessence,theequationcanbeputintoaspreadsheetandatrailingtwelve monthfigurecanbeproducedfornetincomeandtheretentionratiosothateachquarter theequationisupdatedwithnewinformation.However,asaninvestor,weproducea concurrentgrowthrateandcompareitwithanalysts’estimates.Whenestimates underminethisgrowthrate,weassumedownsideriskandearningspressure.Ifanalysts’ expectationsexceedthisgrowthrate,welooktoitasapossibleinvestmentvehicle.The followingwasanassessmentmadeattheendof2006: Table17-23 PRELIMINARYCOMPONENT YEAR2006 NetIncome 15550 InterestBearingDebt 27134 Stockholders'Equity 82646 TaxRate 45.1% InterestExpense 1087 Table17-24 FUNCTION CALCULATION RESULT A)Payout 2277/15550 0.1464 B)ROA 15550+((1087)(1- .451))/164781 0.09799 C)Debt/Equity 27134/82646 0.3283 D)InterestRate 1087/27134 0.04 E)TaxRate 0.451 0.451 458 Pluggingthesenumbersintotheequationyieldsagrowthrateof0.10495or10.495%. EarningspershareforConocoPhillipswere$9.80persharein2006.Analysts’estimates are$8.89persharein2007foradecreaseof(8.89/9.80)-1=-9.29%.Thus,anyinvestor shoulddoathoroughlongtermassessmentbecauseanalysts’estimatesundercutourown growthfigure.Thereisdownwardpressureonearnings. (BacktoTableofContents) 459 APPENDIX:SELECTEDFINANCIALDATA-CONCOCOPHILLIPS2001-2006 Table17-25 YEAR 2001 2002 2003 2004 2005 2006 EBITDA 4937 4953 12638 18713 28297 36704 SALES 25030 57201 105097 136916 183364 183650 ASSETS 35217 76836 82455 92861 106999 164781 INTEREST 338 566 844 546 497 1087 LTD 8610 18917 16340 14370 10758 23091 EQUITY 14340 29517 34366 42723 52731 82646 CAPITAL 22950 48434 50706 57093 63489 105737 NETINCOME 1661 698 4735 8129 13529 15550 ROE% 11.58 2.364 13.66 19.03 25.66 18.82 RETENTION% 75.74 0 76.62 84.84 87.89 85.36 LTD/CAP 37.52% 39.05 32.22 25.17 16.94 21.84 FIN.LEV* 1.0735 1.129 1.0766 1.03 1.0179 1.0305 *ThefinancialleverageratiowasdeterminedfromEBITDA(earningsbeforeinterest, taxes,depreciationandamortization).ThestandardmethodologyistouseEBITonly. (BacktoTableofContents) 460 18 MICROSOFTVERSUSCONOCOPHILLIPS:COMPARING COMPANIESINDIFFERENTINDUSTRIES Earningspershare,forbetterorworsehasbecometheuniversalcomparative measurementincorporatefinance.Itsoffshoot,theP/Eratio,isoftenusedtocompare companieswithinindustries,indicatinga“valuebuy”whenthefigureiscomparatively low.Bothofthesemeasurementssufferfromwhatbehavioralfinancecalls,“reference dependence”:onedollarandninety-ninecentsisonlymeaningful,whenwedefineone dollarasa“small”amount,andaninety-ninepercentincreaseas“large”.Inanother allusiontopsychology,wealsoneedtodefinewhatistermed,“normal”,andwedosoby makingconstantcomparisons.Ifninety-ninepercentincreasespassfor“normal”,thena ninety-eightpercentincreaseisconsidered“small”andsub-standard.Thus,ourtwo dependablemeasurements,EPSandP/E,giveuslittleinformationtodistinguishthe qualityofwhatismeasured. Thecapitaldynamic/EVAisbynomeanstheperfectmeasurementeither.In theory,itcancomparetwocompanieswithdifferentoperatingandfinancialleveragesand putthemonanevenkeel.Forexample,acompanywithlowoperatingleverageshould haveasteadyenoughincometoaffordmoredebt,andbuildupEPSthroughamore limiteduseofequity.Thatrelationshipshouldincreasethecapitaldynamiceventhough netincomemightbecomparativelysmall.Analogously,acompanywithmoreoperating leveragebutlessfinancialleveragewouldbeabletoincreaseitscapitaldynamicbymaking largeincreasesinnetincome.Sincethesecompanieswouldpresumablybeindifferent sectors,thephaseofthebusinesscyclewoulddetermineafirm’srespectivecostofcapital. Latestagesinthecyclewouldfavortheequityfinancingcompany,whichwouldhavea comparativelysmallreactiontointerestratehikes.Earlystageswouldfavorwealthier companieswhofundwithdebtandcouldtakeadvantageoflowerrates. 461 Severalquestionsneedtobeaddressed:accountingforthesizeofthecompany, woulddividingEVAbytotalassetsbeabettercomparativefigurethanEPS?Whatwould theresultsbeifweputEVAonapersharebasis?Infact,therecanbenouniversal, comparativeindicatorbecausethecomplexityofstockmarketpriceincreasesdoesnot warrantit;atdifferentintervals,pricesmaytracksales,dividends,earnings,capital expenditures,andeventheholidayseason.Ourcontentioninthischapteristhatthe adaptationsofEVAoffermoreinformationthaneitherearningsgrowthorP/E calculations,andthattheycanbeusedtocomparecompaniesinentirelydifferent industries. APPLESANDORANGES:MICROSFTVERSUSCONOCOPHILLIPS FewcompaniesareasdifferentasMicrosoftandConocoPhillips.Onecompanyspends billionsofdollarscombingtheearthforaresourcethatliterallyfuelstheworld’s productionlines-physically.Theotherspendsbillionstofuelthosesamelines intellectually-producinga“virtual”symbolicvariantofthe“real”.Naturally,thetwo companieshaveentirelydifferentcapitalstructuresinplace.ConocoPhillipsis comparativelywellbalanced,generatingincomewithequalpartsofprofitmargin,asset turnoverandequitymultiplier.Ontheotherhand,Microsoftdependsalmostentirelyon profitmarginforitsnetincome.Itsshunningofdebtfinancinghasbeenlegendary.Both ofthesestructuresarerepresentativeoftheirrespectiveindustries,characterizedby differentamountsandvolatilityofcash-flow.Thefollowingtablesdisplaythecomparative fundamentalsfortheyears2005and2006: 462 Table18-1 MICROSOFT YEAR 2005 2006 Sales 39788 44252 OperatingIncome 15416 17375 Assets 70815 69597 NetIncome 12254 12599 Stockholders'Equity 48115 40104 CAPM%CostofEquity 11.15 10.11 Growth* 25.47%x0.7345=18.7 31.42%x.0.7345=23.07 SharesOutstanding 10710 10062 DividendsPaid* 36112 3345 *Microsoftpaidaspecialdividendin2005.Thegrowthrateofretentionmultipliedby ROEisadjustedforthispayment. Table18-2 CONOCOPHILLIPS YEAR 2005 2006 Sales 183364 183650 OperatingIncome 28297 36704 Assets 106999 164781 NetIncome 13529 15550 Stockholders'Equity 52731 82646 CAPM%CostofEquity 6.82 8.25 Growth 22.55 16.06 SharesOutstanding 1416.65 1609.73 DividendsPaid 1639 2277 AsdetailedinthechaptersonCapitalDynamicsandAnalyticalTools,theEVAforboth companiesisdeterminedasfollows:NetIncome-[(CAPM%CostofEquity)x(Equity)]= EVA. ConocoPhillips 2005:13529-3596=9933 2006:15550-6818=8732 463 Microsoft 2005:12254-5365=6889 2006:12599-4055=8544 COMMONGROUND Whilewebelievethatthepercentagegaininthecapitaldynamic/EVAwillmirror ashifttowardanoptimalcapitalstructure,weareunsureofthemagnitude.Forexample, weneedtoaskourselvesthequestion,isatwentypercentgaininasmallEVAbetterthana tenpercentgaininanEVAthatistwiceaslarge?ToputtheEVAsonacommonbasis,we candividebybothassetvalueandthenumberofsharesoutstanding. Table18-3 ASSETS CONOCOPHILLIPS MICROSOFT 2005 9933/106999=9.28% 6889/70815=9.73% 2006 8732/164781=5.30% 8544/69597=12.28% Table18-4 PERSHARE CONOCOPHILLIPS MICROSOFT 2005 9933/1416.6=7.01 6889/10710=0.64 2006 8732/1609.7=5.42 8544/10062=0.85 Ultimately,wefaceasimilardilemmathatwehadwithearnings;weareunsure howthemarketwillvaluethemeasurements.Microsoftisthesuperiorcompanybasedon valueperdollarofassets,butConocoPhillipsisbyfarthebettercompanybasedonEVA pershare.Thefundamentaldifferenceisderivedfromthequalityoftheirrespectiveassets andthemethodbywhichtheyarefunded.Microsofthascomparativelylittletangibleasset valuepersharebecauseitproducesitsincomefromintellectualpropertyandplacesa premiumonmanagerialandprogrammingtalent.Ontheotherhand,ConocoPhillipshas manytangibleassetsbecauseoilexplorationproducesaphysicalcommodityrequiringa heavyinvestmentinmachinery.Theexaggeratedgrowthcycleintechnologyduringthe 464 1990srequiredMicrosofttoissuealargeamountofequityforwhichithasbeenpenalized; theybuybackshareswithretainedearnings.Alternatively,ConocoPhillipshasabalanced approachtowardfinancing,keepingsharestoaminimumandexpandingwithretained earningsandsomelong-termdebt.Atonetime,theinvestorwouldchoosebetweenthe moreriskygrowthofMicrosoftandthesteadydividendofPhillipsPetroleum.Afterthe Conocomergerin2002,theriseofoilcompanieshaveprovidedbothrapidgrowthand dependabledividendincome,whiletechnologystockshavemirroredthepeaksandtroughs intheeconomy. Thetwocompaniesfacepolarizedrisksaswell.ForConocoPhillips,therisksof legislation,politicalturmoilandscarcityaffectthepotentialsupplyoftheirproduct.For Microsoft,amarketgluttedwithinnovativesoftwarecanreduceeffectivedemand.To illustratethisdichotomy,considerthefollowing:mostindividualswillneverseeanoilwell intheirhometowns,butthelocaluniversityprobablyhasquiteafewindividualswhoare capableofproducingacommercialsoftwareproduct.This“democratization”of technologyhasafflictedboththepersonalcomputingandsoftwareindustrieswithover production.Althoughthedemandforaqualityproductremainsstrong,competitionhas producedenoughsubstitutes(Linux,OpenSourceCode)tounderminethepricingpower ofmajorsoftwaredevelopers. Additionally,thecapitalintensityratioismuchgreaterforMicrosoft,requiringa higherpercentageoffixedassets.Althoughoilproductionisassociatedwithdrillingand machinery,softwareproductionrequiresalevelofexpertiseandtechnologythatproduces morefixedcosts.SinceMicrosoftcarriesnofinancialleverage,thebetaforitsstockis affectedalmostentirelybyoperatingrisk.And-althoughitislowandstablein comparisontomosthightechfirms,Microsoft’sbetaisabouttwiceaslargeas ConocoPhillips’,andmoremarketdependentaswell.Infact,theuniquepositionofoilas ascarceandvaluedcommodityhaschangedthedynamicsofthecostofequityformostoil companies.Thecorrelationwiththemarket,the“Rsquared”componentoftheregression, 465 hassteadilydeclined,whilethenon-systematicriskelement,the“alpha”componenthas risen.Ineffect,ConocoPhillips’stockhastakenona“lifeofitsown”.Itriseswithout beingdependentonstockmarketvolatility,evenasitbecomesmoredependentonthe commoditiesmarket.Moreover,anexaminationofeachfirm’sregressionlinesfor2005 and2006revealssomeofthetruthbehindtheircapitalstructures;whileahigherbetais notsynonymouswithequityfinancing,moredebtraisesrisk,andwillpropelahighbeta stockfurtherupward: Table18-5 CONOCOPHILLIPS YEAR ALPHA BETA RSQUARED REGRESSION 2005 0.657499 0.505491 0.1030512 Y=0.657+0.505 x 2006 0.343046 0.690769 0.0766277 Y=0.343+0.691 x Table18-6 MICROSOFT YEAR ALPHA BETA RSQUARED REGRESSION 2005 0.633924 1.373933 0.32198 Y=0.634+1.374 x 2006 -0.28049 1.112178 0.380048 Y=-0.28+1.112 x WhilestockpricesmirrorEPSincreasesintheshort-run,itisthesizeandstability ofearningsthatwillhelpmaintainthosegains.Theconsiderableadjustmentforthe amountpershareandthenumberofsharesoutstandingmakesacomparisonbetweenthe twofirmsanexerciseinprobability;givenasetofvariables(marketcap,beta,economic outlooketc.)howwillanincreaseinEPStranslateintoanincreaseinmarketprice?The greateramountofinformationusedinEVAcalculationseliminatessomerandom 466 variation,butthevalidityofpredictionisstillbasedonthedynamicsofthemarketplace. Themarketdecideshowitwillvalueeachsecurity,pricingintheriskfromanarrayof economicfactors.Ineffect,itplacesavalueonprospectivegrowth.Duringtherecovery andexpansionphasesofabusinesscycle,somearbitrarystandardwillbeset:themarket willdecidethatgrowthinsomeindustriesismoredesirableandsustainablethaninothers basedonpatternsofdemandandthecostofcapital. AlthoughthesurgeinConocoPhillips’stockmirrorstheperformanceofMicrosoft’s adecadeago,theinvestorneedstolookpasttheimmediatedemandforoilaswellasthe overproductioninsoftwareproducts.Microsoftisinfinitelymoreflexiblethan ConocoPhillips.Liketheauto-makersinDetroit,oilcompaniesaretiedtotheproduction process.Achange-overtoalternativefuelswouldrequireextensiverisk-takingand investmentthatwouldnotimmediatelytranslateintoprofits.Ontheotherhand,Microsoft caninvestinbothnewsoftwareandthephysicalworlditrepresents-innovative applicationsmultiplythedemandforevenmoresoftware.Consider,forexample,theneed fordiagnosticsoftwarethatiscontingentonthedemandforanewoperatingsystemor- chartmakingsoftwarethatcantieintoanExcelspreadsheetThus,thefutureprospects forthesefirmsmaynotbeanextensionoftheirpresenteconomicrealities. Giventheexigenciesoftheirrespectiveindustries,eachfirmiswellmanagedand mademovestooptimizetheircapitalstructures.Inthisregard,theanalystneedstousethe EVA/capitaldynamicasasinglearbiter;withoutanyfinancialleverage,Microsoft’s optimalstructurewilldependonthequalityofitsoperatingleverageandthemanagement ofitsequity.Moreover,useofacomparativemarginalbenefitsfunctioniseliminatedfrom theanalysisbecauseMicrosofttakesnotaxbenefits.Formyriadreasons,theirtarget structurecontainszerodebt,butrequiresthemtorestricttheapplicationofretained earningstoperiodswhenthecostofequityisrelativelylow.Sincethatconstraintis tantamounttocapitalrationing,Microsoftmayfacemorevolatilityandadilemmanot encounteredbycompanieswithfinancialleverage:itcandoallprojectswithapositivenet 467 presentvalue-evenduringperiodswhenthecostofequityishigh-andfacediminished returns,oritcanrestrictprojectstoperiodswhenthecostofequityislow,andeffectively “ration”capital.Ineitherscenario,Microsoftwillmoveawayfromitsoptimalcapital structureandEVAwillbediminished. AMARKETDISCONNECTANDEVENTUALRECONCILIATION Table18-7 CONOCOPHILLIPS YEAR 2005 2006 PERCENTAGE GAIN PRICE 116.36(split) 71.95 18.71 EVA 9933 8732 -12.09 Table18-8 MICROSOFT YEAR 2005 2006 PERCENTAGE GAIN PRICE 24.18 22.98 -4.96 EVA 6889 8544 24.02 Adirectcomparisonbetweenbothcompanies’EVAsandstockpricesdisplaysa marketreactionthatisdiametricallyopposedtocapitalstructuretheory.While ConocoPhillips’EVAwentdownto8732from9933,a12.9percentdecrease,theirstock wentupanothernineteenpercent.Inthemeantime,Microsoft’sEVAwentupabout twenty-fourpercentto8544,buttheirfiscalyearstockpricedecreasedfrom$24.18to $22.98,adeclineof4.96percent.Normally,EVAisaconcurrentindicatorofstockprice, butinthesecases,twounusualscenarioswereencounteredthatcreatedadelay.Inthe firstscenario,ConocoPhillipsboughtBurlingtonResources,increasingthesizeofthe companybyalmosttwothirds;speculationabouttheacquisitioninaneconomic 468 environmentofrisingoilpriceswouldmomentarilytrumpEVA.Inthesecondscenario, Microsoftgaveinvestorsaspecialdividendof$3.40ashare.Whennetincomeroseonly 2.8percentin2006,investorswereblindedtothefactthatMicrosoftwasdecreasingequity andbuyingbackshares.The36billiondollarpay-outcompensatedforaslightlyreduced sharepricein2005.However,in2006,thereaderwillnoticethatMicrosofteffectively reduceditsbetato1.1from1.3.Thecombinedforceofalowercostofequitypercentage, coupledwiththealreadyreducedstockholdersequityproducedahigherEVA. Additionally,the”expected”rateofreturnfromretentionandROEroseto23.07percent froma2005figureof18.7percent(modifiedforaspecialdividend).Whencomparedto thelower“required”rateofreturnderivedfromtheCAPM,Microsoftseemedlikeagreat bargain;anytimegrowthrisesatthesametimethattheCAPMpercentagedeclines, upwardpressureisplacedonthestock.Andyet-fiscalyear2006closedwithaslight decrease.Theanswertothisconundrumisthatanalystswhoonlyfollowedearningswere unimpressedbythepaltryincreaseinnetincome.Nevertheless,thestockbeganclimbing inadelayedreactionthatisaprimeillustrationofastockthatisoutofequilibriumwith themarket.Themovementtoamoreoptimalcapitalstructurewascoupledwithbetter earningsprospectswiththereleaseofthenew“Vista”operatingsystem;thestocksoared, andgainednearlyfiftypercentbeforethemarketpeakedin2007. AsevidencedinthepreviouschapteronConocoPhillips,thelargeequityissuethat helpedpayforBurlingtonResourcesdiminishedEVA.Fromacapitaldynamic perspective,toomuchequityattoohighapriceproducedmovementawayfromanoptimal capitalstructure.However,by2007,ConocoPhillipswasalreadybuyingbackshares.The nineteenpercentstocksurgethatoccurredinspiteofthedeclineinEVAwasatriple-play combinationofmomentum,speculation,andlow-riskleverage;therewassimplynobetter placeforinvestorstobethan“BigOil”mid-waythroughthedecade. Thecomparisonbetweenthetwofirmsillustratesthreeimportantprinciples: 469 • 1.AlthoughEVAismostlyanindicatorthatisconcurrentwithstockprices,anEVA thatrisesonthebasisofalowercostofequitymaybeapreludetoariseinthestock sincepricesmoreimmediatelyrespondtoearnings. • 2.Wellmanaged,“trusted”companieslikeConocoPhillipscantemporarilymoveaway fromtheiroptimaltargetstructuresandmakeupthedeficitwiththe”“promise”of prospectiveearnings.Thenineteenpercentpricegainmadein2006wasnotnearlyas largeasthethirty-ninepercentgainin2005,butstillbeatthemarketbyawidemargin. • 3.LikeEPS,theEVA/capitaldynamicmustbeassignedariskpremiumbythemarket; anygainsmustbecomparedtotheindustryoutlookandtheperformanceofafirm’s peers. INDUSTRYCOMPETITION:CHEVRONANDTHECOMPARATIVECAPITAL DYNAMIC(CCD) Excludingfirmsinthefinancialsector(insurance,banks,brokeragesetc.)thatare capitalizeddifferentlyfrommanufacturingandservicecompanies,wecanexamine differentfirmsindifferentindustriesusingthecomparativecapitaldynamic(CCD). UnlikeEVA,theCCDisameasureofmagnitude;itisaratiothatmeasuresrelativesize, ratherthanabsolutesize.Wemerelydividenetincomebythetotalcostofequity(Net Income/TotalCostofEquity);implicitintheterm“total”istheproductoftheCAPM percentageandstockholders’equity. Asariskbasedindicator,theCCDispotentiallymoreusefulthanP/Ebecauseit encompassesmoreinformation.Bybringingintheconceptofan“opportunitycost”inthe formofthepercentagecostofequity,itcomparesamarketbenchmarktothefundamental earningpowerofthecompany.UnlikeP/Ewhichisvulnerabletospeculativeexcess,the CCDisgroundedinthemechanicsofcapitalstructure.SimilartotheP/E,however,the CCDcanbeusedtocomparecompaniesbetweenindustriesandwithinsectors.For example,ifourcurrentresearchshowsonlyahandfulofcompanyhaveaCCDover3.5, thenwewouldperceivethatafirm’sriskincreasesatthatpoint.Ineffect,thefirmwould 470 “retrench”beforebuildingahigherEVAagain.AnotheranalogytoP/Emechanics,are thedynamicsofanincrease.AhigherEVAcoupledwithalowerCCDsignifiesadifferent relationshipthanifbothincreasetogether.Justaslowerearningsorahigherprice increasesP/E,theCCDcanbeincreasedwithalowercostofequity,ahighernetincome, orsomecombinationofboth.ThefollowingdataforChevronwillhelpillustratethemany usesoftheCCD. Table18-9 CHEVRON YEAR 2005 2006 Sales 198200 210118 OperatingIncome 25679 32427 Assets 125833 132628 NetIncome 14099 17138 Stockholders'Equity 62676 68935 CAPM% 7.11 8.41 Growth% 16.31 18.4 SharesOutstanding 2155.81 2197.17 DividendsPaid 3876 4456 Table18-10 CHEVRON YEAR 2005 2006 PERCENTAGE CHANGE STOCKPRICE 56.77 73.53 29.52 EVA 9643 11341 17.6 471 Table18-11 CHEVRON YEAR ALPHA BETA RSQUARED REGRESSION 2005 0.980456 0.563904 0.17575 Y=0.98+0.564 x 2006 1.050042 0.721824 0.213414 Y=1.05+0.722 x Chevronisawellknown,equallysizedcompetitortoConocoPhillips.Itfollowed the“sectorimperative”,paringdownlong-termdebttocapitalandexpandingwith retainedearningsbeforeithadatwenty-ninepercentgaininthestockin2006.Like ConocoPhillips,therunupinstockpricehadbeensorapid,thatitsplititsstock-in2004. IncontrasttoConocoPhillips,itexpandeditsEVAin2006to11341.The“expected”rate ofreturn,surpassedthe“required”rateofreturn,derivedfromtheCAPM,by approximatelyninepercent,whichimpliedthatthestockhadupwardpressureonit. AlthoughChevronhadahigher“alpha”intheirregressionthandidConocoPhillips,they werealsomoremarketdependentwithahigherRsquared.Howeverbothoilcompanies hadacomparablebetawhichwouldbeexpectedforfirmswithsimilarcapitalstructuresin thesameindustry.TheEVAdelineated“riskpremium”,thenaturalcorrelationofstock pricetofundamental,wasnothigh.WhileChevronpushedtheirEVAupapproximately seventeenpercent,theirstockroseonlyslightlyaboveConocoPhillips’,whohadactually decreasedEVA.Again,theimperativewastobegrowingwithinthesector;ConocoPhillips wasperceivedbyinvestorsasacompanywhowasrapidlyimprovingmarketshare-inthe “right”market. Thecomparativecapitaldynamicswereverysimilarforbothoilcompanies.Each rosefromaround“2”toover“3”andthenbackagain.LikeP/E,theCCDseemstostay withinparticularboundariesforeachmarket.Ifthemarketisrapidlydeteriorating,the CCDshoulddeclineaswell:.Netincomewoulddecrease,andthetotalcostofequitywould 472 beoverlyhigh-notonlybecausetheCAPMpercentageincreased,butbecausethe formerlyhighnetincomeattractedmoreinterestinequityissues.Inthefollowingdata,a sidebysidecomparisonismadewithMicrosoftincluded.Noticehowthemarketseemedto rewardeachcompanyforreachinga“3”inCCD. Table18-12 CHEVRON YEAR NetIncome TotalCST. Eqty EVA %Change EVA CCD 2005 14099 4456 9643 3.16 2006 17138 5797 11341 17.6 2.956 Table18-13 MICROSOFT YEAR NetIncome TotalCST. Eqty. EVA %Change EVA CCD 2005 12254 5365 6889 2.28 2006 12599 4055 8544 24.02 3.107 Table18-14 CONOCO- PHILLIPS NetIncome TotalCST. Eqty. EVA %Change EVA CCD 2005 13529 3596 9933 3.76 2006 15550 6818 8732 -12.09 2.28 BymaintainingarelativelyhighCCDofnearlythree,ChevronincreasedEVAandstock pricesimultaneously.Aspreviouslymentioned,Microsoftencounteredadelayedreaction whentheysurpassed“3”,notseeingasubsequentriseinthestockuntilthefiscalyearwas overandthefinaltotalsweretallied.AndwhenConocoPhillipshadaCCDof3.76in2005, theyhadtheirbestyeareverwithathirty-ninepercentgaininpriceaswellasastocksplit. 473 Inthesomewhatobscureterritoryoccupiedbyfinancialindicators,theCCDstakes outapositionbetweenP/EandROE.Itisrelativetobothfundamentalsandthemarket, butitneedstobegaugedbybothitsownperformanceandcommonstandardsforthe industry.Arestaurantchain,forexample,willhaveadifferentoptimalCCDthanasteel company.AnindustrywithahistoricallysmallROEwillhaveamuchlowerCCDthan high-ROEindustrieslikesoftware.Theemphasiswouldagainbeontherisk/return differenceofmean-variance,maintainingthehighestpossibleCCDwiththelowest standarddeviation.Anycomparisonbetweenfirmsinanindustryshouldrevealthatthe competitorwiththehighestCCDisoptimizingcapitalstructurebetterthanothers,and maximizingthepriceofthestock. PERCENTAGEOFNEWRETAINEDEARNINGS IfthecomparativecapitaldynamicisdecliningandEVAisincreasing,therateof growthofthecostofequitysurpassesthatofearnings.Thatscenarioisanalogoustoafirm thatmaintainsitsstockpricedespitelowerearnings;theeffectiveresultisahigherP/Eand morerisk.Whenwedecomposestockholders’equityintoitsconstituentparts,wewillfind oldretainedearnings,recentretainedearnings,oldpaidincapital,recentstockissuesand preferredstock.Eitherweeliminatepreferredstockaltogetherandusetheaggregate figureforcommonequity,orweintegratepreferredstockintoourcalculationbyweighting itwithitspercentageandcost.Thelatterapproachismoreconservativeandexacting,and emphasizestheweightedcostapproachtothecostofequity.Infact,itwetreatequitylike theweightedaveragecostofcapital(WACC),wecanusethepercentchangeinnew retainedearningsasanindicatorofpotentialperformance.Notonlywillmoreretained earningsimplythatnetincomehasincreased,buttheycanpotentiallypropelthefirminto aleveragestateinwhichriskisrapidlyreducedbyloweringbeta.Thus,itwouldbe helpfultocalculatenewretainedearningsasapercentageofthetotalcostofcapital.In thisillustration,weassumethatpreferredstockiszero. 474 Thecalculationis((1-(Dividendspaid/NetIncome))x(NetIncome))x(Percentage CostofEquity)/TotalCostofEquitywhichwillsimplifyintoNewRetainedEarnings/ Stockholder’Equity.Therationalebehindviewingnewretainedearningsinapositive lightisthattheyareimplicitinnetincomeandwillnotdecreasethatsideofthefunction. Thus,forexample,ifnewretainedearningsweresubstitutedforastockissue,therewould beanetgaininEVA,becausenetincomewouldincrease,andthenonlyapercentageof retainedearningswouldbesubtracted.Paidincapital(astockissue)wouldnotincrease netincomeandwouldfullydetractfromEVA. ThefollowingexamplewillemphasizethecorrelationwithEVA.Inthechapteron Kimberly-Clark,wesawthatmoreretainedearningscandetractfromEVAwhenthecost ofequitypercentageishigh(suchasattheendofabusinesscycle).However,eveninthat extremecase,thedecreaseinEVAcanbeconsideredthe“leastharmfulpossible”because capitalisineffectbeingraisedwithoutanyaccountingcosts.Alternativesourceslike equityissuesordebtwouldbeevenmorecostly-ineitherinterestpaymentsorflotation costsanddilution. Table18-15 CONOCOPHILLIPS YEAR 2005 2006 DividendsPaid 1639 2277 NetIncome 13529 15550 RetainedEarnings 11890 13273 Stockholders'Equity 52731 82646 RetainedEarnings/ Equity 22.55% 16.06% CCD 3.76 2.28 EVA 9933 8732 (BacktoTableofContents) 475 SECTIONIV:CORRELATIONANDPROBABILITYSTUDIES 19 OPERATINGINCOMECORRELATIONSTUDIES Operatingincomeisthelinchpinofcapitalstructure.Aswesawinourobservations ofFed-ExandStaples,ithasbothriskandreturncomponentsthatsometimesoffseteach other.Ouranalysiswasfullofcountervailingbalances,andwecametotheconclusionthat theincomestreamfromFed-ExwasmoredesirablethanthatfromStaples-althoughit entailedmorerisk.Thevalidityoftheargumentwasbasedonthesizeofthereturn,but alsoonanimportantqualitativefactor-Fed-ExhadpricingpowerthatStapleslacked. Rarelyshouldriskanalysisbepursuedinisolation.Operatingincomeinterfaces withsales,thedemandvariable,butitalsoaffectscapitalstructurebydeterminingthe source,thecostandthetypeoffunding.Infact,operatingincomeissocrucialtocapital budgeting,thatratingsagencieslikeStandardandPoor’sandMoody’suseitasaprime componenttodeterminecreditratings: (NEXTPAGETABLE) 476 Table19-1 RATIO(Percentages whereAppropriate) AAA BB PretaxInterestCoverage 16.66 1.59 PretaxFixedCharge Coverage 6.39 1.33 FundsFlowInterest Coverage 22.98 2.75 Fundsfrom Operations/TotalDebt 134.7 17.9 FreeOperatingCash Flow/TotalDebt 49.2 0.9 PretaxReturnonTotal Capital 24.2 9.1 OperatingIncometoSales 22.5 10.6 long-termDebt/Capital 11.7 51.1 TotalDebt/Capital 23.2 56.3 Therefore,ifCompanyXwantstofloatabondissueandreceivean“investment quality”ratingofAAAinthegivenyear,theywouldneedtocoverinterestbyanaverage of16.66timesintheprecedingthreeyears.Notethatsixofthesetopninekeyratios containsomeadaptationofoperatingincomeasamaincomponent. Thefollowingstudyisapragmaticexplorationoftherelationshipbetween operatingincomeincreasesandlong-termdebt.Originallyculledtogetherasone investor’sattempttocapitalizeonabullmarket,whatitlacksinscientificlegitimacy,it makesupintimelyeffectiveness.However,thestudentshouldbeawareofmethodologies bywhichthestudycouldhavebeenimproved.Attheendofthechaptersomekey statisticalconceptswillbediscussedtobothcriticizethisstudyandproposealternative hypotheses. NAME:ComparisonEffectsofChangesinOperatingIncomeandLong-termDebton NextYear’sMid-rangeStockPrices PREMISES:1:Inaneconomicenvironmentofbothrisingmarketsandrisinginterest rates,equitybecomesproportionallylessexpensivethandebt.2:Asoperatingincome 477 rises,companiescanuselessdebtandmoreretainedearningstolowerthecostofcapital. 3:Acombinationofhigheroperatingincomeandlessuseofdebtencouragesafirmto maximizeitsstockprice. DATAPOINTSANDSTRUCTURE TwentyfourcompanieswerepickedfromtheS&P500atrandom,butany duplicateindustriesweredeleted.Bothfinancialinstitutionsandutilitieswereleftoutof thestudy,becausetheirrespectivecapitalstructuresareuniqueandnotrepresentative,i.e., themeasuredoperatingandfinancialriskswouldskewanymeaningfulresult.Sevento eightyearsoffundamentalincreasesweretabulatedforeachcompany,creatingone hundredandeighty-twoseparatedatapoints.Whilethestudywasconductedduringa periodofrelativeeconomicprosperity(1990-1999),nomacro-variablessuchasinterest ratechangesorriskpremiumspreadswereincluded Uniquetothestudywasthemeasurementofchangesinstockprice.Thesubsequent year’smid-rangepricewascomparedtoseveralfundamentalvariablesthathadbeen segmentedaccordingtowhetheryeartoyearoperatingincomewasincreasingor decreasing.Mid-rangepriceismerelytheaveragebetweenthehighandlowpriceforthat yearwithoutregardfortiming.Consequently,itmeasuresvolatilityasmuchasit measuresperformance,butalsocapturesanyinherenttrendwhenusedcollectively. Companiesthathadnochangeinthelong-termdebttocapitalratioorhadnolong- termdebtatallwerecountedasdatapointsthatdecreasedlong-termdebttocapital. Althoughthese“all-equity”datapointsmightskewresults,itwaslaterdeterminedthat theywereactuallyantagonistictotheresearcher’spremise,loweringtheoverallprice effectsintheirrespectivecategories.Inasimilarmanner,ifoperatingincomedidnot changeatall,itwascountedasadecrease. CATEGORIES Datapointswereseparatedintosixdistinctpopulations,basedonincreasesand decreasesinbothoperatingincomeandlong-termdebttocapital,hereuntoreferredtoas 478 “LTD/CAP”.Thepopulationswereidentifiedas:1)Operatingincomeincreases2) Operatingincomeincreases,LTD/CAPincreases3)Operatingincomeincreases, LTD/CAPdecreases4)Operatingincomedecreases5)Operatingincomedecreases, LTD/CAPdecreases6)Operatingincomedecreases,LTD/CAPincreases. Eachpopulationwasanalyzedintermsofdescriptivestatisticstoenable comparisonsbetweenpopulationgroupsandSpearmanrankcorrelationsthatextracted relationshipswithineachgroup.Sincethesamplesizewaslimited,mediansratherthan meanswereusedtoidentifyaverageswithinthepopulation,andstandarddeviationsof eachfundamentalwerecalculatedtoprovideascopeoftheskewofthedistribution.The Spearmanrankcorrelationsareassociationsmadebyregressingtherankedorderofthe dataagainsteachother.Dataissimplyrankedfromlowesttohighestandastandard linearregressionisperformedontheranks;thisordinalmethodwillproduceassociations thatarecurvilinearandrevealrelationshipsthatmightnotbeapparentinnormalleast squares-typeregressions.However,theprobabilityassociatedwitheachcorrelationmust bereadas“proofofthenullhypothesis”thatnoassociationexists;thatis-thelowerthe probability,themorevalidtheassociation.Allvalueswithaninety-fivepercent significancearereported,butthereadermustrecognizethatthestudyisinvalidatedby smallsamplesizeandismeanttoidentifytrends,notabsolutes.Analogoustolinear regression,thecorrelationvalueisscaledfrom-1to1withgreaterassociationrepresented attheextremes. 479 COMPANIESINTHESAMPLE Table19-2 EnescoGroup(ENC) CeridianCorp.(CEN) BCE,Inc.(BCE) ClaytonHomes(CMH) TenetHealthcare(THC) McKessonHBOC(MBK) CompaqComputer(CPQ) McGraw-Hill(MHP) Halliburton(HAL) Diebold,Inc.(DBD) PhelpsDodgeCorp.(PD) HomeDepot(HD) EmersonElectricEMR) AdvancedMicroDev. (AMD) BlackandDecker(BDK) H&RBlock(HRB) CVSCorp.(CVS) HiltonHotels(HLT) PharmaciaCorp(PHA) NucorCorp.(NUE) GeneralElectric(GE) ElectronicDataSys. (EDS) CendantCorp.(CD) CharlesSchwab(SCH) FUNDAMENTALVARIABLES VariablesderivedfromfinancialstatementsincludefivevariablesthatencompassDuPont -typeanalysisthatwouldgivesomeindicationofeffectsofthereturnonequityorROE. Theseareyearlyincreases/decreasesinoperatingincome,sales,profitmargin,thecapital multiplier(assets/capital),andassetturnover.Sevenvariablesthataffectthecostofcapital arealsoincluded.Theseare:yearlychangesinassets,capital(thesumofequityandlong- termdebt),totaldebt,long-termdebttocapital(LTD/CAP),long-termdebt(LTD),the financialleverageratio(EBIT/EBIT-Interest)andtheinterestrateontotaldebt(interest expense/totaldebt).Eachvariableisanincreaseoradecrease,andwasregressedagainst themid-rangestockpricechangeofthenextyeartoobserveanydiscernibleeffects.Other variablesincludedinthestudyareyearlychangesin:cash,capitalexpenditures,operating momentum,EPS,andthenumberofsharesoutstanding.Descriptivestatisticsareincluded fortheDuPont-typeROEvariablesaswellasforthecostofcapitalvariableswiththe exceptionoftheinterestrateontotaldebtwhichisdeemedbeyondafirm’sdirectcontrol. Eachvariablewasregressedagainstthechangeinnextyear’smid-rangestockpriceand reportedifaprobabilityunder.05(95%confidence)wasencountered.Allyearlychanges arereportedonapercentagebasis.Ifapercentagechangewasjudgedtobeoverlyskewed 480 (suchasachangeinvolvinga10,000percentdecrease),itwaseliminatedatthediscretion oftheresearcher. ALLVARIABLES Table19-3 NextMid-rangeStock Price InterestRate(onTotal debt) CapitalMultiplier (Assets/Capital) OperatingIncome TotalDebt TotalCost(Sales-OpInc.) Sales AssetTurnover OperatingMomentum OutstandingShares ProfitMargin Long-termDebt InterestExpense CapitalExpenditures Long-termDebt/Total Debt Assets Cash TotalLeverage Capital LTD/CAP TotalLeverage-absolute EPS FinancialLeverageRatio OperatingMomentum- Abs Thelastratiosareconcrete,absolutemeasurementsthatindicateascalarnumberrather thanachange.Thesewereincludedtoobservewhetherthesizeoftheseratiosaffected stockprice.Theterm“leverage”in“totalleverage”isderivedfrommultiplyingthe financialleverageratiobyoperatingmomentumandnotindicativeof“true”operating leverage,whichisnotmeasured. 481 STATISTICALRESULTS(PercentageIncreases) INCREASINGOPERATINGINCOME Table19-4 Variable Median StandardDeviation Sales 15.58 25.84 OperatingIncome 21.15 67.6 ProfitMargin 3.67 254.05 CapitalMultiplier -0.0268 20.98 AssetTurnover 1.7 20.11 Mid-rangePrice 16.6277 30.66 Assets 14.25 46.85 Capital 13.64 70.58 LTD/CAP -7.09 740.05 LTD -0.0359 2268.56 TotalDebt 16.5 40.11 FinancialLeverageRatio 0.8413 8.34 INCREASINGOPERATINGINCOME,INCREASINGLTD/CAP Table19-5 Variable Median StandardDeviation Sales 14.39 25.59 OperatingIncome 17.83 51.38 ProfitMargin -2.83 215.48 CapitalMultiplier -1.73 29.7 AssetTurnover -0.57 24.32 Mid-rangePrice 8.56 22.28 Assets 16.22 75.41 Capital 20.06 110.26 LTD/CAP 39.2 1296.78 LTD 69.39 4034.9 TotalDebt 17.033 47.6 FinancialLeverageRatio -0.1613 7.09 482 INCREASINGOPERATINGINCOME,DECREASINGLTD/CAP Table19-6 Variable Median StandardDeviation Sales 16.1 25.99 OperatingIncome 22.49 73.93 ProfitMargin 4.46 267.57 CapitalMultiplier 1.16 15.82 AssetTurnover 2.83 17.44 Mid-rangePrice 21.28 32.62 Assets 13.63 24.92 Capital 12.06 42.46 LTD/CAP -17.17 26.66 LTD -4.14 34.1 TotalDebt 16.5 36.52 FinancialLeverageRatio -1.38 8.72 DECREASINGOPERATINGINCOME Table19-7 Variable Median StandardDeviation Sales -2.58 17.51 OperatingIncome -11.6 20.8 ProfitMargin -22.45 173.25 CapitalMultiplier -1.44 32.5 AssetTurnover -7.68 15.8 Mid-rangePrice 5.47 29.21 Assets 1.34 20.97 Capital 2.96 25.34 LTD/CAP 0 47.99 LTD -1.43 66.29 TotalDebt 1.71 30.87 FinancialLeverageRatio 0.0992 13.8 483 DECREASINGOPERATINGINCOME,INCREASINGLTD/CAP Table19-8 Variable Median StandardDeviation Sales -5.23 19.32 OperatingIncome -14.67 22.53 ProfitMargin -14.37 239.12 CapitalMultiplier -1.097 47.51 AssetTurnover -8.5 17.85 Mid-rangePrice 7.82 35.53 Assets 3.09 26.8 Capital 6.35 34.69 LTD/CAP 55.17 131.68 LTD 29.18 80.43 TotalDebt 5.07 32.32 FinancialLeverageRatio 1.75 16.3 DECREASINGOPERATINGINCOME,DECREASINGLTD/CAP Table19-9 Variable Median StandardDeviation Sales 0.0192 16.27 OperatingIncome -11 19.87 ProfitMargin -29.78 113.53 CapitalMultiplier -1.52 14.24 AssetTurnover -2.35 14.48 Mid-rangePrice 2.78 24.41 Assets 1.31 16.38 Capital 2.55 16.76 LTD/CAP -8.02 25.78 LTD -7.37 27.82 TotalDebt 1.61 30.12 FinancialLeverageRatio 0 11.05 484 SPEARMANRANKCORRELATIONS:NEXTYEAR’SMID-RANGEPRICE INCREASINGOPERATINGINCOME(N=130) Table19-10 Variable Correlation Probability Assets 0.172891 0.049177 LTD/CAP -0.238255 0.006339 FinancialLeverageRatio -0.211257 0.015833 LTD/TotalDebt -0.187093 0.033053 INCREASINGOPERATINGINCOME,INCREASINGLTD/CAP(N=40) NOSIGNIFICANTCORRELATIONS INCREASINGOPERATINGINCOME,DECREASINGLTD/CAP(N=90) Table19-11 Variable Correlation Probability Assets 0.282562 0.006968 Capital 0.261687 0.012722 FinancialLeverageRatio -0.242514 0.021278 DECREASINGOPERATINGINCOME(N=52) Table19-12 Variable Correlation Probability InterestRateonTotal Debt -0.351063 0.010717 OperatingMomentum (Abs.) 0.340562 0.013494 TotalLeverage(Abs.) 0.338513 0.014102 485 DECREASINGOPERATINGINCOME,INCREASINGLTD/CAP(N=21) Table19-13 Variable Correlation Probability InterestRateonTotal debt -0.484416 0.026056 OperatingMomentum (Abs) 0.57013 0.006966 TotalLeverage(Abs) 0.554545 0.009083 OperatingMomentum% 0.616883 0.002895 TotalLeverage% 0.631169 0.002153 DECREASINGOPERATINGINCOME,DECREASINGLTD/CAP(N=31) NOSIGNIFICANTCORRELATIONS INTERPRETATION Thereareseveralcaveatstoobserveinthisstudy.Itisnottheintentionofthe researcherto“implicate”long-termdebtasthemainculpritinprofitabilityloss.Debt issuesarenotdestructiveinandofthemselvesunlesstheyarebeingsubstitutedforless costlysourcesofcapital.Sincecorrelationshowsassociationandnotcausation,itis importanttolookforundocumentedvariables(likerisinginterestrates)thatoccuroutside thedomainofthestudy.Probabilityandeconomicsmakestrangebedfellowsifonly becauseoftheexistenceofsomanyothervariablesthatcanrefuteahypothesis.For example,manydebtissuesgotofundlargeacquisitionsthatmustbeintegratedintothe corporatestructureinsubsequentyears.Ifsuchintegrationprovesrisky,slow,and ultimatelyunprofitable,thetypeoffundingcannotbeheldliable;foresightwaslackingon thepartofmanagement.Ifthissituationoccursregularly,isitthefaultoflong-termdebt, ordoesitstemfromthesizeofthecapitaloutlayprecludingtheuseofequity?Attimes,a largeissueofstockcandodamagetosharepricemerelyonthebasisofsupplyand demand,andretainedearningswouldnotusuallybelargeenoughtocoverthedeficit. 486 Moreover,ourstatisticscanbemisleadingbecausetheyoccurredduringaperiodofrising interestratesandrelativeprosperitywhichwouldhavecreatedapropensityforequity- typefinancing.Inthehighlyinflatedmarketofthe1990s,companieswouldhavedesired toexercisestockoptionsandissuenewequitybecausetheycouldhavedonesowithout payingthepenaltyfordilutingtheshares;manysuchissuesrose“automatically”.Thus, ourpremisesarelargelyunprovedandourresultsareinconclusive,althoughthestudy raisedmanyquestionsandpointedusinthe“rightdirection”. Thedataitselfdisplaysaclassicargumentforminimizingthecostofcapital.The increasingoperatingincome,decreasingLTD/CAPcategoryinparticularexhibitedamid- rangestockpricedifferenceofasmuchaseighteenpercent.Additionally,theDuPont-type fundamentalsthatcompriseROEaremuchhigherforthatcategory.Whileweareunable todrawfirmconclusionsfromthedata,wecanuseseveralknownpropositionstosupport it: • Long-termdebtobligatesacompanytopayinginterestexpense. • PayinglowerinterestexpenseautomaticallyraisesEPS,allothervariables beingequal • Raisingtheproportionoflong-termdebttoequitywillincreasethecostofthat sameequity • Inanormalmarket,long-termdebtwillbemoreexpensivethanshort-term debtbecauseinvestorsneedtobecompensatedfortheinflationriskofalonger maturity. • Moreoperatingincomeandlessinterestexpensecanleadacompanytodo moreequityfinancingthroughretainedearnings. • Greateroperatingincomehasthepotentialtobedistributedasahigher dividend,compensatingshareholdersandraisingthestockprice. • Operatingincomeincreasescanleadto“momentum”,whichisanacceleration ofearningsafteraninitialincrease. 487 Statedasfact,thelastpropositionisquestionable.TheresearcherJosephMurphy, whostudiedthreehundredandforty-fourcompaniesintwelvedifferentindustriesfound nodiscerniblecorrelationbetweenyeartoyearearningsdata.Ontheotherhand,Manown KisorandVanMessnerfoundapatternthattheytermed“increasedmomentuminrelative earnings”whichiscomposedofanacceleratedrateofchangeofearningsincomparisonto theperformanceoftheentiremarket.FurthersubstantiatedbybothRobertLevyand RobertHagin,thisphenomenonisespeciallyapparentduringrisingmarkets.Itmerely encompassesasixmonthincreaseinearningsonceearningsacceleratesfasterthanthe marketintheprevioussixmonths-anditisessentiallyprobabilistic,notdeterministicIn thisstudy,anautocorrelationofoneyear’searningsregressedagainstanotheryear’s provednocorrelationbetweenthetwosetsofdata. Besidesrejectingthenullhypothesisfortheincreasingoperatingincome,decreasing LTD/CAPdata,afewothercomparisonswerepotentiallysignificant.Ifonecomparesthe datasetsofpureoperatingincomeincreasestopureoperatingincomedecreases,onewill findalargedifferenceinthereturnonequityfactorslikesalesandprofitmarginaswellas substantiallydifferentcapitaloutlays;13.64%vs.2.96%.Observingthelong-termdebt increaseinbothdatasetsrevealsthatthereislittlesignificantdifference;bothdecreased long-termdebtbyaboutonepercent.Evidently,morelong-termdebtdoesnotdrivethe decreaseinoperatingincomeortheconsequentROEfactors,butmayriseproportionally becausetherearelessretainedearningstobufferitinthecapitalstructure.Anydecrease inearningsisderivedfromaninefficientuseofcapital,andactualizedbyacombinationof lesssalesandhigherproductioncosts.Itseemsthatwhileloweringcapitalcostswill furtherenhanceanalreadyprofitablecompany,itwillnotpropeladistressedcompany into“theblack”. Withfurtherdataseparation,wefindthatcompanieswhoincreasedbothoperating incomeandLTD/CAPhadadequatesales,earnings,andcapital,butperformedlesswell thancompanieswhoincreasedoperatingincomeanddecreasedLTD/CAP.Onefinds 488 similarincreasesinassets,andyettheROEfactorsfortheincreasedLTD/CAPset significantlyunderperform.Ifmorelong-termdebtdoesnot“drain”acompanyofprofit, thensomeotherfactormustbeworkingtolimitperformance.Thedatadoesnotrevealthe peculiaritiesofacquisitions.SinceS&P500companiesneedsubstantialcapitaltogrow, retainedearningsandcashproveaninadequatesourceoffundswhenpurchasesare literallyinthebillionsofdollars.Suchlargepurchasesarenormallymadewith combinationsofequityexchangesandbondissues,alongwithcash.Acquisitionsoftarget companiesneedtimetobeintegratedintotheacquiringfirm’sassetstructurewhichputsa strainonexistingresourcesandmayleadtosomeunderperformance.The“synergistic” effectsofanacquisitionarenotautomatic.Infact,itistypicalthatspeculativecashgoes intoanacquiringfirm’sstockifthepurchaseisconsidereda“smart”move,onlytobe removedinthenextyearbecausethecapitalturnoverwasnotrapidenough.Therefore, thequalityofassetsmustbeexamined.Alargeamountofcashonthebalancesheetisfar differentfromtheintegrationofacompetitorintoone’sassetstructure.Sincethisstudy waslimitedtonumericaldifferencesandnotqualitativeones,wecanonlyspeculateonthe natureofthereasonbehindthem. METHODOLOGICALCRITICISM Althoughastudyofthismagnitudecanposevaluablequestionsandgive researchersleadsthatmayyieldsomeanswers,itlacksstatisticalvaliditybecauseitfailsto proveordisproveapremise.Unfortunately,financialdataisheavilyskewedasevidenced bythelargestandarddeviationsinthedescriptivestatisticssection.Suchdistributions requireamuchlargersamplesizeandamuchnarrowerfocus-suchasonlyonesegment ofoneindustry-tobevalid.Whileseparatingthepopulationintouniquecategorieshelpsto accentuatethedifferences,thesamplecomprisedonlylargewell-knowncompaniesleaving outsmallcapsandprivatebusinesses.Howdoweknowwhetherdifferenceswereobtained onthebasisofsizeorcapitalization?Researcherscompensateforthisbiasbysettingupa “control”groupandcomparingthepopulations-andyetthattoomayleadtodifficulties 489 becausecompaniesmayprovetobedissimilar.Unlike“twin”studiesinthebiological sciences,fewfactorsarecontrollableintheeconomicenvironment. Characteristicsofmeasurementwerealsoproblematic.Sinceseventoeightyearsof dataareextractedfromeachcompany,thedatamaybeseriallycorrelated:Acompany mayhaveastrategyoveraperiodoftimethatskewsthedatainonedirection.For example,ifonesellshomesatthebeginningofanexpansion,thatsectormaybeprofitable forseveralyearsinarow,biasingsalesandincomeinanupwardsdirection.Tomakesure themeasurementsarenotbiased,itwouldbebettertohaveasmanycompaniesaswehave datapointsinthesample.Analogously,theperiodinwhichthesampleislocatedmaybias thestudy.Anyrecommendationsshouldbederivedfromafewbusinesscycleswhichmay haveaprofoundeffectontheresults-especiallywhentheycompriseincomeanddebt. Lastly,percentagechangesareinferiortologarithmicchangesbecausethebaseofthe percentagewillchangeandskewtheresults.Forexample,anickelchangeonanEPSof tencentsisafiftypercentimprovement,whichisthesamemeasurementasanEPSthat goesfromfourdollarstosixdollars.TheSpearmanrankcorrelationswouldhavefurther emphasizedthisanomalybecausetheyareordinalbasedmeasuresofmagnitudewithno referencetoastablebase. Inthepursuitofmoreinformation,bothstudentsandinvestorsshouldbe encouragedtoexperiment.Whilestatisticallyvalidstudiessometimeshavetoonarrowa focustobeofinteresttothelayperson,farreaching,buttechnicallyincorrectstudiescan proposehypothesesthatmayhavelaterlegitimacy.Althoughitisnotourobjectiveto defend“badstatistics”,ideasthat“breaktherules”mayoffersolutionswhenoldmethods becomestagnant.Sincemosteconomichypothesesmustadapttoachangingmarket,they oftenloselegitimacywhenrepeatedlytestedagainstreality.Thehypothesiswasnotwrong atthetimeofformation,butwasreplacedbyanewdevelopment.Forexample,our hypothesisaboutdebtandtaxdeductionswouldbeobsoleteifthenationwenttoanational salestaxandawayfromanincometax.Companieswithsteadyincomes,whocanmakethe 490 bestuseofdebtfunding,wouldbepenalized.Gradually,somenewsocialoreconomic constructwoulddeveloptoreplaceit.Inessence,scholarshipthatproducespractical, usefulideasshouldbeencouraged,andinfinance,thatentailsobservingallpossibilities. Fourtenableexplanationsforourdatacanbeproduced: 1.Thestudywasflawedandnoideaswerelegitimate. 2.Companiesthatproportionedtheircapitalstructurescountercyclically tothechangesininterestratesprofitedmorethanthosewhodidnot. 3.Arisingmarketmadeequityfundingmorefeasibleandprofitable becausemostcompanies’stockwasinhighdemand. 4.Largedebtissueswillraisethecostofcapitalandbeunprofitable inthenear-term. Thefirstandlasthypothesescanberejectedonahistoricalbasis.Many researchershavefoundthatcapitalstructureisfirmlyintegratedwithtaxandinterest rates,stockpricesandprofitabilityofthecompany.Whenmoredebtisincurredatalower rate,capitalcostsareoftenminimizedbecausethereislessofarisktoequityandthe inherentcostofequityremainsstable.Consequently,thereisnorealcorrelationbetween typeofcapitalstructureandstockpricebecausethemarketdetermineswhichstructure willbefavored.Whentheriskpremiumbetweendebtandequitybecomesverylow, investorssellstockandbuybonds.Whentheriskpremiumishigh,investorsaremore rewardedbyequities.Sincetheinterfacebetweenthemarketandthecompanyisthe interestrate,capitalcostsaredeterminedbytheeffectsofinterestratechangesontheasset structureofthecompany.Companieswithlowoperatingriskwillbeinabetterpositionto takeadvantageoflowratesandincurmoredebt,whilethosewithmoreoperatingrisk mustfinancethroughretainedearningsorequityissues.Whenratesrise,astheydoin prosperoustimestocombatinflation,thosecompanieswhoneedlessdebt,orwholower theirlong-termdebttocapitalratios,willberewarded. (BacktoTableofContents) 491 20 CHANGESINCAPITALSTRUCTUREANDTHEIREFFECTON STOCKPRICES Thischapterexaminesthesamedatathatfueledtheoperatingincomestudyinthe lastchapter.Itattemptstoestablishahypotheticalrelationshipbetweenchangesinstock pricesandchangesincapital.Whileittendstoconfirmconventionalwisdomonthe subject,italsoopensupavenuesofdiscussionintwootherareas:1)whetheralarge capitaloutlayinoneyearcreatesdiminishedcapitalfundinginthenext(capitalrationing), and2)whetherfinancialleverageratioshaveanypredictivevalueconsideringthatmore riskmayimplyahighercostofcapitalinthenearfuture.Asstatedpreviously,the statisticalviabilityofsuchastudycanbequestionedbecauseofsmallsamplesizeandserial correlation,butthepracticalvalueistoobservethebehaviorofcapitalstructurevariables overanentirebusinesscycle. THEVALIDITYOFLEVERAGEFACTORS Ifindeedfinancialleverageratioshavenopredictivevalue,itisbecausethemarket issupposedtofactorinsuchinformationinan“efficient”manner.Likeearnings momentum,theremaybebriefperiodswherethereare“arbitrage”opportunitieswhen inefficienciesoccur;atthesebriefjunctures,theinvestorcanuse“stockscreens”, momentum,orfinancialleverageinformationtomakeextraordinaryprofits.However,the marketchangesandthefactorsthatcreatedsuchopportunitieswillnolongerbeoperable. But-leveragefactorsareaninherentpartoftheeconomicsystemanddefychange;return riseswithrisk,butifriskbecomestooexpensive,asitmayduringaninterestratehike, thenreturnisdiminished.Thus,therearesystemicfactorsthatvalidatethepossibility. Moreover,onecanarguethateconomicfactorslikeinflationbehaveinapredictive,albeit unstablefashion..Inflationmaydisappearforyearsonlytocomebackrampantly,andit isalaggingindicator,notaforwardlookingone.Thecasewemakeforleveragefactorsis thattheyarespecificallyactionable,andaffectprofitsandassets.Thereisalwayssome 492 combinationofleveragefactorsthatwillyieldthehighestamountofreturnwiththeleast amountofriskinanymarket.Ontheotherhand,theresponsetoeconomicfactorslike inflationisreactive,anddifficulttodevelopintoacoherentstrategy;wemaychoosean assetclasslikegoldorrealestatetohedgeagainstinflation,butwillneedtomoveaway fromitoncetheprospectofhigherpricesdiminishes. Onereasonthatthebullmarketofthe1990swasagoodlaboratoryforstudying capitalstructurewasthestabilityofearnings;atwotermpresident,apeacedividend,the endoftheColdWar,andlowinflationallcontributedtostockmarketincreases.Excess speculationinthelatterthirdofthedecadehadaskewingeffectontheresultsofthisstudy, butthesamplepopulationismadeupoffirmswithprimarilylargecapitalizations; absolutesizetendstobuffersomevolatility.Whilethereisalwaysthetendencytointerpret correlationas“causeandeffect”insteadofmereassociation,ourobjectiveinthischapter isnottoproveatheory,buttogetthereadertothinkintermsofpossibilities-and-even ifheorshesochooses,waystorefutethisresearch. CONNECTINGTHEDOTS:EARNINGSANDDIVIDENDGROWTHANDTHECOST OFEQUITY Throughoutthistext,wehaveemphasizedourbeliefthatstockappreciationisa functionofearningsacceleration.Whenonefirm’searningsrisesfasterthanotherfirms’ inthemarket,thatfirm’sstockwilloutperformthoseofsimilarrisk.Secondly,wecan inferthatearningsarerisingfasterthanthecostofcapital(specificallythecostofequity) whenthissituationoccurs.Thirdly,wehavestatedthattheenvironmentforsuchearnings increasesiscreatedbyanoptimalcapitalstructure.Thetheoreticalunderpinningsforthis scenariowouldbethechangeineachcomponentpartofthecomparativedynamic,%∆ ∆∆ ∆ NetIncome/%∆ ∆∆ ∆CostofEquity.Infactahypotheticalcomparativedynamiccanbe constructedfortheentireeconomywith%∆ ∆∆ ∆GDP/%∆ ∆∆ ∆RiskFreeRate(10YearBond) Inessence,thegainfrominvestmentmustoutpaceitscost.Wheninterestratesriseand stocksascendtoorapidlyincomparisontorisk,thecostofequityoutpacesearningsand 493 thestockmarketwithers.Therelationshipbetweencomponentsofthecapitalassetpricing model(CAPM)isestablishedbythisbalancebetweenearningsandthecostofcapital. Analternativetothis“topdown”theoryistomicro-managecapitalstructure throughthedividend.Thosewhoespousedividendtheorybelievethatallstockvaluationis inherentindividendpaymentsandthattheproportionofequityincapitalstructure dependsonthemanagementofretainedearnings.Theresidualtheoryofdividends proclaimsthatunlessprojectedearningsmeetaspecificlevelofreturn,theyshouldbepaid outtoshareholders.Inthiscase,theopportunitycostofretainingearningsisgreaterthan payingthemout,asmeasuredbythecostofequity.Thattenethasmademany corporationsreluctanttocutdividendsintheworstoftimes,becauseofthenegative informationthatsuchamovesignals.Notonlywillsuchacompanyfailtocompensate shareholders,butitwillneedtofinanceinthecreditmarketsforlackofsufficientretained earnings.Thistypeoffundingraisesthecostofcapitalevenasearningsaredepleted,and theresultisadiminishedstockprice.Thus,thedynamicsareamicrocosmofthe“top downapproach”;thechangeinearningsisdecreasedwhilethechangeincapitalcostsis increased. TheGordonmodeloffersanestimateofthecostofequity,butissubjectto distortionindetermininganadequategrowthrate.Nevertheless,thedynamicsbehindthe modelareinformative,becausetheydirectlyrelatethecostofequitytothepriceofthe stock,albeitataconstantgrowthrateofdividends.Whilewebelievethatamoreaccurate costofequityisobtainablefromtheCAPM,itisentirelypossibletomanagecapital structurefromtheperspectiveofdividendgrowthandretention.Infact,without knowledgeofbeta,aneffectiveuseofdividenddiscountmodelsshouldprovidethesame outcomeastheCAPM,becausebothmethodsconcentrateontherelationshipbetween earningsandthecostofequity. EARNINGSACCELERATION 494 ToreiteratetheGordonmodel:Thecurrentpriceofthestock=NextDividend/ (Ks–G).KsisequaltothecostofequityandGisthegrowthrateforwhichwehave formedthegenericequivalent,((ROE)(Retentionratio)).WhiletheGordonmodelisonly applicabletostockswhosedividendsgrowataconstantrate,wedonotuseitforvaluation. Weobtaina“ballpark”figureforthecostofequity,whichiscomparabletoanIRRfor thatstock.IfitishigherthanthatobtainedthroughtheCAPM,thenweobservewhether thatstockisagoodcandidateforinvestment.Inthatcase,theinternalreturnforthe companyisoutpacingstocksofsimilarriskonthemarket. Toobtainthis“internal”costofequity,wealgebraicallyturnthemodelaroundand seethatKs=(NextDividend/StockPrice)+Growth.Sincewehavemadegrowth contingentuponearnings,boththeamountretainedandthereturn(ROExRetention),we caneasilyobservehowthecostofequityhasadirect,positiverelationshipwithearnings. Infact,throughoutthebusinesscycle,thecostofequityrisesasearningsriseexceptthatit lagsbehindtherateofearningsgrowth.Thislagisadoubleedgedsword:asearnings outpacethecostofequity,stocksrise,butwhenadownturnerupts,earningsdeteriorate farmorerapidlythanthecostofequity.WhiletheFederalReservedesperatelycutsrates, stockscannotrecoveruntilearningsagainoutpacethecostofequityintheopposite direction-usuallyasixmonthprocessinrecentdecades,buttotallydependentonthe severityofthedecline. ThedynamicsofthislagareimplicitintheGordonmodel.Ifacompanydelaysin cuttingitsdividend,orcommunicatesexpectationsthereof,theproposeddividendyield component,(NextDividend/Price)willgrowbecausestockprices(thedenominator) decrease.Atthesametime,thegrowthcomponentisshiftingdownwardbecauseretention andearningsarebothlower.Thus,earningsoutpacethecostofequity.Thesame phenomenonoccursduringanexpansion:companiesarenotsurewhethergrowthisstable enoughtowarrantadividendincreaseandtheproposeddividendyieldshrinkstocausean upwardlagbehindearnings. 495 IntheCAPM,thelagappearstobecreatedbydelaysincuttingorraisinginterest rates.Aparadoxicalrelationshipbetweeninterestratesandstockpricesoccurs;stockswill bepositivelycorrelatedwithinterestratesuntilratesarehighenoughtodiscourage investment.Atthatpoint,stocksbegintobenegativelycorrelatedwithinterestrates,and theFedbeginsacampaignofratecutting.Sinceinflationisalaggingindicatorinitself,the Fedmustbewaryofstymieingbusinessgrowth.Itletsearningsgrowthoutpacerate raisinginitiativesandthestockmarketrises.Duringadownturn,theFedcannotlower ratesfastenoughwithoutcausingdislocationsinthecapitalmarkets.Inessence,the connectionbetweenthetwomodelsliesinexpectations:boththeriskfreerateandthe proposeddividendsoffirmsarefunctionsofthecollectiveexpectedeconomicoutlook-but -theyarereactionaryandnotforwardlookingsimplybecausechangingthementailsso muchrisk. STATISTICALVALIDITY Thepropositionthatchangesinleveragefactorssuchasthelong-termdebtto capitalratio(LTD/CAP),orAssets/Capitalshouldbecorrelatedwithstockpricesisevident fromtheimplicationsofthesemodels.IntheGordonmodel,theamountofdebtisimplicit intheretentionratio,whichregulatesgrowth.Ahigherpercentofretainedearningswill raiseafirm’sequityposition,butmayunderminedividendgrowthifitisexcessive. Analogously,intheCAPM,ahigherbetaiscreatedfromtakingonmoredebtwhichwill potentiallyraisethestockpriceinarisingmarket.Together,thetwomodelscreatethe foundationforcorrelation,andyetthereareseveralequivocalvariables.Bothinflation andrisinginterestratescanskewtheresults.Forexample,duringtheearlypartofan expansionwheninterestratesarelow,morelong-termdebtmaybecorrelatedwithhigher stockprices.Ontheotherhand,atamarkettop,thehighercostofdebtmaycausethose variablestobenegativelycorrelated.Similarly,ifinflationishigh,earningswillnotputas muchpressureonstockprices;themarketfactorsintherealrateofinflation.Inthatcase, moredebtmightbecorrelatedwithhigherpricesbecausethemarketfactorsinthe 496 repaymentofloansindepreciateddollars.Thus,thefutilityofacontrolledexperimenton leveragefactorsshouldbeevident.Inordertoachievestatisticalvalidity,theresearcher needsalarge,unbiasedsample.Butwhenitislargeenough,factorsoutsideofthe experimentwillchangetherelationshipbetweenvariables.Thisparadoxhascanblind researcherstotheinherentvalueofsomefundamentals.If,forexample,depreciationis highlypositivelycorrelatedfortwoyears,andthennegativelycorrelatedforone,the resultswouldbetermedstatisticallyinconclusive.Infact,onehastogaugethevalueof thatfundamental(depreciation)intermsoftheeffectofsomeoutsidefactor-achangein thetaxlaw,forexample.Fromamathematicalperspective,thereisnolongterm correlation,butfromaninvestor’sperspective,itmayendupbeingvaluablebecauseitis viewedwithinaspecificcontext.Incapitalstructuralism,weneedtoviewtheleverage factorsfromthecontextofchangesinthecostofcapitalwhicharespecificforeach company.Ifthecostofdebtisridiculouslylowforonecompany,thatfirmwillprobably incurmorelong-termdebtsimplybecauseitminimizesthecostofcapitalevenasinterest ratesareclimbing.However,whentheyieldcurveisascending,inflationisincheck,and earningsarerising,aneconomicenvironmentwillbeconducivetoformingtrendsamong someleveragefactors;whatisconsidereda“normal”businesscyclecreatesan equilibriuminwhichriskispricedcorrectly.Mostcompaniesreacttothesamemarket conditionsandactsimilarlywhenitcomestochangingcapitalstructure.Therefore, researchersinthissubjectmusttreadafinelinebetweentheneedformathematicalproof, andthepracticalobservationof“real”economicbehavior.Inthiscase,anystudythatisso narrowlyfocusedastobestatisticallyvalid,mayalsobeirrelevant. ABRIEFSTUDY Usingthesamedataasintheoperatingincomestudyofthepreviouschapter, changesinmid-rangestockpriceswereregresseddirectlyagainstchangesinleverage factorsonapercentagebasis-exceptwithoutfurthercategorization;eachvariablewas freelyregressedagainstmid-rangepriceswithoutreferencetoanysubcategory.Boththe 497 changeinthecurrentmid-rangeprice(theyearoftheleveragechanges)andthechangein thenextmid-rangepricearedescribed.Thisdivergenceallowedtheresearchertoobserve theimmediateeffectsonpricesandalsotheforwardlookingeffects(ifany).Thefollowing tabledelineatesthevariablesusedinthestudy: Table20-1 NextMid-rangeStock Price InterestRate(Totaldebt) CapitalMultiplier (Assets/Cap) OperatingIncome TotalDebt TotalCost(Sales-OpInc) Sales AssetTurnover OperatingMomentum OutstandingShares ProfitMargin Long-termdebt InterestExpense CapitalExpenditures Long-termdebt/Total Debt Assets Cash TotalLeverage Capital LTD/CAP TotalLeverage-absolute EPS FinancialLeverageRatio OperatingMomentum- Abs. CurrentMid-rangeStock Price Thosevariablesthataresignificantatthe95%level(a95%chanceofassociationwiththe stockprice)aredisplayed.Additionally,anyvariablethathasexplanatoryvalueinterms ofcapitalstructuremaybeexhibited;someofthevariableswereregressedagainsteach otheriftheyverifiedsometenetofcapitalstructuretheory.Thesampleincludesthesame onehundredandeighty-twodatapoints,representingthesamenumberoffundamental changesasinthepreviousstudy.Thesamecaveatsalsoapplybecauseseveraldatapoints willrepresentonecompany(twenty-fourcompaniesinall)andtheremaybeserial correlationinwhichonedatapointwillinfluenceasubsequentdatapoint.Thusthestudy mayevidencestrategiestakenbyafirmoveraspecificperiod.Thisqualitativeaspecthas pragmaticvalue,butinvalidatessomestatisticalconclusions.Ineffect,wecanobservehow capitalstructurechangesaffectedstockpricesoveraneconomiccycle,butwecannot attributecauseandeffect;therearesimplytoomanyuncontrolledvariables. 498 THREEASSUMPTIONS Weinterprettheresultsgiventhreeassumptions: • 1)Weassumethatthereisnotransitivityofcorrelation-thatis:-ifonevariableis correlatedtoanothervariable,andthatvariableiscorrelatedtoathirdvariable,wedo notassumethatthefirstvariableisalsocorrelatedtothethird. • 2)Secondly,weassumethatthereisconsistencyincorrelation.Ifassetsarecorrelated tosalesinyearone,forexample,,weassumethatthecorrelationisvalidinyeartwo, unlessprovenotherwise.Thatassumptionisonlyrealisticifthetwovariablesare deterministicfunctionsofeachother;thatis-onevariablesuchassaleswill mathematicallydetermineanothervariablesuchasoperatingincome.Inthatcase,the variableswillalwaysbeconsistentlycorrelatedbecausetotalcostissubtractedfromone toobtaintheother.However,inprobabilisticassociations,suchaswithstockprices andleverage,thatconsistencydoesnotalwayshold. • 3)Weassumethatconventionalfinancialwisdomisvalid.Stockpricesrisewhen earningsrise.Variablesthatincreasethecostofcapitalwilldecreasethepriceofthe stock.Weusethestudytodefendthesearguments,butalternatively,theremaybe connectionsthatinvalidatesuchfinancialtruisms.Eacheconomiccyclehasperiodsof completeincoherencywhen“normal”correlativevaluesaresuspended. AllregressionsinthestudyusetheSpearmanrankcorrelation.Dataisranked fromhighesttolowestbeforelinearregressionisapplied.Thismethodcapturesthe associativevaluebetterthanstraightlinearregression,butdoesnotcapturethestrengthof theassociationaswell.Nevertheless,itisthepreferredmethodbecauseittendsto emphasizecurvilinearrelationshipsthatmightnototherwisebeapparent.Correlations arescaledfrom-1to+1justasinlinearregression,buttheexhibitedprobabilitiesarethe probabilitiesofnoassociation.Thehighertheprobability,thelessthetwovariablesare associated.Thus,the95%probabilitylevelisstatedas0.05orless.Thisprocessconfuses 499 studentsandinvestorswhohavebeentaughtthat“more”connotesalargernumber.In thiscase,weareactuallyfindingtheprobabilitythatthe”nullset”(noassociation)istrue. EXPECTATIONS Itisexpectedthatcurrentyearfundamentalswillbemorestronglycorrelatedwith stockpricechangesofthesameperiodthanwithstockpricechangesofasubsequent period.Moreover,itisexpectedthatnocurrentperiodvariablewillbehighlycorrelated withearningsinthenextperiod.Theargumentforan“efficient”marketstemsfromfutile attemptstopredictmarketbehavior;capitalstructuralismacknowledgesthetransientand unstablerelationshipbetweenfundamentalsandstockprices.Whenthecostofcapitalis minimizedatvaryinglevels,thecombinationoffundamentalsthatachievesthisoptimum keepschanging.Therefore,no“exact”amountofsales,incomeorcapitalwillyieldalarge changeinpriceunlessitcontributestoloweringthecostofcapitalfirst.Atthatpoint,any magnitudeofchangewouldrestupontheamountofdiminishedrisk,usuallyimplyinga movementtowardacostreducingcapitalstructurecoupledwithanappreciablechangein income. Athirdexpectationisasimpleextrapolationofdeductivelogicwhichmostfinancial professionalswouldrecognize:pricesareaffectedbyearningswhichcontributetoretained earningswhichwouldraisetheproportionofequitytodebtwhenearningsarehigh.In EVAanalysis,wheretheflowofincomeisbalancedwiththechangeinequity,thereislittle explicitawarenessofthequalityofequitychanges.Weoftenassumethatthecostofequity willriseifafirmissuesstock,butthatsucharisemightnotbemalignantiftheequity buildupwasderivedfromretainedearnings.Wealsoassumethatthecompanywillpaya penaltyifitretainsearningsinsteadofdistributingthemasdividendswhenthecostof equityisparticularlyhigh.However,inthemajorityofobservedcases,whennetincomeis greatenoughtowarrantlargeincreasesinretainedearnings,thestockrespondspositively: netincomeismorethanenoughtoabsorbtheincreaseinequityandthecostofequityis keptlowbecausemoreretainedearningsdiminishbeta.Thus,weshouldexpectatleasta 500 concurrentlystrongcorrelationbetweenearningsandstockpriceandanegative correlationbetweenlong-termdebttocapitalandearnings;anincreaseinLTD/CAPmight implythatretainedearningshavedecreased,althoughbothcanriseatthesametimeunder theproperconditions. DATA(SeeTables) INTERPRETATIONANDRESULTS Thebehaviorofthelong-termdebttocapitalindicator(LTD/CAP)defied expectations.Whileithada76.76%probabilityofnoassociationwiththecurrentmid- rangeprice,whenregressedagainstnextyear’sstockprice,ithadaprobabilityofno associationof0.2%.Infact,threeothercostofcapitalfactors-long-termdebttototal debt,thefinancialleverageratioandtheinterestrateontotaldebt-displayedthesame “leadingindicator”characteristic.Theyeachshowedsignificantcorrelationwhen regressedagainstnextyear’sprice,butlittleassociationwhencomparedtocurrentprices. Thereadershouldnoticethatwhentotaldebtalonewasregressedagainstmid-range prices,itdisplayedastrongercorrelationwiththecurrentmid-rangepricethanthenext; thisindicatorexhibitsthesamebehaviorasmanyoftheearningsassociatedfundamentals likesales,assetsandoperatingincome.However,withtheadditionoflong-termdebtinthe numerator,itdisplaystheleadingcharacteristicoftheothercapitalcostcomponents. Whilelong-termdebtandcapitalinisolationbothexhibitstrongercurrentprice correlations,oncetheyarepairedinaratio,thecorrelationbecomesstrongerwithnext year’smid-rangeprice.Thestudent/investorshouldbecognizantofthetaxbenefitsof morelong-termdebtbecausethecorrelationbetweenlong-termdebtandstockpriceturns frompositivetonegative.Oncethedebtiscarriedovertothesubsequentyear,thetax benefitsmayremainthesamebutnowtheprobabilityofdefaultrises-unlessthedebt beginstoenhancereturns.Infactthelagissopronouncedforanyvariablethatdirectly affectsthecostofcapital,thatwecanseparatevariablesinto“capitalcost”components and“profit“componentswhicharemorecorrelatedwiththecurrentprice. 501 Table20-2 COSTOFCAPITALCOMPONENTS:MORECORRELATEDWITHNEXT YEAR’SPERFORMANCE Long-termdebt/TotalDebt InterestRateonTotalDebt FinancialLeverageRatio:OperatingIncome/OperatingIncome-Interest Long-termdebt/Capital Table20-3 PROFITCOMPONENTS(MAIN):MORECORRELATEDWITHCURRENT YEAR’SPERFORMANCE OperatingIncome Sales Assets Capital ProfitMargin Long-termdebt NumberofSharesOutstanding Fromacapitalstructureperspective,wecaninterpretthelagasanincreaseinriskthatis notfactoredintothestockuntilitfullyaffectsthecostofcapital.Moreover,sincethedata describesa“normal”marketwithrisinginterestrates(the1990s),wecannotpinpoint whethertheproportionoflong-termdebttocapitalraisedthecostofcapital.Again,there isacombinationofmarketfactors(interestrateandequitymarketincreases)that contributedtothelag.However,suchamarket“inefficiency”wasthemechanismthat poweredthisresearch;inanalreadyinflatedstockmarket,gainsofoneandahalftimes theaveragecouldbeexpectedsimplybychoosingcompaniesthatwereloweringtheLTD/ CAPratiothroughmoreretainedearnings.Withasmallerriskpremium(thedistance betweentheriskfreeinterestrateandequities),theeffectonLTD/CAPmighthavebeen lesspronouncedandlong-termdebtwouldhavebeenlessexpensive.Alternatively, 502 increasesinlong-termdebttocapitalratioscouldevenindicateanexpansivemarketif interestratesarelowenoughandearningsprospectsarehigh. HYPOTHETICALCAUSATION Whilecorrelationisnottobeconfusedwithcausation,evenstatisticalinference amongallthevariableswillnotgiveasatisfactoryexplanation.Onlythefinanciallogic inherentincapitalstructure,alongwithobservationoftheinvestmentprocess,canyieldan answer.Thereason?Thereisnoconcretemeasurementthatcananticipatearapid turnoverofcapitalandassets:thesoonerafirmcanturnthepurchaseofanassetintoa profitableventure,themoreinvestorswilldemandequityinthatcompany.Sincemost largedebtissuesareforthestrategicpurchaseofprofit-makingassets,thereissomelag betweenpurchaseandthegenerationofincome.Inthisinterim,firmswilloftentake restructuringcharges,giveoutexecutivecompensation(goldenparachutes)and consolidateoperations.Attimes,theconfluenceofoperationsbringsunanticipated problemsandtheneedforanother(albeitsmaller)loan.Forthesereasons,itistypicalto seetheproportionoflong-termdebttocapitalriseformorethanoneyear,andtoobserve atleastsomestagnationinthestock. Sometimes,theincurringofdebtisnotastrategicplay.Companieswhoareina sectorthathasnotbeenfavoredbythemarketmayuselong-termdebttofundfixedassets thatmaintainexistingoperations.Inthesecases,thestockisnotindemandandlong-term debttocapitalrises.Again,theincreasemaybepartofapatternthatlastsatleasttwo yearsifnotmore.Noretainedearningsareavailable,butthecompanymusthaveasource offundingorfacetheprospectofdownsizingandsellingassets.Long-termdebtbecomesa tooltoforestallinsolvencyuntilthecompanycanincreaseincome. Ifwecoordinatethebehavioroflong-termdebttocapitalwithourknowledgeof EVA,thereadercanobserveanotheravenueofapproach.Whenweacceptthepremise thatbothearningsandEVAarehighlycorrelatedwithstockprice,weneedtoexaminethe 503 capitalcomponentsoftheEVAequation.Accordingtoourcorrelations,thereisno variablethatisstronglycorrelatedwithanincreaseinnextyear’searnings.Bytheprocess ofelimination,ifcurrentLTD/CAPisnothighlynegativelycorrelatedwithnextyear’s earnings,thenetincomepartoftheEVAequationisunaffected.Thatleavesariseinthe costofequityasthereasonforadiminishedEVAandstockprice.Suchariseinthecostof equitywouldbeafunctionofahigherbetacausedbymoreleverage-theincreasein LTD/CAP. THEHAZARDSOFPLAYINGDETECTIVE Theconnectionbetweenthecostofcapital,stockprices,andtheproportionofdebt toequityisallencompassing.Unfortunatelywemaytendto“read”athreadintoa scenariowherenoneexists.Theinsistenceoninterpretationthroughcapitalstructure changescanblindustootherexplanationsthatmaybejustaslegitimateifnotmoreso. Forexample,thecostofcapitaldoesnotalwaysrespondtoincreasedrisk.Considershort- termdebtasanillustration.Short-termdebtmayexposeafirmtotheriskofdefaultmore easilythanlong-termdebt(morefrequentpaymentsatvaryingrates),andyetitisless expensive;thecorrelationbetweencapitalcostsandriskisnotalwayspositive.Inessence, itisdifficulttoobtainmathematicalproofofcausationbecausethesamplecontinually changesaseconomicconditionsarise:inonemarket,astockwillrisewithmorelong-term debttocapital,andayearlater,thesameindicatorwillbenegativelycorrelatedtoit.Such amarketefficiencyhascreatedthe“randomwalk”theorybecausefewcansuccessfully coordinateavailableinformationtomakeaccuratepredictions. Intheexhibitofearningscorrelations,thestudent/investorshouldnoticethe strengthofcorrelationforconcurrentfundamentalslikesalesandassets,andthencontrast thesewhentheyareregressedagainstnextyear’searnings:fundamentalshaveno predictivevaluewhatsoever.Thereisnofactorthatsomemanagercan“grabaholdof” andimprovenextyear’searnings.Infact,thefollowingtablerepresentsastraight 504 autoregressionofoneyear’searningsincreasesonthenext,andthereadercanobserve thatnomomentumwasexhibited. Table20-4 AUTOCORRELATIONWITHNEXTYEAR'SEPSPERCENTAGE YIntercept=22.9816 Coefficient=-0.04087 RSquared=0.001671 Y=22.9816-0.04087(X) Theminusculecoefficientofdeterminationrevealsthatthereisnopredictivelinear connectionbetweenyeartoyearearnings.However,severalresearchershavefoundthatif earningsaregreaterthanthemarketoverasixmonthperiod,theywilltendtoriseinthe nextsixmonthsaswell.Inthelate1960s,ManownKisorandVanMessnerworkedwith earningsacceleration,andRobertLevyofAmericanUniversityparalleledthisapproach withstocks.Bothfoundsixmonthmomentumtrendsfollowingaccelerationincreases. Eveninthedepthsofthedepression,nolessaresearcherthanAlfredCavellostatedthatif stockswentupforayear,theywerelikelytofollowwithanothergain. Muchoftheupwardtrendinstocksthatoccurredafterloweringlong-termdebtto capital,maybeattributedtoearningsmomentum.Whenearningsarehigh,moreretained earningsarecreatedthatlowersthelong-termdebttocapitalratio.Theproportionof equitynaturallyrisesbecausemoreearningsattractdemandforthestockaswell.Thus,it wouldappearthatthisleverageratioisinverselycorrelatedwithearnings-whichthedata substantiates:LTD/CAPhasanegativecorrelationwithearningsof-0.195179anda probabilityofnoassociationofonly0.008280or99.2%chanceofbeingassociated. However,thecorrelationofcurrentLTD/CAPincreasestonextyear’searningsincreasesis almostinsignificantat-0.039136witha0.625401probabilityofnoassociation.Obviously, morelong-termdebttocapitalisnotdepressingnextyear’searnings,althoughitmaystop earningsmomentum.Whilelong-termdebttocapitalisheavilynegativelycorrelatedwith currentearnings,therewillbesomeresponsefromtheinvestmentthatwillpropelearnings 505 slightlyhigherinthenextyear.Theintensityofthat“response”seemstodetermine whetherthestockwillbecomeattractiveandthefirmcanbeginraisingequity. Perhapsthebestexplanationofthedynamicsofproportionalityisfoundinthe marginalbenefitsequation.Weknowthatontheleftsideoftheequation,moredebtwill increasetaxbenefits.IfoneexaminesatypicaldefaultprobabilityalgorithmlikeAltman’s, thestudent/investorwillfindthatmanycomponentshave“assets”inthedenominator- retainedearnings/assets,operatingincome/assets,sales/assets,etc.-whichmakesupthe majorfactorontherightsideoftheequation.Asweraisedebt,wealsoraisetheasset portionofthedefaultprobability.Inthenextyear,withoutaninfusionofmoredebt,we needtoincreasetheperformancevaluesthatarecontainedinthenumeratorpartofthe defaultprobability-thesales,workingcapital-thebasicreturnonassets.Whenthese variablesdonotincrease,theprobabilityofdefaultrises,andthemarginalbenefits equationwilldecrease.Thus,thesoonerafirmcanintegrateassetsintoaprofitable scenario,themoremarginalbenefitswillrise.However,anylargeacquisitionormergeror evenachangeinfixedassettechnology,requiresa“periodofadjustment”beforethese assetscangenerateaprofit,nottomentionthepossibilitythattheacquisitionmaynot matchthefirm’soperatingcharacteristics.Alloftheseriskswouldbereflectedbysome stagnationifnotdepreciationofthefirm’sstockprice. THEARGUMENTFORCAPITALRATIONING Byfarthemostcuriousrelationshipwasfoundbetweencapitalandtheinterestrate ontotaldebt.Regressionswereperformedbetweencurrentcapitalandnextyear’sinterest rate,aswellasthecurrentinterestrateandnextyear’scapital.Althoughweassumefrom capitalstructuretheorythatmorecapitalwillincreaseriskandleadtohigherrates,we werenotpreparedtoobserveadecreaseincapitalforthesubsequentyearwheninterest rateswereraisedduringthecurrentyear.Thisnegativecorrelationoccursinacontext wherecurrentlong-termdebtincreasesarehighlycorrelatedwithcurrentincreasesinthe interestrateontotaldebt.Paradoxically,currentincreasesincapitalhavenosignificant 506 associationwithcurrentyearinterestratesatall.Accordingly,wecanhypothesizeatleast threereasonsforthisoccurrence. • 1)Overtheeightyearstudyperiod,anomaliesoccurredinthedatathatcreateda negativerelationshipbetweennextyear’scapitalandthecurrentinterestrateontotal debt. • 2)Thegreaterriskthatisimplicitinhigherinterestratesencouragesfirmstofund projectsinlargeblocksofdebt,especiallywheninterestratesareanticipatedtorise. • 3)Firmsrationcapitalandcutbackonprojectswhenalargeacquisitionismadein ordertoconcentrateonmakingitaprofitableventure. Innormalcapitalbudgeting,afirmacceptsallprojectsthathaveapositivenet presentvalue.Inlayperson’sterms,thatpropositionobligatesafirmtotakeonall profitableprojects,givenaspecificcostofcapitalandrecognizingfuturecashflowin currentdollars.Fromeconomics,weknowthatafirmshouldproducetoapointwhere marginalrevenueisequaltomarginalcost,whichisgenerallyanalogoustoaccepting projectswithapositivenetpresentvalue.Ifafirmdoesnotfundallprojectsthatwould exceeditscostofcapital,itdoesnotmaximizeitsvalueandpartakesinwhatistermed “capitalrationing”.Whilefinancialprofessionalsrecognizecapitalrationing,thescenario islessfamiliartotheaverageinvestor.Inshort,itistradingtheviabilityoffutureprojects inordertomeetsomeshorttermobjective-whichareoftennumerouswhenlargemergers areimplemented.Theshiftinprioritiesisoftentermedagencyfactorsor“agency friction”,becausemanagementactsasan“agent”fortheshareholder.Thesefactorsmay includethefollowing:1)thebeliefthatraisingmorecapitalwillimmediatelyraiseits inherentcost;2)therecognitionofgreaterdefaultriskoncealargeblockofdebtis incurred;3)thefearoflosingvotingcontrolbyissuingmoreequity;4)thebeliefthat currentprojectsareeitherveryriskyand/orunderstaffedandrequiremoremanagerial attention.Bynomeansanexhaustivelist,anyofthesebeliefswillundermineafirm’svalue whencarriedtofruitionbecausetheywillfrustratemovementtowardanoptimalcapital 507 structure.Under-fundingprospectiveprojectswillleadtodiminishedearningsandthe potentialtomisallocatetheproportionofdebttoequity. Companiesthatincurlong-termdebtareessentiallygivingupfinancialcontrolto creditors.Therightsofabondholderarealmostalwaysprioritizedoverthoseof shareholderswhichsetsupanenvironmentconducivetocapitalrationing.Indentures oftenrestricttheraisingofnewdebttoaspecifictimesinterestearnedratio(TIE)andif retainedearningsarenon-existent,thefirmwillhaveproblemsfinancingnewprojects; equityisnotattractiveunlessthefirmisgeneratingsufficientoperatingincomewhichis difficulttoachievesosoonafteranacquisition.Thealternativeistorationcapitaland maximizeprofitgiventhoseconstraints. Withouttransitivityofcorrelation,wecanalsoobservethatlikethecurrentinterest rate,currentlong-termdebttocapitalincreasesarenegativelycorrelatedwithnextyears capitalincreases.Evidently,moredebtraisesinterestrateswhichraiserisk.Sincemany longtermprojectsarefundedwithlong-termdebt(cashflowmatching),thesecapital intensiveprojectsmustbefollowedbysmaller,moremanageableprojectsinsubsequent years,andthuschangesincapitalbecomenegativelycorrelatedwithcurrentrisk However,thehypothesisofcapitalrationingcannotbeproved.Whileitisalwaysa possibility,thepremiseisbasedonknowledgeofafirm'sspecificIRRand“project pipeline”towhichfewindividualshaveaccess.Thisalternativeconcept(adequatefunding forsmallerprojects)isareliableassumptiongiventhestrongcorrelationbetweenincreases incurrentstockpriceandincreasesincurrentcapital(acorrelationof0.418036witha probabilityofnoassociationof0%).Ifindeednewprojectsrequirelesscapital,whileat thesametimeoperatingincome(ROA)istemporarilydepressedbytheneedtointegrate newassetsandpayadditionalinterestexpense,themid-rangestockpricemightdecline. Thequotientofeconomicprofitisthecomparativedynamic,NetIncome/Costof Equity.Inthisratio,thecostofequityisthefullcostoftheCAPMpercentagemultiplied bystockholder’sequity.Inthecapitalrationingscenario,withincreasingleverage,net 508 incomeisslightlydiminishedwhiletheCAPMpercentagerisesbecauseofanincreased beta.Sinceallprofitableprojectsarenotbeingfunded,netincomemaynotincrease enoughtoraiseeconomicprofit.TheresultwillbeadiminishedEVA,andadeclining comparativedynamic,eventhoughequityfundingwasnotextensivelyused.Thefirmfails tomovetowardanoptimalcapitalstructuresimplybecauseresourceswerechanneled towardlastperiod’scapitalallocationandawayfromcurrentprofitableprojects.In industriesthatrequirebothahighleveloftechnologyandahighlevelofdebt,project defermentcanleadtoobsolescencebecauseafirmcanonlycompetewiththenewest innovationswhichtendtohaveashort“shelflife”.Suchcompaniesarehitwitha“double whammy”;theyneedtodecreasethecostofcapital,butbythetimetheydo,theprojects theywouldhavefundedarenolongerlegitimate.This“hidden”costofdebtisseldom discussedwhenfirmsintegratetheiroperatingandfinancialrisks,becauseoperating leverageonlymeasurestheamountoffixedcostsandnottheirquality.Forexample,the PCindustrywaschangingsorapidlyinthelate90sthatevenwithaloweroperating leverage,theriskofcreatingunsoldinventorybecamefartoogreat. Thelackofasignificantcorrelationbetweenanycurrentfactorandincreasesin nextyear’searningsleadstotheconclusionthatraisingthecostofcapitalistheprime reasonforalowerstockpriceinthesubsequentyear.Bydefault,whenweexaminethe nextEVA,wefindthat:1)netincomeisbasicallyunaffectedbylastyear’scapital allocation;2)capitaloutlaysmaybedepressediflong-termdebttocapitalwasincreased inthelastperiod,and3)theincurringofdebtraisesbeta,whichincreasesthecostof equity.TheonlyfactorthatcouldpossiblyoffsetthisdecreaseinEVAwouldbean immediatedecreaseinLTD/CAPthatwouldbederivedfromretainedearningsorequity issues;suchanoccurrencewoulddiminishbeta,butwouldonlybeachievableifoperating incomewerebolstered,orifthestockwasespeciallyattractiveatalowprice. Withoutspecificreferencetoanumericcostofcapital,ourconclusionsaremerely speculative.Toobtainstatisticallyviabledata,thisstudywouldhaveneededtonotonly 509 measurethedifferentcostsofcapitalforeachperiod,buttodeterminetheamount attributabletomyriadfactors-theFederalReserve,thestockmarket,theinternal dynamicsofthecompany.Therefore,thereadermustbewarnedwiththecaveatthatthis studycannotbeoverlaidonotherperiodsandbeexpectedtoberelevant.Itwasthe workablepatternofthe1990s,butmaynotrelateto2015.Levelsandchangesoftherisk premium,technologyshifts,andotherviablealternativeinvestmentsmaychangethewaya firmmovestowardanoptimalcapitalstructure.Othershiftsincludechangesin bankruptcylawsandtaxlegislation.Forexample,taxingdividendsatthesamerateas capitalgainsmakesthosestocksmoreattractive.Greaterdemandforequityinthese companiesmayshiftthecostofcapital,andraisinglong-termdebtmayactuallybenefitthe stockinsubsequentyears.Similarly,theshifttoalowerlevelofinterestratesinthenew millenniumalsoshiftedtheoptimalcapitalstructure.Theprobabilityofdefaultwas differentforthesameamountofdebtheldindifferentyears.Thus,thestudent/investoris encouragedtodoresearchonwhichleveragepatternisworkingforthetopsectors,andto understandwhyitisprofitablebyobservingthecomponentsofthecapitaldynamic:(Net Income-CostofEquity). 510 SPEARMANRANKCORRELATIONWITHPERCENTAGEINCREASESIN CURRENTANDNEXTYEARMID-RANGESTOCKPRICES Table20-5 VARIABLE CURRENT MID-RANGE NEXTYEAR MID-RANGE Most Significant Correlation Probability Correlation Probability Operating Income 0.414726 0 0.230556 0.001741 Numberof Shares 0.190681 0.009926 0.102052 0.17042 Assets 0.415331 0 0.159849 0.031123 Capital 0.418036 0 0.140295 0.058892 InterestRate -0.06787 0.36263 -0.153022 0.039177 Sales 0.443389 0 0.150517 0.042544 TotalDebt 0.222137 0.002579 0.155414 0.036176 EPS 0.209008 0.004632 0.197081 0.007661 Capital Expend. 0.275874 0.000164 0.072362 0.331668 ProfitMargin 0.166469 0.024704 0.149717 0.043669 LTD/Total Debt 0.038071 0.609878 -0.172137 0.020146 LTD/CAP -0.022051 0.767631 -0.220211 0.002816 Financial Leverage -0.171175 0.020864 -0.251781 0.000606 Capital Multiplier -0.135234 0.068727 -0.041765 0.57561 Long-term debt 0.149316 0.044241 -0.106845 0.151118 Eachiteminthevariablecolumnrepresentsapercentchangeinthatfundamental. Allitemswithaprobabilityofassociationof95%arelisted,butifanitemhascomparable valuee.g.,capitalmultiplier,thatfundamentalislistedaswell.Additionally,ifanitemis significantlycorrelatedinoneyearbutnotinthenextorpreviousperiods,itisalsolisted. Severalitemswerenotlistedsimplybecausetheywerenothighlycorrelated-suchas- 511 assetturnoverandendofyearbalancesheetcash.Theirexclusiondoesnotmeanthatthey areanylessimportantthanthelistedvariables,justthattheylackedcorrelationwithstock priceincreasesforthespecificyearswiththesetwenty-fourS&P500companies. SPEARMANRANKCORRELATIONONCHANGESINTHEINTERESTRATEON TOTALDEBT Theeffectofchangesincurrentfundamentalsoncurrentinterestrates: Table20-6 CURRENTYEAR VARIABLE CURRENTYEAR INTERESTRATE Variable Correlation Probability LTD/CAP 0.297846 0.000044 Capital(Equity+LTD) 0.037876 0.611708 Theeffectofchangesinthecurrentyearinterestrateonchangesinnextyear’s fundamentals: Table20-7 NEXTYEAR CURRENTYEAR INTERESTRATE Variable Correlation Probability OperatingIncome -0.045664 0.56886 Sales -0.067203 0.401484 Capital -0.167101 0.035859 Assets -0.09137 0.253549 TotalDebt 0.045173 0.573031 CapitalExpenditures -0.10865 0.17418 NumberofShares -0.082255 0.304204 Outofseveralvariables,onlychangesincapitalhaveasignificant(95%)chanceofbeing associatedwithincreasesinthecurrentinterestrate.Correlationisnotcausation,butit appearslogicalthatlesscapitalwouldbedemandedifthepriceofithasrisen.Thispointis 512 evenmoreemphaticinlightoftheprevioustable’sindicationofnosignificantcurrent associationbetweenthetwo. Theeffectofchangesincurrentfundamentalsonchangesinnextyear’sinterestrateon totaldebt: Table20-8 CURRENTYEAR VARIABLE NEXTYEAR'S INTERESTRATE Variable Correlation Probability LTD/CAP 0.290095 0.000278 Capital 0.165464 0.037738 FinancialLeverageRatio 0.065025 0.416949 CapitalMultiplier 0.068533 0.392213 Noticethattherelationshipbetweencurrentinterestexpenseandincome(financial leverage)isnotsignificantlyassociatedwithnextyear’sinterestrate,butassoonaslong- termdebtisincurred,thereisaneffectonnextyear’sinterest.Evidently,theincurringof long-termdebtobligatesthecompanytoaseriesofinterestpaymentsthatraisestherisk associatedwiththecostofcapital.Moreover,greatercapitaloutlaysinthecurrentyear, tendtoforcetheinterestrateupwardinthenextyear,butthishypothesisisstatistically inconclusive:theeffectmaybeduetoFederalReservehikesandnottheimplicitincrease incorporaterisk. SPEARMANRANKCORRELATIONBETWEENCHANGESINTWOCURRENT FINANCIALLEVERAGEMEASUREMENTSANDCHANGESINNEXTYEAR’S FUNDAMENTALS 513 Theassociationbetweenthechangeinthecurrentfinancialleverageratioandthechange inthecurrentLTD/CAPmeasurement: Table20-9 CURRENT LONG-TERMDEBTTO CAPITAL Variable Correlation Probability FinancialLeverageRatio 0.29166 0.000065 Theassociationbetweenchangesinthecurrentfinancialleverageratioandchangesinnext year’sfundamentals: Table20-10 NEXTYEAR’S CURRENTFINANCIAL LEVERAGERATIO Variable Correlation Probability OperatingIncome 0.05538 0.489493 Sales 0.034869 0.6636 Capital -0.185999 0.019291 Assets -0.103109 0.197328 NumberofShares -0.104523 0.191216 TotalDebt -0.00365 0.963699 CapitalExpenditures -0.119273 0.135525 Onlycapitalhasastronginter-periodrelationshipwiththechangeinthefinancialleverage ratioanditisnegative.Wecanonlyassumethatfirmsaredecreasingthecostofcapitalby minimizingnextyear’scapitalallocation.Itisnotproved. Theassociationbetweenchangesinthecurrentlong-termdebttocapitalratioandnext year’sratio: 514 Table20-11 NEXTYEAR’S CURRENTLONG-TERM DEBTTOCAPITAL Variable Correlation Probability LTD/CAP 0.034729 0.664871 Whilethisauthorbelievesthatincurringdebttendstoencouragemoresubsequentdebt, thisstudydisputesthatassumption.Thereislittleassociationbetweenchangesinthe amountofdebtinjuxtaposedyears,althoughwedonottestforthestatisticalviabilityofa debtincreasebyitself.Long-termdebttocapitalmaystillriseinthenextyearwhendebt isincurredinthecurrentyear,buttheamountofchangeisnotrelatedtopreviouschanges. Theeffectofchangesincurrentlong-termdebttocapitalonchangesinnextyear’s fundamentals: Table20-12 NEXTYEAR’S CURRENTLONG-TERM DEBTTOCAPITAL Variable Correlation Probability OperatingIncome -0.003137 0.968793 Sales -0.104414 0.191686 Capital -0.221585 0.005142 Assets -0.224585 0.004555 NumberofShares -0.138209 0.083311 TotalDebt -0.17629 0.026711 CapitalExpenditures -0.173104 0.029626 Thistableshouldconvincethereaderthatgrowthinthesubsequentyearwasstymiedby anincreaseoflong-termdebttocapitalinthepreviousyear.SinceWallStreettendsto prizegrowthaboveallelse,thestockpricewouldnaturallydecline.Noticethatbothassets andcapitalinthenextperiodaresignificantlyassociatedandthateverycorrelationis negative.Noticealso,thatoperatingincomeisbarelyaffectedwhichwouldprovidethe 515 foundationfornoassociationwithnextyear’sEPS.Wecannotprovethatassetsand capitalareintentionallykeptlowtoimprovethereturnoninvestment.Sincelargeblocks ofcapitalaremorecostefficientthanthesameamountspreadoverseveralyears,we assumethatcapitalrationingisnotbeingimplemented. SPEARMANRANKCORRELATIONBETWEENCHANGESINEARNINGSPER SHAREANDCHANGESINFUNDAMENTALS WedisplayonlythecapitalcomponentsinthecorrelationwithcurrentEPSchanges becausemostfundamentalshaveastronglypositivecorrelationwithcurrentearningsand toexhibitthemwouldberedundant.Itisrighttoassumethatassets,sales,capitaland operatingincomewillriseconcurrentlywithEPS,butthechallengefortheinvestoristo determinethetimeandquantityofthatrise.Whenafirmincreasesitslong-termdebtto capital,thevalueofthecompanyimmediatelyincreasesbecauseoftaxbenefits,andyetthis riseinstockpriceisdisconnectedfromearnings.Itimmediatelyappearsthatinvestorsare speculatingaboutnewassets,butweknowfromourstudyofcapitalstructurethat marginalbenefitsareincreasing.Thisdisconnectcanbeobservedinthetriadrelationship betweenlong-termdebttocapital,earningsandstockprice.Currentlong-termdebt increaseshavelittleassociationwithcurrentstockprice,butcurrentEPSincreasesare heavilyassociatedwithit.Ontheotherhand,currentlong-termdebttocapitalchanges haveastrongnegativeassociationwithcurrentEPSchanges.Weconcludethattheremust besometimeswhenthecurrentstockpricerisesstronglyinthelightofalong-termdebtto capitalincrease;theseriseswouldencompassalargegainintaxbenefitsasopposedtoa smallerincreaseindefaultprobability,theexactrecipeforanimprovementinthe marginalbenefitsequationand(mostlikely)inEVA. 516 TheeffectofchangesincurrentfundamentalsonchangesincurrentEPS: Table20-13 CURRENTVARIABLE CURRENTYEAR’S EPS Variable Correlation Probability LTD/CAP -0.195179 0.004616 Capital 0.202482 0.00612 FinancialLeverageRatio -0.14146 0.0568 CapitalMultiplier -0.069514 0.351096 InterestRateonTotal Debt -0.017927 0.810172 CurrentMid-rangeStock Price 0.209008 0.004632 Theinformationthatwouldbemosthelpfultocorporateexecutiveswouldbeforward lookingindicatorsofearningsgrowth.Whatstrategicactionscanwetakethisyearthat willleadtoincreasedearningsgrowthinthenext?Asrevealedinthisstudy,thereis certainlyno“magicwand”.Allfundamentalsmustbesmoothlycoordinatedforsuch growthtooccur.Therecanbenoleadingstatisticalmeasurementthatwilldetermine whetheracompany'ssaleswillsuddenlyescalate,andsuchresearchisnotintherealmof investorfinance,butisinthedomainofcorporatemarketing.Infact,partoftheproblem withforecastingearningsandtheneedforcompanyguidanceisattributabletothislackof correlationasevidencedinthefollowingtable. Theeffectofchangesincurrentfundamentalsonchangesinnextyear’sEPS: 517 Table20-14 CURRENTVARIABLE NEXTYEAR'S EPS Variable Correlation Probability LTD/CAP -0.039136 0.625401 Capital -0.013119 0.870044 FinancialLeverageRatio -0.086882 0.277712 CapitalMultiplier 0.134557 0.091873 Ifthereweresomeassociationbetweenanycurrentfinancialstatementitemandchangesin nextyear’sEPS(outsideofordinarymaintenancelikecapitalexpenditures),itwouldhave beenexploitedlongago.However,this“analyticalconfusion”hassetupthecapital structuralistwiththeopportunitytocoordinateseveralvariablesandattemptto distinguishthepathofcapitalcosts.Whileitmaynotbeassimpleasincreasingor decreasingoneortwovariables,itfollowsachainoflogicthatmostinvestorsarecapable ofimplementing.Firmsmustminimizecapitalcoststoreachanoptimalproportionofdebt toequity.Inthisstate,profitswillbenearlymaximizedwhilethepriceofthestockwillbe maximized (BacktoTableofContents) 518 21 PROBABILITYANDCAPITALSTRUCTURE Inthemid1960s,astatisticianbythenameofEdwardThorpwroteabookthat completelyrevolutionizedthegameofblackjack.Thebookwascalled“BeattheDealer” andtoutedthepremisethatiftheplayercountedtheamountoftencountcards,heorshe wouldhaveadistinctadvantageoverthehouse.Formerly,goodblackjackskillsmeant “hitting”or“standing”onthebasisofthedealer’supcard,astrategythatgavemost casinosathreetofourpercentadvantage.Assoonasprofessionalgamblersbegan applyingThorp’sprinciples,thecasinosbeganlosingalotofmoneyandeitherbeganusing multipledecks,ormorelikely,threwthecardcountersout. AlthoughmanyinvestorswilltreatFederalReserveratehikeslikeadealer’sup card,thesimilaritieswithblackjackarenotextensive.Theprobabilitiesincardgamesare alwaysconcrete,stableandcalculable,whilethoseinthefinancialmarketsarenot.The market,unlikethehouse,willgivetheinvestorafavorableupwardbiasovertime-the yearlyaveragegainapproachestenpercent-butvariablerelationshipsandtheir consequentprobabilitieschangedramatically.Relationshipsbetweenvariableslike inflation,GDP,foreignexchange,andtheleveloftechnologyareconstantlyevolving, leadingtoaninherentunpredictability. Anexampleofavariablethattypifiesthischaoticfluxisthecaseofearningsin1999 to2001.Bylate1999manyinvestorswereputtingmoneyintoInternetstockswhichwere mostlystart-upswithlittleornoearnings,oratbest,increasingearningsthatwerestill negative.Someofthesestocksshotuptofiftyoronehundreddollarsashare.“TheNew Economy”wastoutedbythefinancialpress,andinvestorsanticipatedbigpaydaysjust aroundthecorner.Bylate2000,evencompanieswithgoodearningsbegantoseetheir stockpricesplummet.Thetechnicalanalysisthatallowedinvestorstoanticipateandprofit 519 offtrendsbegantofail,andinvestorsbegancuttingtheirlosses.Thecorrelationbetween thelastperiod’searningsandthenextstockpricewasnolongervalid.Whathappened? Thestoryofthetechbubbleishistoricallycomplex,butitissafetosaythatthe marketbegancorrectingitselfbeforearecessionwasimminent.Thebloatedmarketcould notwithstandinvestorperceptionchanges,higherpricesandlowerexpectationsforGDP. Inessence,the“NewEconomy”wasoverbecausethemarketrevertedtoitsmean; projectedfutureearningscannotbethefoundationforaninvestmentiftheylackthe productioncapabilitiestobackthemup.Withoutanassetbasethatrepresentedlower operatingrisk,the“pieinthesky”InternetcompanieswentthewayofFloridalanddeals andTexasoilwells. THEEFFICIENTMARKETSHYPOTHESIS Analystshaveattemptedtopredictcyclicalshiftsformanyyearswithoutmuch luck.Infact,anentirebodyofliteratureexiststhatcategorizesourabilitytopredictprices basedonpriorinformation,anditisfoundthatthemarketalmostinstantaneouslyprices securitiesbeforethosewithpubliclydisseminatedinformationhaveachancetoreact. Whileillegal“insider”tradingisalmostalwaysprofitable,tradingonthebasisof fundamentalsgleanedfromaprospectusisnot.Toexplainthisseeminginjustice,financial academicsproposedtheefficientmarketshypothesiswhichcategorizesreactionto informationinthreebasic“forms”:weak,semi-strongandstrong.Intheweakform,the hypothesisstatesthatallpastpricingdataisfullyreflectedinastock’spriceandcannotbe usedtopredictfutureprices.Thesemi-strongformstatesthatallpubliclyavailable information,frompastpricingdatatonewlypublishedsalesfigures,cannotbeusedto predictfutureprices.Finally,the“strong”formstatesunequivocallythatnopublicor evenprivateinformationisusefulforcreatingavalidforecast.Sinceinsidersdomake moneyoffofillegaltrading,mostfinancialprofessionalsadoptthesemistrongformasthe mostrationaltenet. SCREENS 520 Stockscreenshaveanunreliablehistoryfortheveryreasonsdiscussedabove:The marketwillchangeenoughtomakethemobsoleteaftertheyappeartobeworkingfora while.Theillusionofbeatingthemarketstemsfromtheupwardspressureofearnings thattendstocreatemomentumwhentheeconomyisinhighgear.Threeoutoffourstocks willriseinabullmarketandnineoutoftenwillfallduringabearmarket.Thecorrelation betweenearningsandpriceishighandvolatileinthebullmarket,lowandvolatileinthe bearmarket,andstrongandstableinthelongrun.Whenthemarketistrending,screens likethePEGratiocanworkverywell,butwillstoptheir“magic”assoonascapitalcosts outpacetheaccelerationofearnings.Cyclicalstocks,asanexample,willtypicallydowell rightafterarecession,butthenstagnatefortherestofthebusinesscycle.Theseinterest rate-sensitivestockswillhavehighpriceearningsratioswhentheyshouldbebought,and lowratioswhentheyshouldbesold,whichistheexactoppositeofconventionalwisdom! Thereasonforthisanomalyisthatearningsincreasethroughoutthebusinesscycle,but investorsmoveontomoreprofitablesectorsasinterestratesrise,depressingtheprices. Anotherexampleofafailedearningsscreenoccurredduringthe1970s.Withdoubledigit inflationandunemployment,“reading”theeconomywasdifficultatbest,becausehigher earningsdidnottranslateintohigherstockprices;onehadtogaugetheindustrytoseeifit weretobecomesuddenlyobsolete,oroversensitivetotheriseincommoditypricesthat occurredsimultaneouslywithanoilshortage. THERETURNONCAPITAL OuradaptationoftheDuPontequationyieldedareturnoncapital,ratherthana returnonequity.Thefullreturnofcapitalequationwascomposedof(EBIT/Sales)x (Sales/Assets)x(NetIncome/EBT)x(EBT/EBIT)x(Assets/Capital).Wefurther decomposedAssets/Capitalinto(LTD/Capital)x(Assets/Short-termDebt)x(Short-term Debt/LTD).WeknowfromoptimizationmodelsthatROEwillbemaximizedatacapital structureofalldebt,whileROCwillmaximizeatastructureofallequity:givena constantoperatingincome,ROEdirectlytradesequityfordebtwhileROCmaximization 521 hasnointerestexpense.Ifweexaminetheequationbyitsfiveparts,weintuitivelydiscover thatthelastthreepartsconcerncapitalinvestmentthatbecomesthefoundationfor increasingthefirsttwoparts,operatingmarginandassetturnover.Whenwelookfor changesintheleverageratios(EBT/EBITismerelytheinverseofourfinancialleverage ratio,EBIT/EBIT-Interest)wewillfindthattheysometimesoffsetthe“profit”ratiosof operatingmarginandassetturnover.Infact,thereisultimatelyatendencytoincreasethe profitvariablesbyagreaterdegreewhentheleverageratiosaredecreased.Ofcourse, therearemanyinstanceswhenbothleverageratiosriseandprofitratiosrise,but identifiablepatternsemergeinwhichassociationstendtobenegativelycorrelated.There isalwaysalagtimebetweencapitalinvestmentandprofitability,andthestockreflectsthis risk.Differentcombinationsofchangesintheratiosseemtohaveafavorableor unfavorableeffectonrisk/returnfactorsandultimatelyonthecompany’sstock.Whilethe patternsseemliketheoldscreens“dressed”inprobabilisticgarb,theymerelyofferamore elaboratemethodofobservingrisk.Theyarebothactionableonacorporatelevel,and viewableonaninvestorlevel,andofferaninterfacebetweentrendsinthecapitalmarkets i.e.,lowerinterestrates,andintracompanyfundamentals.Theyarenotmeanttoforecast stockprices,becausethemarketreactstotheproportionofdebttoequitybasedonthe priceofriskofeachrespectivecomponent.Forexample,atthebeginningofanexpansion, debtmayberelativelyinexpensiveandleverageratiosthatentailincreasingdebtmight flourish.Ontheotherhand,afterinterestratesrise,leverageratiosthatimplyequity financingmightbemoreprofitable.Inessenceeachcombinationformsa“leveragestate” thathasacharacteristicprofitabilityandforeshadowsnear-termcapitalchanges. LEVERAGESTATES Considerthenatureofcapitalinvestment.Companiesdonotwanttoissuestockeachyear becausesuchactionsdilutethepriceofoutstandingsharesandhaveprohibitivelyhigh floatationcosts.Yet-companiesdowanttomakepublicissueswhenthepriceofthestock ishighenoughsuchthattheissuednumberofsharesiskepttoaminimum.Secondly, 522 onceacompanyeitherstartstopayoffitsdebtorreplacesitproportionatelywithequity,it tendstodosoforaslongasitssectorisfavoredbythemarket,Inessence,ahighermarket willmeanbothmorestockissuesandmoreretainedearningswhichwillboostthe proportionofequitytodebt,aswellasthestockprice.Rarelywillapatternofalternating yearlydebtincreaseswithequityincreasesbeprofitable,becausethereisalwayssomelag time,andmomentumthatwillmakechangesineachcapitalcomponentfallintoatrend. Whenmanyincreasesinleverageoccuroveranumberofyears,itisindicativeofan insolventcompanyandthestockfalls.Iftoomanyequityincreasesoccur,thenthe companyisviewedasnottakingonenoughrisktobecompetitive,andagainthestockfalls. Thekeytostockappreciationistomoveawayfromtheoptimalcapitalstructureand followitupbyanumberofsmalljumpsbacktowardsit.Companiescanmoveawayfrom thetargetwitheitherdebtorequity,butmostmovementswillentailtheacquisitionof productivecapacitypurchasedwithadebtissue.Inthiscase,themovementbacktowards theoptimaltargetstructureencompassespayingoffdebtwithhigherearningsand retainingthosesameearningstoimprovethecompany’sequityposition.Successful companieswilltrytominimizetheyearsthatleverageincreases,andmaximizethenumber ofyearsaninvestmentpaysoff.Usually,themomentumofthecompany’srespectivesector willdetermine,howlonghigh-demandforitsproductswilllast.Atthe“omega”point,a companywillretrench,andhopefully“recharge”thecyclewithresearchanddevelopment, small,riskreducingacquisitions,andevenmoredebtissues.Inessence,aleveragestate willreacttothemarketintermsofpriceandrisk,andchangesintheROCfactorswillbe derivedfromtheintegrationoftheproductioncyclewiththecapitalmarkets.Ontheother hand,thelevel(proportion)ofdebtinthecapitalstructurewillbederivedfromthe steadinessoftheproductioncycleandthetypeofassetsitneeds.ROCfactorscantellusif weneedtoincreasedebtorequity,buttheycannotsettheoptimaltargetproportionby themselves. INDUSTRYAVERAGES:LEMMINGSVSLEADERS 523 Thetypeofindustryrespondstomarketdemandbyestablishingaveragevaluesfor eachROCfactor.Theamountoffixedassetsandtheinherentriskstheyentail,establishes apatternofprofit,costs,andcapitalneeds.Whilenoindustryaverageis“setinstone”, ROCfactorsgenerallyreverttotheirmean,allowinganalyststomakepredictions,and establishingdefinitivecorporatebehaviorforeachleveragestate.Infact,manycompanies withinasectorwillsharethesameleveragestate,simplybecausetheirresponsetoproduct demandissosimilar.Depreciationratesaresimilar,andeachcompanymustmakea requiredamountofcapitalexpendituresatthesametime.Thus,entireindustrieswillbe takingondebtorbuildingequity,dependingonexistingmarketfactors,andthecostof capital.Whencapitalstructureisinvolved,acompanymustexamineitsexisting prerogatives–especiallyifitfindsitself“outofsync”withthefundingactionsofthe competition.Althoughitispossibletotakeriskswhenotherfirmsare“restingontheir laurels”,themarketwilldeterminewhichcapitalincreasesaremostefficient.Infactifa majormacro-factorlikeinterestratesisanticipatedtobeatacertainlevelforalongtime, theoptimaltargetstructuremaychangetoconformtoit. THELEVERAGESTATERATIOS WemodifytheROCfactorstomakethemmoreuseable.Thecostofdebt, EBT/EBIT,becomesthefinancialleverageratio,EBIT/EBIT-Interest.Wechange operatingmarginintoadynamicratio,operatingmomentum,whichis%∆ ∆∆ ∆EBIT/% ∆ ∆∆ ∆Sales.Wemeasurethechangeinlong-termdebttocapital,LTD/CAP.Lastlywe measurethechangeinassetstocapital,Assets/Capital.Alloftheseratios,including operatingmomentum,aretreatedasanincreaseoradecrease;inthecaseofoperating momentum,weareactuallymeasuring“achangeofachange”.Althoughthetaxretention ratio,NetIncome/EBT,iscrucialtothemarginalbenefitsequation,itisnotincludedinthe leveragestatebecauseitistheleastsubjecttostrategicchange;firmsmustmanagetheir taxstructureswithintheparameterssetbythegovernment.Operatingmomentumandthe financialleverageratioestablishaninterfacebetweenproductionandthecostofdebt. 524 Long-termdebttocapital(LTD/CAP)establishesnotonlytheproportionofdebttoequity, butimpliesthecostofequity,whichriseswiththeproportion.Lastly,theratio Assets/Capitalestablishesthedegreeofcurrentliabilitiesemployedbythecompany;an increaseinthisratiocanimplythatthecostofcapitalisdecreasingbecauseofalternative short-termfinancingorontheotherhand,thatdebtissopervasivethatsolvencyis endangered.Itthereforerequiresmoreextensiveanalysistointerpretcorrectly,andmay revealdynamicsnotencompassedbyothermeasurements. COMBINATIONS Thestudent/analyst/investorisencouragedtosearchformeaningfulcombinations ofchangesintheseratiosthatprecedefuturegrowth.Thisauthor’scombinationsareonly suggestedbecausetheyworkedovertheentiredecadeofthe1990s.Whilesentimentswere “bullish”duringthoseyears,someoftheseleveragestatesproducedpaltrygainsifany. Othersbeatthemarketbyasubstantialmargin.Moreover,someofthesecombinations occupiedahighpositiononthe“efficientfrontier”,suchthattheleastamountofriskwas combinedwiththegreatestamountofreturn.Sincethemarketchanges,eachofthese stateswillbevaluedsomewhatdifferentlyineachbusinesscycle.Asanexample,consider whatmighthappenifoneincreasesAssets/Capitalwhentheyieldcurvebecomesinverted andshort-termborrowingratesareabovelong-termrates.Anyalternativefinancingdone withshort-termdebtactuallybecomesmoreexpensiveandraisesthecostofcapital.The marketthenendsuprevaluinganyleveragestatebasedonthisratio.Inessence,investing onthebasisofaleveragestatewithoutdoingpriorresearch,andevaluatingthecontextof themarket,willcreateanuncoordinatedresponsetocurrentconditions. 525 Table21-1 LTD/CAP FINANCIAL LEVERAGE RATIO OPERATING MOMENTUM ASSET/CAPITAL 1.INCREASE + + 2.INCREASE + - 3.INCREASE - + 4.INCREASE - - 5.DECREASE + - 6.DECREASE + + 7.DECREASE - + + 8.DECREASE - - + 9.DECREASE - + - 10.DECREASE - - - Assets/Capitalbecamealegitimatecategoricalvariablewhenfirmsdecreasedtheir proportionoflong-termdebtandfinancialleverageratiobecauseitindicatedasubstitute methodoffinancing.Infact,withthe1990sdata,itsuccessfullypredictedyearlycapital outlayswhenitwasincreasedwhileoperatingmomentumandtheleverageratioswere decreasing.However,extrapolationintootherperiodsrepresentsaspecificdanger;assets tocapitalwillactdifferentlywhentheyieldcurvechanges. THEMECHANISMOFLEVERAGESTATES Whenwecangrowacompanybyincreasingbothdebtandequityand simultaneouslydiminishtheprobabilityofdefault,wewillmaximizethetaxbenefitsof debt,withoutincreasingthecostofbankruptcy.Suchascenarioischaracterizedbyan increasingEVA,andisperpetuatedbylowerinterestrates,alowercostofequity,orboth. Thefoundationsforthismechanismarethefollowing: • CHANGESINBETA:BothHamadaandtheMiller/Modiglianiteamhadestablished thatbetachangedbecauseofincreasesintheproportionofdebttoequity. Additionally,MandelkerandRheedecomposedbetaonthebasisofoperatingrisk. 526 Thusfinancialriskandoperatingriskwillchangetherisk/returncharacteristicsofa stock. • EARNINGSPRESSURE:Anytimeearningsacceleratesfasterthanthatofotherfirms inthemarket,thestockrises,usuallybecauseoneofthetwocomponentsofoperating momentum,EBITorsales,isacceleratingaswell. • CHANGESINTHECOSTOFCAPITAL:Diminishingbetawilldecreasethecostof equity.MoreearningsretainedatthesametimethatLTD/CAPisdecreasingwillgive thefirmamorefavorableinterestrateinadditiontoprovidinginternallygenerated fundsthatwillbelessexpensivefromanaccountingperspective.Fromaneconomic standpoint,thesefundsmustbegaugedbythecomparativerisk/returnprofileofthe market,aseventhis“free”sourceofcapitalcanbecostlywhenthemarketisespecially overheated. • TRENDSANDTHEOPTIMALCAPITALSTRUCTURE:Whenafirmisclosetoits targetcapitalstructure,itsprofitsarenearlymaximizedagainstthecostofcapital, whichismeasuredbychangesinEVA.Reversionoftheleveragestatecomponentsto theirmeansisless,becauseafirmneedstotakelessactiontoachieveanoptimum. Consequently,astrategythatbringsacompanynearitstargetwillbealmostself perpetuating;thegenerationoflargeprofitswillhavediminishedtheneedformore debt,andthefirmwilltrytotakeadvantageofitsequitypositionaslongasdemandfor thestockisbeingupliftedbymoreearnings. Marketinefficienciesseldomoccurforanylengthoftimeandwhentheydo, popularityquicklyeliminatesthem.Forexample,ifItradeonlineandfindthata“double stochasticcrossover”isespeciallyprofitable,itismostprobablethatothershave discoveredit,willbeginusingit,andthatitwillbegintoworklessandlessprofitably;the marketfactorsintheinefficiency.WhenMichaelMilkienstudiedthejunkbondmarketin the1980s,herealizedthatdefaultprobabilitiesdidnotcorrespondtoprices.Hetookthat knowledge“allthewaytothebank”untilpriceanddefaultwere“insync”-which 527 occurredassoonastheinvestmentbecamepopular.Similarly,HarryMarkowitzdid convertiblebondarbitragebeforeinvestorsrealizedthatsuchanopportunityexisted. Again,popularityledtoitsdemise. Sincetheseleveragestatesinvolvechangingdynamicsbetweenleverageandthe market,theyarenotexploitinganinefficiencyperse,onlythedefinitionofthefactors leadingtothehighestcombinationofreturnandrisk.Theyneedtoberedefinedwhenever themarketreachesaturningpoint,usuallybecauseofanimminentdownturn,but sometimesbecauseoftechnicalinnovation(popularizationoftheInternet)andeven becauseofcatastrophicevents.Eightsuchmarket-changingfactorsareasfollows: • 1.InterestRateLevelsandChanges • 2.YieldCurveChanges • 3.LiquidityPreference • 4.Legislation-Newtaxlaws,SarbanesOxley,etc. • 5.ForeignMarketRisk • 6,ExchangeRateRisk • 7.InflationRisks:BothcurrentandExpected • 8.ConfidenceinEquities Asanexample,afteradownturn,moredebtforacompanywithastrongbalance sheetmightlowercapitalcostsbecauseinterestrateswouldhavebeenloweredbythe FederalReserve.Inthatscenario,aleveragestatethatincreasestheproportionofdebt andyetincreasesoperatingmomentumaswell,mightbebetterthanonethatincreasesthe proportionofequity.Stockwouldnotbeinhighdemandatthispoint,norwouldretained earningsbehigheither.Infactthemarketmightevendowngradeafirm’sprospectsfor nottakingonenoughrisk.Wheredoesthatleavecompaniesthatmustfinancewithall equity?Becauseoftheirhigherbetas,suchfirmsmustdependmoreonearningswhich wouldberelativelylowimmediatelyafteradownturn.Thesefirmsmustdoalltheycanto eliminatehigheroperatingriskbyexpandingcustomerbasessothatthereislesssingular 528 dependenceonasmallgroupofcustomersorvendors.Suchactionswilllowerriskwithout changingthefundamentalnatureoftheirbusinesses-usuallyahighprofitalbeithighrisk andlowdebtscenario. MATCHINGTHELEVERAGESTATETOTHEBUSINESSCYCLE Thefollowingtableindicatestendenciesandnotabsolutes.Therearefirmswho takecontraryactionstothebusinesscycleandalsodowell. Table21-2 ECONOMICSTATE RECOVERY EXPANSION MARKETPEAK CAPMVariables CAPMVariables CAPMVariables Market-Low Market-Medium, Increasing Market-High RiskFreeRate-Low RiskFreeRate-Medium, Increasing RiskFreeRate-High Beta-Low Beta-Increasing Beta-Highand Decreasing Debt-Medium Debt-Highand Increasing Debt-Highand Decreasing FavoredIndustries FavoredIndustries FavoredIndustries InterestSensitive SecondaryGoods CapitalGoods ConsumerDurables Transportation OperatingRisk-Low OperatingRisk-Medium, Increasing OperatingRisk-High, Decreasing FinancialLeverage FinancialLeverage FinancialLeverage High MediumtoHighand MediumtoLowand Increasing Stabilizing Decreasing Thecapitalassetpricingmodel(CAPM)offersasimplisticbutsometimesvalid explanationfortheconformityofsomefirms’capitalstructurestothebusinesscycle.Ifwe 529 canacceptthatbetaincreaseswiththeproportionofdebtinafirm’sstructure,andthat theFederalReserveraisesinterestratesastheeconomyimproves,wecansurmisethatthe greatestamountofdebtshouldbeincurredwhenbothinterestratesandbetaareatlow points.Whenthefirmdoesthegreatestamountofdebtfinancingatthebeginningofan expansion,itisincurringtheleastamountofriskforthegreatestamountofpotential return.Asbetariseswithmoredebt,thefirmcantakeadvantageofrisingstockprices.At thispoint,earningswillbeacceleratingfasterthanthecostofequityandcapitalwillbe flowingintothecompany.AstheFedraisesrates,debtfinancingbecomeslessandless attractiveandthefirmbuildsequitythroughhigherretainedearnings,andsometimesnew stockissues,whichhavebecomemoreattractivewithincreasedearningspershare. Decreasingtheproportionofdebtinthecapitalstructurewilllowerbeta,whichisexactly whatthecompanyneedstodo:bythistime,theeconomyhaspeakedandbeginsto plateau,andearningshavebeguntodecelerate.Althoughthestrategicallyastutefirm’s betahasdecreased,boththerisk-freerateandthemarketratehaverisen,creatingmore risk.Eventually,thecostofcapitalrisesfasterthanthevelocityofcollectiveearningsand theeconomygoesintoadownturn.FromaCAPMperspective,themarketriskpremium willbegintofall,asmoremoneyflowsintobonds;stockswillbegintoofferbothmorerisk andlessreturn,althoughdividendyieldswillrise.Themarketactslikeabarometerfor earningsexpectationsastherateofchangeofearningsbecomeslessthantherateofchange forthecostofequity,andstocksbecomeoverpriced. Itisasdifficultforanycompanytofallintothisstrategicimperative,asitisfor investorstoprofitfromit.Theideaofmoving“fund“moneyintoutilitiesandconsumer staplesduringadownturn,isdirectlyreflectiveofthosesector'scharacteristicallylow betas.Thesefirmshavelowoperatingleverageandsteadydemandeveninabadeconomy, andofferareturnthatexceedsthatoftreasuries.Acting,likeamicrocosmofthelarger company,theinvestorprofitsbyshiftingassetsintoareasoflowerrisk.Whilethat 530 techniqueoffersbothentities(thecompanyandtheinvestor)animmediatereward,itisnot asconducivetoloweringriskasisdiversification. PROBABILITYANDDIVERSIFICATION Mostacademicshaveahealthyskepticismaboutbeta.Infact,thisbookoffers severalalternativecalculationsforthecostofequitysimplyforthepurposeofanalytical diversification;ifallmethodspointusintherightdirection,theprobabilityofbeingwrong ismuchless.OnecanneverforgettheBlack,JensenandScholesstudyofthelatefifties, earlysixtiesthatfoundhighbetastockstobenegativelycorrelatedwithreturn,whilelower betastockscompensatedinvestorsmuchmorethantheirinherentriskswarranted.Since thetheoreticalidealencompassesa“riskreturntradeoff”,betawasdroppedbymanywho wereontheCAPMbandwagon.Mathematically,eachcompanyhasadifferentcorrelation coefficienttothemarketindex,whichmakesbetaanissueofindividualresponseandnot collectivedynamics-althoughtheCAPMsometimesmakesperfectpredictionsaboutthe market.Whenthecorrelationcoefficientissmaller,betawilltakelesspreeminencethan “Alpha”,whichwillexplainsucheventsaslawsuits,favorablelegislation,andeven industrycollusion. Thedownwardbiasofhighbetastocksshouldbeviewedasbothawarningandan opportunityforinvestorsandcorporationsalike.Mostofthestocksinthetechbubble werehighbetastocksandyetthelowbetasstocksroseaswellduringthatperiod.Whena largecompanyisgrowingbyacquisitions,itcanminimizeitsmarketvolatility,but increasethe“Alpha”componentandrespondwithlargepricejumpsatthetimeofthe purchases.Suchactionswillkeepbetatoaminimumandyetpotentiallyincreasethestock priceatarategreaterthanthemarket.Secondly,whenacompanydiversifiesits operationsintootherrelatedareas,itcanexpanditssalesandlowerits“unlevered”beta, creatingalowprobabilityofdefault.Infact,acquisitionsareoftenimplementedforthis veryreason:Thereductioninriskisderivedfromalowercovariancebetweensalesand themarket-whichisatheoreticaldeterminantofbeta. 531 SALESANDBETA Ahighermeanincreasecoupledwithalowerstandarddeviationshouldbethesales goalofanycorporation.Steadiersaleswillincreaseprofitsallowingmoredebttobe incurredwheninterestcoverageincreases.Forallequitycompanies,moreretained earningswillbeproduced,andfundsflowingintothestockwillfollowasteadierpattern. Ineffect,acompanycansacrificesomeprofitmarginforassetturnoverandproducethe samereturnonassetswhileloweringrisk.Ofcourse,mostcompaniesdonotwanttogive uptheinherentpricingpowerthatcomesfromhigherprofitmarginsandsotheysettle themselvesintoa“niche”wherethereislittlecompetition.Sincetechnologyisagreat “equalizer”andoftenremovesniches,theimportanceofsalesdiversificationisself-evident. ThecornerstoneofMandelkerandRhees’sdecompositionofbetawasthestrong relationshipofacompany’ssalestothestockmarket.Toreviewthatconnection,recall thatthreemainfactorswereinvolved: • 1.Thehistoricalreturnonequity,ROE. • 2.ThecurrentperiodTotalLeverage,determinedbymultiplyingthepureform operatingleverage(operatingmomentum)bythepureformfinancialleverage:(% ∆ ∆∆ ∆OperatingIncome/% ∆ ∆∆ ∆Sales)x(%∆Ν ∆Ν ∆Ν ∆ΝetIncome/ % ∆ % ∆ % ∆ % ∆OperatingIncome) • 3.Themostrecentcovarianceofthereturnonthemarketwithcompanysales,divided bythevarianceofthemarket:COV(Sales,Market)/Marketσ σσ σ 2 ThefinalequationismerelyROExTotalLeveragexCOV(Sales,Market)/Market Variance. Thelastfactor(3)istheprinciplecomponentofbothsectorrotationinthegreatermarket anddiversificationwithinthecompanyitself.Forexample,inthelate1990s,techsales wereprizedmorethannaturalresourcesalesandproducedbothagreatercovariancewith themarket,andlargeamountsofcollectiveoperatingleverage.DuringtheY2Kfallout, theeconomysoaredbecauseoftech“upgrades”,butbecamemoreriskybecausethemore volatiletechsectorbecamealargerproportionoftotalGDP.Onelookatthebasic 532 structureofcovariancecantellusasmuchasmostjournalistichistories:Allthree componentsofcovarianceroseatthesametime. • 1.“R”:Pearson’sCorrelationCoefficient::Themarketcorrelationwithcompanysales soaredastechbecamemoreprominentandtheInternetbecamepopular. • 2.Thestandarddeviationofthemarketreturn:Asinterestratesrose,thefargreater returnonthemarketproducedahighriskpremium.Moneyflowedintoequities,but wasofteninvestedina“promise”ratherthanrealassets.IPOs,startups,andbiotechs thathadnoearningshistoryreceivedlargeamountsofcapital. • 3.Thestandarddeviationofthechangeincompanysales.Mostfirmsenjoyedsteady salesthroughoutthe1990s;theirriskremainedlow.Unfortunately,theprominenceof techsalesinducedmoreoperatingriskintotheeconomyandacollectivelyhigher standarddeviationforsales. Thecovarianceisthuscomposedof“R”multipliedbythestandarddeviationofthechange incompanysales,multipliedbythestandarddeviationofthemarketreturn.Thestrategic implicationsofthisequationarevast.Althoughfewcompaniescantimethemarket, financialmanagementdoestrytotakeadvantageoflowinterestratesoranaboveaverage demandforcompanystock.Duringanexpansion,thereislessriskfromraisingboth leveragecomponents,andbetawillrisewiththemarket.Asthemarket“overheats”, companiescanexpandcustomerbases,andsimultaneouslylowerbothoperatingleverage andthestandarddeviationforsales.Atthesametime,theycanlowertheirproportionof debtinthecapitalstructure.Betawillnaturallydropwithlessrisk,butweneedto rememberthatitwillalwaysbewithintheboundsoftheindustry.Ahighbetabiotechwill neverdropdownto0.5,iftheinherentbetais1.5.Thus,unlessthecompanywantsto changeitsbasicnatureandbecomeaconglomerate,itwill“alwaysbeatthemercyofthe market”. Oneotherstrategythatismoreesoteric,wouldbetoexpandthealphacomponent intheregressionequationandminimizethecorrelationcoefficient,“R”.Forexample,a 533 fertilizercompanythatisregisteredontheNewYorkStockExchange,butprimarilymines batguanoinAfrica,andsellsittotheMiddleEast,mightbesomewhatbufferedfromthe shocksoftheAmericanmarket.Althoughmarketsaremoregloballyconnected,unusual circumstancessuchasprotectivelegislation,willimmunizeacompanyfrombeta.Another exampleoflowcorrelationoccurswhencommoditypricesincertainindustriesrise contrarytotheeconomy;executiveswilljugglethestrategicadvantagesofthese associationstoinsulatetheircompaniesfrommarketvolatility. Theconnectionbetweentheindividualinvestorandcorporatefinanceiscemented withbeta.Themarketfavorsasectorbecausesalesgenerateincomeatamorerapidrate thantheoverallmarketcanproduce.Individualinvestorsdirecttheflowofcapitaltothese companiesbecausetheyofferthehighestreturnfortheleastamountofrisk.Ona corporatelevel,theindustriesinthefavoredsectorsareoptimizingtheircollectivecapital structuresbychangingtheriskcharacteristicsofleveragetolowercapitalcostsandraise profits.Betaintheseindustriesmaybequitelow,buttendsto“overreact”topositive changesinthemarket.Ultimately,salesaccretionthroughinvestmentinprojectsthat balanceriskanddiversifyoperations,willexpandthecustomerbase,andcreatea favorablebeta. PROBABILITYANDANTICIPATION “Beta”andtheCAPMhaveatheoreticalbasethatisnotalwaysapplicable.The marketcanchangeinunanticipatedwaysthatcanleaveanalystsscratchingtheirheads. Withmorevolatilitycomesincreaseduncertaintyandfunctionsthatmeasurepast performancearesimplyinadequateasforecastingtools.Eventhemarginalbenefits equationandEVAareex-postmeasurementsofmovementtowardtheoptimalwithout referencetothefuture.Whilewecancreateanongoingcompendiumofcorporateresults andchecktoseeifthecompanyisdoingwhatitneedstodo,dataisreceivedafterthefact. Undeniably,asystemthatanticipateslargecapitalmovementswouldbesuperiorto “chasingearnings”.Ifanalysiswereaseasyassightinglargecapitalinflowsand 534 calculatingpayofftimes,boththeriskandreturninthemarketwouldbeminimized; stockswouldbeliketreasuries,andprofitswouldbelikeclockwork.Althoughleverage statescanpointusintherightdirection,themarketwillvalueeachstatedifferentlyat differenttimes,andweneedtogaugethelargereconomyaswellastheindividual company. Bycheckingtheleveragestatesofthemostfavoredsectorintheeconomy,wecan gainsomevaluableinformation,especiallyifthesectorfallsintoapattern.Forexample,if mycompanyistakingonequitywhilethemostfavoredsectoristakingondebt,Ineedto lookatmycompany’sresponsetointerestrates.Ifmycompanyhasanall-equitycapital structureandeschewsdebt,Iobservetheoperatingmomentumofthefavoredsectorand askquestionssuchas,“Isoperatingmomentumdecreasingonthebasisofmoresaleswhile operatingmarginsincrease?CanIanticipateahigherEVAbecauseequitygainsaremade fromretainedearningsandnotfromstockissues?Insomemarkets,leveragestatesinthe topsectormayformnodiscerniblepatterns:companiesthatincurlargeamountsofdebt maybeincreasingtheirstockpricesjustlikethosewhoissuelargeamountsofequity.In suchachaoticsector,itwouldbebesttoavoidinvestmentunlessoneintendsto“getout fast”,becausethesefirmswillbe“outofsync”withthecapitalmarkets,andasurgein profitsmaybefollowedbyacaseofclassic“buyer’sremorse”. PRINCIPLECOMPONENTSANALYSIS Thediscoveryofleveragestateswasbaseduponpatternrecognitionfromseveral setsofdata.Inthesamewaythatagamblerchecksoffhorseracingcharts,leveragedata wasfilteredthroughaseriesofplusesandminusesbeforeitwassubmittedtomore sophisticatedtesting.Thedatasetexhibitedabalancedarrayofpatternsthroughwhat statisticianscall“principlecomponentsanalysis”.Thistechniqueestablisheswhichlinear combinationsofdatawillcreatethegreatestamountofvariationandidentifiescomponents usinganEigensystem.Eachcomponentisgivenaweightasapercentageofthewhole. Thegreatestweightisnotthemost“important”component(althoughitoftenis),butthe 535 elementcausingthemostvariationinthedata.Forexample,thefollowingchartappliesto thedatainthecorrelationstudies: Table21-3 VARIABLE EIGENVALUE PERCENTAGE OperatingIncome 2.4335 60.8% Assets 1.0555 26.4% Capital 0.39864 9.97% Sales 0.1124 2.81% Weknowthatnooperatingincomecanexistwithoutasale,sothepaltry2.81% contributedtothevariationbysalesdoesnotdiminishitsimportance.Wemerelyclaim thatanycategorizationincludingtheoperatingincomevariablewillbeheavilyinfluenced byit.Infact,withmostcombinationsofthreetofourvariables,onevariablewillbe preponderant,andtheotherswillnotcontributeanequitableshareofthevariation. However,thefollowingtableappliestotheleveragestatevariables: Table21-4 VARIABLE EIGENVALUE PERCENTAGE LTD/CAP 1.1687 29.2% FinancialLeverageRatio 0.99438 24.9% Asset/Capital 0.95529 23.9% OperatingMomentum 0.88154 22.0% Theseratiosareveryclosetoeachotherandaddsimilaramountsofinformation.They workwelltogetherasa“team”becausebothnumeratoranddenominatorcontain interconnectedpiecesofdatathatcontributetothewhole.LiketheROC/ROEDuPont equationsfromwhichtheywerederived,theleveragefactorsareflexibleandeasyto manipulate,containingcommondependenciessuchasfixedcosts,interestexpenseand operatingmargins.Whileequalityineffectivestrengthdoesnotimplythattheseratioscan 536 categorizeleverage,theydoformameasuredinterfacebetweenprofitabilityand investment,albeitanimperfectone. STATICVSFORWARDLOOKINGRATIOS Makinginvestmentdecisionsbasedonpastdataisalmostalwaysdoomedtofailure. Infact,theefficientmarketshypothesisevolvedforthisveryreason-enoughmistakes havebeenmadewhileobservingsalesandearningshistoriestodecipheracausation.We termex-postmeasurementstobe“static”,notasanegativeconnotation,buttodeclarethat moreinformationisneeded.Forexample,thatnetincomeincreasedtwentypercentlast quarterisa“static”measurement.Nowiftheanalystproclaimedthatnetincomeis anticipatedtochangeatarateofthreetimesthemarketinthenextquarter,thatwouldbe aforward-lookingmeasurement.Unfortunately,suchproclamationsmustbeaccompanied byacaveat,andaresometimesoutrageouslywrongaswell. Theleverageratios,particularlyoperatingmomentumandthefinancialleverage ratio,areinherentlyforward-looking.Thederivationofoperatingmomentumincreasesis basedonthreeyearsofdata,andimpliesahighprobabilityofaconsistentchangein operatingmargin,especiallywhenmomentumincreases.However,operatingmomentum byitselfistoovolatiletouseasapredictivetool.Bothincreasesanddecreasescanleadto moreoperatingriskandthemeasurementcanincreasetenfoldwhenacompanyturns aroundfromabadyear.Nevertheless,operatingmomentumhasastrongtendencyto reverttoitsmeansimplybecauseitisindicativeofthetypeofcoststructureinherentinthe industry.Ifitdecreaseswellbelowitsmeanfortwoyearsinarow,theprobabilityofan increaseisfairlyhigh.Naturally,ifanalystswillgivetheinvestorasalesforecast,then operatingincomehasatleastsomepredictability.Infact,asmentionedinprevious chapters,operatingmomentumwillmimictrueoperatingleverageinthenextyear,aslong ascostsarestable;thatis,variablecostsremainastablepercentageofsalesandfixedcosts donotchange.Analogously,thefinancialleverageratiowillmimic%∆ ∆∆ ∆NetIncome/%∆ ∆∆ ∆ OperatingIncomeaslongasinterestexpenseremainsconstant.Onatheoreticalbasis,as 537 longascostsandinterestexpensewerestable,netincomewouldbetotallypredictablefrom changesinsales.Weknowthataprecisenumberisnotdeterminablebecauseofvolatility inbothcostsandleverage,butthattheflowofchangetendstoreverttoitsmean.Thus,if weviewtotalleverageastheproductofoperatingmomentum,andtheoreticalfinancial leverage,ourequationwouldbe:(% (% (% (% ∆ ∆∆ ∆OperatingIncome/ // /%∆ ∆∆ ∆Sales)x(%∆ ∆∆ ∆NetIncome/ %∆ ∆∆ ∆OperatingIncome)=%∆ ∆∆ ∆NetIncome/%∆ ∆∆ ∆Sales.Inevitably,ahighlyvolatile electronicscompanywithnodebtcanhavethesametotalleverageasastaidchemical companythatchoosestofinancewithtoomuchdebt;thetotalriskwillbeimplicitineach company’sstockprice.Theconnectionbetweentotalleverageandthefirm’scapital structurewillbereflectedbythreeaspectsoftheratio,%∆ ∆∆ ∆NetIncome/%∆ ∆∆ ∆Sales: • 1.Thesize(level)oftheratio • 2.Themagnitudeofitscomponentparts(netincomeandsales) • 3.Thevolatilityofnetincomeandsalesasmeasuredbytheirrespectivestandard deviations. Toexaminereversiontothemean,wewilluseEnescoasanexample:Thisbranded giftcompanyhadastormyeconomicrelationshipwiththe1990s,andwasnotamenableto prediction.Tobegin,weobtainafiveyeargeometricaverageofbothsalesandnetincome usingthe“growthtrend”technique.Thisisourtarget“guesstimate”,whichdoesnothave tobeprecise;infact,wecanevendropayearthatwesightas“notinthetrend”andadd anotheryearinitsplace.Thefiguresinthetableareabsolutevaluesfromtherespective yearandnotpercentagechangesbecausethegeometricaverageisimplicitlydefinedbythe changebetweenperiods. 538 Table21-5 ENESCO 1996 1997 1998 1999 2000 2001 NET INCOME 38.4 10.5 -22.4 26.9 15.1 1.1 SALES 845 476 451 384 325 269 Ratherthanchoosetherecessionyearof2001astypical,weoptfor1996to2000becauseit ismoreintrend.Separatespanscanevenbechosenforeachratioaslongaswekeepthem intrend.Thegeometricgrowthratewillbetheratioofthelastyearinthetrenddivided bythefirstyearandthen"exponentiated"bytheinverseofthenumberbetweentheyears -four.Fornetincomethemethodyields,(15.1/38.4) .0.25 =0.791883.Notethattheinverse offour,thenumberofyearsbetweenthefive-yearinterim,is1/4or0.25whichisthe exponent.Finally,toobtainapercentagerateofgrowth,wesubtractonefrom0.791663to obtain-0.2081or-20.81%. Forsales,theratiois(325/845) 0.25 =0.787511.Subtractingoneyields-21.25%. Thefinaltotalleverageratiois-20.81/-21.25whichisequalto0.97929.Itisthe“nature” ofEnescotomovetowardthisnumberfromeitheraboveitorbelowit,mostlythrough changesinoperatingmomentum. AlthoughEnescowasanextremelyvolatilecompany,wewilluseanearlierperiodto checkourcalculations,andadditionally,wewillusethefinancialleverageratio (EBIT/EBIT-I),andnotthetheoreticalratiotoderivetotalleverage. 539 Table21-6 Years FinancialLeverage Operating Momentum TotalLeverage 1990 1.0607 -1.59 -1.6865 1991 1.0614 -1.67 -1.7679 1992 1.0388 1.11 1.1549 1993 1.0227 -1.05 -1.075 1994 1.0232 -0.406 -0.4154 1995 1.0845 2.642 2.8658 1996 1.1058 -5.035 -5.568 1997 1.2305 1.375 1.692 Exceptfor1991,eachyearsawamovetowardthemeanof.979,althoughonlyoneyear, 1992,cameclosetoit.Thequestionfortheinvestorandanalystis:givenalevelof financialleverage,whatoperatingmomentumwouldachievethemeanleveloftotal leverage?Toanswerthisquestion,wewilluse1992asanexample.Thecompanyachieved aratioof1.11withanoperatingincreaseof5.32%versusasalesincreaseof4.79%.The entirebasisofcapitalstructurestrategyhypothesizesthatthestockcouldhavemovedup onthesepaltryfiguresaloneprovidingthattheywereconsistentenoughtoallow incrementalincreasesinfinancialleverage.In1992,hadanalystsgivenaperfectsales forecastof4.79%,ourbestextrapolationwouldhavebeen.979/1.0388=0.9424 Multiplyingthatfigureby4.79%equals4.51%.Theactualoperatingincomeroseby5.32 %.Partoftherisk/returnstructureencouragesinnovativemanagementtotakechances- suchasincreasingfinancialleveragewhentotalleverageiswellaboveitsaverage. However,innormalriskenvironments,theimperativemovementisreversiontothemean 540 simplybecausethismeasurementdefinesthenatureofthecompany.Sincevendors provideaspecificcoststructureandcustomersexpectaspecificpricestructure,companies mustworkwithinlimitationsandtheresultwillbeanattempttokeepthisratiostable. Inevitably,thebestruncompanieswilltrytostructuretheirleveragetostayasclosetothe meanaspossible,whichlendsanairofpredictabilitytothisvolatilecalculation. THEQUICKPAYOFF Certainly,themorerapidlycapitalisturnedintoprofit,themorequicklyinvestors getpaid.Whilewecanmeasurethehistoryofcapitalturnover(Sales/Capital),andeven salestolong-termdebtratios,eachmarketisdifferent.Often,thebestprospectwillbea companythatiswellabovetheindustryaverageinLTD/CAP,butmovingbacktowardsit. Averageanalystswillnotpickuponthemovementandwilldismissthefirmas“debt ridden”-whichitis-onpaper.Atothertimes,anincreaseofallfourleverageratioswill provideahighriskleveragestatethatisworthyofinvestment.Inthelattercase,whenan investorisputtingcapitalintoafirmthatisincurringdebt,heorsheistakingahigher risk,andishopingforaswitchtoaleveragestateinwhichthecompanyistakingonequity fromhigherearnings.Thus,morereturnisoftenculledfromthese“debtstates”,simply becausetheoddsof“losingone’sshirt”aregreater.Unfortunately,thereisnomagic formulathatwillguaranteeaquickpayoff,butiftheinvestoroperatesinthecapacityofa quasi-ratingsagency,andusesdefaultprobabilitiestoevaluateafirm,downsideriskwill bereduced.Inessence,theinvestordifferentiatesbetweena“weak”companywhomust takeondebttomaintainbasicproduction,anda“strong”companywhousesdebtasa strategicimperativetofundprojectsandoptimizecapitalstructure.Besidesexamining marginalbenefitsanddefaultprobabilities,severalothertenetsneedtobeadheredto: • Operatinghistoryshouldbestrongenoughtogetapreferredinterestrate. • Salesneedtobeconsistent. • Thedebtissueshouldbelargebecauseitislessefficienttogotothecreditmarkets frequently.Moreover,asmallamountofdebtismoreindicativeofafirmthatis 541 lookingtofundexistingassetsandnotpurchasenewassets.Additionally,small amountsofdebtcanindicatecostoverruns. • Increasingoperatingmomentumwillleadtoahigheroperatingmargin-whichis neededtooffsethigherinterestexpenses. • TakingondebtwhentheLTD/CAPratioisrisingwhilethefinancialleverageratiois decreasingleadstoriskyimbalancesthatmustbecloselyexamined.These“transition states”canbeprofitablebutshouldbeavoidedifalargeinterestexpensewillbe incurredinafewyears. • Increasesintheassets/capitalratioaregenerallyapositiveindicationwhenlong-term debtisincurred.Suchanincreasecanbeasignofgreatervendoractivity,alternative financing,tradecredit,andagreaterreturnoncapital. Thestudentwillrememberfromearliersectionsthatmostfirmswillmatch expectedcashflowswiththetypeoffunding.Thisriskadverseapproachallowssteady inflowsofincometobebalancedbysteadyoutflowsofexpenses-includingsinkingfund paymentsandinterestexpenses.Therefore“long-termprojects”areoftenfundedbylong- termdebt,andthatmaybeaeuphemismforanacquisitionthatisnotexpectedtopayoff immediately.Synergiesinoperationsmaytakeafewyearstoaccomplish,buttheinvestor doesnotwanttowait.Usually,alargepurchasewillbeaccompaniedbyenough “hullabaloo”toincreasespeculationinthestock,whiletheinfusionofdebtgivesan immediateboosttotaxbenefits.However,subsequentyearsmaybelean,withunder performanceofthemarket.Atthispoint,theinvestorhasoneoftwochoices:check analysts’opinionsabouttheacquisitionandcoordinatethemwiththefirm’spresent positioninthebusinesscycle,orleavethestockentirelyifitappreciateswildlyinthefirst yearafteramajorpurchase;sucha“run-up”maybeaccompaniedbyasell-off. BARRROSENBERGANDRESPONSECOEFFICIENTS RiskmanagementwasrevolutionizedbyBarrRosenberg.Thefounderofthefirm, Barra,Rosenbergformulatedwhathetermed“relativeresponsecoefficients”whichwere 542 weightedvaluationsbetweenanindividualstockandnumerousmacrovariablessuchas inflation,oran“energycrunch”.Asanexample,ifGeneralMotorsrespondstoanoil embargofivetimesasmuchastheS&Pindex,theirresponsecoefficientwas“five”.Barr Rosenbergtheorizedthatbetawasessentiallycomposedoftwoelements: • 1.Theproportionalcontributionsofmajoreconomiceventstomarketvariance • 2.Therelativeresponsecoefficientoftheindividualsecuritytotheevent Whenthecoefficientswereweightedbytheoverallimpactofeachmajorevent,a “fundamentalbeta”wasderived.Themeasuretowhichtheeconomicimpactofeventscan beexplainedisthesamemeasuretowhichrelativeresponsecoefficientscanbeinterpreted -changesinbeta.Therefore,toRosenberg,thetheoryoffundamentalbetaentails changingcapitalstructureasaresponsetochangingeventsintheeconomy,whichfurther reflectglobalconditions. Changesininterestratesandthepotentialforcreditdefaultweighheavilyonmost companies.Inthecaseofdefault,evenpsychologicalfactorsmayintertwinewithcapital fundingtodetermineatargetstructure.Asanexample,acompanymaynottakepartin financingwithlowerinterestratesafterarecoverybecauseitobservesominoustrendsin fixedassetacquisitioninthenear-term;technologymayberapidlyreplacingthestandards intheindustry,andthefirmdoesnotwanttoinvestinobsoletemachinery.Bynot financingwithlowercostdebt,thecompanycaneitherunderfundprojectsorunder performthemarketbecauseitisnotminimizingcapitalcosts.Thus,extraneousfactors countagreatdeal. Rosenberg’sfundamentallyderivedbetasdidoutperformhistoricallyderivedones inpredictionmodels,buttheinformationanddatarequirementswereonerous:oversixty ratiosinsevendifferentcategorieswereused.Whiletheaverageindividualstockisabout thirtypercentcorrelatedwiththemarket(R 2 =0.3),Rosenberg’sderivedbetasimproved thepredictivepowerofthemeasurementtoforty-fivepercentwithnomarketinformation 543 whatsoever.Whenmarketinformationwasadded,thepredictivepowerwentupeighty-six percent. Rosenberg’s“fundamentalbetas”aresimilarto“leveragestates”becausetheyare drivenbyaprobabilisticresponsetoworldevents.However,theresponseoftheleverage statesismoredependentoncapitalmarketsandismorederivative.Additionally,leverage statesarecharacterizedbytheireconomy,while“fundamentalbetas”arecomprehensive. Asinthecapitalassetpricingmodel(CAPM),thenumberofvariablesusedinforminga leveragestateshouldbekepttoaminimum.Onlywithafewkeyvariablescanaleverage statebebothflexibleenoughtocoveravarietyofconditions,buteconomicalenoughto specifyacapitalstructuretype.Sincecorrelationsinanymarketchangefrequently,the investorneedstobeawarethattheinteractionbetweenvariablesalsochanges.Leverage statesattempttoestablishacategorythatisstrictenoughtokeepthatchangetoa minimum,whiledefiningabroadertrendincapitalstructure.Analyzedinthecontextof thethreeCAPMvariables,eachtakingonaleveloflow,mediumorhigh,leveragestates wouldyieldmyriadcombinationsSincechangesintheCAPMandtheoveralleconomy aresomewhatimplicitintheleveragestates,theaverageinvestorshouldnothavetoresort tosuchextensivecrossreferencing. (BacktoTableofContents) 544 22 TECHNICALANALYSISANDCAPITALSTRUCTURE Academicshavepilloriedtechnicalanalysisformanyyears,claimingthatitwasthe financialversionof“snakeoil”.Theyportrayitasmathematicalillusion,amethodof rationalizingaprobabilitythatisalreadyheavilyinthemarket’sfavor.Nevertheless,Wall Streethascontinuedtoshelloutmillionstoresearchfirmsthatproclaimtheabilityto chartandpredictthefuturedirectionofastock.Thegistoftheargumentrevolvesaround whetherchartingandtechnicalindicatorsmerelygiveafalseinterpretationtotherandom variationthatoccurswhenalargeforcepropelsanobjectinaspecificdirection.For example,wecanthrowamarbleacrossahardsurfaceandwecancontrolthespeedand directionofthetoss,butwewillnotknowtheexactlocationofwherethemarblewillend up.Ontheotherhand,technicalanalystswillclaimthatwithenoughinformationabout thevarioussubvectorsandvelocityofthemarble,amathematicalconclusioncanbe drawnaboutthespecificendlocation.Academicswillcounterthatsuchinformationis concurrentwiththemovementandisimpossibletoobtain-apriori.Ifthisseemslikea religiousdebatewhereweneedtochooseonesideortheother,considerthefactthatWall Streetemploysmanytechnicalanalysts,andthatsolidfinancialdecisionshavebeenmade withtheirrecommendations.Infact,bothargumentshavemeritsimplybecause“the market’makesroomtoaccommodatethem. Attimes,thestockmarketcantakeonaneeriecoherence;stockswillfollowa smoothtrendwithoutmuchdeviationandinvestorschooseasellpointthatyieldsatwenty percentprofitbeforeamovingaveragebeginsdescending.Justasfinancialmarketsbegin to“makesense”,anuglythinghappens:theDowdropsbytwoorthreehundredpoints. Thevolatilityindicatorthat“workedsowellbefore”isnolongeroperable.Thestock breaksthroughits“supportline”andkeepsmovingdownward. 545 Academicsliketocomparerandomvolatilitytoacointoss.Ifwetossacoin hundredsoftimes,andputanuptickmarkonacharttoindicate“heads”andadowntick markfor“tails”,wewillcreateapicturethatlookslikeafacsimileofastockchart.It trends,itrapidlyrises,anditmakesesotericpatternslike“breakouts”and“triangles”. Thosewhobelieveinthe“efficientmarketshypothesis”believethatpredictionfrompast performanceisimpossibleandalludetothe“randomwalk”ofthiscointoss;theybelieve thatthemarketfactorsinallinformationconcerningasecurityasitmovesupanddown. Infact,thereiscredencetothistheorybecauseevenasfarbackasthelate1950sandearly 60s,computerscientistswereshowingthatrepeatedpatternscouldnotpredictthefuture directionofprices. Severalquestionsneedtobeaddressed.Iftheacademicsarecorrect,howcanmy back-testingofamovingaveragebesosuccessful?HowcanWallStreetchartistsmakethe rightdecisionswithmillionsofdollarsontheline?Ifstockpickingismerelyfollowing trendsbetween“resistance”and“support”lines,whywouldn’titbestandardpracticefor everyone?Theanswerstothesequestionsinvolveacombinedknowledgeofthereasonsfor statisticalvariation,andthefinancialforcesthatmoveastockinoneparticulardirection ortheother. MAJORFORCES Whenacombinationofhighproductdemandcoalesceswithlowercapitalcosts,a competitiveadvantageexiststhatisusuallysectorwide.Ordinarily,acompanywillmove towarditsoptimalcapitalstructureatthistime,andcomparativelyhigherearningsare generatedfromsixmonthstotwoormoreyears.Capitalintheformofademandfor equitywillflowintothecompanyaslongasthecompanymaintainsitscomparative earningsadvantagewiththemarket. Technically,theriseinastockpricewillbemirroredbyariseinashort-term movingaverageabovealong-termaverage.Thatisamathematicalfact.Whatis uncertainisthelengthofthemovingaveragesthatdetectsthestockmovement.Ifthe 546 stockmovementisslight,itwillbedetectedbyashortinterval-evenbyafewhours.Ifthe upwardspressureisgreat,afiftydaymovingaveragewouldbetheshort-termindicator anditwouldmoveabovealongermovingaveragesuchastwohundreddays. AnotheruncertaintyliesinhowlongtheseforcescancontinueIftheupwards pressureisgreatenough,andwedetectastockwhosefiftydaymovingaverageisrising aboveitstwohundreddaymovingaverage,canwepredicthowlongitwillstayaboveit? Theanswerisaresounding,No!Whathappensisthattheupwardspressurecausedby earningscreatesamathematicalillusion:webegintomeasuretheperformanceindicator, thedistancebetweenmovingaverages,andmistakenlybelievethatawiderdistancewill giveusimmunityfromaquickdescent.Infact,moststockswillmakeseveral“runs”well abovethemovingaverage,declinetoapointrightaboveit,andthenmakeanotherquick ascentupwardsAslongasearningsoutpacethemarket,thevariousrunsamountto statistical“noise”.Whenthis“noise”trendsinasmoothpattern,tradersbelievetheycan makemoneyby“gettingout”whilevolatilityislow.Thattechnicalindicatorswillworkin theseperiodsoflowvolatility,addstotheillusionofpredictability,anditappearsthatthe traderis“outthinking”themarket.However,therealmoneyismadewhenthestockis belowitslong-termmovingaverageandisbeginningtoriseupwards.Atthistime,“smart money”,thosewhohaveaninsider’sknowledgeoftheindustry,areanticipatingan investment“payoff’ingreaterproductdemandandhigherearnings,andareinfusingthe firmwithcapital.Thetechnicalequivalentofthisriseisillustratedinaseriesofmoving averagecross-overs,startingoutveryshort-termandgraduallygainingstrengthtowhere thetrendissustainedbylongerperiodlengthssuchasfiftyandtwohundreddays.Atthis juncture,therewillstillbealotofmoneytobemadeifthefinancialforces(sales,earnings, lessdebtetc.)aregreatenoughtopropeltheshortertermaverageabovethelongerterm forasustainedperiodofsixmonthsormore. Obviously,theimperativefortheinvestoristodetecttheforcesthatmovethese averages.Ontheotherhand,tryingtopredictshort-termvolatility,andtakingadvantage 547 ofpricerun-upsonthebasisofhistoricalpatternsisfruitless.Considerthebanksthat sufferedinthe“creditcrunch”of2007:thesebankshadlowbetasbecausetheygenerally madeconservativemovementsbasedonthe“yieldcurve”,thedifferencebetweenlong- termratesandshort-termrates.Banksneedtoborrowshort-termandlendlonginorder tomakemoney.Whenthecurveinverted,bankstockswentupanddownbyasmuchas tentotwentypercentinoneday.Thoseinvestorswhomademoneyduringabrieftwo weekuptrend,wouldsubsequentlyloseitinaone-dayslide. THEBANEOFVOLATILITY Ifresearcherscouldpredictthechangesinthestandarddeviationforasecurity,the marketwouldbefullypredictable.Infact,atmajorfinancialinstitutions,valueatrisk calculations,alongwithsoftwaresuchasRiskMetrics,trytopredictvolatilityusingan exponentiallyweightedmovingaverage-usuallyforoneday.However,suchtechniques areappliedtoassetvalueandnotindividualstockswhichareinherentlymorevolatile.For stocktraders,thebaneofvolatilityisthecauseofuncertainty.Althoughtechnical indicatorscancorrectlyrevealpricingpressureineitherdirection,theyfailmiserablyin tellingthetraderhowlongitwilllast.Forexample,aninvestorcanbuyonehundred sharesofIBMbasedonashort-term“relativestrength”indicator,andthestockcandrop tenpointsafewdayslater.The“new”indicatorwillnowdisplayadownwardbias.Inthe interim(thosefewdays),thestockdidindeedgoup,butthetraderwasn’t“quickenough onthedraw”tomakeaprofit. Asareactiontothevicissitudesoftechnicaltrading,WallStreetdevelopedan adaptive,“survivalist-type”techniquetocircumventvolatility:entertheday-trader.With aframeworkofonlyoneday,thedaytradertriestoeliminatevolatilitybycatchingastock ontheupswing,andthenquicklysellingitbeforeanyuncertaintyarises.Theobjectiveis tocaptureoneday’sgain.Althoughtheday-tradermaymissoutonanyfurther appreciationofthestock,heorsheissatisfied.Thereason?Onlinetradingsitesallow traderstoexchangehugeblocksofstocksforasmallcommission,attemptingtoprofitfrom 548 the“assetturnover”.Thus,iftradershaveenoughstock,theycantradefrequentlyandnot tieupmoneyincommissions.Thistypeofenterprisecanbebothprofitableandnerve- racking,andthequestionneedstobeasked“sowhyisn’teverybodydoingit?”.Forthe samereasonthatpeopledon’ttakeoutamortgageontheirhouseandgoplayblackjackin LasVegas:theriskformostpeopleisfargreaterthanthereward.Awealthytradercan takeasmallbitoftheportfolioandgodaytrading.Ifheorsheloses,thelossisrelatively small.Ontheotherhand,ifmiddleclassAmericagambledwiththesizeoftheportfolio neededtomakemoneyindaytrading,itwouldbeconsideredasickness.Whenthechance ofloss(risk)isgreaterthantheaveragereturn,aninvestmentbecomesagamble. SELF-FULFILLINGPROPHECY Theadventofthe“informationage”hascreatedmorerandomvariationbecause responsetoworldeventsisinstantaneous.Atonetime,severalvariableswereweighed beforeafinancialdecisionwasmade,andinvestmentswerefirmlyconnectedtotheir foundation-asetoffundamentalsthatdeterminedearningspotential.Inrecentyears, however,economicvariableshavebecomemoreinterconnectedonagloballevel,suchthat a“meltdown”intheChinesestockmarketwillmoreimmediatelyaffecttheU.S.market. Thatinterconnectivityhasdemandednotonlymoreimmediatereactiontoevents,buthas createdmoreriskbyinducingmassivemovementsofcapitalintoandoutofcertainsectors. Whenallfinancialmanagersactinconcert,thereislessdiversificationofassets,andso thuswehavethe“techboom”ofthelate1990sorthe“housingboom”oftheearly millennium.Onamicrocosmiclevel,rumorsaboutspecificstocksandcompanieshave nearlydestroyedsomeinstitutions(BearSterns),anditisnocoincidencethatonline technicalanalysiscanmoveastockwhenenoughtradersfollowit.Infact,mostavid tradershaveknowledgeofaspecificportfolioofstocksthatareespeciallyconfiguredfor largemovementsupwardsordownwardsandconducivetotechnicaltrading. Likegolferssearchingforthe“perfect”driverthatputstwentymoreyardsontheir shots,technicaltradersbecomeobsessedwithfindingtheone“tradetool”thatwillgivean 549 edgeoverthecompetition.Thus,theywillbacktesttechnicalindicatorstoseehowwell theyworkedinthepast,pagingthroughhistoricaldataandnotrealizingthateachmarket isdefinedbyadifferentsetofparameters;adifferentyieldcurveisgoingtoproduce differentresults.Lowinflationrateswillaffectthoseindicatorsdifferentlyfromhigh inflation.Theobjectivebecomestofindacohesive“exitstrategy”thatenablesaninvestor togetoutofaninvestmentbefore“othersseethelight”However,thatterritoryismapped outbythoseinvestorswhohavethemostmoneyandtheshortesttimeframe.Aninvestor whomakesalargethree-dayprofitandleavesthestock,willhaveanadvantageoverthe “oddlotter”whowaitsfourdaysandseesabigvolumemovementonthatthirdday.Small investorssimplymustwaitlongertomakeaprofitbecausetheyhavesolittleinvested. Thus,ifasmallinvestoristechnicaltrading,heorshemayseearapidriseandfallbefore anymoneywasmade.Onatechnicalbasis,thestockdidnottrendlongenoughtomake money,butthosewhospottedthetrendinashorttimeframeendupmakinga“killing”. Thesamedynamicthatmakeseconomiesofscaleanadvantagetothelargestfirmsinan industry,giveslargetradersanabsoluteadvantageintechnicaltradingbecauseoftheir potentiallyshortertimeframe. EX-POSTPERFORMANCE Onahistoricalbasis,mosttechnicalindicatorsareexcellenttoolsformappingout thepastperformanceofastock.Afinancialexecutivecanmatchthedemandschedulefor aproductlinewiththesupportandresistancelinesforasecurityandviewanycorrelation. Sincemoststockstrackearnings,theaccelerationofearningscaneasilybecoordinated withthevolumeflowofcapitalintothestock.Thissamehistoricalcorrelationcanbedone betweentechnicalindicators,andtheleveragestatesofcapitalstructure.Choosingone technicalindicator,howdoesitreacttothetransitionofonestatefromanother? Thesimple“stochastic”indicatorisfamiliartomosttechnicaltraders.“Stochastic” isawordfromstatisticsthatsignifiesrandomvariation,andalthoughtherearemany complexadaptationsofthistool,thesimplestversionwaschosentodisplay.Ineffect,the 550 indicatorisaratiothatshowshowcloseastockistoitshighestlevel:((ClosingPrice-Low Price)/(HighPrice-LowPrice))equatesrangewithpricelevelandcanbedoneforany timeperiod.Valuesof80andoversignifythatthestockmaybea“riskybuy”becauseitis “overbought”,whileafigureof20representsa“buy”signalbecauseithasbeen“oversold” andwillriseagain.Sincetherange(distancebetweenhighandlow)cangiveanestimateof thevariancewhenitissquaredandmultipliedby1/16,theindicatorhassome mathematicalsubstancetoit.Forourpurposes,afiscalyearwaschosenasthetime period,andtheleveragestateforthatyearwasmatchedtothestochasticthatwaspresent atthebeginningoftheperiod,aswellasthestochasticthatwaspresentattheendofthe period. Thereaderwillberemindedthatthetypeofmarketandlevelofinterestrateswill determinetheleveragestateswiththemostpotentialprofit.Itwasourhypothesisthatin thismarket(onewithrisinginterestrates)companiesthatincurredmoredebtwouldhave alowerstochasticbecausetheywouldcloseloweronDecember31whenthemeasurement wasmade.Companieswhowereearningmorewouldincreasetheiramountofretained earnings,andreducebetaandlong-termdebttocapitalsimultaneously;thesecompanies wouldclosehigheronDecember31,andalsohaveahigherstochastic.Theparadoxshould beapparent:ifreducingriskbyloweringbetaissoimportant,whywoulditleadtoa higherstochastic?Theanswerhastodowithtwoconcepts:concurrentperformanceand momentum.Whenamarkethasmomentum,investinginastockthatoutperformedthe otherswillpayoffbecausethatstockwillriseagain.Ifthereismomentuminamarket, stockswithhighstochasticswillcontinuetooutpacestockswhicharemiredinalower stochastic.Manyrecentmarketshavemimickedthat“winnertakesall’scenario. However,anystockthatoutperformsthemarketforanylengthoftimeisalsoatagreater riskofdropping.Incapitalstructureterms,whenafirmmovesawayfromitstarget capitalstructure,itspricewillfallandthestochasticalongwithit.Suchaneventoften occurswhenproductdemandislow,andthefirmisalsoincurringmoredebt.Thusthe 551 stochasticisasmuchameasureofperformanceasitisofrisk.Itdoesnotdelineate betweenthosefirmswhowillcontinuetoperformwellandthosewhoareabouttodrop. Nevertheless,investinginalowerstochastic“debt”statewouldalsobeapreludetoa higherpayoffandareduceddebtratiowhenearningsacceleratedandraisedthepriceof thestock.Ineffect,alowerstochasticwouldpreface“gettinginearly”beforeastocktook off.Thefollowingtablesshowtheresultofthisbriefstudy: Table22-1 DEBT STATES STATE Fin.Lev. LTD/CAP OpLev. Asset/Cap. Start Stoc. End Stoc. % Change 1 - + + N/A 66.67 47.83 3.29 2 + + + N/A 45.54 74.51 6.96 3 + - + N/A 35.42 36.23 -14.18 4 - + - N/A 84.29 74.47 -17.21 5 + - - N/A 54.8 55.98 -4.65 6 + + - N/A 74.39 45.64 -37.35 Althoughwehavealludedtotheimportanceofraisingtheasset/capitalratiowhendebtis incurred,itisnotacategoricalseparator.Intheabovedata,number2andnumber5show themostpromiseasaninvestmentvehicle,whilenumber6showstheworst.Thestochastic andpercentagesareskewedeventhoughmediansandnotmeanswereusedtoaverage them.Thedataisalsomarketdependentandisnotmeanttoguidetheinvestorintoa decisionbecauseitcanchangeastheeconomychanges.However,sincethesearestates wherethefirmeitherraisedtheproportionofdebttoequity,orraisethefinancialleverage ratio,orboth,thelowerstochasticmayimplythattheleveragestatehasmorelong-term potentialasaninvestment.Theburdenremainsontheinvestortoknowwhichsectorwill befavoredinthebusinesscycle,andtomatchanalysts’estimatesofearningswiththe leveragestate.Inordertoprofitfromthetransitiontoanequitybasedleveragestate, 552 earningsmustbeacceleratingfasterthanthecostofequity.Atthisjuncture,betamay decrease,EVAwillriseandthestochasticshouldriseaswell. Table22-2 EQUITY STATES (PROFIT) Leverage Fin.Lev. LTD/CAP OpLev. Asset/Cap Start Stoc. End Stoc. % Change 7 - - + - 80.79 26.7 -54.99 8 - - - - 78.78 81.81 17.93 9 - - - + 72.22 77.73 -0.17 10 - - + + 77.24 80.78 -0.32 Fromtheabovedata,states8,9,or10seemtoofferthemostmomentumandstability whilestate7losesitspricestrength.Beforethereaderbeginssellingthisleveragestate, recognizethatthisdatawasfromaspecificera(1990s)andthatthisstatewasactuallya “favored”statethreeyearslaterin2002.Leveragestatesareneversetinstonebecausethe economychangesaroundthem.Theyrespondtoshiftsinthewaythatriskisbeingpriced bythemarket.Asavvyinvestorwillseethepotentialinatransitionfromadebt-oriented statetoanequity-orientedoneandinvestwhenthestochasticislowenoughtomakealarge profit.However,thereisnoguaranteethatafirmwithadebtladenstatewillnottakeon evenmoredebtandwallowinalowstochasticforthreeorfouryears.Similarly,ifadebt statetransitionsintoanequityorientedstatewithoutsufficientdemandforthecompany’s products,thestockcanfallprecipitously.Forexample,whenashiftincapitalproportion iscausedbyassetsales,thecompanymaynotbegrowingandgeneratingsufficient operatingincome.Thus,theEVA/capitaldynamicremainsthebestmeasurementoftrue investmentworth.NotonlycanEVArevealthemismanagementofequity,butitcan measuretheearningspowerofafirminthemiddleofadownsizingordivestiture. Onesignoffinancialstrengthisarisinglowstockoveranumberofyears.When thehighstockisrelativelystable,andthelowstockkeepsrising,therangeofthestock 553 narrowsanditbecomeslessrisky.Ineffect,thebetaofthestockisloweredrelativetoits industrywhichwillbeactualizedbyalowercovariancewiththemarket;thestandard deviationofthestockdeclines.Ontheotherhand,astockthathasrisenwithalower stochastic,maybeonitswaydownbecauseaspeculativehighmayhavebeenreached.The twomostriskyleveragestatesinourstochasticreviewbothhadpricerunupsbefore falling-whichisindicatedbytheirrapidlyfallingstochastics.Noticethemechanicsinthe followingexample. Close-Low/High-Low=70-60/80-60=0.5or50%.Nowifthehighpriceranupto $100.00:75-60/100-60=.375or37.5%.Thestockwouldhaveincreasedto75whichis a7.14%gainandyetthestochasticwouldbegoingdownto37.5%,adecreaseof25%. Obviously,theriskofthisstockgoingevenlowerisveryhighinspiteofthelower stochastic.Itseemsthatwemeasuredthestockasithitaspeculativehighandthen descended,butstillhadaprofit“attached”toit. STOCHASTICCONFORMANCE Competitioninanindustrytakesplacewithintheframeworkdictatedbysuch variablesasdemand,earningspotentialandthecostofcapital.Whileitisnotsacrosanct thatcorporationsmustactinunisontobesuccessful,theopportunitiespresentedtoeach aresimilar.Itwouldnotbehooveacompanytotakeonmassesofdebtwhenthetop competitoris“rakinginthedough”andpayingoffitslatestbondissue.Thus,stockstend toriseandfallasasector,andthatistrueforstochasticsaswell.Oncecapitalisinfused intoafirm,theturnaroundtimedependson:technology,management,andtheresources withwhichtheprojectsareendowed.Theprocessissimilarineachfirminthesectorwith slightdifferencesbetweenserviceandmanufacturingorientations,targetedcustomersetc. Twonotableexceptionsexisttostochasticconformity.Thefirstexceptionisthat manycompanieswilloccupyauniqueproduct“niche”.Theyhavediversifiedandthen specializedinanareainwhichtheyhaveparticularcompetencethatgivesthema competitiveadvantage.Nucor,forexample,specializesinsteelalloysthathavemadethem 554 lessdependentonsteelasacommodity.Whencommoditypricesgodown,variablecosts arenotasaffectedastheywouldbewithothersteelproducers.Thesecondexceptionisin theareaofacquisitions.Manycorporationshavebecomesolargethattheironlyhopefor growthliesinbuyingothercompanies;theyneedtogrowattwentypercent,buttheircore industriesmaybegrowingatthesamepaceastheeconomy-fourpercent.WallStreet dictatesthatthefirmneedstoexpandfasterthantheindustryitself.Thesecompanieswill generallyseeadecliningstochasticasinitialspeculativerun-upsbegintofalter,andthe firmmustworktoloweritsprobabilityofdefault.Fromacapitalstructureperspective, taxbenefitsarerewardedintheyearoftheacquisition,andmarginalbankruptcycosts mustdecreaseinthenextperiod,orthestochasticwillbemuchlower.Sinceitisdifficult tointegratealargeamountofassetsrapidly,turnaroundtimesforcapitalwillusually dictateaslightlyhigherbankruptcycostinthenextperioddependingonthecompetenceof theintegration.Iftheintegrationissmooth,operatingincomewillbeginrisingalmost immediately,higherbankruptcycostscanbeavoided,andthestockwillcontinuetorise. CAPITALSTRUCTURALISM:QUASI-TECHNICALANALYSIS? YearsagowhentheCAPMwasfirstprofferedbyacademia,betawasaccusedof beinganotherversionoftechnicalanalysis,exceptthatithadauniversitypedigreeand camewithasetof“whitepapers”.Thepoliticsofoppositionareespeciallyaptfortherift betweenWallStreetandacademiabecauseeachpanderstoextremistpositions.Adopting oppositesidesofanissueallowsdebatetooccur,andwithconflictcomes(hopefully) profitableresolution.Thesearchtomakefinancemoreobjectiveandscientifichas progressedintoalinearityofthoughtwhereaspecificpostulateisstakedout,promulgated andthendefended.Thus,weseemonetaristssquaringoffwithKeynesians,freetraders andprotectionists,supply-sidersversusrationalists,growthstockinvestorsversusvalue stockinvestors,adinfinitum.Finance,howeverisprobabilisticandnotdeterministic,and thereisalwaysroomforanotherlegitimateidea. 555 Capitalstructurecanbetechnicaltoolwhenitismatchedwithanindicatorlikea movingaverage,butinisolation,itisalegitimatemethodofoptimizingthewealthofa company.Itprofessesnottopredictthepriceofastockbuttopredicttheriskofafirmin thedomainofseveralvariables-theeconomy,taxes,sales,incomeanddefault.Capital structuralismrecognizesthatonlystrategiesthatadapttotheconfluenceofrisksfor economy,industryandcompany,andremainflexible,willultimatelyoptimizewealth. Thus,inoneeconomymoredebtmaximizesthestockprice.Inanother,onlymergersand acquisitionswillwork.Inathirdeconomy,diversifyingoperationsandloweringoperating leverage(andbeta)isthekey.Allofthesestrategieswillhavetheonecommonalityof movementtowardanoptimaltargetcapitalstructure.Whiletechnicalanalysiscanbe adaptedtocapitalstructuralism,optimizingcapitalstructureisundertheaegisofrisk management;eachfieldmakesextensiveuseofthedifferentadaptationsofthestandard deviation,buttechnicalanalysisattemptstomeasureandpredictperformancewhile capitalstructuralismtriestosupplytheimpetusbehindwhatisbeingmeasured. FIGHTINGWORDS:“THEEFFICIENTMARKETSHYPOTHESIS” Thespecterofthe“efficientmarketshypothesis“hashauntedWallStreetfor decades.Anotherproductofacademics,itbrieflystatesthatwithmillionsofbuyersand sellersinthemarket,andeachcompanybeingresearchedextensively,informationtravels sorapidlythatnoindividualcangainacompetitiveadvantage.Evenrumorsofearnings deficienciesbecomefactoredintothemarket.Sincemostfundamentalsarepublishedafter thefact,andsincetechnicalanalystspossessthesamelibraryofindicators,itisnowonder thatmostmutualfundsunderperformthemarketdespiteaccesstomillionsofdollars worthofresearch. Overmanyyearstherehasbeenanupwardbiasinthemarketofapproximately 9.8percentperyear(2008).Theupwardsurgeofabullmarkethasbeenfarmore frequentthananydownwardshifts,andthreeoutoffourstocksriseatthistime. Consequently,someofthetoutedsuccessofbothfundamentalresearchandtechnical 556 analysishasdevelopedfromtheillusionthattheacumenoftheanalysthelpedpickthe “rightsecurityattherighttime”-wheninreality,thestockwouldhaveskyrocketed anyway.Indeed,BurtonMalkiel,inhisclassicbookARandomWalkDownWallStreet,, emphasizedthatifanyspecifictechnique(suchasstochastics)workedforanylengthof time,itwouldcreateitsowndemisebecausethemarketwouldfactoritin.Becausethe marketkeepschanging,andbecausethemathematicalvarietyoftrends,volatilityand correlationaresogreat,thereisalwaysoneindicatoravailabletoexploitaspecific company’ssituation.However,bythetimethis“needleinthehaystack”isfound,the marketmayhavechangedandthecompanymaynolongerrespondtoit. Whataboutcapitalstructure?Wouldn’tanefficientmarketfactorinsuch movement?Infact,itdoes.Evenwhenaninitialinvestmentismade,thetaxbenefitsmay causesomeupwardmovementinthestock.However,neitherthemarketnortheinvestor hasanycertaintyastothelengthoftimetheinvestmentwillbegintoproducemore earnings.Theadvantageofcorporateinsidersisthattheyhaveknowledgeofprojecttime- frameswheninvestmentsareexpectedtopayoff.Thisinsideinformationisagreat opportunityforthecapitalstructuralist,becausetheleveragestatecanbematchedwith publicknowledgeaboutinsidersales.Iftheinvestorobservesastatethatisconduciveto growth,andcompanyexecutiveshavebeenputtinguptheirowncash,thereisahigh probabilitythatthefirmismovingtowardanoptimalcapitalstructure.Infact,eventhose whostronglybelieveintheefficientmarketshypothesiswillclaimthatthistechniquehas beensubstantiated. Anadditional“chink”inthearmorofefficientmarketsincludesmomentum. Severalresearchershavefoundamomentumvariablethatcanbeexploitedinsomestocks. Stocksthatoutperformthemarketinpriceorearningsforasix-monthperiodtendtodo betterthanthemarketforanothersixmonths.Manytechnicalindicatorsmeasurethe forceofmomentumwhichcanlogicallybeappliedtothoseselectstocksthatkeeprising. Whileanefficientmarketspuristwoulddiscountpricingpressureasrandomvariation, 557 twopointsappeartobevalidintechnicalanalysis:1.Atrendcanbeexploitedforaslong asitcontinues,and2.Thereisnopredictinghowlongthattrendwillcontinue.Since manystocksmoveinlargeirregular“jumps”,theapplicationoftechnicalanalysisto exploitmarketinefficienciesappearstobelimited.However,asamanagementtoolthat givesdetaileddescriptionofpastperformance,itcanbematchedwiththeforcesthat affectedstockprice.Losingouttocompetitors,forexample,wouldbemirroredina downturn.Sectordominationandincreasedmarketsharewouldbeobservedinasurge upwards.Bycorrelatingperformancetoactionablepolicy,technicalanalysisisan instructivetool. THEARTANDSCIENCEOFFORECASTING Thechainoflogicbehindforecastingbeginswiththefactthatallcompaniesneeda salesforecasttocoordinatethevariousactivitiesofafirm;planningisessentialtoanyfirm. Analystshavetakenthisneedonestepfurther:byextrapolatingearningsfromsalesand thenobservingthecorrelatedmovementofthestockpriceovermanyyears,ademand forecastmorphsintoastockforecast.However,inthenearterm,theassociationisnot nearlyasstrongasitisoverthelongterm,becauserandomvariationfromoutsideforces hassuchastrongpullonthestock.Evengreatfundamentalswillelicitdownwardpressure inabearmarket.OurownanalysisofEVAconfirmsthisnear-terminabilitytocorrelate earningswithstockprice.Ifearningsareextraordinarybuttoomuchequityisincurredat toohighaprice,EVAwilldeclineandthestockpricewillbediminished.Thathighcostof equityisoneofthose“outsideforces”toconsiderbecauseitispartiallyimposedbythe relationshipbetweenearnings,interestratesandinflation. Studentsoffinanceshouldbeespeciallyfamiliarwithforecastingtechniqueswith thecaveatthatevensteady,lowoperatingleverageproductscannotbeforecastedcorrectly onehundredpercentofthetime.Althoughpredictionofstockpricesisforeversuspect, predictingtheforcesthatmoveastockandcombiningthemintoaconsensusthatsmoothes outinaccuraciesinanyspecificpredictionisstandardforfinancialexecutives.Specific 558 techniqueslikeArima,Box-Jenkinsanalysis,Holt-Winterexponentialsmoothing,and aboveall-movingaverages-canletthestudentobservetheconnectionbetweendaytoday activitiesandthelargerfinancialimperativeofmaximizingwealth.Indeed,capital structurebeginswiththesalesforecastbecausetoomuchortoolittlecapitalraisedwill causeadeficiencyintherelationshipbetweenearningsandthetargetproportionofdebtto equity. Onemajorreasonfordiscrepanciesbetweensalesforecastsandstockpredictionsis thatsalesaregenerallybasedontheneedsofaneconomicentity-consumer,businessor government.Noone“needs”tobuyastock-althoughsomepeoplemaythinktheydo. Sellingtireshasadegreeofpredictability-numberofvehiclesontheroad,season, structuralchangesinrubberetc.Whiletireforecastingcanbeasophisticatedscience,it doesnotapproachtherandomvariationsineconomy,psychology,specificindustryand personalwealththatstockselectionentails. MOVINGAVERAGESTOUSEWISELY Sometoolsoftechnicalanalysiscanbeveryhelpfultoanyonewhoneedstospotan acceleratingtrend.Forexample,youmaydothebookkeepingfora“momandpop”ski lodgeinVermont.Youwanttocontrasttheamountofcoffeesoldtoskiersinthewinter versustheamountsoldinsummer.Acoffeesalesmovingaveragewillshowthedifference. StartinginNovember,theshorttermmovingaverageforcoffeewillrisefasterthanthe long-termaverage.InDecemberorJanuaryitwillbeabovethelong-termaverage,and finallydroptowhereitisequaltothelong-termaverageinMarch.Tomeasurethe acceleration,thebookkeepersubtractsthelong-termaveragefromtheshort-termaverage. Whenthatnumberisrising,salesareaccelerating.Whentheshort-termaverageisabove thelong-term,thereisupwardspressureonsales.“Upwardspressure”doesnot necessarilymeanthatsalesareaccelerating,butthatthereismoretendencytodoso.In fact,thesetrendsareindicativeofthefirstandsecondderivativesincalculus.Thefirst derivative(upwardspressure)willshowwhetheranumberisincreasingordecreasing, 559 whilethesecondderivative(acceleration)willshowifthenumberisincreasingatan increasingordecreasingrate.Thus,understandingtheconceptofcontrastingmoving averageswillgiveeventhesmallbusinesspersonausableformofcalculus.TheWall StreetJournalusesthesetodisplaytrendsandmanyotherbusinessesusethemtotrack salesandinventory.Theexponentialformisself-correcting(recursive),andyetsimpleto implement,andisdiscussedinthenextsection. THEEXPONENTIALMOVINGAVERAGE Thismovingaverageisaveryflexibletool.Unlikemanyspreadsheet-typemoving averages,itdoesnotneedanyleadcellsbeforeitbeginstowork,anditisadaptabletoany timeperiodyouspecify.Thebasickernelof“artificialintelligence”isobservedinthe recursive,orself-correctingbehavioroftheaverage.Likeahumanbeing,“neural networks”aresupposedtomakecomparisonsbasedonlearnedadaptation.Whilethe conceptismoreprofoundthanthepracticalcalculusofanexponentialmovingaverage, thissimplespreadsheetfunctionoffersthestudentanintroduction. Whenformingaworkingmovingaveragesystem,werepeatthesameprocedurefor boththeshortandlong-termaverages.Tocircumventtheneedforaleadtime,wecreate twosmoothingconstantsbasedonthelengthoftheaverage.Thesmoothingconstantis merely,(2/(N+1))whereNisequaltotheperiodthatyouspecify.Forexample,ifyouwant theshort-termaveragetobeninedays,thesmoothingconstantwouldbe:(2/10)or0.2.To illustratetheadvantageofcreatingasmoothingconstant,consideraspreadsheetwherea movingaveragerequiresaspecificnumberofleadcells.Anytimetheaverageisused,it musthavethespecificnumberofcellsasa“lead”:Thisexampleisasimplefiveday movingaverage. 560 Table22-3 EXCELCOLUMN SPREADSHEET EXCELCOLUMN E1 25 F1(BLANK) E2 30 F2(BLANK) E3 35 F3(BLANK) E4 40 F4(BLANK) E5 30 F5=SUM(E1:E5)/5 Noticethatweneeded(N-1)blankspotsbeforewecouldbegintheaverage,andthatwe couldnotspecifyadifferentnumberforperiodlengthlike10,21,or50.Theformulain cellF5wouldbecopieddowntomatchtheremaininglengthofthedata.Inthestepby stepprocessIamabouttogiveyou,onlytwoblankspotsareneeded,preferablyatthetop ofthespreadsheetandsidebyside.Oneblankspotwillbeforinputtingtheshort-term periodlength,andtheotherforspecifyingthelong-termlength STEP1:WewillusecellsG2andH2asblanksforinputtingshortandlongperiods respectively.LabelcellG1“short”andlabelH1“long”. STEP2.Weidentifyacolumninwhichtoplace(copyfromanotherfileorinputmanually) ourdata.OurdatawillbegininE2andwewilllabelcellE1“price”.Wecancopydata fromaCSVortextfilebypastingitintoE2orwecaninputthedatamanually-whichof course,istediousforverylongsets.Thedatacanbesalesdata,stockprices,currency exchange,orinventorydata. STEP3.WeformthefirstoftwoequationsbeginningincellI3.Thefirstequationbegins theprocessofsmoothing,settingitupagainstitself:CellI3=(2/($G$2+1))*(E2-E2)+ E2.AlthoughthenetoutcomeisthecontentsofE2,thelogicoftheprocessisillustrated. STEP4.ThesecondequationisinputtedincellI4,onecellbelow.Itusestheanswerin cellI3tocorrectitself:I4=(2/($G$2+1))*(E3-I3)+I3 STEP5.IncellsJ3andJ4youwillcopycellsI3andI4respectivelyexceptforchanging thesmoothingconstant.Thesecellswillbeyourlong-termaverages,andthebeginning 561 part(theparttochange)willbe(2/($H$2+1))whichwillrefertocellH2(thelong period)ratherthanG2(theshortperiod). STEP6.BothequationsinI4andJ4canbecopieddownforthousandsofcellsifyouwish, andyoucansavethesetupasatemplate.Fromnowonyoucanjustinputperiodsinto cellsG2andH2andthencopydataintoE2.Whenyouarefinishedwithanalysis,you click“donotsave”andyouarereadytoinputanothersetofdata. STEP7.Toestablishtheaccelerationfactor,wesubtractcolumnJfromcolumnIstarting incellK3andcopyingdownforthelengthofthedata.K3=I3-J3.Whenthedistance betweenshortandlongbecomesgreater,thereisupwardspressureonthedata,andvice versa. STEP8.Awordaboutperiodlength.Forstocks,manytradersusea50and200day averageforshortandlongrespectively.Somestabilityoftrendisestablishedwiththis combination.Thestudentshouldbeawarethatthesmallerthetimeperiod,thegreaterthe sensitivitytoacceleration. BRIEFINTERPRETATION Thegistoftechnicalanalysisistofollowanupwardstrendandsellbeforetherisk ofadownwardtrendistoogreat.Whiletheinvestorcanobserveaccelerationveryacutely withsmallaverages(say9and14),heorshewillbetradingfrequentlywithnoguarantee ofmakinganymoney.Infact,useofmovingaveragescanmimicThreeCardMonte:one can“uptheante”forabigpaydayifluckfallstheinvestorsway,andthestockstaysina stabletrend.Inevitably,Gambler’sAnonymousisfullofcertifiedgeniuseswhohavetried severalsystemsandfailed;temptingfatebydependingononemeasurementinvites disaster.Whiletheinvestorcanbeawareoftheoveralluptrendbecausethe50-day averageisabovethe200-dayaverage,thesubtrendscancontinueforweeksatatimein eitherdirectionandwillbepunctuatedbylargeunpredictablejumps.Inthisgame,the brokerage(“thehouse”)getstokeeptheircommissionswhiletheinvestor(“bettor”)barely breakseven. 562 PRIMARYTRENDSANDSECONDARYTRENDS Thecombinedforceofhigherearningsandalowercostofcapitalwillpushafifty- dayaverageaboveatwohundreddayaverage.Thislongertermtrendistheprimary trendandismanifestfromafirm’spayoffonitsinvestment(eitherdebtorequity).At times,speculationcancreatesuchanup-trendsoitismandatorythattheinvestordothe researchandfindoutwhetherthetrendissustainable.Inthecaseofageneraleconomic expansion,manystockswilltrendupwardswithoutmuchimpetus,andtheinvestorwants tobeinastockwhosecompanyisintheprocessofoptimizingitscapitalstructure.The reason?Thelongertheshort-termmovingaverageremainsabovethelong-term,themore strengthandstabilitythepricerisewillhave.Acompanythatisdominatingafavored sectorwillcreatealargedifferencebetweenshortandlongmovingaverages.However, oncethestockhas“jumpedup”(shorttermaveragemovingabovelongterm),muchofthe movementwillbeacombinationofsustainabletrendandstatistical“noise”whichwillbe difficultforeventopmathematicianstodecipher. Thesecondarytrendsthatareacombinationofvolumegeneratedrisesandnoisecanfool aninvestorintolosingmoney.Thetimetobeinthestockwaswhentheinitialforceof earningswaspropellingitabovethelong-termmovingaverage.Inthecaseofaspeculative run-up,thechartistswillimmediatelylookforanexitstrategyleavingslowinvestors “holdingthebag”.Ifoneisaddictedtowatchingstocks,investingonthebasisofcapital structurewillatleastgivetheinvestortimetoleavethestockaftermakingaprofitbecause thesestockswillberisingforaperiodofsixmonthstotwoyearsormore.Investorsmust decidewhentogetoutbybalancingtheprofitalreadymadewithriskaversion.Thosewho arestillgreedyafteratwentypercentrun-upwillkickthemselvesifthestockbeginsto “tank”.Indeed,investinginthesecondarytrendsisabitlikefast,“Vegasaction”,butthis authorwouldbenegligent,ifitwerenotincludedinthistext.Whentherisksfaroutweigh therewards,aninvestmentbecomesagamble,andthestudent/investorshouldunderstand thedifference.(BacktoTableofContents) 563 23 STATISTICSPRIMER ThereisavastmarketinAmericaforbadlyappliedstatistics.Whenevertwo agendascollide,eachadvocatemustproveapoint.Andwhatbetterwaytolendan ambianceofpseudo-sciencetoone’smethodologythanbyexhibitinganaveragethathasno realstatisticalsignificance?Infactwhenwespeakintermsofthe“averageAmerican”,it isanextrememisapplicationofstatisticaltechniquebecausethereissomuchvariation amongthepopulation. Andvariationinfinanceiswheretherootofmostproblemslies.Thelargerthe amountofvariation,theharderitistoproveapremiseistrue;asampletypeandsizemust mathematicallyadjusttotheamountofvariationorthestudywillbeinconclusive. Unfortunately,financialdataisusuallyhighlyskewed:thatis-itisnotheavily concentratedaroundanaveragebuttendstobedistributedwithlargeamountofsample pointsatoneendortheother.Thus,financeandstatisticsmakestrangebedfellows: withoutadequatemeasurement,propercapitalallocationcouldneverbemade.However, veryfewofthemeasurementsarestatisticallyvalidbecauseofsmallsamplesizeandlarge amountsofvariation.Variationamongstockpriceincreases,forexample,canapproach twoorthreetimestheaveragewhichrendersmanypricerelatedstudiesinconclusive. Whenwetrytofitfinancialdatatoa“normal”curve,weareinfact,takingliberties thataremathematicallyinvalid,butneverthelessmakinganattempttodrawactionable intelligencefromthem.Andthatistheessenceoffinancialstatistics-topointusinthe “right”direction.Weconcentratenotsomuchonmeasuringanaveragethatisderivedin theabsenceofvariation,asfocusingonmeasuringthevariationitself. Thepragmaticrealityofinvestingistomakedecisionswithinsufficientdata.For thesmallinvestor,accesstorealtimecorporatedataissimplyimpossible.Theindividual mustmakeseveraldecisionsbasedonfactsthatheorsheculledtogether.Iftheindividual 564 istobeguidedcorrectly,anystatisticallibertieswetakemustbeusedtoavoidexcessive riskratherthantospeculateonlargereturns.Throughoutthisbook,thestudent/investor isencouragedto“diversify”indicatorsandusemanydifferenttypes.Inmostcases,the samplesizewillbeinadequatetodrawfirmconclusions.But-ifseveralmeasurements pointusinthesamedirection,theinvestorcanbesatisfiedthatarationaldecisionwas madegiventheavailableinformation. THEMEAN,MODEANDMEDIAN MEAN-Thearithmeticaverageofapopulationisitsmeanandisthemostused measurementinstatistics.Thesumofanysequenceofnumbersdividedbytheamountof figuresinthesequence(samplesize)willyieldthemean.Thefollowingexampleisa sequenceofnumbers:7,17,14,32,60,3,21.Wecounttheamountoffiguresinthe sequenceandobtainasamplesizeof“7”.Wethensumallofthesequenceanddividethe samplesizeintothesum:(7+17+14+32+60+3+21)/7=22. MEDIAN-Whenasampleofnumbersisarrangedfromlowtohigh,themedianis themidpointwithanequalcountofdatapointsbothlowerandhigher.Inthesample above,wearrangethesequencefromlowtohigh:3,7,14,17,21,32,60.Sincethereare threenumbersbothlowerandhigherthan17,thatisourmedian.Ifthesamplewerean evennumber,sayeightinsteadofseven,thetwonumbersinthemiddlewouldbedivided by2.Inthatcase,therewouldstillbethreenumbersbothlowerandhigher,butthe medianwouldbeanaverageofthetwomiddlenumbers. MODE-Themodeisthemostfrequentnumberencounteredinasample.Thus,in theabovesample,allentriesareuniqueandthereisnomode. Sincefinancialdataissoheavilyskewed,inmanycasesthemedianisafarbetter indicatorofdistributionthanthemean;themedianbetterincorporatesinherentvariation. However,theflexibilityofwhatistermed“parametricstatistics”,definestherelationship betweennumbersthrough“parameters”suchasthemean,anddictatesabalanced approach;any“statisticalsignificance”istestedbytheamountofvariation. 565 Whenapopulationis“normal”,themean,medianandmodewillbethesame number.Infinance,populationsareoften“skewed”andheavilyweightedinonedirection ortheother.Infact,datapointswilloftentakeonthecharacteristicdepictionofa “barbell”andbeheavilyweightedonbothsideswithfewpointsinthemiddle.This “bimodal”distributionisdifficulttointerpretbecauseitslackofcentraltendency eliminatesthestatisticalsignificanceofmeanormedian.Thefollowingdistributionsshow thelocationsofmeaninrelationtomedian. Figure23-1 Figure23-2 Mean Median LEFTOR"NEGATIVE"SKEW Mean,Median,andMode NORMALDISTRIBUTION 566 Figure23-3 Inabimodaldistribution,itappearsthattwoseparatepopulationsexistwhichmakesit difficultfortheanalysttodeterminesignificantrelationships.Inthemarketof2005to 2007,forexample,anyoilrelatedstockwasdoingquitewellwhiletherestofthemarket laggedbehind.Anyrelationalstatisticbetweenthetwosubpopulationssuchastheamount ofassetsorsaleswouldhavebeeninsignificant.Onlythetypeofindustryactually mattered. Figure23-4 THEVARIANCEANDASSORTEDADAPTATIONS BIMODALDISTRIBUTION Mean,Median,Mode RIGHTOR"POSITIVE"SKEW Median Mean 567 THEVARIANCEThevariancemeasuresthedeviationofindividualobservations aboutacentralvalue.Mostdatapointswilltrendtowardeithergreaterorlesserdistance fromacentraltendencyandthevariancemeasuresthisdistance.Ineffect,itistheprime indicatorofriskandusedtocompareuncertainoutcomeswiththe”normalcy”and implicit“certainty”ofthecentralvalue. Tomeasurethevariance,wesumthesquaredvaluesofdistancesfromthemeanof eachdatapoint,andthendividebythesamplesize.“∑ ∑∑ ∑”istheGreekletter“epsilon” whichdesignatesthateachoperationinasequenceshouldbesummed.ForexampleifXis anumberinthesequence(1,2,3)andafunctioncallsfor∑ ∑∑ ∑ (X+2),wewouldproceed: (1+2)+(2+2)+(3+2)=12.Theformulaforthevarianceis∑ ∑∑ ∑(X-µ µµ µ) 2 /N.Inthisfunction, Xisequaltotheindividualdatapointinthesequence(sample),µ µµ µisequaltothemeanof thesample,and“N”isequaltothesamplesize.Thus,eachdatapointissubjectedtothe sameprocess,andthentheprocesseddatapointsaresummed.Inasamplepopulationof (1,2,3,4,5),withameanof“3”,theprocesswouldbe(1-3) 2 +(2-3) 2 +(3-3) 2 +(4-3) 2 +(5-3) 2 / 5whichis10/5or“2”.ThesymbolforthevarianceistheGreekletter“sigma”squared, orσ σσ σ 2 . STANDARDDEVIATION-Thestandarddeviationismerelythesquarerootofthe variance.ItsGreeksymbolistheletter“sigma”,σ σσ σ.Itismorecommonlyusedinfinance thanthevariancebecauseithasacloserelationshipwiththemean,andcanbeusedin predictiveforecasting.Ifthereaderrefersbacktothe“normal”distribution,eachpoint withinthedistributionissomeamountofstandarddeviationsawayfromthecenterwitha specificlevelofprobabilityattachedtoit:thatis-theprobabilityofbeingonaninterval betweenthemeanandacertainnumberisbasedonthenumberofstandarddeviations awayfromthemean.Intheabovesequenceof(1,2,3,4,5),wefoundthatthevariancewas “2”.Nowthestandarddeviationis√ √√ √2,or1.41. THECOVARIANCE-Attimes,theanalystneedstodeterminetheriskofa combinedpopulation,examiningwhetherthecombinationreducesriskorincreasesit.A 568 muchappliedadaptationofthevarianceistermed“thecovariance”Forexample,ifIam anoperationsmanager,Imightwantlessriskbetweenthereturnsfromproductline#1 (1,2,3,4,5),andproductline#3(6,7,8,9,10). Iknowthereturnfromproductline#1alreadybecauseithasthesamemeanasin thestandarddeviationexample,(3).Inowobtainthemeanforproductline#3and determineittobe“8”.Ithenobtainasum(notsquaredthistime)oftheproductofthe deviationsfromtherespectivemeansanddividebythesamplesize.Ineffect,Itakeeach individualnumberinonesequenceandsubtractitfromthatsequence’smean,andthen multiplyitbythesameoperationdoneoncorrespondingnumbersinthesecondsequence. Ithensumeachproductanddividebythesamplesize.Theformulais∑ ∑∑ ∑(X(i)-µ µµ µ(i))(X(j)- µ µµ µ(j))/N.Inthisformula,theparenthetical“i”and“j”representthetwodifferentsamples. Eachsamplepointinonepopulationshouldhaveacorrespondingentryinanother populationanditbehoovestheanalysttodropdatapointsifonepopulationislargerthan theother.Workingthisexamplethrough:((1-3)(6-8)+(2-3)(7-8)+(3-3)(8-8)+(4-3)(9-8)+ (5-3)(10-8))/5=(4+1+0+1+4)/5=2 Thecovarianceisthesameasthevarianceinthiscaseandthetwostandard deviationsarethesameaswell.Althoughthenumbersinthesecondsequenceweremuch biggerthanthefirst,theriskofachievingthemwasthesamesimplybecausethedistance tothemeanwassimilar:eachnumberincreasedby“1”inthesecondsequencejustasthey didinthefirst.Intermsofrisk,onewouldwantthelowestcovariancepossiblewhichis achievableifsomenumbersrisewhiletheircorrespondingnumberfalls.Infact,apositive covarianceindicatesthetendencyforthepopulationstorisetogether,whilelowerriskis indicatedasthecovarianceapproacheszero. DOWNSIDERISK-Byexaminingthepopulationinpartsthatareaboveorbelow acertainvalue,theanalystcreatesathresholdlevel.Downsideriskmeasurements establisharisklevelthatiscustomizedtotheriskaversionofthepractitioner.Thesemi variancesumsonlythosedatapointsthatarelessthanthemean,butusestheentiresample 569 sizeforthedivisor,“N”.Thus,iffewpointsaresummed,thedownsideriskwillbe negligible,evenas“N’,thesamplesize,islarge.Thelowerpartialmomentsindicatoralso sumsthosepointsthatarebelowauserdefinedthresholdlevelwhichissubstitutedforthe meaninavariance-likefunction.However,thelowerpartialmomentsfunctioncanbe carriedtothethirdfourthorfifthpowersatthediscretionoftheanalyst. Tousethesemivariance,weformafunctionthatwillyieldtheriskofbeingbelow themean.Thefunctionissimply∑ ∑∑ ∑(µ µµ µ(i)-X(i)) 2 /N,withX(i)<µ µµ µ(i).Achosentargetlevel canalsobesubstitutedforthemean.Inthedatasequence(1,2,3,15,30,40),15.17isthe mean,andweonlysumthosedatapointslessthanthatnumberintheoperation:Semi variance= ((15.17-1) 2 +(15.17-2) 2 +(15.17-3) 2 )+(15.17-15) 2 /6=(200.79+173.45+148.11+ 0.0289)/6=87.06 Thelowerpartialmomentwillyieldthesameresultsasthesemivariancewhenthe thresholdlevelissettothemean,andthefunctionissquared.Thefunctionis:∑ ∑∑ ∑(X(i)- T) M /N,where“T”equalssomethresholdvaluethattheuserwantsto“beat”,and“M”isa valuegreaterorequalto“2”.Noticethatthemeanisnowheretobefoundandthatthe thresholdvalue,“T”,replacesit.“X”arethosevalueslessthan“T”.Todifferentiatethe twofunctions,wewillsetathresholdrateof“7”whichisnotpartofthedatapoint sequencefromabove:((1-7) 2 +(2-7) 2 +(3-7) 2 )/6=77/6=12.83.Noticealsothatthe differenceissetupforthepossibilityofraisingittothethirdfourthandhigherpowers. Whilethesemivarianceisalwayssquared,thepositionsofthemeanandindividualdata pointsinthedifferencingdonotmatter((µ µµ µ(i)-X(i)) 2 =(X(i))-µ µµ µ(i)) )) )) )) 2 22 2 ) )) ).However,sincethe lowerpartialmomentsfunctionmaybecubed,thedifferencingpositionmattersandthe individualdatapointissetuptosubtractthethresholdamount. Usually,bothofthesefunctionswillbecomparedamongalternativepopulations (portfolios,cashflows,stocks)andwhenthetwopopulationshavesimilarstandard deviations,thedecidingfactorcanbetheirdownsiderisk. 570 ANNUALIZEDVOLATILITY-Often,weneedtoconvertdatawithouthavingthe actualsample.Forexample,wemayhavemonthlydatabutwewantayearlyvalueforthe standarddeviationorviceversa.Byusingthesquarerootoftheperiodoneisconverting to–intermsoftheexistingdata-wecanmaketheseconversionseffortlessly,butweneed torecognizethatanyconversionwithouttheactualdataisa“guesstimate”atbest.The keytoconvertingvolatilityliesinconvertinganexistingperiodtothenewperiod,finding itssquareroot,andthenmultiplyingtheexistingstandarddeviationbythatnumber.The periodtowhichoneisconvertingisexpressedintermsoftheperiodictypeofdataone currentlypossesses.Thus,ifIcurrentlyhaveweeklydata,andIneedayearlystandard deviation,Iconvertoneyearintoweeks,whichisfifty-two.Ithentakethesquarerootof fifty-two,andmultiplytheweeklystandarddeviationbythatnumber.Asanextreme example,ifIwantedtoconvertfrommonthlytomillennial(1000years)data,Iwould expressthemillenniumintermsofmonths:12months=1yr,12000monthsequalone millennium.Thus,Iwouldmultiplymycurrentmonthlystandarddeviationbythesquare rootof12000. Naturally,shortertimeframeswillleadtomoreaccurateestimates,andthelongest timemostanalystswillconvertisdailytoannualdata.Followthelogicbehindthese conversions: Annualtomonthly:=(annualstandarddeviation)(√ √√ √1/12)) Monthlytodaily=((monthlystandarddeviation)(√ √√ √1/30) Dailytoyearly=(dailystandarddeviation)(√ √√ √365) ESTIMATEDVOLATILITY-Thistoolisavaluable,littleknown,quickestimator ofastandarddeviation.Itneedstobeusedcautiously,however,becauseitassumesa perfectlynormaldistributionwhichisrarelythecase. • 1.Subtractthelowestpointfromthehighestpointtodeterminetherangeofthe distribution. • 2.Squarethisrangeandthenmultiplyitby1/16ordecimally,0.0625. 571 • 3.Thisresultingfigureistheestimatedvariance.Wetakethesquarerootofthis numbertodeterminethestandarddeviation • 4.Informulatermsitis:0.0625(range) 2 =variance,thus:√ √√ √(0.0625(range) 2 )isthe estimatedstandarddeviation. Example:Ahighlyskewedpopulation-(7,8,14,21,56). • 1.56-7=49,whichisthehighminusthelow. • 2.0.625(49) 2 -150.0625=Variance • 3.√ √√ √150.0625=12.25=Standarddeviation Sincetheactualstandarddeviationis18.1andnot12.25,thestudent/investorcandetect thediscrepancyfromthehighskewofthepopulationwithtwoclosenumbersatthelow endandanumbermorethantwiceanyotheratthehighend. SAMPLESTANDARDDEVIATION-Insmallsamples,usuallylessthanthirty elements,thesamplesizeneedstobeadjustedtoincreasetheaccuracyofthecalculation. Wesimplyreduce(N),thesamplesizeby“1”andproceedasnormal.Theformulais modifiedas: ∑ ∑∑ ∑(X-µ µµ µ) 2 /(N-1)forthesamplevarianceandtheapplicationofasquarerootforthe samplestandarddeviation:√ √√ √(∑ ∑∑ ∑(X-µ µµ µ) 2 /(N-1)).Thus,ifthesumofsquareddeviations was60,andthesamplesizewas15,wewoulddivide60by14andget4.2857forthesample variance.Obtainingthesquarerootwouldyield2.0701. Notethatthesmallervalueof“N”,pusheduptheexpectedvolatilityofthenumber. Ifthedenominatorwere“15”andnot“14”,thevariancewouldhavebeencalculatedas “4”;thesmallersamplesizeaddedsevenpercentofvolatilitytothecalculation. THEMEAN-VARIANCE-Thisisamajorbutoftenneglectedindicatorofrisk.It isagreatequalizerbetweenhighreturninvestmentsthatarevolatileandlowerrisk, “safer”investments.Inanysample,wemerelysubtractthestandarddeviationfromthe meanandcompareittoanotherpopulation.Thefunctionis(µ µµ µ-σ σσ σ).Thehigherthe number,thegreaterthecombinationofreturnversusrisk.Althoughthisindicatorgivesa 572 veryaccurateassessmentoftherisk-returntradeoffwhenpopulationshaveasimilar magnitude,itislessaccuratewhencomparingasmallreturntoaverylargeone.The followingexampleshowsthisdivergence:XandYaretwodifferentstocks.Xhasamean of20%forareturnandastandarddeviationof24%.StockYhasameanof6%anda standarddeviationof8%.Thecomparisonis:(20-24=-4)<(6-8=-2).Bythisrule, stockYhasbetterrisk-returncharacteristicseventhoughtheinvestorwouldforego14% inpotentialreturns.Thenextindicatorbetterclarifiesthetradeoffwhentwopopulations aredivergent. THECOEFFICIENTOFVARIATION-Toovercomethelimitationsofthemean variance,wemerelydividethemeanintothestandarddeviation.Thefunctionis(σ σσ σ/µ). µ). µ). µ). Thisindicatorwillassesswhethertheextrareturnisworththerisk.Alowervaluewill indicatebettercharacteristics.Intheaboveexample,StockXhasacoefficientofvariation of(24/20)=1.2.StockYhasacoefficientofvariationof(8/6)=1.333.Thus,bythe coefficientofvariation,StockXisabetterchoicewithmorepotentialreturnandlessrisk associatedwithit-incombination.Byriskfactorsalone,anypopulationwithahigher standarddeviationismorerisky. WORSTCASESCENARIOS-Neitherthemean-varianceruleorthecoefficientof variationcoverallriskscenarios.Attimes,theanalystmaywanttomakeprovisionsfor thestatisticallyworstcasescenariothatwouldoccurbychanceonepercentofthetime.By multiplyingastandard,statisticallyderivednumberwiththestandarddeviation,andthen subtractingthatproductfromthemean,theanalystwillproduceanestimateofaworst casescenario.Thestandardnumberscoverspecificpercentagesofa“normal”population andsothistechniqueshouldnotbeappliedtohighlyskeweddata.Thesestandard numbersareasfollows: 573 Table23-1 NORMALDEVIATES PercentageoftheCurveCovered Standardnumber(NormalDeviate)x STD.DEV. 90% 1.64 95% 1.96 98% 2.33 99% 2.57 Figure23-5 Thesestandardnumberscoverthecurvefromthemiddle,leaving(1-Percentage/2)on eachside. Thus,asanexample,ifIhaveastockwhosemeanyearlyincreaseis10%witha standarddeviationof20%,andIwantedtofindouttheworstcasescenariothatwould happenonly5%ofthetime,Iwouldmultiply0.2by1.64whichisthenormaldeviateat 90percent,andthensubtractthatproductfrom0.1.Thefunctionyields,0.2x1.64=0.328, (0.1-0.328=-0.228)or-22.8percent.Foreveryfiveoutofonehundredperiods,Iwould 5% 5% 90% 1.64 574 expectalossof22.8percent.Noticealsothatwhilefivepercentofthecurveisuncovered onthelowerside,theuppersidehasacompanionareaoffivepercentthathasamirror probability.Likevolatilityestimations,theaverageinvestorshouldbewaryofsuch forwardlookingstatistics.Althoughanentirebranchofriskmanagementisdedicatedto thistypeofanalysis(Valueatrisk,betterknownas“VaR”),itisusedincombinationwith severalotherindicators(exponentiallyweightedmovingaveragesandMonteCarlo analysis)todiversifyawayitsmajorpremise-thatthepopulationisconsidered“normal”, ararityintheworldoffinance. ACCOUNTINGFORADDITIONALRISK:COMBININGSTANDARDDEVIATIONS Whenunitsaresimilar-dollars,widgets,boardfeet,butnotpercentages-wecan addstandarddeviationsandsometimesreduceriskbycombiningtwoormorepopulations. Thecovariancebetweenthepopulationsbecomespartofthecombinedstandarddeviation anddeterminestheamountofriskreduction.Ineffect,thisreductionformsthefoundation for“modernportfoliotheory”aconceptthatisstillasvalidtodayaswhenHarry Markowitzfirstproposeditmanyyearsago.Infact,almostalldiversificationhasthe mathematicalconceptofcombinedstandarddeviationsasitspremise.Riskisreducedby theapplicationoftwoprinciples:1)Thesamplesizeisincreasedand2)Therelationship betweenthepopulationsisopposed;asonemovesup,theothermightmovedown,oreven notatall. Toformacombinedstandarddeviation,westartbyformingavariancebyadding theseparatevariancesofthepopulationsandthenaddingthesummationofallpossible combinationsofcovariances,multipliedbythenumber“2”.Thefunctionisdefinedasσ σσ σ 2 (i)=varianceofpopulation(i),andσ σσ σ 2 (j)=varianceofpopulation(j).Theseare combined: σ σσ σ 2 (i)+ ++ +σ σσ σ 2 (j)+(2)(∑ ∑∑ ∑COVij)Justaswithanyvariance,wemerelyapplythesquareroot tothefunctiontoobtainastandarddeviation.Thisnextexampleformsaportfolio 575 betweenstocksX,YandZ.Inthiscase,wewillhaveasumofthreeseparatevariancesand covariances. Table23-2 Three Populations X Y Z Portfolio Year1 1 12 30 43 Year2 2 7 41 50 Year3 3 14 49 66 Population Means 2 11 40 53 StandardDev. (Sample) 1 3.605 9.54 11.79 Table23-3 Combinationsof Covariances COV(X,Y) 0.9998 COV(X,Z) 9.5006 COV(Y,Z) 6.499 FormulafortheVariance:σ σσ σ 2 (i)+ ++ +σ σσ σ 2 (j)+(2)(∑ ∑∑ ∑COVij) =(1) 2 +(3.605) 2 +(9.54) 2 +(2)(0.9998+9.5006+6.499)= 139.006,=√ √√ √139.006=11.79=Thesameastheportfoliostandarddeviation. Thereaderwillobservethatthelargestamountofvariancewasderivedfromstock Z,inadditiontothelargeamountofcovarianceamongthecombinations.Hadsomeofthe numbersdeclinedasothersrose(asinthecaseofY)thestandarddeviationmighthave beensmaller.Obviously,whenmorecovariancesareused,aspreadsheetisthebesttoolto calculatethem.Infact,usingthisformulawillshowtheimpactofanyportfoliodecision fromaperspectiveofrisk. 576 PREDICTIONANDVARIANCE Statisticianshavefoundthatwhenasamplepopulationislargeenough,afairly symmetricalcurvedevelopswiththemeaninthecenter.Eachstandarddeviation representsaspecificpercentageofareaofthatcurve.Multiplestandarddeviationscanbe matchedwiththecoverageofastandardareawhichwasthebasisfortheformationof “normaldeviates”inthe“worstcasescenario”analysis.Whilewecanusethis“normal” curveforestimationsandpredictions,wemustbecognizantthatfinancialdataisusually highlyskewedCreatinga“ballpark”estimateisbetterthanhavingnoneatall,onlyifwe haveothercorroboratingmeasurementsthatdiversifyandvalidatetheanalysis. Evenmoretenuousthantheshapeofthecurveisthefactthatwearemaking determinationsfromhistoricaldata.Incaseaftercase,historicaldatawillmisdirectthe analystunlessitisweightedwithmorerecentdata;hencetheevolutionoftheEWMAor “exponentiallyweighted”movingaverage.Althoughthecapitalstructuralistavoids makingpredictionsandconcentratesondetectinganenvironmentthatisconduciveto earnings,heorshewillneedtomakethembecausecustomers-aCFO,abiginvestor,ora pensionfundmanager-cravecertaintyandexpectthem.Thus,theanalystshouldavoid anydependenceonforecasts,butwelcometheirinformationalcontent-cautiously. CONFIDENCEINTERVALS-Onceameanisdetermined,itmayormaynotbea goodestimatorofapopulation,dependingonthesizeofthevarianceandbydefault,the standarddeviation.Thenormalcurvetheoreticallycoversanentirepopulation.The mean,plusonestandarddeviationcoversapproximately68%ofthenormalcurve. Moreover,95.5%ofthepopulationlieswithintwostandarddeviationsofthemean,and 99.7%lieswithinthreestandarddeviations.Thusonanormalcurve,ifIhaveameanof “7”,andastandarddeviationof“2”,99.7%ofthepopulationwillbebetween(7-3(2))and (7+3(2))orthenumbers“1”and“13”. The99.7%leavesjust0.3%or0.003(decimal)intotalareaontheendsofthecurve thatareuncovered.Oneachend,therewillbe(1-(percentcovered/2))or(1-(0.997/2))= 577 0.15%ateachend.Someconfusionexistsamongstudent/investorsastothecoverageof standarddeviations,andonemustrealizethatthecoveragegoesinbothdirections: Figure23-6 Rememberthattheliteraturealwaysregardsstandarddeviationsaseitheraddedtoor subtractedfromthemean.Thuswhenwespeakoftotalareacovered:,2σ σσ σcoversatotalof 68%,and4σ σσ σcoversatotalof95.5%.Finally,6σ σσ σcoversatotalof99.7%. Statisticianswilloftenrefertothepositivevariantofastandarddeviation(thearea greaterthanthemean)withoutregardtoitsopposite,thenegative,andthetotalarea covered.Weoftensay“within”suchadistancebutinactuality,wemeanthatthedistance coverstwicethereferencebecauseitappliestobothsidesofthemean.Thus,ifa statisticiansaysthatacertainnumberis“within”2.5standarddeviationsofthemean,the totalcoverageistwiceasmuch-orfivestandarddeviations. Anothermisconceptioncanbemadeabouttheareanotcoveredbythestandard deviations.Manyfalselybelievethatthepercentagenotcovered(0.3%intheabove example)liesatoneendofthecurveortheother.Infact,weneedtodividethisareaby twoandrealizethatthispercentagewillexistatbothendssinceweexpandfromthe 1 13 7 99.7% 3Std.Dev. 3Std.Dev. 0.15% 0.15% 578 centrallylocatedmean.IfIam“within”twostandarddeviationsofthemean(orcovering fourstandarddeviationsintotalarea),itmeansthat95.5percentofthecurveiscovered, butitdoesnotmeanthattheleftover4.5percentisatoneendofthecurve.Infact,2.25 percentareatbothends. Ifthestudentwillrecall“theworstcase”analysis,heorshewillobservethatthisis merelyaspecificcaseofdeterminingthe”negative”effectofastandarddeviationgoing belowthemean.Thusa“confidence”intervalismerelyaspecificnumberofstandard deviationsthatarematchedwithapercentageonthenormalcurve-1.64for90%,1.96for 95%etc.Weincreasetheaccuracyoftheintervalbydividingthestandarddeviationby thesquarerootofthesamplesize:Thefunctionis:µ µµ µ± ±± ±σ σσ σ/√ √√ √N.Thuswearemakinga predictionthatanumberwillbetweenacertainintervalaspecificpercentageofthetime. Table23-4 NumberofStandardDeviations(above mean) PercentageoftheNormalCurve Covered 1.64 90 1.96 95 2 95.5 2.33 98 3 99.7 Example:withameanof“4”andastandarddeviationof“5”,whatarethe confidencelimitsat98%ifthesamplesizeis“60”?Thisisequalto4± ±± ±(2.33)(5/√ √√ √60)=4 +1.5or5.5and4-1.5or2.5.Thusthe98%confidencelimitsare5.5and2.5. THENORMALDEVIATEOR“Z”SCORE-Itisoftenhelpfultodeterminethe probabilityofaneventhappeningornothappeningbyexaminingthenumberofstandard deviationsthatacertainnumberisawayfromthemean.Sincethenumberofstandard deviationscorrespondstoaknownareaofthecurve,wecandetermineprobabilitybyhow muchofthecurveiscovered.Wemerelysubtractthemeanfromthenumberanddivide 579 bythestandarddeviation.Thenumberofstandarddeviationsproducedcanbematched withastandardstatisticaltabletodeterminethepercentageofareacovered.Ineffect,the functionis:(X-µ µµ µ)/σ σσ σ. Forexample,supposeourmeanis“20”withastandarddeviationof“5”,andwe wanttofindouthowmuchareawouldbecoveredfromthemeantothenumber“7”. Thus,(7-20)/5=-2.6andtheabsolutenumberofstandarddeviationsis2.6,exceptthis numberis2.6standarddeviationsbelowthemean.Onatablethatisappropriatelycalled the“normaldeviate”whichisfoundinallstatisticaltextsandonsomecomputers,Ilook up2.6andfindthatitcorrespondstothenumber“0.4953”.Thisistheareabetweenthe meanof“20”andthedatapointof“7”.IfIdoublethe0.4953,Ihavethedecimal equivalentofthetotalareathatwouldbecoveredonbothsidesoftheentireinterval,both aboveandbelowthemean.Inthiscase,thenumber“33”wouldbethesameareaabove themean(20+13=33,20-13=7)andwouldalsocorrespondto0.4953.Thefunctional representationofthisassumptionis:(33-20)/5=2.6.Thustheintervalbetween“7”and “33”wouldcover0.4953x2or99.06percentofthepopulationTheprobabilityofa numberinthepopulationthatisbetween“7”and“20”,is49.53%justasitisbetween “20”and“33”.Toobtainanumberthatisgreaterthan“33”orlessthan“7”inthis distribution,wecansubtractthe0.9906(99.06%)from100percent(1)andthendivideby twotoobtain0.0047or0.47percent. Theprobabilityofobtainingaspecificnumberunderthenormalcurveiszero percent,becausetheareaunderanyoneunitiszero.Tofindouttheprobabilityofanyone number,twochoicesareavailable:wecanfindthefrequencyofthenumberinthe populationanddivideitbythesamplesizeofthepopulation-or-wecanapproximateit byfindingtheareabetweenthatnumberandtheclosestpossiblenumberandinterpolate. Tointerpolate,wefindtheareaunderthecurveforthenumbersthatareadjacent tothenumberofwhichwewanttofindtheprobability.Again,thisistheareathatcovers thedistancebetweenthemeanandthoseadjacentnumbers.Wethensubtracttheareasto 580 producetheintervalbetweenthenumbers;itmaybeanintegeroritmaybecontinuous, dependingonthesample.Ifitiscontinuous,thereareaninfinitenumberofvaluesinthe interval,andtheprobabilityofgettinganyoneofthemisstillthedifferencebetweenthe adjacentareas.Toillustrate,usingthesamedataasinthelastexample,“7”isbetween“6” and“8”.TheZdeviatesare(6-20)/5=-2.8and(8-20)/5=-2.4.Thesecorrespondto thedecimalareas,0.4974and0.4918,respectively.Wesubtract0.4918from0.4974and obtain0.0056.Thustheprobabilityofhavinga“7”inthepopulationwhenthemeanis “20”andthestandarddeviationis“5”,is0.56percent UPDATINGTHEMEANANDSTANDARDDEVIATION:MOVINGAVERAGES- Manybankswillusethe“worstcase”scenarioanalysistotrackthevalueoftheirassets (especiallyderivatives)onanygivenday.Thistypeofvalue-at-riskcalculationattemptsto compensateforitsnormaldistributionwithatechniquethatgivesmoreweighttorecent databycontinuallyupdatingitwithamovingaverage.Whilea“mean”givesthe arithmeticaverageoveranentirepopulation,amovingaverageupdatesittoreflectthe mostrecentdata.Ineffect,itshowstrendmovementbetterthanastaticaverage.Thus, whena“short”movingaverageiscomparedtoarelativelylongermovingaverage,it exhibitspressureonthedatatomoveupordown;thistypeofvectoringmakesmoving averagesthemainstayoftechnicalanalysisbutwillnotpredictvolatilejumpsorreactions torandomevents. Themethodologybehindamovingaverageisthatitismeasuredbydroppingthe lastdatapointinthesample,andupdatedwiththeadditionofanewdatapoint.Thetime spanbetweenfirstandlastdatapointsprovidesthesamplesize.Inasixmonthmoving average,wewillhavesixdatapoints,let’ssay,(20,41,32,21,33,30).Theaverageinthisset is29.5.Whenwegointoanewmonth,wedroptheverylastdatapoint(20)andaddanew datapoint-let’ssay“52”,andthenewmovingaverageis(41,32,21,33,30,52)/6=34.83 Themovingaveragehasincreased,showingthesignificanceofthemostrecentdata(52). Thetrueaverage(mean)oftheentirepopulationwouldincludealldatapointsinthe 581 sampleandnotjustthesixthatwerepartofthemovingaverage.Althoughtechnical analystsusemovingaveragestotrackvolatilepopulationslikestocksandcommodities, movingaveragesaremoreusefulwhentheytrackrelativelystablepopulationslikemonth tomonthchangesinassetturnover;theabilitytoforecastatrendismoresignificantina lessvolatilepopulation. ACCOUNTINGFORSMALLSAMPLESIZE-Analogoustothenormaldeviateis the“studenttdistribution”,namednotforstudents,butforitsinventor,W.S.Gosset,who hadtopublishunderthenomdeplume,“Student”,becausehisemployerdidnotallowhim topublish.Thestudenttdistributionisverymuchlikethenormaldeviateexceptthatit appliestosmallsamplesizesunder30.Itsstatisticaltableisbasedonthesamplesize minusone,andacorrespondent“uppertail”area.Thestudent/investorwillremember thattheuppertailrepresentsonehalfoftheuncoveredarea(theotherhalfbeinginthe lowertail),andsoanuppertailareaof0.025signifiesatotaluncoveredareaof0.05anda coverageof95%.Ifoursamplesizeis“30”,wesubtract“1”andobtain“29”.Now,we needacorrespondingconfidencelevelwhichwecanpickarbitrarily(say,90%).The uppertailareaisalsocalled“alpha”andhasasymbolof“α αα α“andisequalto((1-0.90)/ 2).Wematchthe“29”(alsocalledthedegreesoffreedom)withthe0.05,andobtainthe number,“1.699”,fromthechart.Thiswillbeournumberofstandarddeviationsaway fromthemean.Itisusedinconfidenceintervalsjustasthenormaldeviateexceptthatit willbemultipliedbythesamplestandarddeviation(withN-1weighting)andnotthe largerpopulationstandarddeviation.Again,thestandarddeviation(sampleinthiscase)is dividedbythesquarerootofthesamplesize,“N”,whenusedinaconfidenceinterval. Example:mean=7,samplestandarddeviation=3,samplesize=30,confidencelimitsof 90%=1.699.Calculation:7± ±± ±1.699(3/√ √√ √30)=7.9305or6.0695. REGRESSION Theproblemwithregression-asappliedtofinancialdata-isthesameasthatof usingparametric(normalcurve)statistics:ouranalysisfitsnormaldatabutlosesaccuracy 582 whenhighervariabilityisencountered.Sincefinancialdataisvariedandanythingbut “normal”,statisticsbasedonanormaldistributioncangiveonlya“guesstimate”andnota trueprediction.Ineffect,wecanderive“ballpark”estimatesfromregressionbutmust neverbedependentonitfordecisionmaking.Usedincombinationwithmanyother indicators,itoffersavaluabletoolforcomparison,butattheotherextreme,itcanmislead theanalystintothinkingthathighlyassociatedvariablesareproofofa‘cause”. Regressionhasseveraldifferentforms,butthetypemostencounteredissimple linearregression.Itmayhelpthestudent/investortoviewitasanotheradaptationofthe standarddeviationbecauseineffect,itusesallofthemechanicsofcovarianceandvariance butplacestheminanassociativecontext.Itattemptstorelateonesetofdatatoanotherset byfittingthemtoastraightline,andmeasurestheirassociationbythe“correlation coefficient”or“R”.Thecloserthedatacomestoformingastraightline,themorerelated arethetwosetsofdata.Onascalefrom-1to1,thefartherfromzeroisthecorrelation coefficient,themorethedataisrelatedbecauseittendstoriseorfallatthesamerate. Correlationmustbeexaminedcloselybecausethereisatendencytoconfuseitwith cause.Thepractitionerhastorememberthatonlytwovariablesarebeingassociatedand thatsomeothervariable(onethatisnotbeingmeasured)maybeattherootofcausation. Forexample,thenumberoffootballgamesplayedisperfectlynegativelycorrelatedwith thetemperatureinautumn;thatis-moregamesareplayedasthetemperaturekeeps dropping.However,playingfootballinfallweatherhasnothingtodowiththe temperature,althoughitmayhaveahighcorrelationwithit.Therefore,regressionisa moreusefultoolwhenthereissomedeterministicrelationshipbetweenvariablessuchas thereisbetweensalesandoperatingincome,forexample.Operatingincomeisformed whentotalcostsaresubtractedfromsales,andthedifferencewillbedependentoncosts. Graphically,onepopulationisrepresentedontheverticalaxis,calledthe“Y”axis, whileanotherpopulationisrepresentedonthehorizontalaxis,calledthe“X”axis.Each pointonthegraphrepresentssomecombinationofsingle“X”and“Y”variables.The“Y” 583 populationchangesasweinputdifferentvaluesfor“X.Thus,wecallthe“Y”variablethe dependentvariableand“X”theindependentvariableRegressionwillattempttofitthe twopopulationsintoastraightlineintheform,Y=A+BX-eveniftheyareperfectly relatedinsomeotherwaysuchasinthefunction,Y=X 2 /2.“A”iscalledtheYintercept andoccurswhenXequalszero,while“B”representsthecoefficientof“X”,calledthe slope,whichisdeterminedbytheamountofchangein“Y”intermsofthechangein variable“X”.Inanexperimentalmode,wecaninferthatsomeactionupon“X”causesa changein“Y”statistically,butnotnecessarilyphysically;onlyanactualmeasurementof observationsinacontrolledenvironmentcanpointusinthedirectionofcauseandeffect. However,closelyassociatedvariablescanleadtofurtherexperimentationandidentifyan hypothesis. Ineffect,regressioncanprovidevaluableinformationbutdoesnotdeterminethe causeofanyevent.Ifweweretodeterminetheamountofkilocaloriesprovidedbyone barrelofoil,wecanproclaimthatthetwovariablesaredirectlyrelated,butwecannot claimthatmoreoilprovidesmorekilocaloriesbecausetheproclamationislimitedtojust twovariables.Eventhoughthehypothesisistrueinthiscase,regressiondoesnotprovide uswiththetoolstomakeanassurance;someothervariable,atmosphericpressurefor example,mayhaveplayedaroleintheexperiment. Moststatisticianswillleadthestudent/investorthroughwhatistermed“theleast squares”approachtoregression.Beforetheadventofpersonalcomputers,statisticians wouldperformregressionswithmultiplevariablesrequiringextensivecalculations;the “leastsquares”methodologycouldbelogicallyadaptedforusewithmanyvariables.This authorpreferstoviewtheanalysisasanextensionofthestandarddeviationbecauseitis simplertorememberandalsointegratesotherareasofriskintoit.Regressioncanbe learnedbymakingratiosofthetechniqueswehavealreadylearned:covariances,standard deviations,variancesandmeans.Thefollowingsectiondelineatessomeoftherelationships inlinearregression. 584 RELATIONSHIPSINLINEARREGRESSION • 1)THESTRAIGHTLINE-Thestraightline,Y=A+BX,haspreviouslybeen mentioned.The“Y”intercept,“A”,isapointontheverticalaxisthatwilloccurifwe inputzerofor“X”.The“B”orslopeisthecoefficientof“X”anddeterminesthe amountofvariationof“Y”intermsof“X”.Infinance,weoftenrefertothisratioas “beta”because“B”hastheGreekequivalentofthatdescription,andweapplyittoa particularstockandthemarket.However,“beta”canrefertothechangeinany populationintermsofthechangeinanotherpopulation. • 2)COVARIANCE-Wedefinedcovarianceas∑ ∑∑ ∑(X(i)-µ µµ µ(i))(X(j)-µ µµ µ(j))/N. .. .Inthis format,wecansubstitute“Y”for“X(j)becauseitdesignatesadifferentpopulation, andthereareonlytwovariables.Ifweknowthecorrelationcoefficient,“R”,weneed onlytomultiplythestandarddeviationsofXandYbyittoderivethecovariance: (R)(StandarddeviationofX)(StandarddeviationofY)orRσ σσ σ x σ σσ σ y . • 3)CORRELATIONCOEFFICIENT-Inmostcases,wewilldeterminethecovariance usingthestandardmethodology,whichallowsustodeterminethecorrelation coefficient“R”.“R”issimplythecovarianceofXandYdividedbytheproductofthe standarddeviationsofXandY:(Covariance.X,Y)/((Standarddeviationof X)(StandarddeviationofY)).Thismeasurementisusedtodeterminethedegreeof variationbetweentwopopulations;thecloserthemeasurementgetsto“1”or“-1”,the moreonepopulationvarieswithanotherpopulationwheneachdatapointmovestothe next.Thetwopopulationsareclosely“correlated”,becausethepatternofmovement betweendatapointsissimilarforeach. • 4)THESLOPEORBETA-Again,weusethecovarianceinthenumeratorandthis timewedivideitbythevarianceof“X”:(Covariance,X,Y)/(VarianceofX).Another wayofdetermining“R”isthroughtheuseoftheslope.Ineffect,ifwemultiplythe slopebythestandarddeviationofXanddividebythestandarddeviationofY,we 585 againobtainthecorrelationcoefficient,“R”.R=(Slope)(StandarddeviationofX)/ (StandarddeviationofY). • 5)THE“Y”INTERCEPT-“Alpha”,orthe“Y”interceptismerelytheproductofthe meanof“X”andtheslope,subtractedfromthemeanof“Y”:(MeanofY)- ((Slope)(MeanofX).for“purists”whopreferterminology,itis( (( (∑ ∑∑ ∑Y/N)-(∑ ∑∑ ∑(X(i)- µ µµ µ(i))(Y(j)-µ µµ µ(j))/N) )) )/σ σσ σ 2 x )(∑ ∑∑ ∑X/N)). THEBASICFIVE-Itshouldbeevidentthatregressioncanbeperformedonthe basisofonlyfivecomputations:1)Themeanof“X”.2)Themeanof“Y”.3)The standarddeviationof“X”.4)Thestandarddeviationof“Y”.5)Thecovarianceof“X” and“Y”.Mostreaderswillperformtheseoperationsonacalculatororinaspreadsheet. Knowingtherelationshipswillenablethestudent/investortounderstandtheoryandbuild evenmorerelationshipsbecausethebasicmeasurementofrisk,thestandarddeviation,is attheircore. THECOEFFICIENTOFDETERMINATION-Evenmoreimportantthanthe correlationcoefficient,“R”,isthecoefficientofdeterminationor“R 2 ”.Whenwesquare “R”,wecreateaproportionalindicatorthatestablishestheratiooftheamountof explainedvariationtotheamountoftotalvariation.Likethecorrelationcoefficient,the closerR 2 getsto“1”,thestrongeristhepotentialrelationship.However,lowervaluesare discountedatanincreasingratebecausetheactofsquaring“R”createsanexponential distribution.Thusavalueof0.900for“R”,onlygoesdownto0.810for“R 2 ”,butavalue of0.5for“R”willbecutinhalf,oncewesquareit. MAKINGPREDICTIONS-Forecastingwithsimpleregressionismostappropriate forneardeterministicrelationshipsthatarestable.Forexample,salescanberegressed againstoperatingincome,andafairlylinearpatternwouldresultbecauseonlythe varianceofcostswouldseparatethem.However,inanindustrythatisshiftingmargins, costscanvaryagreatdealoveradecade,andsotheresilienceofsucharegressionisonly 586 temporary.Outsideofengineering,andsomelogistics,fewrelationshipsinbusinessare deterministicenoughtoformalinethatcanbepredictedforlongperiods.Thus, regressionismoreamenabletoshortperiodsinwhichtherelationshipbetweenthe populationsisobservedtodecreaseorincreaseandcreatessomesortofcontingentaction formanagement. GROWTHRATES Inanutshell,growthratesrevealmoreaboutarelationshipbetweentwovariables thananyothermeasurement.Itistheessenceofcalculus,logarithmsandmeasured regression.However,somegrowthratesmaynotbemeaningful,whileothersaresovolatile thattheydefymeaningfulmeasurement.Wecancombinearegressiontechniquewith standarddeviationstocaptureboththemagnitudeandstabilityofagrowthrate;the student/investormustrememberthatthecoefficientofvariation(σ σσ σ/µ µµ µ)or(Standard deviation/Mean)appliestopercentagegrowthratesaswellasstaticunits.Therefore,all growthratescanbemeasuredintermsofthesamerisk-returncharacteristicswithwhich wemeasurefundamentalslikenetincome. GEOMETRICAPPROXIMATION-Averyquickandeasymethodofcalculating growthratesistoobtainthegeometricmeanbyusingtheintervalsbetweenperiods.In effect,thismethodestablishesatrendlineandpredictsthenextsequentialentry.The techniqueisbestusedwhenaquickcomparisonisneeded,orwhenthedataissostable thatmorecomputationisdeemedunnecessary.However,itwillnotbeasaccurateasthe techniquedescribedinthenextsection,logarithmicregression,andshouldnotbeusedto meetthestandardsofprofessionalanalysis.Weusetheseratesfreelywhentheevidence theyexhibitiscorroboratedbyotherindicators. Thebasefigureforthecalculationisaratioformedfromtakingthesame measurement,Xnumberofperiodsapart;thenearestperiodmeasurementisthe numerator,andthemeasurementtakenXnumberofperiodsoutformsthedenominator. Themeasurementcanbeafundamentallikesalesorincome,oranymeasurementthatcan 587 bephysicallycounted.However,ifthenumberisnotrepresentativeofatrend(itisheavily skewedinoneofthemeasuredyears),itshouldbereplacedbyamorerepresentative number,evenifthenumberisfartheroutintermsofthenumberofperiods.The judgmentoftheanalystisparamount,becauseheorshemustdecideifaparticular numberispartofatrend.Forexample,ifwehadasequenceof(7,10,11,14,9,17,18),we wouldthrowoutthe“9”andusethe“17”instead,simplybecause“17”ismore representativeofatrend.Thus,thistechniqueismoreaccuratelyappliedtostable populationslikedividendsthanitistovolatilemeasurementslikestockprices. Thenumberofperiodsisdeterminedbytheneedsoftheanalyst.Wefindthat measurementsoverfiveperiodsconfirmatrendbetterthanotherintervals,andindeed fiscalperformanceisdeterminedbyaperiodlengthoffive(December31toDecember31) withfourperiodsbetweenthatperiodlength.Theperiodnumberbetweenthelengthisthe determiningfactorinthisanalysisbecausewearemeasuringgrowth.ThusifNisequalto theperiodlength,then(N-1)isequaltothenumberbetweenperiods-whichiswhen growthoccurs. Oncewehavearatioofmeasurementstakenatperiodlength,weraiseittothe inverseofthenumberbetweenyears.Iffiveisourperiodlength,and(N-1)orfour,isour numberbetweenperiods,thentheratiowillberaisedtothepowerof1/4or0.25.For example,intheabovesequence,(7,10,11,14,9,17,18),wewilldictate“7”asthecurrent year,and“17”asthemeasurementfiveperiodsago,dropping“9”fromthesequence.In thiscase,thereis“negativegrowth”because7islessthanthe“17”fromfiveperiodsago. Thegrowthfactorisdeterminedbymakingaratioofthenumbers,(7/17)andraisingitto the0.25power,whichwillrequireacalculator:(7/17) 0.25 =(0.41176) 0.25 .=0.80105.When wesubtract“1”weobtainadecimalincreaseofthesequence:0.80105-1=-0.19895or- 19.89percent.IfIwantedtoknowthenextpredictednumberinthesequence,Iwould multiply0.80105by“7”andobtain“5.607”. 588 Onamorepositive“growth”note,if“17”werethecurrentyear,Icouldusethe numberfromfiveperiodsago,“10”,becauseitisintrendsequence.Theratiowouldbe(17 /10)andthegrowthfactorwouldthenbe(1.7) 0.25 or1.14186.IfIsubtract“1”,Iobtain a14.19percentgrowthrate,andifImultiplythegrowthfactorbythecurrentperiodto predictthenextperiod,Iobtain(1.14186)(17)=19.412. ACCURATEGROWTHRATESFROMLOGARITHMS Aslongasdataisapositivenumber,wecanobtainaccurategrowthratesfora sequenceevenwhennumbersarenot“intrend”.Inthesequence,(7,10,11,14,9,17,18), wewilldoafiveperiodgrowthrate,againusing“7”asourcurrentdata,butusing“9”as thedatafromfiveperiodsago. Thetheorybehindusinglogarithmsisthattheywillformastraightlinewhen appliedtoagrowthrate.Thismakesthemamenabletolineartechniqueslikeregression. AnycurrentmeasurementisequaltotheproductofthemeasurementXperiodsagoanda growthratethatisraisedtothepowerofthenumberofperiods.Thus,2008SALES= (2004SALES)(1+GrowthRate) 5 .Todeterminethegrowthrate,wemerelyusetheruleof logarithmswheremultipliedproductsturnintoaddedlogarithms.Wetakethelogarithm ofbothsidesoftheequation: Log(2008SALES)=Log(2004SALES)+5Log(1+GrowthRate) Inthisconstruction,ifwetakethelogarithmsofeachnumberinthesequence,and thenmultiplyitbythedeviationoftheperiod,wecanformasummationthatactsasa numeratorforabetatypevariable.Thedenominatorwillmerelybelikeavariance,the squareddeviationsoftheperiods.Toseehowthisworks,wesetupachartthatdesignates T=(totalnumberofperiods).Eachnumberinthesequenceismatchedwithitslogarithm, andaperiodtoidentifyit.Thelastnumberinthesequence(theperiodfarthestawayfrom thecurrent)isdesignatedperiod1,thesecond,period2..Thecurrentperiodwouldbe designatedasperiod“5”.Wethenformameanoftheperiods((1+2+3+4+5)/5)=3Ifwe designatethelogarithmofthemeasuredfundamentalas(X(i)andtheperiodasT,wewill 589 refertotheperiodmeanas(MeanT)andwecanformthefunction:∑ ∑∑ ∑((X(i)(T-MeanT)) /∑ ∑∑ ∑(T-MeanT) 2 .Notethattheperiod,“T”,andMeanTarenotlogarithms. Table23-5 X(i) T Sequence Number Logarithm Period (X(i))(T-TMean) 7(Current) 1.946 5 3.892 10 2.303 4 2.303 11 2.398 3 0 14 2.639 2 -2.639 9(Beginning) 2.197 1 -4.364 STEP1:Formthenumerator.Sumthelastcolumn,(X(i)(T–TMean) STEP2:Formthedenominator.Σ(T–TMean) 2 . STEP3:Dividethenumbersandputtheminanalogform(EXP),andsubtract“1”. Toformthenumerator,wesumthelastcolumnandobtain-0.808.Thedenominatoris formedby(1-3) 2 +(2-3) 2 +(3-3) 2 +(4-3) 2 +(5-3) 2 =10.Theresultingfigureistheratio- 0.808/10whichneedstobeinanalogform.-0.808/10=EXP(-0.0808)=0.922378=(1+ Growthrate),0.922378-1=-0.07762ornegative7.76percentgrowth.Theactualnegative growthwasmuchlessthandictatedbythecurrenttrend. SPEARMANRANKCORRELATION:NONPARAMETRICSTATISTICS Althoughparametricstatisticsdominatesthefinancialworld,nonparametric methodsareoftenmoreappropriate.Emphasistendstobeplacedonlinearsolutionswhen thevariablesactinacurvilinearfashion;therelationshipsbetweenvariablesrarelyforma straightline.Nonparametricstatistics,themethodologythateschewstheformalizeduseof meanandstandarddeviation,cansometimesfindcorrelationwhenregressionfails.IfX risesinanexponentialpattern,nonparametricstatisticswillcapturetherelativepattern betterthanparametricstatistics. Theprimarydifferencebetweenparametricandnonparametrictechniquesisthat thelatterwillusuallyrankdatafromlowesttohighestbeforeperformingoperationsonit. 590 Inthisregard,themethodologybecomesordinaland“monotonic”,lessresponsivetothe amountofincreaseandmoreresponsivetotheincidencethereof.Forexample,parametric statisticswillemphasizeanaverageincreaseinapopulationwithoutreferencetothe amountofdatapointsthatsignificantlycontributedtoit;itwillsummarizethe contributionwithacorrelationcoefficient,“R”.Ontheotherhand,nonparametric statisticswillrefertotheamountofdatapointsaboveacertainlevelwithlessreferenceto theaverage,andmoreregardtothelevel:highmediumorlow.Thus,extremevalues contributetothecorrelationasmuchasvaluesthatarerightaroundtheaverage. IntheSpearmanrankcorrelation,wedothesamelinearregressionthatisdonein normalparametricstatistics,exceptthatweperformitonrankeddata.Asanexample, considerthecorrelationbetweenearningsandstockprices.Ifoneperformsregularlinear regression,onearningsandmarketpricechanges,onefindsthecorrelationislowbecause thevariabilityofthechangesissogreat.Alternatively,intheSpearmanrankcorrelation, theassociationishighbecausethereislessregardtotheamountofchangeandmore regardtoitsoccurrence.Stockpricechangestendtoincreasewithearningschanges,and thefactisfirmlyestablishedbySpearmanrankcorrelation.Ineffect,ifonedidlinear regressionbyassigning“1s”and“0s”tothe“Y”variableforincreasesanddecreases,one wouldaccomplishasimilarobjective. ThesumofsquaresmethodologyappliestoSpearman.Whenwederivethe“R”, correlationcoefficient,weareactuallymakingaratioofthesumofsquareddeviations amongthedifferentpopulations.Thus“R”isproperlyreferredtoasthe“Pearson product-momentcoefficientofcorrelation”.Inmathematicalnotation,itisSS(xy)/ √ √√ √(SS(x))(SS(y)).Inextended,texttypenotation,itis:SS(xy)=∑ ∑∑ ∑(X(i)-MeanX)(Y(j)- MeanY),SS(x)=∑ ∑∑ ∑(X(i)-MeanX) 2 ,SS(y)=∑ ∑∑ ∑(Y(j)-MeanY) 2 .Innormallinear regression,thisformulaappliestotheunaltereddatapoints.InSpearman,“X”refersto thenumberedrankofdatafromlowesttohighestand“Y”istherankingfromitsdata fromlowesttohighest. 591 Naturally,manyoftherankingsmaybetiedwhentherearefrequentduplicates amongdatapoints.Tocircumventthisproblem,tiesaredividedproportionatelyamong therank(withadecimal)andthenextrankupiseliminatedinfavorofthedividedrank. Thenumberofrankswillmatchthetotalnumberofelementsinthesample(including repetitions).Forexample,thefollowingisanincidenceinthenumber“7”rankofa repeatednumber,“34”. Table23-6 RANKINGSFORONESEQUENCEOFNUMBERS(XorY) NORMALRANK NUMBER SPEARMAN RANK NUMBER 1 14 1 14 2 16 2 16 3 17 3 17 4 22 4 22 5 25 5 25 6 27 6 27 7 34,34,34 7.33 34 8 41 7.33 34 9 44 7.33 34 9 41 10 44 Ifthereareveryfewrepeateddatapoints,thepractitionercangetbywiththe followingequationwhichsimplifiestheprocess:R=1-[(6∑ ∑∑ ∑(D) 2 )/(N(N 2 -1))].Inthis function,“D”representsthedistancebetweentheranks,whichis(X(i)-Y(i)),andN representsthesamplesize.Notethatthesamplesizedoesnotmatchthenumberof rankingswhentherearerepeatedentries. Onceweobtain“R”,wehavetotestwhetheritissignificantgiventhesample numberandprospectiveconfidencelimits;thesmalleristhesamplesize,thehigher“R” mustbeatanyconfidenceleveltobesignificant.Weuseappropriatestatisticaltablesora computerizedprogramtomatchtheobtainedcorrelationcoefficientwithcriticalvalues 592 giventhesamplesize,atanyconfidencelevelthatwechoose.Thisisanexampleof “hypothesistesting”;theprobabilityof“noassociation”orthe“nullset”liesbetweenthe criticalvalues.Ifweobtainaverylowprobabilityof“noassociation”,weinterpretitasa significantcorrelation.An“R”valuethatliesinthemiddleofthecurvehasahighvalueof notbeingassociated,whilepointsattheendsofthecurvehavesignificantvalues(alow probabilityofnoassociation).Likethenormaldeviatechart,aconfidencelevelof90% wouldhave.05atbothends.Thusifthechartdisplaysthe“alpha”area,wemultiplyitby twoandsubtractfromonetodeterminetheconfidenceintervaltowhichitrefers. Toillustrate,thissimpleexamplewillsuffice: Table23-7 Rankings X RankforX Y RankforY 1 1 66 3 2 2 65 2 3 3 64 1 4 4 67 4 5 5 68 5 BythePearson-Spearman“R”,SS(xy)=6,SS(x)=10,SS(y)=10,SS(xy)/ (√ (√ (√ (√(SS(x))(SS(y))=6/√ √√ √100=6/10=0.6.Welookonthechartfor90%confidencelimits (.05“alpha”),andasamplesizeofN=5,anditdisplays0.900.Thecorrelationcoefficient correlationof0.6doesnotexceed0.900andsothesampleisnotconsideredsignificantly correlatedatN=5,R=0.6,90%confidencelimits. SAMPLESIZE Designinganexperimentwithasamplethatreflectsthecharacteristicsofthe greaterpopulation,isamethodologybeyondthescopeofthistext.However,mostflawed studieshaveaninadequatesampleattheircore.Asamplebecomestaintedbythebiasor 593 lackofknowledgeofthepractitioner.Asanextremeexample,considertheattemptto determinethepercentageofbrowneyedpeopleonthecontinentofEurope.Ifwedecided totakeoursamplefromthecountryofSweden,anddisregardedSpain,Italy,andFrance, ourresearchwouldbehighlyskewedbecausesomanyScandinaviansareblue-eyed. Obviously,designingasignificantsamplecanbeadifficulttaskbecausetheremaybe informationorknowledgethatthestudydoesnotconsider. Ineffect,thereisoftenatemptationtomakethesamplefitadesiredobjective,such ashappenswhendemocratspollonlydemocrats,orrepublicanspollasectionoftheir constituency.Analogously,whenweinterpretdata,itisoftensimilartosamplebias becauseweviewitfromanarrowperspective-theneedtocoordinatenewinformation withwhatwealreadyknow.Thus,awellrunstudywithanadequatesampleshould ideallybeinterpretedbyadispassionateobserverwhomayknowlittleaboutthe expectationsanddetailsoftheresearch.Armedwithaknowledgeofstatistics,theobserver canverifythehypothesisbycoordinatingnumbersalone.Thisistheessenceofthe“blind” studiesfavoredinthephysicalsciences. Thevariablesthataffectfinancialdataareoftenuncontrollable.Apremisethatwas truethirtyyearsago,establishedbyacauseandeffectrelationship,mayhaveno correlationatalltoday.Taxlegislation,technology,demographicsandforeigncompetition canaltertheinteractionsbetweenvariables.Itisthereforedifficulttoestablishany deterministicrelationshipsamongthem;thetransienceofanystatisticalconclusionis confirmedeachtimeaprobabilityislessthanonehundredpercent.Whileitisthe objectiveofthefinancialresearchertodeciphertrendsandgleanactionableconclusions, thesearchmustbetemperedbytherecognitionthatitisprimarilyafunctionoftimewith resultsthataresubjecttochange. Withthegreatvariationamongdatathatoneencountersinfinance,thesamplesize mustbelargertobesignificant.Thestatisticianmustalwaysattempttobuildthesample 594 basedonexpectedvariation.Toestablishanestimate,thepractitionerneedsthreepieces ofinformation: • 1.Anormaldeviatethatestablishesaconfidenceinterval,say95percentwhichisequal to1.96 • 2.AvariancethatcanbeestimatedfromasmallersamplefromthepopulationwithN- 1weighting. • 3.Theamountoferrorthatwillbetolerated. Theamountoftolerableerror,alsocalled“boundonerror”isequatedwithtwo standarddeviationsdividedbythesquarerootofsamplesize,or2σ σσ σ/√ √√ √N.Obviously,this function“begsthequestion”becausewedonotknowthesamplesize.However,whatever leveloferrorwechoosewillhaveaneffectonsamplesizeandattemptstomimicthis constraint Forthevariance,wecanalsoobtainanestimatefromthepreviouslydiscussed “rangeestimationmethod”.Therangefromlowestnumbertohighestnumberisusually aboutfourstandarddeviationsina“normal”population.Wemultiply1/16bytherange andthensquarethatnumbertoobtainanestimatedvariance.Thus,theestimatedvariance is1/16(range) 2 . Wearrangethethreepiecesofinformationinafunction((Normaldeviateatagiven confidencelevel) 2 (Variance))/Tolerableerror 2 ,orZ 2 σ σσ σ 2 /E 2 .Avariationofthisfunction existsinsomestatisticstexts:Nσ σσ σ 2 /(N-1)(D)+σ σσ σ 2 .Inthisinstance,“N”representsthe sizeoftheentirepopulationfromwhichthepractitionerwantstogarnerasampleand“D” isequalto E 2 /4.Whentheentirepopulationisverylarge,Nσ σσ σ 2 /(N-1)(D)+σ σσ σ 2 reducesto,(4σ σσ σ 2 /E 2 ) anditisbesttocheckallthreecalculationsandchoosethosethatagree. Sinceeachofthesefunctionscanyieldadifferentanswer,thejudgmentofthe practitionerisparamount.Alwayschoosethelargestsamplesize.Asanexample,suppose Iwanttoestimatethemeanreturnforcompanieswithovertenbilliondollarsindebtand 595 myresearchshowsthat100suchcompaniesexist.Asamplestandarddeviationshowsa20 %variationamongreturnsandIset5%asmytolerable(boundon)errorata95% confidencelevel.Aquickcalculationshows:(1.96) 2 (.2) 2 /(0.05) 2 =61.46.Inextdothe( 4σ σσ σ 2 /E 2 )calculationandobtain0.16/0.0025=64.Finally,Idotheothervariant,100(.004) /(99)(0.000625)+0.04=4/0.101875=39.26Sincethefirsttwocalculationsarein agreementandaremuchlarger,Iconcludethatthepopulationsizeistoogreattousethe (Nσ σσ σ 2 /(N-1)(D)+σ σσ σ 2 )variant.Ichoosethelargeststandardwhichisasamplesizeof “64”. (BacktoTableofContents) x SELECTEDREFERENCES Altman,EdwardL.“FinancialRatios,DiscriminantAnalysisandthePredictionof CorporateBankruptcy.”JournalofFinance.September,1968. Brigham,EugeneandMichaelC.Ehrhardt.FinancialManagement:TheoryandPractice. n.p.:HarcourtCollegePublishers,2001 Byoun,Soku.“WhyDon’tSomeFirmsUseDebt?.” http://www.fma.org/SLC/Papers/zero_debt2006_01.pdf.(April13,2008) Byoun,Soku.“HowandWhenDoFirmsAdjustTheirCapitalStructuresToward Targets?”http://www.afajof.org/afa/forthcoming/3724.pdf Cheng,LouisT.W.,GayB.Hatfield,andWallaceN.DavidsonIII.“TheDeterminationof OptimalCapitalStructure:TheEffectofFirmandIndustryDebtRatioson MarketValue.”JournalofFinancialandStrategicDecisions.8(Fall1994):1-14 Crosby,PhilipB.QualityisFree.NewYork:Mentor,1980. Damadoran,Aswath.“DebtandValue:BeyondMiller-Modigliani.”PowerPoint Presentation. http://pages.stern.nyu.edu/`adamodar/pdfiles/country/levvaluehort.ppt.(August13, 2008). Dolan,EdwinG.andDavidLindsey.BasicMacroeconomics:PrinciplesandReality. HinsdaleIllinois:TheDrydenPress.1974. Downes,JohnandJordanElliotGoodman.DictionaryofFinanceandInvestmentTerms. 6thed.Barron’sEducationalServices,Inc.,2003. Duett,Edwin,AndreasMerikasandManolisD.Tsirijakis.“APedagogicalExaminationof theRelationshipBetweenOperatingandFinancialLeverageandSystematicRisk.” JournalofFinancialandStrategicDecisions.9(Fall1996):1-8. Ehrbar,Al,G.BennettStewart,JoelM.Stern.EVA:TheRealKeytoCreatingWealth. JohnWileyandSons,Inc.,1998. xi Ehrbar,AlandG.BennettStewartIII.TheEVARevolution.JournalofApplied CorporateFinance.12(Summer1999):18-31 Fama,Eugene,F.andKennethR.French.“TheCapitalAssetPricingModel:Theoryand Evidence.”JournalofEconomicPerspectives.16(Summer,2004):25-46. Fisher,PhilipA.CommonStocksandUncommonProfits.NewYork:JohnWileyand Sons,Inc.,2003. Garrod,NeilandDusanMramor.“OnAccountingFlowsandSystematicRisk.”Feb.4 1999.http://miha.ef.um-lj.si/-dokument/wp/garrod1.doc.(August8.2008). Grieder,William.SecretsoftheTemple:HowtheFederalReserveRunstheCountry.New York:Touchstone,1987. Hagin,RobertL.TheDowJones-IrwinGuidetoModernPortfolioTheory.Homewood, Illinois:DowJones-Irwin,1979. Hagstrom,RobertG.TheWarrenBuffettWay.JohnWileyandSons,Inc.,1995. Hahn,Carsten,AndreasPfingstenandPeterWagner.“AnEmpiricalInvestigationofthe RankCorrelationBetweenDifferentRiskMeasures.” http://papers.ssrn.com/sol3/papers.cfm?abstract_id=301982.Feb20,20021-48 Hakim,SamRamseyandDavidShimko.“TheImpactofaFirm’sCharacteristicsonJunk BondDefault.”JournalofFinancialandStrategicDecisions.8(Summer1995):47- 55. Hamada,R.“TheEffectofFirm’sCapitalStructureontheSystematicRiskofCommon Stocks.”JournalofFinance.May1972:435-452. Hendricks,Darryl.“EvaluationofValue-At-RiskModelsusingHistoricalData.”Federal ReserveBankofNewYork:FRBNYEconomicPolicyReview,April,1996):39-70. Hunt,LacyH.DynamicsofForecastingFinancialCycles:Theory,Techniquesand Implementation.Greenwich,CT:JAIPress,1976. Hunt,LacyH.ATimetobeRich.Macmillan(RawsonAssociates)1987. xii Johnson,RaymondM.,JosephA.NewmanandA.LeeGurleyIII.“CapitalStructure Optimization:AConstrainedMinimizationModel.”JournalofAccountingand FinanceResearch.13(April,2005):177-189. Jorion,Philiippe.ValueAtRiskSecondEdition,McGraw-Hill,2001 Lapin,Lawrence,L.StatisticsforModernBusinessDecisions.3rded.NewYork: Harcourt,Brace,Jovanovich,Inc.1982. Levy,HaimandMarshallSarnat.CapitalInvestmentandFinancialDecisions.2nded. Prentice-HallInternationalInc.,1982. Lintner,John.“SecurityPrices,RiskandMaximalGainsfromDiversification.”Journalof Finance.(December,1965). Malkiel,BurtonG.ARandomWalkDownWallStreet.SecondCollegeEdition.New York:W.W.NortonandCompany,1981. Mandelker,G.andS.Rhee.“TheImpactoftheDegreesofOperatingandFinancial LeverageonSystematicRiskofCommonStocks.”JournalofFinancialand QuantitativeAnalysis.March1984:45-57. Mendenhall,WilliamandJamesE.Reinmuth.StatisticsforManagementandEconomics. 3rded.NorthScituate,Massachusetts:DuxburyPress,1978. Modigliani,FrancoandMillerM.H.“TheCostofCapital,CorporationFinance,andthe TheoryofInvestment.”AmericanEconomicReview.(June1958). Modigliani,FrancoandMillerM.H“TaxesandtheCostofCapital.”AmericanEconomic Review.(June1963). Myerson,RogerB.ProbabilityModelsforEconomicDecisions.Belmont,CA: Thompson/Brooks/Cole,2005 Petit,Justin,JuanGolic,andAprilPark.“TheEquityRiskMeasurementHandbook.” SternStewartResearch:Evaluation.3(March2001):1-12 Porter,MichaelE.OnCompetition.HarvardBusinessSchoolPress,1998 xiii Prasad,Dev,GarryD.BrutonandAndreasMerikas.“LongRunStrategicCapital Structure.”JournalofFinancialandStrategicDecisions.10(Spring1997):47-58 Ramesh,Ram.FinancialAnalyst’sIndispensablePocketGuide.McGraw-Hill,2001. Saretto,AlessioA.“PredictingandPricingtheProbabilityofDefault.”August4,2004. http://personal.anderson.ucla.edu/alessiosaretto/default.pdf.(August8,2008) Sivy,Michael.MichaelSivy’sRulesofInvesting.NewYork:WarnerBooks,Inc,1996. Sharpe,WilliamF.“CapitalAssetPrices:ATheoryofMarketEquilibrium.”Journalof Finance.(September,1964). Stovall,Sam.StandardandPoor’sGuidetoSectorInvesting.NewYork:McGraw-Hill, 1995 Swanson,HuntingtonS.andRobertEdWoolsey.OperationsResearchforImmediate Application:AQuickandDirtyManual.NewYork:HarperandRow,Publishers, 1975. TexasInstruments,Inc.BAIIPlus:Guidebook.TexasInstrumentsInstructional Communications,1991. Thomas,LloydB.Money,BankingandEconomicActivity.2nded.EnglewoodCliffs,New Jersey:Prentice-Hall,Inc.,1982 UnitedKingdom,OfficeofFairTrading.Part1MainReport:EmpiricalIndicatorsfor MarketInvestigations.CrownCopyright,2004. (BacktoTableofContents) %' ' ' $ ' ' ' ' #% ' ' '' % '' ''# ' ' ' # ' ' ' ' ' '' # ' $ ' ' ' '' '' '' ' ' '' '' ' ' # ' '' ' # ' ' ' ' ' '# ' ' '' # ' ' ''' ' ' ' ' '' '' # % ' ''' ' '' ' ' ' ' '' ''# ' ' ' '# '' #' ' ' ' ' ' ' ' # # # '' ' ' ' '' '' ' ' '' ' ' ' ' '' ' ' ' '' ' ' ' ' ' ''' ' %' #' ' ''' '' % ''' # ' '' ' ' ''' ' ' ' ' ' ' '' ' # '' '' ' ' ' '#' ' ' '' # '' ' ' ' '' '' ' ' ' ' ' ' ' ' '' ' ' ' ' ' '' ' '' '' ' '' ' ' ' '' ' TABLE OF CONTENTS i 1. - Introduction: A Tale of Two Companies, or three, or four... 1 SECTION 1: THE THEORY OF CAPITAL STRUCTURE 11 2. - Leverage 11 Business Risk, 12 Financial Risk, 13 Leverage: A Definition, 13 Basic Risks and Proportions, 14 Balancing Leverage, 16 The Anatomy of Financial Leverage, 17 The Anatomy of Operating leverage, 20 Total Risk, 22 Leverage Measurement, 23 Theory vs. Reality: Financial Leverage, 24 Theory vs. Reality: Operating Leverage, 24 Theory and Reality: Total Leverage, 25 Leverage Management, 27 Sources of Variation, 28 Return on Equity, 30, The Forces Behind Leverage, 31 Appendix: Streamlining?, 33 3. - Capital Structure 35 The Cost of Capital: Market Orientation and Practical Application, 36 The Advantages of Debt, 38 Risk and Debt, 40 Mathematical Optimization: Theory vs. Reality, 43 The Modigliani - Miller Propositions, 45 The Optimization Problem, 46 The Proposed Ideal and its Inherent Problems, 48 Capital Structure and the Cost of Capital, 50 The Concept of a Weighted Average Cost of Capital, 53 Earnings and Capital Structure, 55 Capital Structure Logic, 56 Changes in Capital Structure and Stock Prices, 57 Four “Postulates”, 59 Share Limitations, 60 Adapted Measurements, 61 Explicit vs. Implicit Costs, 62 TABLE OF CONTENTS ii Implicit Costs of Debt, 64 The Implicit Cost of Equity, 64 The Moment of Truth, 66 Appendix: The Net Operating Income Approach to Stock Valuation, 68 4. - The Cost of Debt 71 The Problem of Short-term Credit, 71 Interest Expense Inequalities, 73 Risk, Return, and the Significance of Short-term Credit, 74 The Corporate Cost of Debt, 76 The Nominal Cost of Debt and the Cost of Bankruptcy, 78 The Cost of Bankruptcy, 80 The Probability of Default, 80 Commercial Ratings Systems, 82 Types of Bankruptcy, 84 The Amount of Loss, 85 The Role of Debt in Capital Structure Optimization, 90 The Optimal Amount of Debt, 92 Long-term Debt and the Amount of Loss, 93 Graphic Depiction, 93 Appendix: Making Sure That What You See is What You Get, 95 5. - The Cost of Equity 99 Modifications, 100 The Reinvestment Rate and Opportunity Costs, 103 Evaluating the Cost of Equity: Methodology, 104 The CAPM as a Building Block for Capital Structure Analysis, 110 An Estimate of Risk and not Prediction, 117 Anomalies Pertaining to the Use of the Cost of Equity in Capital Structure, 119 Adaptive Expectations versus Rational Expectations, 120 The Decomposition of Beta, 122 Balancing Leverage, Risk and the CAPM, 127 Beta and Leverage, 128 A Simple Equalizer, 129 Concluding Comments, 131 Appendix: The Mathematical Relationship Between Levered and Unlevered Companies in Terms of the CAPM, 133 6. - The Capital Dynamic and Other Tools 138 The Percentage Trap, 138 The Weighted Average Cost of Capital, 139 The Theoretical Shape of the Weighted Average Cost of Capital Curve, 141 TABLE OF CONTENTS iii Acceleration Rates, 142 The WACC and Risk, 143 The Mechanics of the WACC: Risk Adjustment, 143 Corporate Method, 144 Risk Adjusted Method, 145 The Marginal Cost of Capital, 145 Decision Making and the Marginal Cost of Capital, 147 Economic Profit, EVA, and the Capital Dynamic: Utilizing the Opportunity Cost, 149 Elements in an EVA Calculation, 150 The Capital Dynamic, 151 The Relationship between EVA and the Capital Dynamic, 152 Economic Profit and Correlation, 152 Management of Economic Profit, 154 Component Movements of the Capital Dynamic, 157 The Comparative Capital Dynamic, 159 Appendix: The Efficiency of EVA versus ROE, 160 Appendix: The Cost of Capital and What The Investor Needs to Know, 161 7 - Fundamentals and Capital Structure 168 Du Pont Analysis, 170 Comparing ROE Components, 175 Modifying and Enhancing Du Pont Analysis, 181 The Return on Capital Ratio, 186 8. - Capital Structure and the Business Cycle 189 Common Elements of Business Cycles, 189 The Yield Curve and Interest Rate Behavior, 190 Graphs That Unite the Two Theories, 192 Strategic Considerations, 198 The Business Cycle and the Cost of Equity, 200 The Capital Asset Pricing Model and Sensitivity Analysis, 202 Circumventing the Optimal Capital Structure, 208 The Game of Capital Structure “Gothcha”, 209 Idealized Trends, 210 Sector Rotation, 212 Sector Logic, 213 Industry Response to the Business Cycle 213 Economic Signals, 217 9. - Operating Risk 219 Fixed Costs and Economics, 220 The Case of Compaq Computer, 223 TABLE OF CONTENTS iv The Nature of Costs and Margins, 224 Fixed Costs and the Breakeven point for Sales, 227 Compaq Computer: The Rest of the Story, 229 Operating Leverage and Prediction, 231 Characteristics of Operating Leverage, 232 Operating Trends and Reversals, 235 The Quality of an Operating Margin, 235 A Risky Proposition: Confidence Intervals, 237 Operating Beta, 240 The Unlevered Beta Equation, 241 The Ardco Barbell Company: An Example of Unlevered Beta, 241 “Mom and Pop Store” Betas: Companies Who are Not on the Market, 244 Operations Research for the Investor, 245 Two Masters: Fisher and Buffett, 247 A Brief Operating Analysis of Fed-Ex and Staples for the Year 2000, 249 Staples and Fed-Ex Operating Histories: 1994-1999 Analysis and Statistics The Confidence Interval Tool Operating Margin 10. - Operating Momentum 258 Reasons for Study, 259 Operating Momentum Sensitivity, 264 Regression, 269 The General Electric Solution, 271 Classical Microeconomics and Operating Momentum, 275 11. Strategic Capital Requirements 279 The Realities of Funding, 279 The Proper Amount of Capital, 280 The Debt / Equity Tradeoff and EVA, 282 EVA / Capital Dynamic Based Improvement, 285 Incremental Equity Improvement, 287 The Irrationality of Rationing Capital, 287 Incremental Debt, 288 Dividends and Retained Earnings, 288 Capital Funding From EVA: Two Methods, 288 Method 1: Solving for the Optimal Equity Method 2: Solving for the Optimal Net Income Determining Capital Proportions and Requirements, 290 Projected Analysis, 291 Financial Engineering: Setting Capital Requirements From EVA, 291 ConocoPhillips 2005-2006: a Real World Example, 291 TABLE OF CONTENTS v Method 1: Capital Proportions from Optimal Equity Method 2: Capital Proportions from Optimal Net Income Financial Engineering Method The Additional Funds Needed Equation, 296 The Modified Additional Funds Needed Equation, 299 Degree of AFN Logic, 303 The Need for Qualitative Assessment, 306 The Problem with Optimality, 306 Merger Mania, 308 Merger Growth Illusion and EVA, 311 SECTION 2: BUILDING CAPITAL STRUCTURE MODELS 315 12. - The Economic Profit Laboratory: Computer Applications 315 Set Up, 315 Section One, 315 Section Two, 316 Sample Data, 318 Sensitivity versus Optimization, 319 Proving the Capital Dynamic / EVA Hypothesis, 320 Setting the Constraints, 320 Trigger Points, 321 Setting the Trigger Point Module, 322 The Earnings Solution, 323 Optimization and Correlation, 324 Raising Capital Effectively, 324 The Connection Between Capital, Stock Price, and EVA, 326 Sensitivity Analysis: The Effect of Changes in Operating Income and Capital, 327 Establishing Guidelines, 331 Appendix: Spreadsheet Examples, 333 13. - The Marginal Benefits Equation: An Experimental Model 335 The Modeled Concept, 336 The Marginal Benefits Equation, 336 Default Probability and Bankruptcy, 338 The Interest Benefits Mechanism, 339 Checking Results Against a Viable Standard, 340 Default Mechanics, 340 Strategic Implications: Financial Leverage, 341 Strategic Implications: Operating Risk, 343 Spreadsheet Constants, 344 TABLE OF CONTENTS vi Spreadsheet Logic, 346 Dynamic Variables, 347 Model Set Up, 348 The Process: Entry Variables, 349 The Process: Optimizing with Solver, 350 The Results: Three Examples, 350 EVA Discrepancies, 354 Appendix: List of Formulas and Spreadsheet Construction, 355 14. An Introduction to Residual Economic Profit Theory: Using a Constant Dividend Discount Model 359 An Introduction to Residual Economic Profit Theory, 359 Opportunity Cost, 360 Valuation Models, 361 Dividend Theory, 362 Residual Economic Profit, 364 The Dividend Trap, 365 Model Optimization, 366 Model Background, 366 Model Set Up, 368 Model Adaptations, 369 The Case of: Can You Top This?, 369 Comparing Spreadsheets, 370 Appendix: Three Spreadsheets, 373 SECTION 3: REAL WORLD CASES 376 15. - Analytical Tools: Practical Application 376 The Toleration of Imprecision, 376 Erring on the Side of Conservatism, 378 Brief Methodologies for Determining the Cost of Equity, 379 The Hurdle Rate, 382 The EVA / Capital Dynamic, 383 The Weighted Average Cost of Capital, 383 Comparing Risk: Justification for Two Costs of Equity, 386 Changes in the CAPM, 386 The Comparative Capital Dynamic, 388 The Marginal Benefits Equation, 388 Leverage State Analysis, 391 The “Look Ahead” Bias, 395 Micro Analysis: Quarterly Observation, 396 Naive Extrapolation, 397 Earnings Pressure, 399 TABLE OF CONTENTS vii Appendix: Dividend Discount Models, 402 16. - Kimberly-Clark-”Too Much of a Good Thing”: Economic Profit and Marginal Benefits Analysis 404 Underpinning 1: Position in the Business Cycle, 404 Underpinnings 2, 3 and 4: Opportunities for Analysis, 405 The Leverage State, 406 Changes in Economic Profit, 411 The Extreme Consensus Method, 411 Economic Profit, 416 Too Much of a Good Thing, 417 Marginal Benefits Analysis, 418 Basic Methodology, 419 Tax Benefits for Kimberly-Clark, 420 Amount of Loss for Kimberly-Clark, 420 The Probability of Default, 421 The Cost of Bankruptcy, 422 Marginal Benefits, 423 Confirmation, 424 Altman’s Z Score: Book Value Version, 424 Kimberly-Clark’s Z Score, 426 Investment Conclusion, 427 Appendix: Extrapolated Risk: When “Normal” is Too Risky, 428 17. - “Full Steam Ahead”: An Analysis of ConocoPhillips, 2002 2006 431 The Context, 431 The Decision, 434 Price Performance, 435 Anticipating Performance: Leverage States, 436 Quarterly Leverage Results, 438 Interpreting Regression, 439 Establishing a Comparative Cost of Equity, 441 Contrasting the Required Return with the Expected Return, 445 The EVA / Capital Dynamic, 446 Naive Extrapolation, 448 The Marginal Benefits Function, 451 The Comparative Capital Dynamic, 434 Earnings Pressure, 456 Appendix: Selected Financial Data for ConocoPhillips, 2001 – 2006, 459 TABLE OF CONTENTS viii 18. - Microsoft Versus ConocoPhillips: Comparing Companies in Different Industries 460 Apples and Oranges: Microsoft Versus ConocoPhillips, 461 Common Ground, 463 A Market Disconnect and Eventual Reconciliation, 467 Industry Competition: Chevron and the Comparative Capital Dynamic, 469 Percentage of New Retained Earnings, 473 SECTION 4: CORRELATION AND PROBABILITY STUDIES 475 19. - Operating Income Correlation Studies 475 Name, Premise, Data Points and Structure, 476 Categories, 477 Companies in the Sample, 479 Fundamental Variables, 479 All Variables, 480 Statistical Results, 481 Spearman Rank Correlations: Next Year’s Midrange Stock Price, 484 Interpretation, 485 Methodological Criticism, 488 20. - Changes in Capital Structure and Their Effect on Stock Prices - 491 The Validity of Leverage Factors, 491 Connecting the Dots: Earnings and Dividend Growth and the Cost of Equity, 492 Earnings Acceleration, 493 Statistical Validity, 495 A Brief Study, 496 Three Assumptions, 498 Expectations, 499 Interpretation and Results, 500 Hypothetical Causation, 502 The Hazards of Playing Detective, 503 The Argument for Capital Rationing, 505 Spearman Rank Correlation and Individual Interpretation, 510 21. - Probability and Capital Structure 518 TABLE OF CONTENTS ix The Efficient Markets Hypothesis, 519 Screens, 519 The Return on Capital, 520 Leverage States, 521 Industry Averages: Lemmings vs. Leaders, 522 The Leverage State Ratios, 523 Combinations, 524 The Mechanism of Leverage States, 525 Matching the Leverage State to the Business Cycle, 528 Probability and Diversification, 530 Sales and Beta, 531 Probability and Anticipation, 532 Principle Components Analysis, 534 Static vs. Forward-Looking Ratios, 536 The Quick Payoff, 540 Barr Rosenberg and Response Coefficients, 541 22. - Technical Analysis and Capital Structure 544 Major Forces, 545 The Bane of Volatility, 547 Self-fulfilling Prophecy, 548 Ex-Post Performance, 549 Stochastic Conformance, 552 Capital Structuralism: Quasi-Technical Analysis?, 334 Fighting Words: “The Efficient Markets Hypothesis”, 555 The Art and Science of Forecasting, 557 Moving Averages to Use Wisely, 558 The Exponential Moving Average, 559 Brief Interpretation, 561 Primary Trends and Secondary Trends, 561 23. - Statistics Primer 563 The Mean, Mode and Median, 564 The Variance and Assorted Adaptations, 566 The Standard Deviation, 567 The Covariance, 567 Downside Risk, 568 Annualized Volatility, 569 Estimated Volatility, 570 Sample Standard Deviation, 571 The Mean – Variance, 571 The Coefficient of Variation, 572 Worst Case Scenarios, 572 Accounting for Additional Risk: Combining Standard Deviations, 574 TABLE OF CONTENTS x Prediction, 576 Confidence Intervals, 576 The Normal Deviate, 578 Updating the Mean and Standard Deviation: Moving Averages, 580 Accounting for Small Sample Size: The Student “t” Distribution, 581 Regression, 582 Relationships in Linear Regression, 584 The Coefficient of Determination, 585 Making Predictions, 585 Growth Rates, 586 Geometric Approximation, 589 Accurate Growth Rates from Logarithms, 588 Spearman Rank Correlation: Non Parametric Statistics, 589 Sample Size, 592 SELECTED REFERENCES When individuals with similar tolerances for risk pool their resources. he is “spat” out onto a beach. there is life after consumption. Like the whale. “the Big Fish” swallowing a smaller creature is only part of the story. The following story relates a scenario that is quite characteristic of competition between modern corporations. Wall Street creates mutual benefits by sharing risk. the embodiment of this concept is the inherent limited liability of the corporation itself. and ends up borrowing . seeks risk in its manner of funding. This “symbiosis” is the product not of some hierarchical structure that eliminates competition. or the chance for an unacceptable outcome decreases.. and begins preaching to the kingdom of Nineveh. brutish and short”. Despite its reputation for a “bottom line” mentality.1 1 “A TALE OF TWO COMPANIES” . nowhere is the Darwinian concept of “survival of the fittest” championed more. or four or. Another company whose revenue stream is more diversified. it is trite and myopic enough to divert our attention from the truth: the concepts of mutual benefit and shared gain are more prominent than any vestiges of unilateral conquest. One company’s risk. if any allegory is especially applicable to characterizing “the Street”.or three. it is the Biblical story of Jonah and the whale. Belief in miracles notwithstanding. the financial community must continue to grow to remain competitive. In fact. the potential return rises. However. like Jonah. but a mathematical process that combines risk and return in the most efficient manner possible. Jonah turns the crisis into an opportunity. The comparison between Wall Street and a “jungle” was never valid. While consumption of small fish by a larger one seems an appropriate metaphor. If life in a state of nature can be described as “nasty. or both. changing their misbegotten ways..seeking marketing strategy narrows its focus to a point where its cash-flow is compromised. Leverage is merely the measured manipulation of these risks. while an optimal capital structure points to their successful use. Curtin Matheson’s executives became disciples of Philip Crosby. larger. The marketing initiative was steadfast: the firm would attempt to carve out a niche for itself based on a high level of service and fastidious product knowledge. Curtin Matheson Scientific was a quality distributor of scientific products for over twenty-five years.2 money to take over the first company. corporate share prices will maximize when these two risks are perfectly reconciled. After the famed statistician. By the time a recession hit in the early 1990’s. however. one of the early and premier adherents of the concept of quality leadership. any miscalculation in either risk by a smaller “Jonah– type” company will have a far-reaching impact because of the relative size of the firm. Edwards Deming. reported on the precision quality control of Japanese firms. In effect. competition had altered the health care landscape. This tension between two types of risk – one that affects revenues and one that affects funding – forms the crux of all capital structure decisions. Less risk in their revenue stream allows them to take more risk in other areas – especially in their sources of funding. Ultimately. larger corporations can swallow-up smaller ones because they have a wider array of options when combining risk. Selling to the rapidly consolidating HMOs (health maintenance organizations) would ensure cost effectiveness and high profitability. On the other hand. and Curtin Matheson shifted its focus away from industrial laboratory equipment and toward the burgeoning diagnostic testing field. Cut-backs became even . Like the proverbial “whale”. A shift to higher volume and lower prices necessitated the closing of distribution centers and the consolidation of customer service. A booming market in health care products in the middle 1980s produced high profits and some pricing power. “predator” companies can exploit an imbalance between risks in smaller companies usually a matter of timing. Dr W. . the floundering Curtin Matheson was ripe for a takeover. building profitable vendor relationships that had produced a long track record of consistent sales. However. Although profit margins had been shrinking.blood analyzers that sold for $200. In fact. approximately 350 million dollars. Fisher moved in the opposite direction. revenues are stable. which was considered paltry in the hyper-inflated market of the late 90s. there was nothing “romantic” about specialty chemicals and analysts maintained a low key coverage on the company. from sharps containers to beakers and test tubes. Fisher Scientific International was listed on the New York Stock Exchange. the price of the inventory was buffeted by technological scarcity . Management became jittery about losing the company they had so adeptly built. even in a downturn. for example. Their stock sold for about eleven dollars a share. By the middle 1990s. Fisons’. Moreover. even as a tell tale sign quashed any rumors about maintaining the Curtin Matheson Scientific (“CMS”) brand integrity: the Fisher logo began appearing on every product. but they were not a “big” player on Wall Street. Fisher garnered a loan from a Canadian bank and paid Curtin Matheson’s English holding company.3 more rampant when fears of nationalized health care gripped the industry. the firm had been around since the late industrial revolution of the 1800’s. Curtin Matheson Scientific became vulnerable in the one area on which they concentrated . Price control would revert to a government entity and sales reps would be competing for contracts that would yield almost nothing.health care Health care product distribution has two characteristics that make it especially attractive for acquisition by larger firms. Fisher Scientific was an old nemesis of Curtin Matheson.000. When Curtin Matheson began concentrating on the health care market. Fisher had a solid reputation in the scientific community. the demand for health care products is steady enough to cushion other risk taking ventures. significant cash-flow was channeled through very high revenues. focusing on specialty chemicals and industrial products . Fisher spotted an opportunity.a move that limited their exposure to an industry with declining margins. “Japanese-like” commitment and unity between team mates. By the onset of the new millennium. carrying over 250. which of course. called Hampton. By early 2001.4 The integration of Curtin Matheson into Fisher went smoothly. The dichotomies posed more questions than they did answers. In fact. but all executive decisions were conveyed from a small town on the coast of New Hampshire. While the two entities were closely integrated in Curtin Matheson. the various . As an interested participant. health care. like brokering a commodity. Stock was never issued for purchases. health care products had become.000 in a field that was not particularly capital intensive . The funding for acquisitions came from debt a lot of it. the firm was extremely well-diversified. Curtin Matheson was just one acquisition that fueled this diversification. electronics and even had a supply center for radioactive material at the Los Alamos nuclear facility in New Mexico. Fisher could not generate the type of internal funding that supported both existing operations and a program of diversified growth simultaneously. Fisher had divisions in safety. I could not help but notice some oddities. and its revenue base was maintained.000 per employee by the year 2000. the Fisher approach was very clinical. there seemed to be some disconnect between sales and operations. The main operations center was in Pittsburgh. the risk of Fisher’s cash-flows were decreased. and retained earnings were insubstantial. Fisher carried negative equity.000 items. albeit the largest at the time. Although we laughed at how rapidly Fisher put their brand name on “our” products. emphasizing an emotional.distribution. chemicals. Fisher had a stream of about $340. margins were just too low. Fisher Scientific was rapidly becoming a strategically-run financial powerhouse that expertly negotiated risk. Although revenues were high. On the other hand. By having at least one division that would react favorably to a changing economy at any one time.tech companies were struggling with revenues of approximately $250. While most high. they also had the last laugh. The stock barely moved at all in fact. a company that had no long-term debt whatsoever. remained very high. Fisher began to pare down its debt and issue equity. however. They ended up calling themselves. Fisher had two types of risk that were in potential conflict: business risk sometimes called economic risk or “operating risk”. but it did not move on the basis of sales or profits. It works together with business risk through the variability of operating income. financial risk is almost entirely selfgenerated. even during the recession that began in 2001. In the mean time. Creditors took one look at the size and variation of Fisher’s revenue stream and gave them the “green light”. The firm responded by renewing loans at lower interest rates. except for a single situation: when any news or rumor of an acquisition occurred. courtesy of the Federal Reserve. the stock would jump out of its usual stable dormancy and take off like a rocket. Ultimately. On the other hand. Business risk is the variation in revenue. Equity was kept to a minimum. and other economic factors differently than others. its only volatility was self generating. costs and operating income that stems from the type of industry. recessions. the stock was a good addition to any portfolio. And inevitably. In a nutshell. They bought biotech suppliers in Sweden and test equipment companies in the United States. they had merged with ThermoElectron.they seemed poised to start the whole “process” over again. In essence.5 integrations of acquisitions were expensive even as Fisher’s long-term debt to capital rate approached eighty-five percent. foreign competition. financial risk is expressed as the potential for defaulting on interest payments and principal. and stems from the variation in net income from the decision to use debt. the story of Fisher Scientific provides a valuable lesson in managing capital structure. . By 2006. Cash-flow. and financial risk. “Thermo-Fisher Scientific” (TMO). when acquisitions began to pay off. causing the stock to soar even higher. Since some small acquisition occurred at least twice a year. some industries react to inflation. the stock soared. the decision by Curtin Matheson to focus on health care to the exclusion of other divisions made the company a takeover prospect. and 110. but simultaneously decreasing the risk of its operating income. However. Even as margins declined. They never have the type of sensational results that makes them the darlings of speculators. Fisher configured the risk-return tradeoff in its favor. and the company was left with a riskier and depleted cash-flow. While the average in Company A is much greater than in “B” (79 vs. Imagine a cash flow for Company A of 60. In that period. By financing with debt. the risk is far greater also. Companies like Fisher Scientific are quite ordinary. 65. Fisher Scientific treated its operating cash-flows like a portfolio. the difference is small. adding and dropping product lines that would make it less risky. Which would you prefer? Most people would pick the first because the chance of getting a high number is greater. margins declined. it was only about half as volatile as the rest of the market. One misconception that students and investors share alike is that a business is suppose to “maximize” profits: the “bottom line” mentality is almost an endemic archetype and yet rarely occurs in economic behavior. Alternatively. With few barriers of entry. Capital structuralism is not about directly maximizing profits through programs like a new marketing campaign or “zero base budgeting” or the implementation of new . 67). 90. the industry invited intense competition. Now compare it to the cash flow of Company B: 60. and 75.6 an adequate and steady operating income can keep financial risk very low because there is less probability of default. 75. And yet . 70. the flow from Company B is much steadier and by several mathematical gauges of risk has a better risk-return characteristic than Company A’s. 65.here was a company whose stock was selling at $11 a share in 1997 only to rise to a peak of $77 eight years later. its return on equity (ROE) increased because it never funded with its own money. They rarely make the evening news. In fact. but may be compelling enough for investors to choose Company B as an investment. 60. What about variation? Random fluctuation is the bane of any analyst. By identifying and investing in firms who repeatedly move toward their optimal capital structures. the cost of the mixture of debt to equity will determine the plausibility of each project because of the necessity of exceeding capital outlays with returns. the greater the number of projects that will be potentially profitable. On the other hand. time is the essential component in all risk-return distributions . The lower is the cost of capital. As an example. current and future behavior of an investment seems to defy formulation. time is a unifying factor between risk and return and encourages their interdependence. Risk and return are so intertwined that it is proper to refer to them as a statistical “distribution” with two parameters. rather than as separate categories. It chooses a course of action from several alternatives that balances the risk of different types of funding with returns that exceed their cost. Although the investor is not encouraged to “time” investments over the short term. No matter how precisely one measures the deviations in past performance. On the one hand. The market keeps changing and the response to world . long-term viability is never certain. which are dependent on the state of the economy. it takes a far more subtle approach. the marketing of more shares of stock to raise additional funds. some awareness of the correlation between sector performance and the business cycle is essential. The investor is left in a precarious position. the goal of minimizing the cost of capital is implemented through the capital budgeting process. he or she is encouraged not to time the market because it is not successfully done over a long period. The characteristics of risk and return for such an issue are much different at the beginning of a recovery than at the end of a bull market . consider an equity issue. Therefore.for both the issuing company and the investor. While Wall Street prizes certainty. capital structuralism resolves some of this conflict.from investment horizons to the choice of which investments to make. Capital structure is dependent on the relationship between interest rates and the equity market.7 technology. Thus. Since capital structure is dependent on the business cycle. it responds to societal trends.8 events is embedded in corporate gains and losses.This is the cost of an action with economic conditions factored in. In effect. then my effective tax rate was reduced and I will have to make a much greater net income in the next year. In capital structure. While the market responds to information instantaneously. Others can compete with firms that have two or three times its profit margins simply because they know how to use debt judiciously. then my “real” net income is probably only $95. Capital structuralism is flexible enough to encompass change because it never defines optimality as a rigid set of conditions. “the going market price”. It is most related to what can be termed. If I have tax “look backs” of $20 figured into that $100. • 3) The Risk Adjusted Cost . If my net income is $100 and the inflation rate is five percent. creating either a gain (opportunity gain) or a loss (opportunity loss). our analysis attempts to put a dollar price on risk. capital structure reflects the reasons why a certain entity is in business in the first place: to grow and make a profit. Some industries have better risk-return characteristics without any debt at all. I will be penalized for not putting more money into the market. The risk adjusted cost is the comparative cost of taking one action over another. we attempt to measure its content before it becomes meaningful. this risk adjusted.If I keep all of my money in a checking account when the market is rising by fifteen percent a year. We can define cost in three different ways.This is the “up front”. all of which are used to evaluate risk: • 1) The Nominal Cost. demographic changes and political risks better than the various “systems” that have made their way into the investment literature. “opportunity cost” is more important than any other . accounting cost of an action which will be reported in financial statements • 2) The “Real” Cost . Each industry has a particular response to economic factors that produces a different optimum level of proportional debt and equity. Ultimately. the difference between student. Correlation Studies: Examination of the relationships between stock price and capital structure variables gives insight into the behavior of some major corporations. and the greater economy. Each of the spreadsheet models has been used to evaluate corporate behavior. investor. The flow of the text is as follows: • 1. the investor and the financial manager from the same viewpoint: he or she gauges the risk of operating cash flows and balances that observation with the choice of alternative sources of capital. The integrative approach of this text is to position the analyst. There is a chapter dedicated to statistics. many of the costs we incur in capital structure are not representative of a physical asset and passed on from a previous owner.9 because it looks at an array of alternatives and chooses a course of action that attempts to create the largest possible opportunity gain. Model Building: A mathematical conception of capital structure is built through computer models and the adaptation of existing formulas. Therefore. and most spreadsheets have step by step instructions. the sector. • 3. The text requires some familiarity with statistics and computer spreadsheets but not an extensive background in either. tendencies that support capital structure theory are examined using the Spearman rank correlation. analyst and manager is clouded because each perspective is directed by the need to seek and discover optimality. Under the premise that capital structure is the interface between comparative accounting and the macro economy. While no definitive conclusions are drawn. In effect. . • 2. but are the result of a choice of actions with which we have comparative information. repeatedly making comparisons between the industry. the student receives an overview of corporate finance through the attempted reconciliation of risk and return. Theoretical Background: Capital structure theory is examined through previous research with an emphasis on integrating the evaluation of risk and return. William James. for the executive. For investors. the author hopes that this book will help them see beyond the superficiality of conventional wisdom with the knowledge that the cash value of any idea is almost always found in its underlying structure. might have appreciated the personal computer revolution. the author hopes that this book will unify financial thought into a comprehensible “whole” and encourage the actualization of “just theory”. Case Studies: Application of capital structure analysis to Kimberly-Clark and ConocoPhillips. (Back to Table of Contents) . For students. as well as Microsoft and Chevron display the effectiveness of the techniques. Finally. at least “the information age” has made them available. which is certainly “half the battle”.10 • 4. One of the great philosophers of the early twentieth century. the imperative is placed on innovative thinking: a time-tested solution is the outgrowth of a new perspective. While we often lack the political framework to implement ideas. He who championed “the cash value of ideas” and the philosophy of pragmatism might have found solace in a machine that tested the viability of theory. the risks incurred by the plan are often imposed from above: government regulations. business risk and credit availability put restrictions on all available choices. While the ideas that generate high returns often come from a detailed marketing plan that forms a foundation for the business. This friction between marketing strategy and absolute risk is reconciled by a strong adherence to the principles of leverage. and the fluctuations inherent in a typical business cycle. capital structuralism often seems to deny the need for greater returns by focusing on risk. a choice to add a new product line may not come to immediate fruition for a company who pays high interest rates and has excessive debt on its balance sheet. capital structuralism is a “top down approach”. However. Since it makes comparative choices from a macroeconomic perspective. competitor’s actions. The gist of capital structure analysis is to resolve this conflict between what can be produced and what will be produced. uneven streams of income that might upset “corporate equilibrium”. the increase and acceleration of sales is paramount. and secondly. even to the exclusion of large.11 SECTION I: THE THEORY OF CAPITAL STRUCTURE 2 LEVERAGE In a fundamentals based “bottom up” analysis system. the tools to manage this resolution are encompassed by two distinct measurements: operating leverage and financial leverage. Accordingly. . The axiom. “nothing in business happens without a sale” appears self evident. Capital structure analysis adjusts for both the amount and variability of sales by first evaluating operating income as a function of sales. by reconciling risk with return. by choosing the amount of funding from several alternative sources based on the risk of this evaluation. For example. adding fixed costs to any operation raises the breakeven point for sales. when an operation has a higher proportion of fixed costs. This single kernel of corporate risk. When competitive pressure demands that specific quality standards are met. even when the total cost is the same. storage. and transportation. Because fixed costs must be paid regardless of the level of demand. However. foreign competition and the availability of substitutes. planning must be totally contingent on the unexpected. the farmer must be concerned about the cost of seed. Many economists believe that business risk is a reflection of the level of technology in an industry. but necessary to meet these competitive pressures. Once the percentage of fixed costs is increased. irrigation. Little growth will occur in such an environment because no investor wants to commit capital without confidence in a minimum return. To stay in business. Fixed assets that have long depreciable lives are very costly.12 BUSINESS RISK Operating leverage is one measure of business risk also known as economic risk. When suppliers’ prices also vary. the double edged sword creates an environment of high business risk Without any idea of how much to pay vendors or how much to charge customers. higher fixed costs imply that more business risk is incurred. Consider for a moment. the result is a higher operating profit. more sales must be generated to cover them. Demand for his or her crop is dependent on weather. more units of production will be spread among the same amount of costs. those standards are an outgrowth of the level of technology required by the industry. the shrink wrapping on a CD. an immediate barter-like negotiation where uncertainty prevails. affects all other elements in . High fluctuations in demand often cause large swings in the prices that a farmer can charge. The variability of inputs (costs) and outputs (demand and quantity produced) form economic risk. Would a customer buy a hand wrapped CD when the industry standard is to wrap it “as tight as a drum”? Moreover. and sales are adequate. consider the attributes of farming. As an example. the risk entailed by the cost of higher interest payments is much less than the probability of new cash-flow. and away from the potential for higher dividends. the student/investor will turn this decision into a profit-making venture by determining the crucial difference between strategy and obligation. 2. earnings.13 the chain: demand schedules.a complex interaction between sales. the probability of default and the methods and sources of funding projects. FINANCIAL RISK Financial leverage is one measure of financial risk. The probability of loss . . commercial paper). which is the risk incurred by a firm for its decision to use debt financing. common stock and preferred stock) or debt (bonds. In return. bank loans. In fact. variability of income. the firm increases the risk to existing shareholders because earnings become partially channeled toward creditors in the form of interest payments. the “prime rate” is set low enough to attract the best customers without burdening them with worries of insolvency. Firms that are obliged to increase financial leverage in order to cushion poor demand have radically different characteristics from those who optimize capital structure. many well-run firms lower their overall risk because of the choice to use more debt. Companies face a choice of funding projects with equity (retained earnings. Indeed. the variability of income is increased. LEVERAGE: A DEFINITION If the choice to take on debt sounds dire. When deciding to increase the amount of debt. Consequently. This risk can be decomposed into two basic elements: 1. shareholders receive the possible reward of higher earnings on a per share basis because fewer shares will be outstanding when debt is used in place of equity. the firm increases its chance of bankruptcy when it incurs more debt.the claims that creditors have on a firm. The amount of potential loss . it can default on interest payments if earnings are not high enough to cover them. and liabilities that determines solvency. we add the element of connotative risk: we look for other associations that the level of leverage may affect. since each remaining person is more responsible for total production. few display the integration of risk and return better than the balance between financial and operating leverages. but in finance. more risk is involved. we put the “derived component” in the numerator. it is simple to observe how a change in one component affects the change in the other. For example. and determine the change in both.14 If we think of leverage as a proportion of two different components of the same risk. Therefore. the adult drops with a thud. and the “source component” in the denominator. if we discover a “new” labor saving method in which two people can accomplish the same amount of work as twenty-two. Of course. and the larger force as the numerator. A child’s see-saw is the classic example of this principle: when a fifty pound child balances a two hundred pound adult and then jumps off. In fact. In mathematical terms. Leverage can only be increased if the risks have been fully vetted. we usually view leverage in terms of input and output. which needs to be identified. each seeking to balance the other. Secondly. the child’s weight was the source component. If we view the smaller force (the child’s weight) as the denominator of a ratio. the method undoubtedly has a lot of “leverage. In a financial context. In economics. which had an exaggerated effect (derived) on the adult’s weight. depending on their relative amount of association. we speak of “leverage” when a smaller amount of one variable has a larger effect on the other. the two laborers would at least have rising expectations about wages. In physics. if not actual demands. In our example. a small force applied at one point can balance or control a much larger force at another point. While there are other measures of risk besides leverage. leverage always implies some risk-return tradeoff. leverage almost always exacts an inherent “cost” and in our example. capital structure theory is founded upon this integration: behind every strategic . we can form a general definition. losing one person may cut production in half. Operating leverage is conceptually measured as % ∆ Operating Income / % ∆ Sales and implies the inherent volatility of a change in sales creating a change in EBIT (earnings before interest and taxes). lays some thread of leverage. The same theme of variability applies to the financial leverage ratio. Cash. however. but that interest payments. Long-term debt to capital and EBT / EBIT which is . but also less profit when a positive shift in demand occurs. which is theoretically defined as. when a market is coherent. “financial leverage”. Lower operating leverage creates lower volatility.15 decision. only to leave a futures owner poorer a few weeks out. At the other extreme are certain commodities that can skyrocket overnight. as well as any tax advantages such payments entail. more risk is incurred. When operating leverage is large. Each time that capital is allocated for any given project. the derived component (net income) is affected by a change in the source component (operating income). Leverage. for example has almost zero volatility. If there is one quality to cultivate in the world of finance it is consistency. Among those proportions are: Assets / Equity. profits are jeopardized. with lower demand. The student/investor will note that the term “debt” is no where to be found in the ratio. are implicit when operating income becomes net income. and it is the magnitude of these deductions that will cause variability. the possibility of high profits is greater in an upturn when sales are large. that changes the price of a stock. and very little return when kept in that form. A reference that confuses both investors and students alike is that some academic literature designates several different ratios by the term. Again. BASIC RISKS AND PROPORTIONS Wall Street does not like uncertainty. implies volatility and it is when two different types of volatility are mixed that a level of return is derived. but in a downturn. Both taxes and interest payments must be deducted from EBIT before it is termed “earnings”. ultimate profitability depends on leverage. % ∆ Net Income / % ∆ Operating Income (EBIT). the financial community can make plans around expectations and predictions. % ∆ Net Income / % ∆ Operating Income (EBIT) because it best expresses the conceptual integration of risk and return when combined and balanced with operating leverage. or maintain its existing debt with less risk. We use the ratio. However. investors would be overjoyed by the stability. When the economy lags and sales abate. and is an integral part of any capital structure analysis system. Demand cycles invariably change. (% ∆ Operating Income / % ∆ Sales) x (% ∆ Net Income / % ∆ Operating Income). the higher leverage firms suffer the most . although some specifically measure interest expense and others measure a particular liability or category thereof.16 earnings before taxes divided by earnings before interest and taxes. The source of income is evaluated for its amount and consistency. while debt is evaluated for the size of interest payments and the amount of principal. especially in relation to assets and net worth. which can be dampened by maintaining a steady leverage figure.total leverage . if the same change in sales is enacted year after year. which is considered high.more fixed costs entail the necessity of higher sales to cover them.as possible without the inherent volatility such a level implies. operating income reacts heavily to changes in sales. the premium for Wall Street is to have as high a level of (% ∆ Net Income) / (% ∆ Sales) . The balance between the comparative amounts of income and debt are then gauged in terms of risk and return. The problem with high leverage is its inability to adjust itself to a changing economy. which conveniently becomes (% ∆ Net Income) / (% ∆ Sales) Obviously. BALANCING LEVERAGE Since incurring debt implies the risk of default (failure to pay interest expense in a timely manner). fixed costs must still be paid despite a slower production cycle. The common point in each of these ratios is that they represent some form of debt financing. For example. if operating leverage is normally 3 / 1. and when they do. Leverage is always a measure of potential volatility. In fact. we can form a measure of total risk if we multiply the two leverages together. a steadier source of income allows a firm to either engage more debt. . they would know exactly what to expect and when. Any downturn in sales would be magnified with a higher operating leverage. which is a function of paying out interest expense. • 1) The probability of default affects the cost of future financing by increasing both interest rates and the cost of floating an equity issue.17 A proper balance of leverage will yield tax benefits because interest is a tax deductible expense. • • 3) Creditors may have a claim on assets and restrict income potential. i. Thirdly. The two risks. more debt can make a company prohibitively expensive for a takeover and even shift control to a supportive investor or “white knight” in case of a hostile attempt. % ∆ Net Income / % ∆ Operating Income . business risk and financial risk. The main risk of leverage is the risk of default. there is the potential for a rise in both earnings per share and share price because funding is achieved with less shares of stock outstanding than if done through an equity issue. creating volatility in earnings and making the firm more likely to default on interest payments. are inseparable: both the amount of debt and the chance of loss are predicated on a steady operating income. What is not apparent from the financial leverage ratio. and restrictive covenants (bond contracts) may limit their growth.. debt can shift the balance of control away from shareholders and toward creditors in a legal fashion. Secondly. which detracts from a firm’s viability in several ways. more income is spread over fewer shares. Keeping operating income constant and changing interest will increase the volatility of changes in net income. THE ANATOMY OF FINANCIAL LEVERAGE The volatility of net income is a natural outgrowth of its dependence on both operating income and the amount of interest expense. 4) The probability of default affects market volatility for the stock Added to the risk of credit default is the risk of income variability. is .e. • 2) Default may endanger dividends. is spread over fewer shares. The ability to limit shares adds a new level of volatility to financial leverage. To illustrate this concept. the variability of net income remains constant as well. consider the following flows of operating income. the first group with no debt. Net income may not vary more than it does without debt.the manipulation of the number of shares outstanding by financial management. net income that is normally derived from the deduction of interest and taxes from operating income.18 that EPS (earnings per share) can change because of exogenous factors . Taxes are 30 % in both groups. and the second group with $20 of interest expense. In effect. When interest and taxes are held constant. Table 2-1 NO DEBT EBIT TAX NET INCOME Table 2-2 WITH DEBT EBIT INTEREST TAX NET INCOME 90 27 63 100 30 70 110 33 77 120 36 84 130 39 91 90 20 21 44 100 20 24 56 110 20 27 63 120 20 30 70 130 20 33 77 . but limiting the amount of shares will make the per share figure more volatile and more profitable. If the mantra of financial leverage is to do “more with less”. increasing both the return and the volatility of that return on a per share basis. it is partially achieved by increasing the potential value of each share by issuing less of them. 11. (11.5 11.63 1. It may be less preferable based on the coefficient of variation which measures the standard deviation divided by the mean (11. but grows at a faster rate: Table 2-4 YEAR TO YEAR CHANGES FLOW 1 .the lower the figure.07 MEAN SAMPLE STD. while five are outstanding for the second.19 We now measure the mean of each flow and use the sample standard deviation as a measure of risk: Table 2-3 ACTUAL FLOW OF NET INCOME FLOW 1 (NO DEBT) 63 70 77 84 91 77 11. the better. we find that the debt laden flow is much more volatile.97 1.07 FLOW 2 (DEBT) 44 56 63 70 77 63 11.853 MEAN SAMPLE STD DEVIATION Now assuming that ten shares are outstanding for the first flow.NO DEBT FLOW 2 .33 9.07 -as an absolute value for net income.09 8. DEVIATION Although the debt laden flow has a smaller mean (63) than the no debt flow (77).07/63) vs. we have EPS calculations for each: .DEBT FLOW 1 (PERCENT) 11. However. if we proceed to observe the four separate changes in net income for each flow.194 FLOW 2 (PERCENT) 14.1 10 11.07/77) .29 12. its essential risk is the same .1 10 9. Higher fixed costs require a higher level of sales to break even. many firms adhere to the creed that “more is better”. % ∆ EBIT / % ∆ Sales. Since creating equity reduces some volatility. causing a higher risk of default and dilution of EPS at the same time.20 Table 2-5 EARNINGS PER SHARE 10 SHARES .NO DEBT 5 SHARES .106 FLOW 2 (DEBT) 9. the potential to increase price is balanced by the risk of volatile movement in both directions. especially if the equity is being used as a bargaining chip for executive compensation or for acquisitions. THE ANATOMY OF OPERATING LEVERAGE Like the financial leverage ratio where a variability factor (interest) causes volatility in the numerator (net income). Any poorly managed capital structure will have too many shares outstanding in addition to debt.3 7 7.fixed costs. In the degree of operating leverage. In essence.8 11. Such a combination moves companies away from the optimal proportion of debt to equity. and many of these firms issue equity without such foresight. operating leverage also carries an implicit and undisclosed variability factor .4 9.7 1.4 12.6 14 15.DEBT FLOW 1 (NO DEBT) 6. the proportion of debt to equity must be ideal in order for the stock to rise.2 12.24 MEAN SAMPLE STD DEVIATION Note that both mean and standard deviation increase on less income in flow 2 than in flow 1.1 7. Since EPS behavior is often a proxy for share price behavior. “fixed” costs contribute more to variability given the same level of sales and operating income. but variable costs are absorbed by the sale of . a financial paradox occurs: steadier. However.6 2. it is the change in fixed costs that contribute to earnings volatility.7 8. by making costs less variable with sales. The variability is as much a function of the type of cost as it is of fluctuating demand. storage. making it more severe. any business fluctuations will have an exaggerated effect on those firms with higher fixed costs. Companies often want to automate remedial tasks to reduce labor costs or to add a special competitive quality to the product. some operating costs must be paid even when nothing is produced (insurance. profits may surge during an upswing but fixed costs must still be paid during a downturn. we can plug an increase directly into the breakeven equation.21 each additional unit. (Price x Quantity). maintenance). we will decompose sales into the product. The fundamental key to understanding operating leverage is to recognize the effects of increasing fixed costs as a percentage of sales. have little debt.6S). We . and see high profits in an expansion but near insolvency when a downturn occurs. An example of how this works in “breakeven notation” is as follows: Table 2-6 VARIABLE SALES PROFIT FIXED COSTS VARIABLE COST BREAKEVEN EQUATION SYMBOL S PFT FX VC PFT= S-VC-FX VALUE 1000 300 100 600 300 In later chapters. a company must increase sales by a specific amount to cover these costs even if the variable cost per unit of production remains unchanged. The true risk of increasing operating leverage stems from the possibility of not increasing the breakeven point for sales. Perhaps the best example of over sensitivity to the business cycle occurs with biotech companies that spend a lot for research and development. Thus. while maintaining a 60 % variable cost proportion (that is: VC remains 0. Moreover. To illustrate an increase in fixed costs. Besides the inherent cost of machinery. that operating leverage changes. It is in the rare times when fixed costs change as a percentage of sales (higher rent. each revenue-generating procedure in a plant has its own unit-dependent operating leverage.(. On a macrocosmic corporate level. any time that a segment of the business is discontinued or changed. PFT=S-VC-FX or 1000600-200 = PFT = 200. The higher profits that are cited will be accompanied by the higher risk of more fixed costs. and sales are a function of the price of the unit multiplied by the quantity of the unit. Notice also that 0. operating leverage is more volatile. we obtain 1250 as the new level of sales needed to maintain the old profit level of 300. The smaller profit can be raised up to its old level by algebraically solving for sales: X . An acquisition of a service type business. A $100 increase in fixed costs instigated an increase in variable costs to $750.200 = 300.6(1250) = 750.75” (750/200). they are independent of the level of production. variable costs are spread over each unit produced. operating leverage is relatively stable because most processes are standardized.22 will increase fixed costs to 200. more salaried employees. as well as the purchase of capital equipment. A one hundred percent increase in fixed costs required a twenty-five percent increase in sales to absorb it. a one hundred percent increase. Fixed costs.) or as new technology is added. etc. are not unit dependent. however. Solving for X. In this case. and also a decrease in the ratio VC / FX from “6” (600/100) to “3.6X) . will probably lower a manufacturer’s risk. Ultimately.6 (X) is variable costs. On a microcosmic production level. there is a risk-related effect involved – some change in operating leverage. At this lowest level. On the other hand. Sales is the symbol “X” and 0. New product lines and processes are constantly changing operating leverage. TOTAL RISK . the switch from distributing an item to manufacturing it requires an entire shift from former methods of ordering and storage. for example. Any process that needs more machinery also needs a reliable source of steady financing to replace and repair equipment and fund necessary shifts in production. If the risk of default can be minimized by spreading a parent company’s fixed costs over more units of production.23 The higher level of fixed costs associated with more operating leverage requires extensive capital funding. In short. we are measuring the ratio of two changes: meaningful measurement is difficult when large variability is incurred through the use of percentage changes. there exists some level of total risk. creditors will only extend loans at high interest rates or not at all. the financial leverage ratio is inherently unstable because like operating leverage. the brokers of a leveraged buyout use less of their own capital (equity) and allow the corporation to assume the “limited liability” of greater debt. However. Additionally. They desire a customer who has steady cash-flow and is not sensitive to business cycle fluctuations. more debt can be incurred (due to lower total risk). the product of operating and financial leverages that determines the source of funding. If total risk is too high. . The addition of companies with less operating risk is a method of diversification that works in tandem with financial leverage. which puts them more at risk during a downturn. LEVERAGE MEASUREMENTS Using conventional methods. and less equity shares issued. creditors do not want to extend loans to companies that evince income variability. it is difficult to obtain a realistic operating leverage figure from the financial statements of a company. or operating leverage by itself is high. but accountants often have difficulty attributing fixed and variable costs to specific units: many costs have both “fixed” and “variable” characteristics. To analyze capital structure in this context. The firms with the highest operating leverages end up financing with retained earnings or equity issues. Thus. The leveraged buyout is perhaps the best example of how the two leverages interact. we attempt to find a concrete proxy. The result will be a higher share price for the “merged” company. Not only will production changes obscure the “true” number. 7)= 56]. The net income calculation is 140-20 = 120.25. operating income (EBIT) is 100. When interest expense remains unchanged.0. For year 2. we subtract 20 from 100 =80. However. interest expense is 20 and the tax rate is 30 %. Thus. The unique characteristic of the financial leverage ratio is its dual capacity: it is not only a comparative rating tool for default. but it predicts the pressure on earnings and the relationship between net and operating incomes. % ∆ Net Income / % ∆ EBIT is ((84/56)-1)/((140/100)-1)= 50 % / 40 % = 1.Variable Costs ) / (Sales .3) = 84. the concrete ratio predicts the exact earnings figures one year hence. say 40%. REALITY: FINANCIAL LEVERAGE The financial leverage ratio. Since investors rarely have access to such specific cost break-downs between fixed and variable. The ratio is: (Sales . it is almost unusable on that level. The increase in net income to operating income. which we multiply by (1-tax rate) to obtain 56 [80(0. the financial leverage ratio is 100 / (100-20) = 1. and 120 x (1 . making year 2’s EBIT equal to 140. % ∆ EBIT / % ∆ Sales also has a concrete counterpart that is calculable when the analyst has full knowledge of assigned costs. has a concrete counterpart that is more amenable to direct measurement which is.24 a figure that will mirror the analytical value of financial leverage while remaining stable enough to buffer the volatility of changes in operating leverage. we pick at random an operating income increase from year 1. by keeping interest expense constant. on a corporate “need to know” basis. it can be . net income becomes fully predictable! THEORY VS.Interest Expense ). % ∆ Net Income / % ∆ EBIT . REALITY: OPERATING LEVRAGE The theoretical relationship of operating leverage.25. EBIT / (EBIT .Variable Costs . To view how this works.Fixed Costs). and it can easily be converted to a times interest earned (TIE) ratio which is used to calculate default ratings for bonds. examine the following: In year 1. The stability of this ratio makes it ideal for comparison. To configure net income in year 1. Thus. THEORY VS. If we pick a sales increase at random. Also note that the only difference between numerator and denominator is in “fixed costs”. The student/investor should note that the numerator (sales variable costs) is referred to as the “contribution” while the denominator is actually EBIT.25 utilized in combination with the financial leverage ratio to form a measure of total risk.1333 % CHANGE IN SALES = (1676 / 1000) . restated with its component parts.4 CHANGE IN EBIT = (570. the ratio will predict % ∆ EBIT / % ∆ Sales exactly one year into the future. Determine next year’s EBIT if fixed and variable costs remain stable. Year 1 Operating Income (EBIT) is Sales .6 % 90.6 FIXED COSTS = 100 EBIT = S-VC-FX = 1676 .33.1333 / 67.Variable Costs .6 .Fixed Costs = 300. Besides measuring operating risk.1 = 90. operating income becomes fully predictable! THEORY AND REALITY: TOTAL LEVERAGE . and to anticipate changes in EPS.676) = 1676 VARIABLE COSTS = 0. no matter the level of sales. and fixed costs remain unchanged.33 Year 2 will yield a change of EBIT over the change in sales of 1.33 IN OPERATING LEVERAGE Thus. say 67.1 = 67. by applying operating leverage to a known change in sales. the concrete version of operating leverage also has a predictive capacity: given that variable costs are a stable percentage of sales. variable costs that are 0.6 = 1.6 (1676) =1005. the Hardseat Bicycle Company has 1000 in sales. the following values apply: Table 2-7 YEAR 2 SALES = 1000 (1. The following example will exhibit this relationship: In year 1.100 = 570.6 times sales and fixed costs of 100. Operating leverage is: (1000-600)/(1000-600-100) = 1.4 / 300) .6 %.1005. predicting the next earnings cycle becomes a remedial equation as long as sales are forecast correctly: (Old Net Income) x (1 + (Total Leverage x % ∆ Sales)) = (New Net Income). The fully converted equation is: (EBIT / (EBIT . Leverage is the precursor to a margin because it exhibits the dynamic movement necessary to change it. When compared to industry averages. if industry standards for operating leverage are traditionally high.26 Since reality dictates a constantly changing interest expense. If multiplied together. many stock run-ups will occur precisely because a company is successfully defying the odds and not succumbing to the negative associations that occur with too much leverage. A firm who increased its financial leverage at the beginning of a recovery and could afford to . analysts are hard pressed to make predictions from leverage ratios. each management team has a unique flexibility in changing them as a response to competitive pressure. more net income can arise from more leverage or greater sales. In fact. profit margin (Net Income / Sales). While meeting industry standards will determine “ball park” figures for the leverage ratios.too much debt for the level of income. margins and leverage are closely related and more leverage will contribute to a larger margin. but the leverage components cannot violate industry standards or dysfunction will occur that upsets a firm’s equilibrium . When management attempts to have a controllable total leverage ratio. but the two should not be confused. Conceptually. lack of capacity utilization etc.Interest Expense)) x ((S . However. For example. which appears to be a dynamic version of the classic ratio. fixed cost percentage and variable cost rate. each ratio accurately predicts the pressure on earnings if the status quo is maintained.VC . the ratios can gauge relative risk and exhibit the pressure to conform to those standards. cost overruns.VC) / (S . diversifying the firm with acquisitions that have a lower operating leverage would buffer the firm from an economic downturn.FX)) = % ∆ Net Income / % ∆ EBIT) x ( % ∆ EBIT / % ∆ Sales) = % ∆ Net Income / % ∆ Sales Indeed. we produce % ∆ Net Income / % ∆ Sales. would be taking advantage of lower interest rates and ultimately increase net income. Management will match the need for capital with anticipated cash-flows. earnings and debt that enables a corporation to limit the amount of shares issued and raise the price of the stock . one cannot manage financial leverage without determining the stability of cash-flow. the risks incurred in managing financial leverage lie outside the components of the measurement. interest expense and taxes. sales. LEVERAGE MANAGEMENT For a firm that is solvent.vendor prices. It is this symbiotic relationship between stable sales. meet the thresholds of the industry.and do so with a minimum of risk. which is derived from operating leverage. variable costs. vendors and the general economy. In essence. From a macro standpoint. and the level of technology in the industry. Only when sales are steady can management use financial leverage as a strategic tool to increase net income. financial leverage is much more amenable to management than operating leverage. a favorable economy. the other factors are greatly affected by the industry. The premium is to find a combination of leverage that will immunize the firm from business fluctuations . five components are the key to controlling leverage: fixed costs. they will not surge suddenly up or down. While tax policy and interest rates trend. sales and costs can fluctuate wildly depending on the state of the industry and economy. On the other hand. financial leverage trends more than operating leverage because it reflects management strategy and must conform to the necessity of raising capital in large increments. Of the five.27 do so. causing dysfunction. and yet strategically contribute to a large increase in sales. tax policy and interest expense are the most controllable. However. costs. For a project that is expected to pay off through a number of years. the relationship between interest expense and operating income are determined by the sometimes “uncontrollable” elements of operating leverage . Thus. it is simply more cost effective to raise large amounts of debt when the conditions are right to do so when the firm does not already have excessive debt and when interest rates are relatively . manufacturing instead of buying a part always increases both risk and potential return. . or it wallows in debt . diversifying one product line into different standards of quality appeals to a greater customer base. there are myriad methods of combining operating leverages to reduce risk or increase return: outsourcing and diversifying into related products often reduces risk as long as core competencies remain strong. The result is a higher profit margin and a lower asset turnover than in other industries.the same high need for capital to fund fixed assets creates pricing power in those industries because the large investment and added expertise act as barriers of entry to the industry. More on this subject is contained in the chapter on operating risk. Moreover. If sales are especially steady. there is a risk-return tradeoff in any industry with consistent sales. consolidating processes in one location increases risk and return. But . operating leverage will be relatively low and more financial leverage can be afforded. However.the grocery chain may be a superior investment if it balances risk and return more adeptly.28 low compared to other sources of funding. In effect. the root of the function lies in fixed costs: firms with higher fixed costs must constantly “up the ante” and increase sales to cover additional investments in technology. the objective of capital structure analysis is to discriminate between these two outcomes and choose the former before it occurs. Within narrow parameters.retained earnings and common stock. but often reduces risk.taking on even more debt. divesting assets etc. larger returns often accompany more risky operating leverage and those companies will trade the benefits of financial leverage to fund mostly with equity . the investor sees a build-up of risk which must be ultimately followed by one of two scenarios: either the firm increases return. By watching the flow of funds into a firm. SOURCES OF VARIATION Variation in sales is the common factor to variation in both types of leverage. Firms who make fifteen and twenty percent profit margins are not funded in the same manner as a grocery store chain making two percent but. this trending characteristic is a boon to investors. Again. This latter ratio is fully available to all investors but suffers from large jumps in measurement. will affect the level of funding for future projects if the rate is going up. % ∆ Sales for example. and represents just a small example of the possibilities available to reduce variation and increase return. When the Federal Reserve cuts or adjusts both the discount rate and the federal funds rate. On the other hand. This sensitivity to both the internal risk dynamics of the individual company and the state of the overall economy makes the financial leverage ratio a prime barometer for changes in both capital structure and stock prices. may indicate upwards earning pressure and is usually accompanied by an increase in operating margin. An increase in % ∆ EBIT / % ∆ Sales. When each ratio is broken into its component parts. % ∆ EBIT / % ∆ Sales. the investor starts to seek out more definitive reasons for the behavior of the ratio and the direction of risk. the combination of changes will indicate a general direction for risk and return in the firm. besides the derived variation in operating income from sales and fixed costs. The financial leverage ratio will not vary by a relatively great amount. which makes it less amenable to interpretation.29 This brief list of techniques is by no means exhaustive. Together. for example. Rather than seek out a threshold number for risk. While interest expense is partially controllable by the amount of debt a firm incurs and the type and source of loans. or alternatively. . The risk of changes in interest rates will affect the need to refinance at a lower rate. it is also a function of the state of the economy. the investor uses the combination to look at changes in both ratios. If it goes up or down by ten to fifteen percent it is considered a large change This relative stability makes changes significant and allows the investor to gauge risk by using it in combination with more volatile ratios like the theoretical construct. those firms who fund with debt are most affected. In the case of financial leverage. lies the variation in interest expense. an increase in the financial leverage ratio usually indicates that more debt has been incurred. 5. is one method of this type of extensive analysis. will not be the same levels . The risk characteristics of maximizing either EPS or ROE. the investor needs to “diversify” his or her analysis and assume a “balanced approach” by evaluating the risk of earnings. for example. yielding some “guesstimate” of optimality when that sector is favored. one must look for the implicit dynamics as well as the obvious interactions. Net Income / Stockholders’ Equity. we must combine that information with several other indicators to form a true picture of capital structure. gives us little information. focusing on one concrete number can be myopic at best and disastrous at worst.30 The variation in interest and operating income tell only a partial story. some paradigm of an optimal proportion is formed: usually a stock will peak at least once during the cycle. If a company finances with no debt at all. The risk of diluting EPS and market price is especially high in those companies who use minimal debt financing and need to issue stock to maintain a level of fixed assets. then the management of income and equity forms a similar risk/return imperative. However. relying on averages is risky: characteristics of corporations and industries change. To understand financial leverage. however. no where in either leverage ratio is this risk defined. We not only need to understand the changes in leverage. also known as the return on equity or ROE. Part of the reason for funding with debt is derived from a desire to keep the number of shares outstanding to a minimum. an absolute leverage ratio of 3. RETURN ON EQUITY Maximizing the return on equity is very close to optimizing capital structure. Thus. Thus. Observing the component parts of the quotient. and the investor is encouraged to coordinate information from several sources. The premium in capital structure analysis is to detect movement toward an optimal proportion of debt to equity. By comparing the industry averages of the largest competitors over the span of a business cycle. However. the risk of maximizing ROE is to grow at too rapid a rate. In effect. The very same elements that reduce risk will sometimes increase return. which is often. THE FORCES BEHIND LEVERAGE In the context of capital structure.31 needed to maximize the share price of the stock. product lines. known as the asset turnover ratio. Instead of measuring the change in sales and EBIT. may be more deleterious than a ten percent increase in earnings that is accompanied by low interest debt funding. EBIT / Sales.Interest. the ROE equation uses the absolute value of the ratio. and so capital structuralism opts to achieve the highest level of ROE with the least amount of risk. the ROE equation turns it upside down and measures the inverse. the objective is to itemize the characteristics of each component and seek some combination of factors that produces an optimum. Assets / Sales. the same components that form leverage measurements. The return on equity has as its foundation. The ROE equation turns this into another inverse. Sales / Assets. which eventually propels the stock downward. EBT/EBIT. For now. the student/investor should realize that risk and return are firmly interconnected on both conceptual and mathematical levels. Fixed costs are often measured by the ratio. also known as “operating margin”. called the “capital intensity” ratio. These ratios will be explained in detail in subsequent chapters. which is formed by strategic movement toward an optimal proportion of debt to equity. Each change in capital structure needs to be examined in the context of changes in the economy. Instead of measuring EBIT / EBIT . Since leverage is the backbone of . the right “infrastructure” needs to be in place. A twenty percent increase in net income accompanied by a thirty percent increase in equity. industry and internal dynamics of the company. To create a sustainable gain in a stock. but also includes integration between scheduled demand and the business cycle. absolute levels of risk and return are less meaningful than the integrated relationship between them. We gauge that relationship by observing their movement. This strategic movement is a balancing act between leverage risk and the cost of capital. In fact. The best choice for the investor may be the first quadrant (upper left corner) because default risk seems neutral. Table 2-8 INCOME RISK ++ (Increase debt and income risks) +(Increase debt. setting up a simple quadrant analysis will exhibit the fundamental forces. and observing the degree of respective increase. Gauging the other three quadrants is a matter of factoring several variables. because financial leverage is a function of operating income whose risk is derived from the consistency of sales.32 corporate risk. By that chain of logic. increase income) Obviously. but the effect on income may be greater. the “+ -” quadrant should have both the most risk and most return.increasing debt risks while reducing income risks appears to be the most dangerous. In this quadrant. We will then delineate the increase or decrease by a “+” or a “-” respectively. and each must be examined for individual risks and returns. the “+ . (Back to Table of Contents) .+” quadrant in the lower right hand corner. We will substitute the terminology “income risk” for operating leverage and the term “debt risk” for financial leverage. default risk (not paying interest in a timely manner) would be the least because income would cover debt better. the least risky quadrant would be the “. reduce income) DEBT RISK -(Reduce debt and income risks) -+ (Reduce debt. None of these combinations are “set in stone”.“ quadrant . Their operating leverage and profit are as follows: OPERATING LEVERAGE) 1000 . systems would go down for days at a time and it was important to have a contingency backup .usually in the form of a machine. operating leverage increased to 1. more efficiency will lead to greater risk. more operating leverage and temporary cost savings would lead to less profit and not more. While computers increased productivity.33 OPERATING INCOME) 1000-600-100 = 300 A new type of rubber is both less expensive.400 . The new operating leverage is: NEW OPERATING LEVERAGE) 1000 . While "naysayers" decry the dehumanization of work. Managers might reply. The Skidmark Tire Company has $1000 (million) in sales. but needs to be applied with machinery costing $100 more. indicating more risk.5 NEW OPERATING INCOME) 1000-400-200 = 400 While this scenario would represent a dream come true for some operations managers.a “paper trail”. Some readers of this book may be old enough to remember when entire industries were computerized in the late 1970s and early 80s. The following scenario should demonstrate that even in the best of circumstances. Technology and risk go together.600 / 1000-600-100 = 400/300 = 1. fixed costs must be paid regardless of the level of sales. a variable cost rate of 0. “But where’s the risk?” In fact. A common method is to consolidate repetitive functions and avoid duplication by adding technology . this project should not be implemented.5 from 1.33. they think of reducing costs and increasing profits.33 APPENDIX: STREAMLINING? When most businesses think of “streamlining” an operation.400 / 1000 . Cost savings would bring variable costs down to 0. without a thorough examination of sales volatility.200 = 1. Any time productive activity is . If sales deteriorate on a seasonal basis (as they do in the tire industry). the business person must ask if the return is worth the risk. and stronger.6 sales and fixed costs of $100.4 sales and improve operating income by $100 at the existing level of sales. even though potential profits increase. (Back to Table of Contents) .34 centralized in one process or location. more risk is incurred . deflation. “the trend is the friend”. by the time the year 1980 rolled around. Students often leave college with a set of ideals that readily flourish in a sanitized. debit vs. Often. only to falter when tested and stressed by real-world random variation. there is eventual disenchantment with “fundamentals” because stocks seem to have a ”mind of their own” and rarely respond to such analysis. However. demand. Keynesian vs. expectations of inflation carried as much mathematical “weight” as any other hypothesized cause. When students become “investors”. Indeed. isolated laboratory. Rarely does one financial thread weave itself through the random chaos of opposing ideas.35 3 CAPITAL STRUCTURE Economics has a dualistic nature. monetarist. if one graduated before 1970. In this case. if the return were as high as touted by the software dealers who sell it. Investors become discouraged because some “magic” combination of sales and earnings fails to beat a competitor who is barely functioning.until the investor realizes that he or she is merely following random patterns made within the context of a rising market. risk vs. but the changes in patterns and transitions are unpredictable. technical analysis . That thread is capital structure. inflation vs. credit and ultimately. the frustration with “chasing earnings” will turn into a penchant for technical analysis. each concept can be defined by its opposite: supply vs. Whether it is a human quest for certainty or simply a need to achieve balance. which at first appears to yield legitimate results . both high inflation and high unemployment occurring at the same time was inconceivable: the Phillips curve professed a tradeoff between these economic states and direct correlation was infrequent. and holistically embrace their reconciliation. most students will adapt them to the vicissitudes of modern finance. inflation and unemployment had been so rampant that economists changed their own concept of causation. For example. Rather than abandon those ideals. return. the cost of capital does not always have an “upfront” accounting cost. the variation of the return has as much import as the return itself. is not so unusual: movement toward a firm’s optimal proportion of debt and equity funding tends to propel the stock upward. In effect. Capital structure is firmly entrenched in academic tradition. and beguiling enough to lose money over. and its only “cost” may be the greater . capital structure analysis concentrates on the context of earnings because it is concerned with both risk and sustainability. Secondly. Without this domain of risk. The ability to forecast is not as important as the ability to coordinate information and identify firms whose leverage is conducive to greater earnings. however. and the need to price all sources of capital – debt. THE COST OF CAPITAL: MARKET ORIENTATION AND PRACTICAL APPLICATION The traditional definition of the cost of capital is conceptually vague. The basic concept. equity and associated variations – at the market rate. It is considered an “opportunity cost” that is comparative in nature. profit appreciation can be both deceptive and transient. but is flexible enough to apply to real world situations. Foremost among these is the gauging of “similar” risk.36 would attract major companies like Boeing and GE who would happily forego their eight or nine percent profit margins in pursuit of “safe” forty percent returns. most analysis concentrates on earnings because it is the most correlated fundamental to stock price. What is more difficult to grasp is the balance between risk and return that allows this to occur. the cost of capital encompasses several implicit factors that complicate its practical use. its study can lead to both revelation and financial remuneration because it is interactive with so many economic disciplines. Defined as the amount of return that a business could make on alternative investments of similar risk. Alternatively. If a firm has a 30/70 percent debt to equity. Similarly. investors are not privy to negotiations with creditors over interest rates nor are .37 risk taken to generate more income. the respective costs of equity and debt are proportioned by their relative weights in the capital structure to form an aggregate cost of capital. then 30 % is multiplied by the most current interest rate and then by the incurred tax advantage of (1 – tax rate) to produce a percentage cost of debt This figure is added to the product of the percentage cost of equity and the proportion of equity in the capital structure (70 % in this example). Stock prices can be quite volatile and anybody who follows the market can testify to the futility of gauging a corporation’s required amount of equity from market values alone. Together. Comparative actions that are not pursued may have as much significance as the actual course of action. By interfacing individual corporate risk with the prices configured by the current state of the capital markets. the scope of the calculation is beyond the purview of most investors. Since the price of a firm’s debt will change exponentially depending on the relative increase or decrease in the newly negotiated rate. all sources of capital are priced at what the market currently dictates and implicit in the analysis of comparative investments is the breakdown of these capital components into relative price levels. Each value can be multiplied by the proportion of its respective component in the capital structure. Problems arise when the investor attempts to corroborate theory with reality. a specific “required rate of return” will be determined. investors will attempt to determine the expected appreciation of stocks with equal risk and attach this rate to the market value of the firm’s stock. For debt. Theoretically. the return is the most current interest rate which is multiplied by a reciprocal of the tax rate and then by the market value of a firm’s debt. For equity. adaptations of the cost of capital are improvised throughout this text to assist the investor in better gauging risk. the history of the industry. Thus.38 market values of debt always determinable. for example. but it needs other financial information to corroborate it. alternatives can be compared and the most cost effective path can be realized While the use of capital implies long-term planning and obligation. THE ADVANTAGES OF DEBT . it may incur what is termed an “opportunity loss” – a measurable outlay of interest expense over and above the cost of equity. determining capital proportions from market values can even be dangerously misleading.such as interest expense –may offer the investor a practical analog. a corporation can issue stock at a high price but chooses to issue high coupon rate bonds instead. or “lower” than competitor’s rates. If. the investor needs a common denominator. and the expected size and stability of earnings must be counterpoised to the current market cost of capital. the market will frequently misprice risk over short periods. for example. As long as a dollar amount can be attached to any strategic action. the absolute size of the cost of capital is less important than its relational value and its context. Moreover. With market rates. In effect. the volatility of market rates makes the cost of capital a relative value: it can be designated as “improved”. the types of assets. Therefore. a corporation can observe the direct gains or losses in following a specific course of action. Why use market rates to price the cost of capital? In order to gauge the risk of alternative investments. and why using some book values . To determine capital proportions. An appendix in chapter six covers some of the theoretical underpinnings in determining a “real cost of capital”. During this lag. Performance indicators like the return on capital are significant. profits are enhanced and risk is diminished if payment is made promptly but not too soon. Like “just in time” inventory systems. most businesses recognize the importance of receiving cash-flows at the moment a bill is due. when these measurements are below that standard. interest may be due on the loan. if a firm’s peers are returning one and a half times as much. When servicing debt. debt is evaluated by comparison to other firms. Not only are returns greater when the payoff from an investment occurs faster. which may be accompanied by a sell off. industry standards are an important element in evaluating debt. but even more indicative are the average ratios of debt to equity. no cash inflows are balancing the outflows. but the risk of the loan is diminished because the firm has adequate cash to service interest payments. Any amount of cash-flow is more valuable in the present than it is in the future because it can be invested and earn interest. When both returns and leverage ratios are hovering around the industry standard.the reception of immediate funds with payment “postponed” until a later date. a net advantage occurs However. or maintenance of the position. risk and return will be commensurate with a firm’s peers. increasing its risk. Thus. Whenever an account returns more on a loan than it costs. the investor can look for one of two outcomes: either a compensating bounce (turnaround) that lowers risk and propels the stock upward. the timing of the inflows is of even greater importance. A loan that returns more than its cost may be considered “ineffective”. However. This latter position can make or break investors because maintenance of the position can be a sign of either faith in an investment payoff.39 Implicit in the definition of debt is its inherent advantage . Besides the time value of money. or . profits received in January and February are much more valuable than the same profits received in November and December simply because the “lag time” is not productive. In effect. and the amount of interest expense relative to long-term debt. Investment with debt may limit the number of shares outstanding . interest expense. the optimal proportion of debt to equity would be less significant.40 uncertainty over the inevitable “cost over runs” that may further scare investors. but interest payments must be timely. the amount of fixed costs in a production cycle regulates . a corporation who does so flirts with default. depending on the risk of insolvency alone.e. Without this tax benefit. sinking funds. and once debt is incurred. Earnings are often cyclical. Since interest is tax deductible. more debt on the books will limit takeover attempts because prospective buyers do not want to be burdened with the obligations of leverage. taking on debt will save the firm an amount equal to the tax rate multiplied by the amount of bonds. • 2. • 3. Interest is a tax deductible expense. A number of mechanisms are in place that confers other advantages: • 1. the probability of bankruptcy increases exponentially. and then plans to make up the difference in new income taxes when the investment generates profits. performance can be enhanced without burdening shareholders if earnings per share (EPS) rise. the government confers tax advantages to encourage investment. which can dilute market price. In effect. A long-term loan (over one year) grants tax advantages until the loan is paid off and may have a stabilizing effect on the firm if cash-flow is especially tight. the cost of debt is almost uniformly less than the cost of issuing equity. Interest expense is routinely deducted from operating income before taxable income is calculated.. the tax savings are the cost of interest multiplied by the tax rate. Less new shares restricts voting power to the largest current shareholders. Issuing debt rather than equity helps maintain existing control of the company. etc. In fact. RISK AND DEBT While it seems advantageous to pile up debt and use the tax savings for other investments. restrictive covenants. • 5. i. Moreover. • 4. In fact. Since all debt funds a specific level of assets. Their risk of bankruptcy is higher simply because their production cycles are more reactive to economic conditions. once a default probability is ready for the market. Once the “generic premise” is accepted. Default probabilities are created from historical distributions. The financial community develops . In effect. and have a higher risk of default during economic downturns. or sales stability changes. it is no longer one hundred percent accurate. the proper confluence between several asset classes and the amount of intrinsic shareholder value forms an amount of loss. there is some cost of bankruptcy that is composed of at least two associated elements: some amount of loss and some probability of default. Thus. Moreover. the probability is no longer valid.41 both operating risk and the amount of financial leverage that can be incurred. Those firms with very high fixed costs may have earnings that fluctuate more than firms with lower fixed costs. The term “bankruptcy” has many definitions and covers a wide breadth of legal states and financial conditions. market value is affected by psychology as much as assets are affected by inflation. There are some assets that can be sold in the event of liquidation and some assets that do not directly affect the operating capacity of the firm. liabilities and market value. the main tenet needs to be examined. these firms with high “economic” risk are poor candidates for financial leverage. which by its very nature is probabilistic. the market value of the stock above the level of assets is based on the expected ability to generate income in the future. One major problem with capital structure analysis is enumerating the cost of bankruptcy. Thus. these form the base of a generic model. there is considerable difficulty in creating a relationship between fixed costs and asset structure to obtain an “amount of loss”. This text approaches the concept from the perspective of the corporate common shareholder and always presumes the loss of shareholder value. When the relationship between typical inputted variables like asset size. It designates a relationship between assets. Although there may be many other factors. However. However. The cost of each source of funding is interdependent on the market for alternative sources. This interaction between risk and the cost of capital is never static and forms the basis of capital structure analysis.42 limits of tolerance that accepts imprecision as a given variable. Since the cost of capital cuts into profitability. the firm’s stock price will be maximized. the reliance on default probabilities is in itself “risky”(the 2007 “credit crunch”) because there is risk that cannot be enumerated until the event occurs. Inevitably. those firms who have greater resources can afford to use more debt in their capital structures which causes less strain on existing shareholders. enhancing market value. However. cost. proportional increases in debt have a fixed tendency to raise the cost of both equity and debt as risk becomes higher. Their inability to maintain cash-flow perpetuates a chain of loans in which one loan retires another with little payment of principal. risk. Ultimately. When projects . low-earners with diminished market values. seeking an optimal capital structure turns into a game of strategic risk. capital structure analysis rests upon this tolerance for a “ball park figure”. allows fewer shares to be issued. they may affect a firm’s performance more than any internally generated variable like sales or assets. In fact. to form a “generic cost of bankruptcy”. Investment banks demand more interest on bonds and higher flotation costs for stocks to mirror the greater risk in a company. contributing to performance on a per share basis. the utilization of debt with recourse. but must have data that is coordinated to alert the analyst to the presence of more risk. but each component part is hypothetical and tentative. Thus. This strategic use of leverage. The economic and political risks that are outliers to any “system”. are very illustrative of this concept. firms with too much debt are usually cashpoor. and the interrelationship between economic outlook and corporate performance will be balanced. With an increase in bankruptcy costs comes a concurrent rise in the cost of capital. When an optimal proportion of capital sources is achieved. we depend on default probabilities and we multiply them by an estimated amount of loss. is not in the concept of the function. some of the variables depend on probability and some are deterministic. the efforts may be wasted. The problem. Consequently. the term “bankruptcy costs” is unique to the realm of the amount of loss. the function is at an optimum and ∆ Tax benefits = ∆ Bankruptcy costs. investors might even demand that the firm take more risk by moving away from the relatively “safe” world of a stock price optimum and engage in mergers and acquisitions. Wall Street rewards companies who “defy the odds” by out performing the market in some special way. they are interdependent nevertheless and create both variation and uncertainty. there is little room for it to move . Since a stock price maximizes when a firm’s capital structure is optimal. however. This equality optimizes the proportion of debt to equity at the point where the change in tax benefits equals the change in bankruptcy costs. If we calculate the first derivative of the function and set it to zero. but the potential outcome of inaction can be devastating. As previously mentioned.up or down. MATHEMATICAL OPTIMIZATION: THEORY VERSUS REALITY The theory of capital structure is predicated on the balance between the tax benefits of debt on one side of a function and bankruptcy costs on the other. but in the definition of variables that interact. the need for mathematical certainty is not as significant as the need to realistically gauge risk and move the firm in the right direction. but resolutely moving toward it. we try to maximize the function (Tax benefits of debt) . In effect.43 become profitable. the potential for price appreciation in a stock is much greater when a firm is a long distance away from its optimal capital structure. To a financial executive. Ultimately. the mathematical situation is akin to forecasting the path of a hurricane and deciding to make preparations. there is less shareholder investment but greater return. In fact. In this case.(Bankruptcy Costs). Moreover. each corporate entity loses something different. and the market price of the stock increases. firms who seem laden with debt but begin to escalate sales and earnings are generally observed to be a long distance from an optimal . and the equity market. and especially any changes thereof. 1 EVA® is the registered trademark of Stern Stewart. The true optimal capital structure is in a state of flux.almost instantaneously.44 capital target. the foundation of corporate risk is the inter relationship between long and short-term interest rates. They can help the analyst to understand the components of corporate risk. These relationships change daily. such exercises will be laden with at least three discrepancies: 1) A lack of adequate variables 2) Artificial constraints and 3) Too many constants. even if imperfect. However. Several useful indicators exist to aid in determining whether the firm is moving in a favorable direction. and EVA®1 (economic value added) can be modified to reflect activity toward an optimal target. Thus. if a firm were shut down for a weekend. Inc. their higher risk of bankruptcy coupled with subsequent improvement substantiates more investment. While leverage factors may determine an estimate of the amount and sources of funding. Fortunately. Although the student/investor is encouraged to create a working model and obtain a “ball park” estimate. In fact. The reason behind this variation is that it is dependent on external relationships outside the internal control of the firm. the analyst does not need a mathematically precise rendition of an optimal capital structure to make effective recommendations. its optimal capital structure would change ever so slightly. such efforts can be informative. . it responds to so many different variables that it perpetually changes . even depending on the performance of foreign markets. Classic measurements like the return on equity (ROE). the cost of debt would always be lower than the cost of equity. the value of a firm is independent of its capital structure. we relax constraints and add variables to derive a realistic hypothesis. By adding bankruptcy costs. (Tax Benefits) .MILLER PROPOSITIONS Most of our present knowledge of capital structure is merely an extension of the research done by the team of Merton Miller and Franco Modigliani in the late 1950s. As long as interest is tax deductible. In their famous equation V(l) = V(u) + TB. In fact. the value of the unlevered and leveraged versions of a firm would be equal. Without such tax advantages. a firm can increase its value simply by incurring more and more debt. we merely regulate the tendency to use lower cost debt. the necessity of covering interest payments increased the cost of equity. in the absence of bankruptcy. we can conclude that the optimal capital structure is composed of one hundred percent debt. In essence. When this proposition is examined in a world of taxes. Since interest is a tax deductible expense. the function. free cashflow would always be greater with the use of debt. proves that increasing the proportional amount of debt increases the cost of equity.45 THE MODIGLIANI . Since the real rate of return on debt must cover interest payments before it shows profitability. which forms the crux of Miller / Modigliani’s Proposition I: In the absence of taxes or bankruptcy. The logic behind proposition II was that the cost of equity reacted to the increase in debt by rising. in the absence of bankruptcy. The second proposition. it is higher than the rate of return on equity by a “risk premium” that pushes the cost of equity upward. this extreme “corner” solution forms the foundation for all further extrapolations of capital structure theory. This “risk premium” was reconciled with equity by a higher required rate of return. for example. proposition II. In fact. which was the amount of the firm’s bonds multiplied by their tax rate. Without . the value of a leveraged firm was greater than the value of one with an all equity structure by a factor of TB.(Bankruptcy Costs) is simply a relaxation of a constraint that they used to determine the incremental value of a leveraged firm over an unlevered one. 46 taxes and bankruptcy. Without a standardized “cost of bankruptcy”. THE OPTIMIZATION PROBLEM While tax benefits are composed of a simple linear function (long-term debt multiplied by the tax rate). the shape may be defined by an exponentially increasing loss or a linear default probability. with a constant amount of loss. Figure 3-1 Cost of Capital Cost of Equity Average Cost of Debt and Equity Cost of Debt Debt / Equity Ratio Each incremental unit of debt pushes up the cost of equity because it increases the required rate of return. . In other models. each manifestation of bankruptcy is fundamentally unique. We assume that the levered company is making at least as much as the value of its interest payment more than the unlevered firm. In some models. it takes on the shape of the default probability. the cost of debt would be a horizontal straight line that fell well below the angled line of the cost of equity. the cost of bankruptcy will adopt a shape that is defined by its variables. or there would be no reason to incur debt in the first place. solving for an optimal amount of debt depends on the combination of variables that make up the function. each additional unit of debt will increase rates so rapidly that financial leverage will no longer be cost effective. Thus. it will match the amount of capital with the boundaries of the default probability. The limiting constraint in any optimization model is the amount of capital. they should show increasing costs for greater risk just as creditors charge higher interest rates for more risky loans. For example. the curve should emulate a risk-return tradeoff and increase at an increasing rate at some point. Secondly. followed by a rapid increase and then a leveling off when actual bankruptcy occurs. the cost of bankruptcy must have a slope that is greater than TB (Tax Rate x Bonds) or tax benefits will exceed bankruptcy costs at every level and the model is not operable. When a model attempts to optimize the amount of debt. through fewer shares issued. many default algorithms will allow a greater amount of debt than the capital limit simply because they are not configured for extreme values: default probabilities are created from averages and may not be accurate within the level of debt that is normally used by the company. The greatest accuracy is achieved within narrow bounds and may not include the given amount of capital. While bankruptcy costs do not perfectly mirror the cost of capital. In reality. . past a specific point. The linearity of the assorted costs of capital in the Miller/Modigliani propositions was a function of the constraints . the cost of capital is regulated by several relationships that create a jagged curve and limit the amount of debt that a firm can incur. Lastly. mimics the cost of bankruptcy curve as creditors charge more interest at an increasing rate. some of the shape of the cost of capital curve should be incorporated into the cost of bankruptcy. It then moves rapidly upwards as interest expense becomes onerous and the firm moves closer to bankruptcy. the theoretical underpinnings of the cost of capital will create a curve that in the long run. The cost of capital curve will at first decrease as more debt creates both tax benefits and a larger EPS.especially the absence of bankruptcy costs.47 At the very least. a modified “S” shape is ideal because it shows a threshold amount of bankruptcy costs. Nevertheless.48 Some models shelve the concept of market value altogether and concentrate on the amount of asset loss that occurs in a typical bankruptcy within the industry. any model will oppose the costs of debt with its benefits. Naturally. and we can satisfy the equation by manually inputting each unit of debt from zero to the capital limit. Alternatively. When the function . Default algorithms may be ‘customized” to meet the needs of firms within an industry and will be more accurate than the generic algorithms used in this text. . we can use a linear programming module like Excels “Solver” to maximize the function. ∆ Tax benefits = ∆ Bankruptcy costs.(Bankruptcy Costs) no longer increases. The graphical display would be thus: Figure 3-2 $ COST TB Cost of Bankruptcy X DEBT .a spin off of the original Miller/Modigliani thesis. longterm debt would be a variable in bankruptcy costs. (Tax Benefits) . an optimum is found. and these models will be configured for specific sectors.. by finding a level of long-term debt that makes the function equal to zero. From twenty to sixty-five percent of assets is a common figure. THE PROPOSED IDEAL AND ITS INHERENT PROBLEMS We can solve the marginal benefits function. The marginal benefits function. the optimal amount of debt is at point “X” because the difference between tax benefits and bankruptcy costs are at their greatest point. a program might optimize debt under the premise that the firm is earning a ten percent return on capital and not the fifteen percent of its competitors. the function. for example. Any optimization model works within a given capital constraint which may not be the optimal one. we produce the optimal amount of equity.Few companies have the ability to match the amount they can raise with actual capital requirements. “Long-term debt”. and yet capital allocation depends on expectations. Among the several problems in such a model are: • 1. but excessive if rates are expected to rise. we produce a differentiated function.Besides the construction of hypothetical variables (amount of loss). ∆ Tax benefits = ∆ Bankruptcy costs is operable when the slopes of each curve are equal. • 3. By linking tax benefits to bankruptcy costs through the variable. The assumption of appropriate capital. In this case. When we subtract this optimal amount of debt from a given amount of capital. For example. No optimization variable is cognizant. some variables must be held constant when in reality. they would change dynamically with the level of debt. or . However. an optimization program might lock a firm into a level of debt that is appropriate at a lower rate. • 2. From the perspective of calculus. a true optimal proportion of debt to equity would bring the return on capital (ROC) up to industry standards.Optimization variables are not formulated in terms of corporate potential. either the amount of capital is excessive. judgment.. Another example would be the expectation of increasing interest rates. and wisdom garnered from historical observations. (Tax Benefits) (Bankruptcy Costs) is maximized when its first derivative is set to zero. Context . Artificiality .49 In this “ideal” depiction. if under performing its peers is considered “normal” performance for a company. of a marketing strategy that will make a firm an industry “front-runner”. Using this faulty line of logic.50 production problems limit the amount of net income . In effect.comparatively favorable rates.a problem that is not part of the optimization function. firms do take advantage of interest rate cycles to incur more debt but do not materially change their optimal capital structures. the major portion of the cost of capital is configured away from the internal dynamics of the firm and is dictated by a confluence of factors in the greater economy. an aggregate of the demand for money will determine the absolute level of interest rates while competitive forces within the company (sales stability. In fact.that the price of debt (interest) determines how much debt a firm can incur. For example. large firms with steady incomes and salable collateral can seek and incur more debt in their capital structures. CAPITAL STRUCTURE AND THE COST OF CAPITAL Both students and investors get the misconception that the relationship between the cost of capital and capital structure is based on affordability . Some of the factors that regulate the proportion of debt to equity affect the cost of capital. Thus. and how well the assets can serve as collateral for a loan. by steady earnings and stable prices. On the other hand. the same performance that they expect when receiving interest payments. or rates that are at the upper most levels.the proportion of fixed assets. The fixed income market is based on dependability. earnings. which is actualized. The decision to use debt is not so affected by the price of debt. . creditors demand from firms’ asset structures. as by the asset structure of the company . a company can have a forty percent debt to equity ratio when interest rates are down and a twenty percent proportion when they rise. lower rates become an incentive to take more risk but are balanced by the diminished earnings outlook that usually accompanies a Federal Reserve rate cut. and assets) will determine the limits of tolerance . Capital structure has an interactive cause and effect relationship with the cost of capital. real estate may serve as better collateral for a bond issue than a more valuable commodity like gold simply because it is less volatile. the optimal capital structure is made up of all debt. The investor can verify this concept by observing the demand for any stock that has a junk bond status.more capital to pay for outdated machinery. When we turn this around to: Assets / Sales. firms who. by price alone. In effect. a direct correlation that is maintained at lower levels as well. operating leverage and capital intensity. The third characteristic. a higher probability of bankruptcy will raise the cost of capital. capital intensity will affect both operating leverage and sales stability and is very dependent on the type of industry. Capital structure. However. the cost of capital responds to the risk of bankruptcy in the capital structure and will adjust to reflect the stability of the company. Risk is created by the propensity to generate fixed costs. capital intensity is merely the inverse of the fundamental. the price of the stock deteriorates as the cost of its debt sky-rockets. In the long run. and the stock becomes a tool for speculators. firms with higher capital intensities will have projects that encourage long-run profitability and require extended funding. however. asset turnover.which often go” hand in hand”.51 Three characteristics of a firm’s asset structure are especially significant: the stability of sales. within those confines. which is: Sales / Assets. it is doubtful that the firm can obtain a level of risk that pushes the cost of debt above the cost of equity. Firms desire to match the timing of cash-flows from projects with the type of funding because such a strategy increases return and reduces risk. By Miller/Modigliani proposition II. Consequently. By itself. the decision to fund with equity or debt still exists. and higher capital intensities will require just that . we obtain a comparative ratio of the degree of fixed assets in a company. Therefore. management salaries and various “overheads”. The risk of holding one thousand dollars worth of stock is much greater than holding a one thousand dollar bond of the same firm because the bond needs to be repaid or the stock will become worthless. and is as much dependent on the risk of physical collateral as it is on operating risk . responds slightly to changes in the . deal with riskier assets will often have the most operating risk. However. moving past its optimal target from the other direction. as indicated by asterisks. changes in rates give incentive to raise more capital but not to make major changes in capital structure.52 cost of capital. the cost of bankruptcy is not always perfectly aligned with risks and calls for executive . The following chart delineates different levels of debt to equity from the perspective of both the cost of bankruptcy and the cost of capital .2003. mostly adjusting through the amount of capital raised. For a firm who funds with all equity. the firm begins to fund with more debt. When rates are lower. shareholders will sell the stock because too much risk is incurred. but may adjust the amount of funding for all projects and move slightly past its target with more equity funding. Such patterns occurred in the late 1970s with massive inflation and double digit interest rates and again with near record low rates in 2002 . firms whose asset structures have changed and are more amenable to using debt. The student/investor should recognize that no level of interest rate will ensure steady repayment of a loan because rates change frequently with the level of GDP growth. the company may not change its target structure to a lower level of debt. When interest rates are high. However. On the other hand. For companies who use debt on a regular basis. The cost of debt is indicated with tax savings and rises with bankruptcy costs. for example. Only when there has been a systemic shift that will raise or lower the interest rate throughout an entire business cycle will the optimal target change substantially. In effect. will receive an initial boost upwards as tax advantages eclipse the cost of bankruptcy. The cost of equity is above the cost of debt and also rises with the probability of bankruptcy. a favorable change in the cost of capital is an impetus to move well past the optimal target or even temporarily move away from it altogether because the risk of doing so is rewarded with a lower cost of capital. taking advantage of “cheap interest” for a short amount of time will merely cause the firm to suffer the consequences of a poor decision. firms realize that such conditions are temporary and usually move back toward the optimal target as soon as possible. 75 11.5 12 14 17 20 24 30 60 90 120 150 180 210 240 270 300 This cost of bankruptcy function is a bimodal distribution. for example. preferred stock and some short-term debt and capital leases.18 5.53 judgment.5 6. .5 6 6.8 Cost of Equity % 7 9 9.8 7.3) percent.5 10.85 4. Many bankruptcy algorithms are only accurate across a narrow range because they are based on average historical distributions. Table 3-1 Percent D/E 10 20 30 40 50 60 70 80 90 100 TB TB .4 8.88 6. the measurement that is formed is called the weighted average cost of capital or WACC. THE CONCEPT OF A WEIGHTED AVERAGE COST OF CAPITAL The standard definition of “capital” comprises several component parts including equity. The challenge to any mathematician is to produce an algorithm that can maintain accuracy over its entire range and yet be flexible enough to solve for capital structure inputs. When the percentage of each component in the capital structure is multiplied by its specific cost and then summed together.76 5.225 9. The judgment of the analyst is paramount and it is obvious that the most cost effective path is at forty percent.4 9. producing two inflection points: one at forty percent debt to equity and another at eighty percent. at twenty percent debt to equity. The cost of bankruptcy is a typical algorithm.75 14 17 After Tax Cost of Debt 3. The calculations are based on a level of capital of $1000 and a tax rate of thirty (0.55 4.85 8. long-term debt. In the chart above.5 9.Cost Cost of Bankruptcy 20 30 40 55 90 130 180 170 1250 1350 10 30 50 65* 60 50 30 70* -980 -1050 Interest Rate % 5.2 4.8 11. but minimizes when the capital structure is optimal .2) (0.04 percent.stability of sales and income in the domain of higher returns. the cost of capital will be uniformly higher for all equity companies. there may be eccentric movement because the market does not always price risk correctly.0804 or 8. it may rise from year to year simply because there has been a systemic shift toward comparatively higher capital costs. In the long run. It funds capital projects to make back its investment as rapidly as possible. Thus. which they need to overcome by both increasing and stabilizing the amount of sales and income.09) = 0.042) by the percentage of debt (0. it does so to preserve systemic equilibrium.8). Therefore. the WACC is a reliable indicator of movement toward an optimal capital structure. Over a two or three year period.042) + (0.8)(0. The total expression is (0. but maximizes tax benefits. In the short run. However.keeping shares to a minimum and lowering the probability of default . by minimizing bankruptcy costs . . the WACC will follow the probability of default either up or down. there will be periods when the firm is using more equity while interest rates for debt are decreasing and vice versa. a balance between growth and inflation.at a point where tax benefits and bankruptcy costs are at their greatest distance. For firms who fund only with equity. the greatest amount of the least expensive capital component is used. producing an optimal capital structure. a firm does not set policy to conform to Federal Reserve decisions. an unlevered firm must minimize its WACC without the luxury of substituting lower cost debt for equity. the probability of default is minimized with the same actions that will boost structural integrity and decrease the cost of equity . In most cases. However.2) and sum it with the product of the cost of equity (. When the government sets interest rates. any decrease would be viewed as favorable.54 we multiply the after tax cost of debt (0.09) and its respective percentage (0.it will also minimize its WACC. However. By using the amount of debt that keeps the cost of bankruptcy at a relative minimum. . they incur the cost of equity. which is the ratio. but let it be stated that the cost of equity is a comparative cost very much related to the return on equity. In this regard. dividend policy is very integrated with capital structure because it determines retention and the ultimate WACC. if the cost of equity is particularly high. In effect. However. it will cost the firm more to retain earnings than to distribute them as dividends or buy back shares of stock. Net Income / Stockholders’ Equity. The full development of the cost of equity is left to another chapter. the amount of earnings is also a major factor in determining capital structure. When earnings are large. large amounts of earnings tend to be unstable because of the risk-return tradeoff. they can be retained. When earnings are retained and become part of stockholders’ equity. Thus.Cost of Bankruptcy Debt / Equity EARNINGS AND CAPITAL STRUCTURE In the chapter on leverage we mentioned that the amount of fixed assets in an industry will determine the stability of earnings and that financial leverage can be increased when operating leverage is lower. but enable a firm to build stockholders’ equity through retained earnings. and outside sources of funding (debt and new shares of equity) can be avoided.55 Figure 3-3 Dollar Value WACC TB . CAPITAL STRUCTURE LOGIC Capital structure theory displays a pattern of alternating risks and returns that are encapsulated in several measurements. high tax benefits will be accompanied by an even higher probability of default because firms are generating less income. Consequently. the marginal benefits function can grow when equity is added to debt but does not replace it. In the marginal benefits equation. more shares.56 Another effect of earnings on capital structure is the minimization of the probability of default. Only when the cost of equity is considered too high will share buybacks and special dividends need to be considered in lieu of retention. and sometimes even the luxury of moving away from the “safe” environs of an optimal capital structure. In expansions. . The business cycle determines when certain industries will have favorable sales and earnings. Consequently.capital appreciation that leads to asset growth. The effect of an optimal capital structure. is to reduce the negative effects of the business cycle and to accentuate the positive. or some combination of both . Inevitably. more income lowers the probability of default allowing more tax benefits. Each category of risk stems from the previous category and has a corresponding set of measurements. the proportion of equity grows through retaining the same earnings that diminished the probability of default. to preserve tax benefits. however. the other risk categories must be properly assessed. the company will have the financial flexibility to take on more risk. capital allocation. they must pare down their debt above all else. In a downturn. In the following chart. an entire sector will exhibit similar patterns of marginal benefits and leverage that will be dependent on the probability of default. during an expansion. to get to the highest level. a smaller probability of default allows more debt funding. The cost of capital is kept low enough during a downturn so that competitive progress can be made. Most default algorithms explicitly define some aspect of earnings as a variable that diminishes this probability. Cost of equity c. Sales b. ROC. any system that . Amount of loss 3) FINANCIAL RISK a. Capital Requirements e. Operating Income CHANGES IN CAPITAL STRUCTURE AND STOCK PRICES “Like a monkey throwing darts”. Equity multiplier = Assets / Equity c Change in Net Income / Change in Operating Income d. Project analysis c. ROE. Operating Margin MEASUREMENT 1) COST OF CAPITAL a. Number of equity shares c. Financial leverage ratio b. Operating Leverage b. Measurements and risks are the same. Fixed and variable costs c. d. Probability of default b. WACC e. Stability of net income and operating income 4) ECONOMIC RISK a. TIE (Times Interest Earned) 4) OPERATING RISK a. Economic Profit 2) COST OF BANKRUPTCY a. has been a common description of the random variation that analysts face when picking individual stocks. Cost of debt b. Capital budgeting d. Indeed. Tax Benefits 3) LEVERAGE RATIOS a. Cost of various other components d. Capital Intensity c. Capital market conditions 2) BANKRUPTCY RISK a. Interest expense b. Proportion of debt to equity b.57 Table 3-2 HIERARCHAL RISK RISK / RETURN 1) CAPITAL ALLOCATION a. for example. the one to three year time frame of capital structure analysis does give it some perspective. most firms will reach an optimal capital structure some time during a business cycle and do so when their entire sector is dominating the market. consider the hypothetical XYZ Company. Although some money is made off of “momentum” in some markets. In fact. earnings have been magnified in comparison to the competition. In fact. and then everyone else.58 purports to comprehend some of the mismatches in risk and return throughout the years. By staving off bankruptcy costs and keeping the cost of capital low enough. they have now increased debt to equity for two years in a . Most improvement in stock prices happens concurrently with earnings improvements. Wall Street tries to separate investors into two categories: those who make good judgments well ahead of time. the premium was placed on minimizing the number of shares outstanding. However. including capital structuralism. the same dynamics that proved successful for IBM in the 1960s have been proven successful for Google in the new millennium. “investing after the fact” would be profitable and simple. It is a cycle of matching the opportunities created by the relationship between interest rates and equities. If the housing sector is dominant. Neither company worried about the “clientele effect” of keeping their share prices too high and not splitting them. While the “devil is in the details”. In effect. appears to be a blind attempt to rationalize chaos by giving it meaning. both companies have worked diligently to optimize their respective capital structures. capital structuralism does not try to forecast stock prices. these firms have improved stock price by increasing the flow of earnings toward shareholders . At the top of a business cycle.mostly through capital gains. the paint companies will not be far behind. with the structure of the firms that can most take advantage of it As an example of a very typical scenario. Although historical patterns seem to repeat themselves “with a new twist”. it only attempts to identify an environment conducive to increasing them. the ability to keep earnings expectations high has been the prime ingredient in stock price appreciation. If this were not so. At this point. How does the business cycle suddenly match a firm’s structure with the pattern of available funding? Each firm has an optimal proportion of debt to equity that responds to the level of interest rates and their relationship to equity. the firm needs to regroup and fund new projects with a higher risk . sales jump twenty-five percent. this type of opportunity can only be realized when the firm is moving toward its optimal capital structure. for example. but now they sense that their investment in Chinese furniture is going to be a “cash cow” and pay off. some of the volatility in the market is attributable to business cycle fluctuations. and investors begin ignoring the stock because returns are not as substantial. and they are well past their optimal capital structure with too much debt. may use more financial leverage and fund projects when interest rates are low at the beginning of a recovery. those firms with the greatest operating risk can only perform for a brief amount of time when their particular sector is favored. Dart throwing monkeys notwithstanding. In the third year of their investment cycle. Consumers have money to spend. After two years of diminishing its debt. followed by weak or even negative results. Investors who entered this game early enough receive the spoils of victory. Their earnings have been tepid. At this point. However. Firms who have low operating risk. the firm may be past its optimum in the other direction. but even low-risk firms will have brief periods of wild profitability followed by below average results. the firm’s earnings will begin to far outpace the cost of capital. Increased earnings move the stock up thirty percent and the firm decreases its debt to equity ratio by paying off some of its loans and retaining earnings.59 row. FOUR “POSTULATES” . and its stock price will soar.which may again mean increasing its proportion of debt. the distance the firm needs to travel actually accentuates both the risk and the return. Lower total leverage (operating leverage multiplied by financial leverage) will buffer some of the effect of the business cycle. The performance of such firms may entail an eighty percent rise in the stock price over two or three years. hard and fast rules “break like twigs” and so we call these observations “postulates” with the knowledge that each will be broken at some time or other: • 1) All variables held equal. However. In fact. given a choice. SHARE LIMITATIONS The effect of financial leverage is to reduce the potential number of shares outstanding by funding with debt instead of equity. firms who increase debt by small amounts on a constant basis may have cost over runs or problems with remaining solvent. • 3) The greatest stock returns occur when earnings are accelerating upward while the cost of capital is accelerating downward. Many industries do not have the earnings capacity to do continual internal funding. The firm receives a tradeoff between the potential amount and variability of earnings per share because net income and market price are not diluted by more shares outstanding. • 2) Companies who simultaneously and substantially increase both debt and earnings may pay out in income taxes much more than they have deducted in interest tax savings. It is no coincidence that companies time their debt issues when earnings are lower. The fact that a firm needs to make a choice between the “lesser of two evils” (debt or more shares) is indicative of the inadequacy of internally generated funds. • 4) Small blocks of debt are prohibitively expensive because there are economies of scale when bonds are issued.60 In the world of equities. more earnings tend to decrease the proportion of debt to equity (D / E). This insufficiency is in no way a pejorative. because the firm will pay off some loans and/or increase retention. most firms would fund with sufficient retained earnings as long as dividend growth were adequate. their ability to increase their stock price rests wholeheartedly on managing a capital . These are well-managed and profitable companies who happen to be in an industry that have historically low margins. Alternatively. Firms who garner large loans usually do so with strategic purpose. This type of control over capital structure allows these firms to compete in markets that have players with profit margins five to seven times as much.. However. On the one hand. While other sources of funding may slightly change the risk profile of the entire firm. Firms who are funded with an all-equity structure face a ‘double edged sword”. . these companies are usually small.61 structure that has higher amounts of debt because operating leverage is so low. the fewer shares need to be issued. any quick statistical survey will find that industries with more debt tend to have less stock price volatility .e. ADAPTED MEASUREMENTS This text accentuates the tradeoff between long-term debt and common equity. and when Wall Street requires them to grow. the need to compete and replace a high level of fixed assets requires a constant source of funding which further requires these firms to issue shares of stock. and even limiting retained earnings by paying a steadily growing dividend.periodically . Thus.and fund their projects internally from retained earnings. they are usually very profitable . It considers all other sources of capital to be adjuncts that attempt to lower the cost of capital.which seems to be an anomaly . Those “diamonds in the rough” that are fortunate enough to have a high operating leverage with stable sales can fund all of their needs with internally generated retained earnings. large increases in sales will translate into small increases in operating income. Since debt is unwarranted given the level of operating risk. In fact. these firms will dilute their EPS and market price with more shares outstanding. i. more shares outstanding. the crucial components are the amount of long-term debt and common equity because these require the most expense and obligation. Our definition of capital may include all of stockholders’ equity and all liabilities for the purpose of theoretical illustration: indeed. the more stable and large is their operating income.until one considers that firms with less operating risk can incur more financial risk. On the other hand. The premium is placed on keeping share issues to a minimum. there will be some tradeoff made between stability and the method of financing. Moreover. For example. we use the ratio. short-term debt is viewed as a major element in the need to fund with long-term debt. categorical risk measurement that is removed from corporate efforts to minimize the cost of capital.those paid in an actual exchange of cash. we concentrate on long-term debt to capital (LTD/ CAP) which is more sensitive to change (mathematically) and better elucidates the tradeoff between capital obligations. we exclude preferred stock and interest bearing debt of less than one year’s maturity. but better gauges the risk of these two components. By narrowly defining capital. the figure is more resistant to false interpretation. we eliminate some of the risk adjusting effects of other sources like short-term debt. By eliminating other variables (preferred stock and other interest bearing debt). when we compare companies and make investments. EXPLICIT VERSUS IMPLICIT COSTS We are already familiar with some of the explicit costs of capital structure . we do want indicative measurements to focus on the tradeoff between long-term debt and common equity and so we adapt measurements to fit this urgency. materials and office supplies. instead of alluding to the proportion of debt to equity (D /E). return on equity (ROE). we in no way discount their importance: in fact. and form a cost that is dependent on long-term debt and common equity. Some of these costs include: wages. simply because in our more narrow definition of capital. it may be a higher figure than the actual. Costs are made up of fixed and variable varieties that together make up the total cost when operating income is subtracted from sales. machinery maintenance. . rent. For both the investor and financial management.62 any corporation must itemize every source of capital when it implements a project. return on capital (ROC) more than the ratio. the focus needs to be placed on evaluating the firm through its long-term capital obligations because the success of the company rests on their viability. However. Another example applies to the weighted average cost of capital (WACC). because we desire a strict. However. some type of comparative paradigm. Since capital structure analysis encompasses decisions about choices among alternative actions. each with earnings of one dollar. but the decision to delay and save money has no corresponding entry. In fact. the primary implicit cost is termed an “opportunity” cost. “implicit costs” . . no less important are what are termed. but if my firm under performs its peers in ROE. “the cost of equity”. The “fine” will become part of the balance sheet. For example.(100/110)) x(110) = 10 dollars. However. In effect. since capital structure analysis involves making choices between competing actions. One of the most familiar implicit costs is dilution. will have an implicit cost of (1 . an opportunity cost so to speak. increasing the number of shares to 110. a gain or loss that occurs when we choose one action over another. For example. there is some adjustment made to the stock which is difficult to predict or enumerate. but does not reduce the fundamentals on the balance sheet. Thus. many of its decisions are based on these inherent “implicit“ costs. Another implicit cost stems from delaying actions. any time that an analyst researches industry averages. A comparative return on equity (ROE) of competitors in an industry exhibits many of the characteristics that we term.costs that have an effect on the price of the stock but are difficult to enumerate because they do not represent a physical asset. my opportunity loss is three percent if I choose bonds over stocks. or I can spend too much for a system that will be both less expensive and obsolete in a few years. is being formed in his or her mind. If the XYZ Company has 100 shares of stock. which has an effect on the price of the stock. Thus.63 Interest expense is indeed a prominent explicit cost that needs to be paid regularly. we often form these costs unconsciously. an opportunity cost implies that we are comparing the cost of two different actions. I can choose not to install pollution control equipment and incur a small fine. The action of increasing outstanding equity by ten percent had the net effect of reducing EPS. there is no physical. “up front” cost. if bonds are paying six percent and stocks are paying nine percent. However. If more leverage raises the return that investors require to invest in a firm’s stock. there is some cost to refinancing a loan. several other implicit costs exist when debt is incurred. • 3) The cost of bankruptcy is always implicit unless the firm is actually bankrupt. High inflation has a varying effect on firms with good credit because assets appreciate while loans are paid off in depreciated dollars. • 1) The interest rate needs to be compared to not only competitors’ rates. but to the risk-free rate of the ten year treasury and the averages in the equity markets as well. Even if interest rates decline. debt needs to be curbed and more earnings need to be retained. If it is too high. The cost to some firms is excruciating. The implicit cost of impairing future financial flexibility must be examined. • 2) The ‘real” cost of interest may not only include tax deductibility but inflation as well. if some threshold amount of return is not surpassed? • 5) The implicit cost of possible asset impairment must be examined.64 IMPLICIT COSTS OF DEBT Naturally. THE IMPLICIT COST OF EQUITY . Some restrictive covenants in bond indentures restrict the use of assets and put other restrictions on the actions of management. • 4) The effect on the cost of equity must be determined. interest expense is the explicit cost of debt and tax deductibility is acknowledged when we use it to form the cost of capital. no firm wants to be laden with debt at the top of a market because this is the point of greatest earnings opportunities for most companies. Similarly. can the firm be profitable enough to warrant the increase in debt? Will demand for the stock actually decrease. The firm needs to determine which assets can be secured as collateral in addition to the immeasurable effect on the stock of changes in the probability of default. • 6. any CFO knows the value of a diversified mix of sources of . the company’s capital structure must be viable enough not to depend on a stock issue for its total funding.such as at the end of a business cycle. This latter tactic raises adequate capital. the firm may have trouble covering it with adequate earnings. Thus. The result is often a large adjustment downward because performance does not meet the over-hyped expectations. If a firm is “maxed” out on its credit and is not earning enough to raise capital through retention. it must meet its funding needs through issuing stock. Alternatively. a firm can issue stock when the price is high. the “cost of equity” is almost entirely implicit and shifts in value from year to year depending on the rate investors will require. When the cost of equity (the required rate of return) is exceptionally high. it will need to issue many shares to achieve it. if the firm has a target level of capital requirements. receiving the most capital per share issued. The final strategy is most preferred by insiders and large investors: issue stock when the price is low enough to appreciate substantially.65 The cost of equity is the rate of return that investors will require to invest in a firm’s stock. this rate naturally rises. When the market has not factored in expected earnings from projects that it knows nothing about. the price received will be diminished because investors will not find the stock attractive. It is used synonymously with the term “required rate of return” and is referred to as an “opportunity cost” because it compares the rate of return of firms with similar risk in a least squares type correlation. Another implicit cost arises in the timing of a stock issue. However. In a “bull” market. there is a threshold point where funding should be delayed or avoided because it diminishes the market price of the stock too drastically. the required rate is applied to all retained earnings (not just the current year’s) as well as all outstanding stock that has been issued by the company. it declines. while in a “bear” market. the only up front costs will be “flotation costs” which are a payment or a percentage of the proceeds to the underwriting firm. the risk and cost of equity is low. but may be undertaken when the cost of equity is very high . When equity is built through retained earnings. Thus. However. When equity is issued. Raising large amounts of capital with an equity issue has many repercussions. however (like pharmaceuticals). debt and equity in a particular direction.Investors will not demand the stock of a company that under performs its peers. The only time it seems justified is when a major merger occurs in a favorable economic environment. Naturally. In fact.The firm must consider the effect on both EPS and market price. will have a long lag time but return more once the pay off occurs. we can put them into one of three categories: • 1) The implicit cost of dilution . It is not a frequent occurrence. Any time that a firm depends too much on a single source. he or she will understand it as a shift in capital structure where accelerated earnings begin to propel the combinations of assets. as well as future dividend obligations. there is more risk of a higher cost of capital. that scenario occurs to any investor who is well diversified and stays in the market long enough. • 2) The implicit cost of the “required rate of return” . Inevitably. To summarize the implicit costs of equity.66 funding. Some industries. THE MOMENT OF TRUTH Ultimately. • 3) The implicit cost of timing. despite its volatile changes. The uncertainty is derived from the timing of that prospect and whether it will occur at all. More often than not. most investors want to time the market so that they are in the early stages of a large pay off. more than one investor has left a firm only to find that more patience would have led to profitability. it is the prerogative of management to keep lag time between investment and pay off to a minimum. A firm that does so will have a very high “opportunity cost” and be unable to cover it with enough earnings. Thus equity issues should be relegated to “executive currency” rather than exist as a major source of funding. If the student/investor observes the corporate side of any investment. . investment screens are designed to fail because the market will change and make sure that they do. the tide may start shifting to more equity financing. • Look for a shift to a smaller proportion of long-term debt to capital as well as a shift to a lower financial leverage ratio (EBIT / (EBIT . • Although financial statements occur “after the fact”. The risk should be accompanied by some confirmation from analysts that earnings are going to improve. For example. At first. when earnings increase. expecting large gains from the housing sector at the top of the market may be wishful thinking. Earnings accelerate when a sector is receiving high demand at the same time that its capital costs are low.Interest Expense)). the two ratios may be “out of sync”. • Know the business cycle. When a company’s own executives are buying stock and are doing so from a leveraged position. the return can be greater if the firm knows how to use debt strategically. (Back to Table of Contents) .there are a few signals that are related to both capital structure and early investment success.67 But . % ∆ Sales / % ∆ Capital or % ∆ Sales / % ∆ Long-term debt • Investing in a company who is increasing its proportion of long-term debt to capital is riskier than investing in a company who is building equity. look for quarterly improvement in capital turnover. • Executive Trades. the financial leverage ratio will begin to drop in harmony with the other ratio. However. 68 APPENDIX: THE NET OPERATING INCOME APPROACH TO STOCK VALUATION Student/investors are advised to take the most conservative approach to valuing a stock. Academicians developed two methods to contrast opposing views. In mathematical notation.(Interest Rate)(Bonds)) / Cost of Capital. the net income approach to valuing a firm’s stock subtracts interest expense from operating income and then divides this difference by the cost of capital.operating income divided by the cost of capital.Bonds. 50 % and then 100 %. consider a firm that is capitalized at $10000 with D/E of 0 %. On the other hand. the net operating income approach determines the value of the company as the ratio of capitalized operating income: that is . The second approach was the “net income” method. reality dictates that returns must be evaluated in the domain of risk. or different rates of interest between firms and individuals. In mathematical notation it is: (X . where X is equal to operating income. the value of the stock is : (X / Cost of Capital) . . bankruptcy costs. and the relationship between its interest rate and the cost of capital. The interest rate on bonds is 5 %. contributing no additional value to the firm. As an example of the two approaches. which proclaimed that a firm’s value was an extension of its degree of leverage. While both methods value a company as the sum of its bonds and its stock. It then takes this value of the stock and adds the value of its bonds to determine the value of the company. Both approaches were developed in a hypothetical world of “perfect competition” .no taxes. It then subtracts the value of the firm’s bonds to determine the value of its stock. While more leverage may escalate the price of a firm’s stock in a world without bankruptcy. The first method was called the “net operating income” method and postulated that the extra return from leverage was balanced out by the extra risk. Operating income is $1000 and is capitalized at 10 %. the net operating income . However. all increases in EPS are immediately transferred into the price of the stock.69 Table 3-3 Net Operating Income Method D / E Percent Net Operating income Capitalization Rate Total Market Value of the Company Market Value of Bonds Market Value of the Stock Table 3-4 Net Income Method D / E Percent Net Operating Income Interest Expense at 5 % on Bonds Net Income Capitalization rate Market Value of the Stock Market Value of Bonds Total market Value of the Company 0 1000 10% 10000 50 1000 10% 10000 100 1000 10% 10000 0 10000 5000 5000 10000 0 0 1000 0 1000 10% 10000 0 10000 50 1000 250 750 10% 7500 5000 12500 100 1000 500 500 10% 5000 10000 15000 By capitalizing operating income and deducting interest. the net income approach adds a substantial amount to both the market value of the company and its stock. the net income approach assumes that there is no risk. Alternatively. and the amount of bonds is subtracted to determine the stock price. there would be no benefit from the proportion of debt to equity in the capital structure. the value of the company is calculated first. there is no difference between the two methods. Miller/Modigliani argued that the only correct approach was the net operating income method.70 approach assumes that the risk of leverage perfectly balances the potential effect on EPS and that risk and return cancel each other out. The student/investor will observe that if the interest rate is the same as the cost of capital. The same amount of earnings would flow to the shareholders regardless of how the firm was funded (Back to Table of Contents) . In the net income method. and the amount of bonds is added to this figure to determine the value of the company. Thus in a perfectly competitive economic environment where individuals and firms can lend at the same rate ( no bankruptcy). no gain would be garnered from leverage because the interest rate would always approach the capitalization rate. the value of the stock is calculated first. In the net operating income approach. indeed the net income method rewards management for keeping both rates as low as possible. Their Proposition I argued that in a world without taxes. we cannot neglect the importance of short-term debt and all current liabilities. buys in volume and never takes discounts at all. and cost has a tendency to adjust to risk rather than the opposite. next to proportional capital allocation itself. in fact. however. debt/assets?” While it is true that some companies will fund long-term fixed assets with revolving credit from a bank.71 4 THE COST OF DEBT While we define capital as long-term debt and equity. one company may take out short-term loans to take advantage of trade credit discounts because the interest on the loan is less than the discount. Part of our solution is to find an interface between cost and risk that reconciles temporary inequalities. We propose that interface in this chapter. Herein lies a major problem in capital structure. it may typically extend payment periods well past due and the vendor tacitly accepts this behavior because the account is so profitable. the difference between long-term and short-term credit is substantial enough to cause the eccentric pricing of risk. If the classic inverted yield curve is a sure sign of imminent economic trouble. they are not the same. and define it as the cost of bankruptcy. “Why not define capital structure in the traditional mode of the proportion. Despite creating productive synergy. Another company may be much larger. There are. the pricing and quantification of risk becomes obscured. THE PROBLEM OF SHORT-TERM CREDIT The significance of short-term debt especially begs the question. Although most industries go through some sector and market volatility. it is working capital management that guides a company through the trough of business cycles and grants it the flexibility to take advantage of the peaks. several characteristics that make each type of debt unique. Which company has the lower risk? The answer is ambiguous because we are comparing an . While the two are inseparable. the strategic use of “working capital” forms the backbone of sustainable profitability. it accentuates the problem of matching cost with risk. In fact. For example. The first company is managing its costs. For example. and as a signal for a potential increase in earnings. a firm may want to wait out the uncertainty in anticipation of lower rates.72 opportunity cost of debt to a business condition. which almost perfectly correlates price (interest). However. While short-term debt may have no effect on market risk at some points in the business cycle. depending on the timing. an inventory build up can either be anticipatory of sales increases or a signal that demand is too low. but the same condition arises during a shortage of cash. as a substitute for long-term debt at the end of a cycle Rather than commit to the higher interest payments of a large debt issue. it is tempting for a firm to continually fund long-term projects with shortterm loans to keep capital costs at a minimum. Shortterm credit encourages this financial flexibility. as more purchases are made. short-term credit is less amenable to risk analysis . accounts payable would display the same large balance regardless of the boom-bust condition. Since short-term debt is normally less expensive than long-term debt. and risk. there is a non-linear relationship between the cost of short-term debt and its risk. The deceptive quality of short-term debt encourages the confusion of profitability with insolvency. Unlike long-term debt. current liabilities often rise dramatically. In essence. As both alternative financing. short-term debt gets priced in terms of the risk to the creditor with less reference to the long-term viability of the company. In fact. at other times it may be a determining factor .on a shareholder level. it must be observed as an adjunct. Although some academic literature argues that short-term debt is only a factor in total risk and does not contribute significantly to market risk.for example. Such a strategy is fraught with two major . while the second company is taking advantage of its absolute size and power. While long-term debt and equity can be gauged in terms of cost and obligation. a rise in short-term credit can help optimize long-term debt in some situations. the cost savings from “floating’ a non-payment may be just as great as the fastidious management of credit. a “rogue” financial executive can finance with short-term debt. if a firm chooses to finance inventory with a vendor. Moreover. investors who need a precise break-down of interest expense attributable to long-term debt must read the financial notes in a firm’s 10K or annual statements. ski lodges and golf courses for example . and the company may not have the cash flow at the time it is needed. firms will “net out” their interest expense with interest gained in selling securities. interest expense can easily be subsumed into “cost of goods sold”.farmers. For example. and income volatility. Each maturing issue will be matched with the date of maturity and an interest rate. INTEREST EXPENSE INEQUALITIES Another anomaly occurs when interest expense at the end of the year does not reflect the activity in the short-term credit market during the year. a sample breakdown is as follows: . and end up funding long-term projects with long-term loans. The result is an interest expense that does not reflect the greater amount of risk during the period. Interest expense may escalate during the second and third quarters but loans get paid off just in time for the annual income statement. and a short-term loan exposes the firm to both a potential rise in rates. Thus. To circumvent these risks.would be more amenable to short-term credit. firms with more seasonal demand schedules .73 risks: 1) The loans need to be frequently renewed. The payoff would be more certain and based on historical repetition. and the higher risk will not be reflected on the income statement. in a company with an interest expense of 9 million and 100 million in long-term debt. exposing the firm to the risk of default. 2) Interest rates vary. but time the cash-flows with “luck” and pay off the loans on time. such obfuscation is sometimes a red herring: transparency can be confounded with both euphemisms and extensive circumvention. In fact. Very little of this activity will be reflected in annual reports and it appears that the executive successfully funded long-term projects and lowered the cost of capital simultaneously. Other times. On the other hand. most companies will match the cash-flows from operations to the maturity of their debt. 933 (0. To conclude the treatment of short-term debt within the capital structure: • 1) Capital structure as defined in this text is a combination of long-term debt and stockholders’ equity. As short-term credit has a major effect on these components.105)= 8.5 % 7.33 33. • 2) Without a component breakdown of interest expense. it may be gearing up for a successful period of revenue generation by building up trade credit.33 MATURITY 2009 2011 2013 INTEREST RATE 5. • 3) Short-term credit has a non-linear relationship with risk.74 Table 4-1 DEBT AMOUNT 33.082 8. while debt is piling up on the “accounts payable” side of the ledger. At the same time. RETURN AND THE SIGNIFICANCE OF SHORT-TERM CREDIT The nature of the risk-return conflict for short-term credit is the timing of cashflows. it approaches what is known as “technical insolvency” such that it cannot meet short-term obligations.075) + 33.33(0.5 % 10. RISK.917933) was attributable to short-term debt.33 (MILLION) 33. it is an adjunct force to the risk of capital.33 (0.055) + 33. engaging vendors through purchases and creating inventory. there may be some computational error if all interest is attributed to long-term debt.33(0.5 The amount of interest expense attributable to long-term debt is the amount of each maturity multiplied by its respective interest rate and then summed together: 33. A balance sheet will show that current sales have not produced adequate accounts receivable or cash for that matter. any attempts at minimization of that cost would be skewed.082 million in interest expense is attributable to long-term debt and just $917. As a firm’s current ratio (current assets/current liabilities) declines. . If it were included in the cost of capital. 75 This is often the time when insiders begin accumulating stock because it may be depressed. An examination of two liquidity ratios and two cash-flow equations will reveal the ambivalent analytical nature of short-term credit. it is rarely indicative of stock movement. The second equation is the capital requirements equation. the need for outside capital diminishes. A higher ratio simply means that short-term obligations can more easily be paid. free cash-flow and the capital requirements equation. Both the assets and liabilities variables are those that increase spontaneously with sales which includes mostly the current type. These classic measurements of solvency do not always indicate profitability. cash-flow decreases and that more current liabilities increase cash-flow. Trade credit. which gives a rough estimate of outside capital requirements in line with the sales forecast. . for example is an internally generated source of funding. but is ready to soar when revenues improve. The average investor perceives a crisis rather than an opportunity and fails to invest. current assets inventories / current liabilities. As Liabilities / Sales increase. The reason is found in two other equations.(Liabilities / Sales)( ∆ Sales) Retained Earnings.Capital Expenditures . The basic unlevered free cash-flow equation is: (EBIT)(1-T) + Depreciation and Amortization . It is: (Assets / Sales)( ∆ Sales) .∆ NWA Table 4-2 EBIT T ∆ NWA Earnings Before Interest and Taxes Tax Rate The Change in Net Working Capital (defined as the change in current assets current liabilities) The student/investor should notice that as NWA increases. however. We have already observed the “current ratio” which also has a modification called the “acid test” or “quick” ratio. the potential return is far greater than the risk of insolvency. Any time that short-term credit can be substituted for capital without undue risk. When the ratio. for example). Depending on the leverage situation.76 Thus. the rise in current liabilities supports investment in long-term debt and equity. THE CORPORATE COST OF DEBT Banks and ratings agencies will determine a corporation’s potential default by analyzing its various leverage ratios and its future prospects. Assets/Capital increases will also contribute to an increase in the return on capital (ROC). A more positive outlook for the issue of long-term debt occurs when this ratio is increasing because business activity is stepped up and the turnover time for the larger investment may be shorter. a return on assets (ROA) is formed by multiplying profit margin (Net Income/ Sales) by asset turnover (Sales / Assets). The primary concern is that operating income is accelerating enough to lower the risk of default. The standard method for evaluating short-term debt is in the context of cash-flow. Moreover. the return from revenues will most likely be greater than the risk of increasing short-term liabilities. as it is one of the three major components of that ratio. Essentially. The difference between numerator and denominator is current liabilities. the ratio. when short-term credit increases and the outlook for earnings is positive. Assets/Capital . A risk premium will be . because those components are purchased at a higher cost. follows a loose chain of logic. the firm moves toward an optimal capital structure. “Assets / Capital” rises at the same time that cash-flow increases as a percentage (ten percent in the last period and twenty percent in this period . First. The resulting ROA figure is then multiplied by Assets / Capital to form a return on capital (ROC). an increase in current liabilities may show the use of alternative sources of capital to keep the cost of capital at a minimum. . When economic conditions dictate an interest rate change for an entire industry. a new prime rate for example. and also on the company’s prior interest expense obligations. A new rate will change the market price of a firm’s debt. “The Real Cost of Capital and What the Investor Needs to Know”. This ” nominal “ cost of debt fails to gauge immediate changes in the cost of capital.77 attached to an appropriate risk-free rate (a ten year treasury yield for example) and this will be the company’s interest rate input for its cost of debt. For the investor. The cost of debt is an opportunity cost and is compared to both competitor’s rates and the rate the firm would actually pay if it chose to incur new debt. Such expediency will derive an interest expense that equates a firm’s book value of its debt with its market value. Since incurring debt implies a tax deduction. the new rate of interest is multiplied by the reciprocal of the effective tax rate (1 – tax rate) and this figure is further multiplied by the market price of the firm’s debt. Thus. but can be used as proxy in other relational values such as the financial leverage ratio and the TIE (times interest earned). A ballpark estimate can be formed by equating each interest rate that a firm pays on its existing debt with the corresponding proportional maturity in the firm’s debt structure and then forming a weighted aggregate. anytime a new loan is negotiated at a different rate. a company’s cost of debt changes. those firms who do not incur new debt will still have a change in their respective costs of debt. it is simply overly “research intensive” to configure a firm’s new cost of debt each time it occurs. The theoretical underpinnings of this process are discussed in the appendix entitled. depending on the amount of decrease or increase in the rate. The computational and informational requirements are not justified by the performance gains in accurately measuring risk. However. is that earnings per share may increase in variability. each maturity level of debt has the potential of costing a different amount in different years because of a changing effective tax rate. For example. in order to functionally calculate the optimal capital structure. and then multiplies by the corresponding interest rate. deferred taxes and tax gains or losses becomes paramount. the nominal cost of debt is the summation of each proportional maturity multiplied by its corresponding interest rate. The effective tax rate applies because it is the rate at which current deductions are considered. several “tax strategies” emerge when a firm can use leverage to its advantage. which allows companies to grow at a faster rate than they would if financed solely by equity. Interest expense is fully tax deductible. Those companies with more volatile earnings. In effect. as pointed out in the chapter on leverage. and multiplied again by the current effective tax rate. 50 of 8 % debt and 20 of 6 % debt with an effective tax rate of 30 %. it is this cost which comprises the debt component in valuation models and capital budgeting.78 THE NOMINAL COST OF DEBT AND THE COST OF BANKRUPTCY When calculating the cost of capital. the analyst uses the real cost of debt which is composed of the interest rate on the debt and a tax deduction. Consequently. where i = the interest rate and T = the effective tax rate. the mix of cash-flow. The main attraction in using debt to finance capital needs is its tax deductibility. and so . Total debt is 100 million. i (1-t). simply cannot compete on this basis. A precise enumeration of the cost of debt may contain ten or fifteen separate maturities but is quite simple to calculate in a spreadsheet. an average interest rate can be obtained. a firm has 30 (million) of 7 % debt. The cost of debt is simply. If one calculates the proportion of each maturity as a percentage of total debt. the government gives an advantage to firms with lower albeit steadier cash-flows by subsidizing growth through tax breaks. who would be in greater danger of default. Naturally. Thus. The tradeoff. we need to incorporate several other costs and inherent risks into the model and we accomplish this by forming a comprehensive “cost of bankruptcy”. 3.2 respectively. creditors will simply “turn off the spigot” and .3(. Alternatively.3 %.08) + 0.8 + 0. banks and underwriters will charge greater interest rates to compensate for greater risk.11 million. Why does the government make the use of debt tax deductible? There are many sides to this controversial question. While the typical shareholder does not normally need the type of precision that is used for capital budgeting.07) + 0. Consider the result of large cuts in the federal funds rate. While it seems to give an unfair advantage to companies who can most afford debt both financially and structurally. In fact.08) + (20)(. firms that fund with equity may reap more benefits in good economic times because income flows directly to the shareholders. If this “symbiotic” relationship seems ambiguous.47 + 2. the cost of debt is 5.07) + (50)(.11 percent. I could also go through the entire calculation of (30)(.11.7)(. more information helps form better decisions. The cost of debt is both a component of the cost of capital and a risk-adjusted precursor to an optimal proportion of debt to equity.5(. When this figure is multiplied by the 100 million in total debt. even as this “cost” is circumvented with greater tax deductibility. At extreme levels of debt. A steady income that is amenable to debt financing encourages steady revenues (for the government too!) and steady employment.5. I now have the choice of multiplying this rate by (1tax rate) to achieve a rate of . the average firm will do more debt financing but not enough to radically change capital structure. The average interest rate calculation is 0.0511 or 5. and 0. as the risk of default becomes greater.. it is because “risk” and “cost” are not always compatible.7)(.06) = 7.79 the proportions are 0. as interest rates decline. The ability to use debt financing is derived from business risk and implicit operating leverage. 0. no level of interest rate will alter the ability to pay interest in a timely manner.7)(.06) = 1. Also. many other advantages are incurred when equity financing is implemented instead: consider the lack of interest rate risk and credit crunches that all-equity funded firms can embrace.2(.84 = 5. Perhaps the biggest reason is to encourage small businesses and those who have small cash-flows to stay solvent. The cost of equity is increased by the use of leverage.if tallied by cost alone This extreme “corner solution” was proposed under the assumption of a zero probability of bankruptcy. most investors are at least familiar with the FICO score . THE PROBABILITY OF DEFAULT While teams of actuaries set loss liabilities in the insurance industry. there exists a “risk premium” that compensates shareholders for additional risk of uncertain income. Therein lies the problem of using the cost of debt to determine an optimal capital structure. which further emphasizes the significance of default. An optimal solution to finding the right capital proportion exists only when the cost of debt is reconciled with the cost of bankruptcy.the credit rating which never seems factually accurate. The word “safely” is a connotative term. tax deductibility will ensure that the cost of equity is greater. but inroads into a solution can be made when we look for common factors in the calculation. each firm is ultimately structured in a unique way with different patterns of risk and return. the correct proportion of debt will determine the optimal proportion of equity. In essence.80 interest rates will stabilize at a high level. In effect. Two of those factors are: 1) The probability of default and 2) The amount of loss. which needs objectification. THE COST OF BANKRUPTCY For a firm who uses leverage. the key to an optimal capital structure is to use as much debt as safely possible. dependent on “niche” or specialty. but haunts everyone who ever got turned down for a loan Therein lies the . Even at the aforementioned extreme levels of interest. and since debt is less expensive than equity. To form a “generic” cost of bankruptcy is to assume a daunting task. Even within an industry. the Miller Modigliani proposition II becomes operative because the optimal proportion of debt in a taxed economy will be one hundred percent . and it is no accident that companies spend millions in risk management to gain a precise definition. the cost of capital will be minimized at any given level of operating income. but the cost of that leverage never surpasses it. types of variables relating to the economy are more volatile. The nominal cost is affected because better bond ratings lead to favorable interest rates. In effect. most rating systems are very specific and apply to a particular industry or economic outlook. they are robust for the very reason that commercial banks cannot use them to set interest rates: they offer a general purpose risk analysis without the specificity needed to gauge a particular economic environment. most analysts settle for the proverbial “ball park” figure. but precise prediction of near-term probabilities remains elusive. . but lack the ability to transition to a changed economy. as complicated as default analysis has become. there is no such schedule available for periods of hyperinflation or various speculative “bubbles”.81 problem. the academic generics have eclipsed them in accuracy over certain periods. Law suits and fires have a historic frequency that can be extrapolated into a loss schedule. probit and multiple regression analysis are valuable tools. for the purposes of capital structure. the generic rating systems promulgated by various academics have held up over many years. This rating is of primary importance to both the rated company and to investors. Both Moody’s and Standard and Poor’s have been rating long-term debt issues for many years. Economic conditions change and using the same methodology with different inputs may lead to inaccurate probabilities. Logit. the extremes of solvency and default are much more predictable than the chance of transitioning to a higher risk category because the latter is dependent on shifting economic variables. bond ratings set the stage for capital structure because they help determine the cost of debt in both nominal and default forms. Tested against commercial algorithms. In fact. While there are many commercial algorithms available. However.the probability of default. In fact. Although specific conditions like house fires or homeownerliability law suits can be determined by probability distributions. while the default form is affected because those ratings are configured in terms of solvency . In fact. For the student/investor. the modified Z Score is 0.4(X2) + 3.82 Edward Altman’s “Z Score” is perhaps the best known of these. a summation is performed on asset-related components and then a mathematical procedure is performed on the sum. 1. it is still 94 % accurate in predicting bankruptcies one year away.71(X1) + 0. while a spreadsheet determines the probability. the procedure involves determining whether the sum is large enough or increasing.999(X5) All entries are made in decimal form. the analyst will merely plug in decimal ratios into a weighted coefficient equation and tally a score.2(X1) + 1. .6(X4) + 0.847(X2) + 3.1(X3 + 0.998(X5) This Z Score will produce a figure from zero to three or more. it offers a common ground with other algorithms because it bases its analysis on an array of performance variables as a percentage of assets.81 is a troubled firm. In that case. logit operations are performed where the logarithm of “the odds” are obtained. In the case of Altman’s Z Score.3(X3) + 0.80 % accurate in predicting bankruptcies two years away. Progress from quarter to quarter can be discerned by forming a running total on a spreadsheet and any large increases may signal stock price appreciation. and one needs to form a Z score from book values. and about 72 . may be an integer. and over three is considered “safe”. X5. but the last figure. In most other algorithms. The product of multiple discriminant analysis. Usually. Table 4-3 ALTMAN'S Z SCORE (X1) = Working Capital / Total Assets (X2) = Retained Earnings / Total Assets (X3) = Earnings Before Interest And Taxes / Total Assets (X4) = Market Value Of Equity / Liabilities (X5) = Sales / Total Assets Sometimes the market will be inflated.42(X4) + 0. The algorithm and its components are as follows: 1. where zero is bankrupt. For many years.83 COMMERCIAL RATING SYSTEMS The Moody’s and Standard and Poor’s systems are updated regularly and use probit analysis among other techniques to create ratings. down to a “D” rating which indicates default probability of one hundred percent. Standard and Poor’s used a combination of probability algorithms and qualitative judgment to rate companies . we see many similar type variables as in Altman’s Z score: operating income must cover immediate obligations. the six most significant variables in probit analysis are: • • • • • • 1) ( EBIT + 1/3 Rent) / (Interest Expense + 1/3 Rent + (Preferred Dividends/0. In most economies. Investors are thoroughly familiar with the lettered rating system of “AAA” standing for high investment grade and a small probability of default. According to Moody’s. The following list entails some of the variables used in this judgment: . working capital) and performance must be matched against assets.which is entirely legitimate because many economic variables can only be measured after their qualitative analogs occur.65)) 2) Adjusted Debt / Adjusted Book Equity 3) Cash and Equivalents / Total Assets 4) Five Year Revenue Volatility 5) Retained Earnings / Adjusted Debt 6) Asset Growth Again. there must be liquidity (cash. sales must be adequate. they are accurate but require a subscription from users. 84 Table 4-4 VARIABLES USED IN DETERMINING DEFAULT PROBABILITY 1) Debt / Asset 2) Times Interest Earned (TIE) 3) Times Fixed Charges Covered 4) Current Ratio 5) "Quick Ratio" 6) Mortgage Provisions (Collateral) 7) Call Provisions On Bonds 8) Other Restrictive Covenants 9) Sinking Fund Provision (Account For Retirement Of Debt) 10) Regulatory Climate 11) Anti.Diversification It is quite difficult to endow a variable like “labor relations” with an accurate probability that enhances credit prediction. Large companies are typically encouraged to file chapter eleven and reorganize rather than liquidate. Vendor Stability 15) Labor Relations 16) Percentage Breakdown Of Sales/Customer . 3) the currency of data. they have assets and obligations . 2) the effect of the economy on the probability of default. Any determination of an optimal capital structure while using default probabilities could not realistically delineate between 30 percent debt and 31 percent debt but should be able to alert the investor to excessive levels . Diversification 13) Environmental Factors 14) Resource Availability.Trust Legislation 12) Overseas Operations. The student/investor should recognize that setting an optimal capital structure is a probabilistic venture subject to: 1) the technical accuracy of the algorithm.such as five or more percent above or below the target. which is why any attempt at precision must be updated frequently. TYPES OF BANKRUPTCY One problem that occurs with generic application of bankruptcy algorithms is the lack of uniformity among legal dispositions. bondholders are entitled to a series of fixed payment and return of principle and nothing else. Indeed. creditors and even the local community have a stake in seeing the operation continue in a different capacity . the type of industry is significant because the asset structure will determine whether it is marketable. one firm requires managerial foresight. Since the decision to reorganize or liquidate rests with trustees. which is the present value of future earnings.100 percent recoverable in a reorganization. THE AMOUNT OF LOSS A model of default will determine the amount creditors receive by reconciling the amount of liabilities with the asset structure. On the other hand. they are 95 . shareholders claim the intrinsic value of the company. any time a firm incurs long-term debt. most models will not differentiate between types of bankruptcy because of the legal ambiguities involved. the stockholders forfeit a claim on the firm’s assets. Firms who have more intellectual property and yet have a good historic track record of meeting obligations.under different circumstances. claims are set by the “doctrine of fairness” and are recognized by legal and contractual priority. the other prospers on the market price of an asset. In essence. they need to analyze two basic issues: 1) Is the firm worth more as an ongoing enterprise or as sold assets?.85 that are great enough to continue operations. Although the science of probability can factor in qualitative decisions.either downsized. or broken up. any future earnings gained from the investment in debt accrues to shareholders. . Government entities. Thus. Obviously. bought out. and 2) Can earnings cover fixed charges in the future?. will be more of a reorganization target than a firm who trades in an appreciating commodity. part of the tension between bondholders and shareholders stems from the establishment of a “risk premium” which is derived from contractual priority. Shareholders usually receive nothing. while recovery of assets may be as low as 20 percent in a liquidation. Bondholders have a claim on an extrinsic value. In bankruptcy. from the extrinsic value of salable assets that creditors possess. If we state that a stock is worthless upon bankruptcy. the assets do not produce present or future income and the stock is worthless. However. Note that shareholders have claims during ongoing operations. During normal operations.interest payments and assets that can be easily sold. The controversy entails the definition and enumeration of bankruptcy costs. and the difference of the present value of bankruptcy costs. once bankruptcy commences. the product of the tax rate and the amount of bonds. Each case is unique based on legal type. Thus. intrinsic value is based on the prospect of future earnings. However. That is. Creditors have a claim to part of those extrinsic assets while shareholders lay claim to what is left over. and even adapt it to a modified version of Miller/Modigliani’s hypothesis: V(l) = V(u) + TB . the true market value assumes some monetary multiple above them. the confluence is no longer viable. we are implying that the cost of bankruptcy is the entire loss of market value. each is a separate entity. In effect. There are no set formulas or series of calculations that enable the analyst to determine a generic amount of loss in a bankruptcy.86 which may be marketable assets or prioritized earnings flow . the value of a levered firm is equal to the summation of the value of an unlevered firm. but will forfeit the greater percentage of them during bankruptcy.in other words . liquidity. .PV. structure of debt and absolute size. Unless the firm is reorganized. we are separating the intrinsic claims on earnings flow that the shareholders possess. and yet implicit in that value are the prioritized claims of creditors. financial theorists are well aware of the significance of potential bankruptcy to capital structure. the “extrinsic” assets provide the framework for earnings. it will be helpful to put bankruptcy costs into perspective by relating the value of various assets with the market value of equity. Both the priority of claims and the type of assets determine the amount of loss. When we consider these claims. Accrued Liabilities d Notes payable e. and shareholders may lay claim to some of them. Salvage INTANGIBLE a. Marketable b. Once bankruptcy proceedings begin. classified as “intangible assets”. priority is given to creditors. Assets are made up of tangible assets and intangible assets. Goodwill Dependent on Prior Claims UNCLAIMED Figure 4-1 Intrinsic value of the stock (ongoing operations) Book Value Market Value Tangible Assets + Intangible Amount of Loss Extrinsic Value True intrinsic Value at Bankruptcy The intrinsic value of the company is equal to market value as long as the company is solvent. can have . and shareholders forfeit their clams on assets. intellectual property like patents. Trademarks c. Patents b. Senior Mortgage Debt b.By prior agreement some creditors may have claims on marketable patents Common Stockholders TYPES a.87 Table 4-5 ASSET CLASS COLLATERALIZED PRIORITY a. Debentures NONE . Tangible assets may or may not have a market value. Administrative Fees c. Analogously. Intangible Assets (including goodwill) . Next. the potential to raise the probability of default was greater. to view the amount of loss as the entire market value of shares is improper. As long as operations were ongoing. Once the contract was near termination. However. agreed to pay a risk premium when they decided to incur long-term debt in an effort to boost earnings and share price. 2) not only were tangible assets given up to creditors. in effect. The reality of most bankruptcy proceedings. Priority of claims goes to senior mortgage holders.88 market value and be claimed by either creditors or shareholders. pension funds and even auctioneers will eliminate the possibility of such compensation. dictates that claims by employees. which is primarily the excess market value paid for acquisitions. which is (loosely): (Total Assets . shareholders “rented’ a certain portion of tangible assets to produce future income. To the shareholder. This premium was a double edged sword of unrealized potential: 1) debt might increase the volatility of EPS before it boosts share price. “tangible book value per share”. the majority of intangibles are made up by the designated asset class “goodwill”. we form a ratio . unless it is used as a bargaining tool for administrators. Subordinated debt holders may receive any where from 20 to 70 % depending on the claim and marketability of assets. However. At bankruptcy. and. To construct a working model of this concept. we need the calculation for tangible book value per share. trustees can point to a company’s historic responsibility to creditors and the industry. Any unclaimed assets belong to the common stockholders and those are designated by the ratio.Unamortized Debt) / (Number of Shares Outstanding). however. it is virtually worthless. It is the amount a shareholder would theoretically receive upon liquidation. Stockholders. those assets reverted to the creditors. and some subordinated debt holders receive nothing at all. the amount of assets a creditor receives is impertinent. followed by administrators who may get 20-25 % of monetary gains from asset sales. he or she is left with a worthless investment. The leeway for claims is quite large. our structurally dependent model can be used in capital structure development. Essentially.which gauges the potential to cover interest expenses . the amount of loss will be some function of these three variables and it is the relationship among the three that can determine the proper amount of debt to employ. A linear format would allow firms to reap the tax benefits of more debt without violating the constraints of bankruptcy.a cost of bankruptcy can be estimated By multiplying the two constructs: (Probability of Default) x (Amount of Loss) we can establish a comparative proxy. The entire expression is as follows: Amount of Shareholder Loss = (1-(Tang. competitors and the equity market. perhaps even separating it from its true asset value. The amount of loss is probabilistic. When these variables interact with the probability of default . book. Val.per sh / Mkt. which is determined by the product of outstanding shares and price per share. penalizing firms at an increasing rate for taking on too much debt. we multiply this value by the market value of the company. Lastly. Lastly. Any optimization model must allow for the primacy of all-equity funding if the asset structure is not amenable to financing with debt. per sh)) x (# shares x Mkt Val. For example. the probability of default needs to form an exponential curve. depending on interaction with the economy. finding a default probability calculation that is also precise at the extremes would be difficult. the cost of bankruptcy model has some serious shortcomings.(Tangible Book Value per share / Price per share)). (1 . a speculative run up in the stock would affect the amount of loss. Since most algorithms are formed from probability distributions that emphasize central tendency.89 between tangible book value per share and the market price of the stock and subtract this fraction from “1”. per sh) Implicit in this loss is all market value above a firm’s extrinsic value net of liabilities. Despite its utility. Val. Although the true cost of bankruptcy must be configured for each unique entity. Secondly. which rephrases the term. the amount of shares . “ tangible book value”. The maximum of a function such as (TB PV) occurs when ∆ TB = ∆ PV. what we have . a theoretically viable model would allow that number to be dependent on necessary funding . The optimal amount of debt will optimize equity as well but it must first be reconciled with the production of income. note the following “paradox” that occurs when a stock has been “beaten down”: Table 4-6 THRIVING COMPANY # SHARES = 100 TANGIBLE BOOK VALUE / SH = 4 MARKET PRICE / SH = 20 LOSS = (1-4/20) x (100 x 20) = 1600 NEAR BANKRUPT COMPANY # SHARES = 100 TANGIBLE BOOK VALUE / SH = 4 MARKET PRICE / SH = 7 LOSS = (1 .PV which states that the value of a leveraged firm is equal to the value of an unlevered firm plus the product of its tax rate and bonds and minus the present value of its cost of bankruptcy. However. $7) allowed the amount of loss to be much less for the near bankrupt company.PV).the one element that would certainly push the probability of default upward. debt and capital already in the capital structure. THE ROLE OF DEBT IN CAPITAL STRUCTURE OPTIMIZATION Debt is the fulcrum between higher tax benefits on the one hand and the chance of loss on the other. Thus the distance in value between the unlevered firm and the leveraged firm is maximized when the distance between TB and PV is also at a maximum (V(l) – V(u)) = (TB . When the marginal tax benefits equal the marginal bankruptcy cost. To observe this discrepancy.90 outstanding is a variable that is determined outside of the model. that smaller number may open the door for more debt financing in an optimization model . Mathematically. If more debt violates the parameters set forth in the cost of bankruptcy.PV will be a maximum. To see how this works. future cash-flows are jeopardized and the value of the company deteriorates. we will revisit the adaptation of the Miller/Modigliani “value equation”: V(l) = V(u) + TB .4/7) x (100 x 7) = 300 The difference in share price ($20 vs. the function TB .the levels of equity. which is a signal for the firm to take action and lower leverage. if enough debt is added. both TB and PV must be a function of debt. were used instead of debt. Therefore. this model will minimize the cost of capital.producing a positive net present value on corporate projects. If equity. indicating that too much debt has been incurred. both TB and PV will change accordingly. At smaller levels of debt. which will lower the cost of bankruptcy. bankruptcy costs will exceed tax benefits. its tax benefits increase. This presumption that capital is set at the correct level is a limiting constraint of the optimization model. in the short run. In that regard. when we change that variable (debt). By encouraging the use of lower priced debt to the extent where it does not materially undermine the price of the stock.91 done is set the first derivative to zero. the appreciation in the stock would be great enough. In the mean time. we know from observation that firms do not always raise the correct amount of capital and that the proper amount is the outgrowth of project analysis and capital budgeting . the left side of the equation (TB) would remain stable.∆ PV = 0 The function will increase until the change in tax benefits equals the change in the cost of bankruptcy. the tax benefits will exceed the cost of bankruptcy. any appreciation of the stock would be countered by a lower default probability. the probability of default increases at a much faster rate than tax benefits. Had the pattern of funding with retained earnings continued. various mispricings occur that will temporarily allow too much or too little debt. However. in the form of retained earnings. Eventually. that is ∆ TB . and the decrease in default . and movement toward an optimal capital structure. Does this model minimize the cost of capital? In the long run. the market will correctly price risk by charging higher interest rates for riskier leverage. and then it should decrease. The result is a higher marginal benefits equation. creating a disproportionate increase in bankruptcy costs. As a firm increases leverage. but the cost of bankruptcy also increases. nor violate default constraints. As more debt is added. stockholders’ equity will be set by subtracting the optimum level of debt from a given level of capital. The judgment of administration. The firm’s capital investment is another constraint that is inputted as given.net income. will yield the optimal amount of equity. THE OPTIMAL AMOUNT OF DEBT From the perspective of a theoretical ideal where the intrinsic value of the stock is equal to market value. Since the model implies that the tax rate and the amount of loss are constants (determined as given). while the amount of capital is imperative to capital structure. we can control our own flow of money into a stock based on numerical aggregates like the amount of debt. At this point. it is a given constraint in optimization. the amount of capital is an a priori figure. what looked like “excessive” capital turns into “adequate” capital. the analysis of various projects. we add (or subtract) this amount from the actual long-term debt that the firm had on its books. and the time frame of management are all variables in the decision to raise capital. Besides displaying the aforementioned shape.92 probability small enough. While our model can deliver a “realistic” optimum. Subtracting the optimal amount of debt from the total amount of capital invested. For example. the equation can reduce to: ∆ Long-term debt = (∆ Probability of Default) x (Amount of Loss / Tax Rate) When we solve for ∆ Long-term debt. Thus. it cannot deliver a “true” optimum because it relinquishes control of the capital . If earnings are large. As investors. In fact. However. we may second guess the amount of capital as excessive or inadequate. we also need an ideal default probability. the company needs an infusion of lower cost debt because the cost of retained earnings is too high. outside the purview of the model. but we need to use it as a realistic constraint. a change in long-term debt will affect interest expense which will affect the amount of taxes. a practical model is consummated when variables in the default algorithm are made dependent on changes in debt. only the change in long-term debt and default probability are the dynamic factors. and ultimately . and this figure will represent the optimal amount of debt. to actually begin increasing the cost of bankruptcy. The higher market value that is implicit with equity funding will be an outgrowth of more shares issued when earnings are up. most of the curvature is derived from the probability of default. but the most onerous constraint is the amount of capital. Growth stocks often have higher earnings than average but more volatility. Since the amount of loss is a constant. firms with a lot of leverage tend to have stocks with a lower price earnings ratio. the cost of bankruptcy would be prohibitively high for a debt issue even without a higher probability of default. Many of them are funded with equity because the instability of cash-flow keeps them away from the credit market. When we optimize debt. GRAPHIC DEPICTION To gain a better understanding of the optimum. a higher stock price discourages the use of debt. Thus the model mirrors the P/E characteristics of debt. • 3 More tangible assets would decrease the cost of bankruptcy and allow more debt to be used. In this model. LONG-TERM DEBT AND THE AMOUNT OF LOSS The peculiarities of the amount of loss mesh with the addition of long-term debt: • 1. income. it is significant to note that the expression TB is a straight line.assets. the rate of . On the other hand. At this juncture.. Higher stock prices encourage the issuance of equity. the cost of bankruptcy will be a curve that increases at an increasing rate to reflect more risk at higher levels of debt. and other liabilities. A business that is founded upon “managerial expertise” is usually less credit worthy than one that has salable assets. we do so in relation to the variables of default probability . From a creditor’s standpoint.93 function to outside factors. because more debt would raise the chance of default and be magnified by the higher price. this is a realistic assumption because tangible assets provide collateral for loans. When the function TB minus the cost of bankruptcy is at a maximum. • 2. Where the two lines cross. Figure 4-2 $ Cost Cost of Bankruptcy TB O D Debt At point O. an optimum occurs which indicates that the distance between tax benefits and bankruptcy costs is a maximum. it is usually because management is gambling on some unrealized stream of income that is outside the necessary parameters of tax incentives.94 change (slopes) will be the same and the amount of distance between the lines will also be at a maximum. (Back to Table of Contents) . When the model is circumvented. an optimal amount of debt is incurred. It should be no surprise that the decision to use debt is founded on classic cost / benefits analysis. The marginal tax benefits equal the marginal cost of bankruptcy because the slopes are the same. most strategies will be linear enough to allow the assumption of some understandable relationship between basic fundamentals. At point D. assets. market value and default. While the probability of default curve can assume myriad shapes and include many complex variables. debt is so excessive that it must be reduced to save the company. 95 APPENDIX: MAKING SURE THAT WHAT YOU SEE IS WHAT YOU GET Most companies and business people are inherently honest. While many companies have several incorporated sub. but as an effort to educate the investor. but many of their shenanigans had tip offs that would have alerted prudent analysts. It is kept separate to maintain core competencies. but there is a growing faction in the investment community that operates on the edge of the law. While this is not a course on business ethics or forensic accounting. Often the cry goes up that “We need to do it because the competition does it. decisions need to be based on transparent financial statements. the obfuscation of debt is pertinent to capital structure. Enron was extensively involved in “entity structuring”. One look at their last 10K showed page after page of sub. A distribution company may have its own incorporated. control expenses and create tax advantages. which may certainly be true. Armed with competent teams of lawyers. The complexity of the modern balance sheet warrants an array of footnotes to accompany it. They should not act as the accounting version of “fine print”.units all over the globe. “in house” transportation company that serves only the distribution business. Investors and analysts did not question the practice because Enron was a “cash cow” that apparently made money and increased its stock price. Enron piled sub-unit upon sub-unit and had more than most analysts could realistically track. However. and fully disclose the complications in the balance sheet.units whose performance is risky enough to separate them from the larger entity. these companies concentrate more on what they can legally garnish than on servicing the customer or investor.”. Enron attempted to use them to shield the extent of liabilities from investors. This author first encountered Enron because they were users of a certain brand of risk management software that did Value-at-Risk calculations on derivatives like oil . Make sure you read these notes carefully. For example. General Motors is in the auto business. but keeps their financial unit separated from producing autos. Enron were the supreme masters of obfuscation. A case in point: A CEO emphasizes current sales and operating . After a little research. “Obligations incurred in risk management activities “ or something to that effect. The income is legally claimed. Although it may seem that senior management is defending the stock against a lower share price. If corporate control is an issue. A larger amount of debt will also immunize a company against a takeover. then a few shareholders at the top can profit by issuing debt instead of equity. most banks profit from these swaps every day but do not detail them on financial statements. Sometimes taking on debt is an emotional issue. where it could conveniently be paid back as soon as cash-flow improved. but the investor knows nothing about the other parties in the contract or especially about the risk involved. cash-flows in one currency for cash-flows in another. On a corporate level. any credit-worthy individual can engage in a forward contract. even if it means a nominal decrease in the stock price. the entire domestic banking system has been under scrutiny for mortgage lending practices and the further securitization of such mortgages.96 futures. the unstated reason is to maintain power. creating derivatives that need to be traced back to the originator. I spied a financial statement item that was termed. When might the largest shareholders refuse to move toward a more optimal capital structure? When issuing more shares takes control away from them. it appeared that Enron was placing money lost in hedging futures into a long-term debt category. ad infinitum. I closed the Internet window and shuddered. a balance sheet must “balance” which allows the investor to play “detective” if there are misappropriations between short and long-term debt. At first glance. The term. because most companies have second thoughts about assuming the obligations of another. and I was curious. assets and equity. “agency friction” was developed out of frustration that management was looking out after their own interests and not the shareholders’. As this chapter was being printed. swapping fixed rates for variable rates. While the futures market is highly regulated. In fact. Although interest rates are at an all time low. A few shareholders will gripe. The reason? There is no indication in financial statements of what the “proper” capital target should be. The old paradigm does not work. and company income is rock steady. debt is not being removed through increased operating income. Shareholders will be left wondering why the current adequacy of sales and income could not buffer the stock from declining. they need to be manifest.another reason to read footnotes. done in anticipation of higher earnings. while allotting the difference between the two amounts to net income . However. a broker buys a firm’s outstanding bonds and trades them to the company for newly issued shares. the capital structuralist recognizes that the cost of capital rose even as income remained steady. the stock had to deteriorate because not enough risk was taken. In brokered defeasance. will wipe out a substantial amount of debt. it issued zero coupon bonds . While these techniques are not implemented to obfuscate financial viability or to confuse investors. In either case. the CEO is leery about “too much” debt. In effect. nonetheless. The broker subsequently sells the shares. Such a move.a win-win. Sometimes it appears that a company negotiated a fantastic deal on interest rates. but is mysteriously disappearing off the balance sheet . lower face value bonds. but most will be “out of the loop”. it is . His reference is a “credit crunch” incident six years ago that almost bankrupted the company. because of tax savings and no regular interest payments. A company will attempt to clean-up its balance sheet by converting its debt to higher interest rate.securities which will be paid off in a lump sum when they mature. but “borrowing from Peter to pay Paul”. Such an issue is advantageous to most companies. even as the stock price diminishes.97 expenses to the exclusion of funding future projects with lower cost debt. when in fact. Long-term projects that might not pay off until years later can be implemented without excessive interest expense. An extreme case of “balance sheet magic” is called “in substance defeasance”. He goes ahead and funds projects with equity. Nevertheless. this is a new phase in the business cycle. as should the practice of refinancing loans. the company decided to net it out with interest gained from investments. any lease to own contracts or any liability lasting over one year should be referred to under the heading “long-term debt”. Lastly. that is too much work and a phone call needs to be made to investor relations. (Back to Table of Contents) . but when it is couched in language that implies such a loss is normal part of business. To keep and establish good relationships with clients. Investors should be aware of the various categories of long-term debt. accounting firms will be generally compliant and go out of their way to make the customer look good . because some companies will state the sub unit without referring to it as “long term debt”.98 important for the investor to treat the amortization of the loan as interest expense in an optimization model. Miraculously. Capital leases. when it is not. The risk of non-payment when the issue comes due is viable and needs to be distributed over the life of the loan.despite its having long-term debt on the books. The increased complexity of corporate finance has led to both oversights and purposeful manipulations Calling a bad hedge a “risk management loss” is not illegal and may not even be purposefully deceptive. no one thought that interest expense was meaningful enough to declare. Clarity is to the benefit of both management and investors. a red flag should be waved. This author once went through a 10K for a well known steel company and could not find any interest expense in the income statement .legally. there is sometimes an unintentional misstatement of debt obligations because of confusion or naiveté of the preparer. If the investor has to match the number of years that a contract lasts with the liability type. These indications are also available in the footnotes. Evidently. 2. The price of the stock. No accounting balance sheet itemizes it. this opportunity cost is simple to fathom: a company can either reinvest earnings or return them to stockholders.in the industry. Some of the factors that affect that potential loss (risk) are the following: 1. The cost of equity is not an accounting cost but an economic one. Future prospects. it is a cost so prominent that it not only affects every multinational business from Cisco to US Steel. Alternative investments . Therefore.given the constraint that all steel companies have the same risk. It is the cost of equity. If the company reinvests earnings. While this example represents an oversimplification. in the bond market. a comprehensive look at this subject would include a model with hundreds . 5. or in the stock market in general. if all steel companies return 15 % and steel company Z returns only 10 %. it needs to make a return at least as great as investors would make in alternative investments in companies with similar risks. In fact. steel company Z has an opportunity cost of 15 % but under performs the industry by 5 % . In fact. This list is by no means exhaustive as myriad other factors affect the cost of equity.99 5 THE COST OF EQUITY There is a cost that hovers above every major business deal like a pesky fly at a summer picnic that no one wants to acknowledge. Current income. It represents the “opportunity” of taking one action over some alternative action. 4. and yet it is omnipresent. 3. the reader should understand that the cost of equity is a comparative cost that is implicit in every profit-making vehicle: each dollar of added income costs the shareholders some amount of unrealized but potential loss. it is an integral part of every mom and pop store as well. Dividend yield. If we consider that the foundation of stock ownership is based upon the flow of corporate earnings toward shareholders. It is rarely mentioned in the Wall Street Journal. we need to bring the theoretical concept. MODIFICATIONS Ultimately. which requires some modification. Now suppose inflation takes hold.100 of variables. Since the cost of any capital component is gauged by its required return. If we use the 8 % rate as the required return as well as the input into the cost of debt. then the company can buy back its own lower priced bonds in the market and earn a yield of 10 %. the market method allows the analyst to balance this equivalence between cost and return. and the price of this company’s bonds falls to 80 million. • But . book values are more controllable and actionable. and shareholders pay an implicit price for such . The book value designates debt of 100 million and interest payments of 8 million.The objective of capital structure is to seek an equilibrium between market risk and company risk that may require techniques that compare market to book values. any comparison between net income and stock price is obfuscated by extrapolation. they begin by paying 8 million a year in interest expense. Essentially. For our purposes. but market values have shifted both the required return on debt and its consequent cost upwards. we need to create only the best estimate possible. Since the market value of a stock is based on projected future growth in earnings. The following issues need to be reconciled: • 1) The cost of equity is based on market values and not book values. Using the lower interest value of 8 % as the required return on debt would lead to a debt laden capital structure because it would under price its cost. The proper method of calculating not just the cost of equity. but the cost of any source of capital is to use market values. given the limited amount of information available to investors. Traditional use of the cost of equity for stock valuation and capital budgeting needs to be expanded. and interest rates are raised to 10 %. Consider a firm who floats a bond issue of 100 million at 8 % interest. If earnings are to be distributed to shareholders. Simply stated. “the cost of equity” into the practical realm of capital structure optimization. but reflects movement towards an optimal capital structure.which may not be any longer than a few minutes! This is a circular argument which assumes that the premise is true. the stock price rises and falls based on the proportion of debt to equity. nor an opportunity cost. present value calculations only discount future earnings. Inc.The target proportion is always moving and it is questionable whether an optimum can be reached. the concept has been trademarked by the firm. Inc. Most theoretical models ignore the fact that shares cannot be issued at a constant price and that the true mathematical optimum is fleeting. capital structure analysis multiplies a market derived percentage cost of equity by the book value of equity. While a company needs to struggle to minimize its cost of equity. Thus. it 2 EVA is the registered trademark of Stern Stewart. Since net income is a book value component. Another tenet of capital structure is that the stock price is maximized when a firm’s capital structure is at an optimum. . when the cost of equity is dependent on a shifting stock price. Therefore.101 earnings. • 2. and is called EVA® or “economic value added”2. net income and equity. which later becomes the book value item. a concrete minimum cost of capital cannot be obtained. it is consistent to use the book value of equity when making direct comparisons . a required return. Stern Stewart. Multiplying the estimated cost of equity percentage by the market value of the stock yields a realistic cost of equity for only as long as the stock remains stable . The investor friendly version is further developed in the chapter called. Every firm has an optimal proportion of debt to equity that increases returns and minimizes risk. • But . both variables. and then compares it to net income. This composite value is neither the cost of equity. but it never remains constant. “retained earnings”. “The Capital Dynamic”.at least when gauging near-term performance. need to be calibrated within the same time frame. once dividends (if any) are paid. In fact. it may well be just a percentage of the cash received from the issue. it is more practical to use this measurement in near-term comparisons when seeking an optimal capital structure.102 may never realize how close it is to its goal. the need for stalwart precision is insubstantial. and each component part of the firm’s equity is implicit in the comparison. because the determinant variable (stock price) keeps changing. we gain in expediency and utility. • But . the variables of trend and comparative magnitude are most significant. such a comprehensive tool exists: the “capital asset pricing model” compares the changes in equity of an individual firm to the general market. With an “opportunity cost” that shifts frequently. it is more prudent to seek out a model that yields the best estimate given the current level of risk. . • 3) The cost of common stock is valued like the cost of retained earnings. except for flotation costs on new issues. observation of long term changes in the cost of equity may indicate how the company is reacting to macroeconomic trends. Nevertheless. A precise itemization would require separate valuations for new common issues and the category of retained earnings and past issues. Therefore. all of which ascend and descend at different rates. Moreover. Since the market can be volatile. Without a cash outlay. the amount of past issues and retained earnings has no immediate cash outlay (disregarding administration fees) and may represent investment in some asset that was amortized long ago. and with inherent volatility in the measurement. What we lose in precision. However. • 4) The definition of stockholders’ equity includes preferred stock. these components represent a substantial investment that must be enumerated.there may be no actual cash outlay on any of the components When an investment house underwrites an issue of stock. Fortunately. the cost of equity for new common stock is higher than for both retained earnings. it charges the issuing company a substantial fee which is deemed a “flotation cost”. and past issues of common stock by this factor. but is in line with companies of similar risk. investors flock to Company A and bid up its price. If Company A can get greater returns than other companies with the same risk. Like short-term debt. In fact. it is valued as a component of the cost of capital. any investor who purchases a rising stock that is beating the market is likely choosing one that is not in equilibrium with it. In this case. This . combining a regular dividend with a set issued price. it is a conservative error that will have few tacit repercussions. it acts as an adjunct to other sources of capital. The regularity of payments gives it some of the risk characteristics of bonds. In a quick comparison. preferred stock needs to be fully accounted for.103 • But .preferred stock is not valued like the other equity components. and pricing it at the cost of common equity. and its cost is as follows: (Preferred Dividend / Net Issuing Price) x (Proportion of Preferred Equity in the Capital Structure). will not be as accurate as if preferred stock were itemized as a component cost. but without the projected growth characteristics of common equity. Preferred stock is a hybrid security. these rates are equal. the market for preferred stock is limited and consists largely of corporate ownership because of the tax breaks involved. less shares of common equity need to be issued when some funding is done through preferred stock THE REINVESTMENT RATE AND OPPORTUNITY COSTS Some confusion may exist between the reinvestment rate on earnings flowing to shareholders and the return that would be received on comparatively risky investments. in a thorough analysis or in capital budgeting problems. If the investor ignores it in his or her calculations. the cost of equity will be higher than it actually is because preferred stock lowers the risk of equity. However. while it maintains the liquidity of stock. Thus. When the market is in perfect equilibrium (which is rare). This upward pressure on the stock price will continue until the point where Company A no longer garners an extraordinary return. making comparisons between two companies that have preferred stock. However. it keeps the cost of equity less risky. Many companies carry no preferred stock at all. Indeed. the student who has compared the difference between “value” investing and “growth’ investing. we always refer to the opportunity cost. Any measurement of change requires some relational values around it in order for it to be meaningful. For purposes of capital structure. may take a leap of faith. “the expected rate of return”. Since company fundamentals tend to determine reinvestment rates and market competition determines required rates. but its opportunity cost may be low if comparatively risky firms are making only 10 %. its reinvestment rate is high. Similarly. We never assume that the market is in equilibrium. there will be enough corroboration from several sources to create our own “in house” tolerance levels. However. any reinvestment rate that is lower than companies with similar risk will end up depreciating its stock price. if Company A makes 25 % on reinvesting shareholders’ earnings. “the required rate of return”. “the required rate of return” as the true cost of equity.a lack of equilibrium between company and market rates of return that work on the same continuum to correct themselves.104 “risk arbitrage” is fundamental to all speculation. and the rate of return on companies of similar risk. Academics often call the reinvestment rate of return. it is time to sell the stock and “take the opportunity” to invest in other companies. nor do we assume that reinvestment rates are similar among companies of comparative risk. can observe that both methods are part of the same phenomenon . EVALUATING THE COST OF EQUITY: METHODOLOGY The attempt to enumerate a theoretical cost requires a tolerance for uncertainty. the forces may overlap each other. To those who work in the accounting or engineering fields. work in concert. or against each other and one may dominate the other. However. In . The “opportunity cost” is the opportunity of not investing in the present company. we submit to the market as the greater of the two forces. basing decisions on a “guesstimate”. Thus. If the opportunity cost exceeds the reinvestment rate. but investing in similarly risky companies. • 2) The Rule of Thumb Method: The rule of thumb method is a modification of market valuation methods that use the present value of dividends received in the future to obtain the “fair value” of a stock. In the case where the dividend growth is constant. present value discounting is implicit in the calculation. “Triple A” (AAA) ratings would justify just two percentage points added. analysts (with green eye shades intact) would add from two to four percentage points on to a firm’s bond yield and call it “the cost of equity”. Realizing that riskier lower rate debt would contribute to more equity risk. analysts set a risk premium that combined the timing of the business cycle with the rating of the firm’s bonds. when information was not so readily available. those relational value always provide a solid foundation over the long-term. while lower rated bonds at the top of the business cycle would warrant tacking on four points. because a firm cannot be in business without adhering to them. the Gordon Model emerges: P = D1 / (K – G).105 capital structure analysis. The terminology is as follows: Table 5-1 SYMBOL P D1 K ROE RETENTION RATIO G EXPLANATION Current price of the stock Dividend received in the next period Cost of equity Return on Equity (net income / equity) 1 . Since the dividend grows at a constant rate. Although this method is subjective. it does give the modern analyst some guidance: if he or she calculates an eight point difference between equity and debt costs. • 1) The Risk Premium Method: Years ago.(Dividends Paid / Net Income) Growth rate that dividends are expected to follow . the figure is most likely too high and needs to be recalculated. However. we rearrange and form K = E / P. the method is simply inadequate. Furthermore. is merely the return on the market price of the stock. it must be gauged with a reference to the market average For example. the estimation of future income is not that formidable a task. The fact that P/Es can tell us so much about a stock’s behavior is partly attributable to this relationship. when “E / P” is used as a proxy for the cost of equity. and when “E / P” approaches “ROE”.5 %. we equate it with the retention ratio multiplied by the return on equity (Retention Ratio x ROE). analysts can scan this number rapidly and combine it with other information to screen for prospective investments. With a stable multinational firm.. then the following equality is met: (1-X)E / K . If we substitute “X” for the retention ratio and use the symbol. we will assume that G = (Retention x ROE). like the P/E ratio. “K”. In actuality. “P / E”. growth stocks do not qualify for consideration using this method. and proceed with deriving “the rule of thumb” method. For the time being. Many of these can be quite elaborate and require foreknowledge of the path of earnings far into the next business cycle. • 3) Valuation Methods: The Gordon Model is only one of many dividend discount models that equate the price of the stock with the present value of cash flows received in the future. Since we now have an equality where P = E / K. For the sake of simplicity.70 for an EPS would assume a ridiculously low E/P of 3. Many growth stocks have very high P/Es. but with untested or volatile . for stocks with much anticipated growth. Obviously. Used judiciously. this inverse of the “age old” financial indicator. the stock may be considered a bargain. a P/E of 29 based on a $20 share price and 0.106 The growth rate “G” is controversial because it is presumptive and extrapolates into the future. the P/E is an indicator of growth. the model is now P = (1-X)E / K-(ROE)X. For stocks that have an average P/E that is close to the market. the method offers a very quick estimation of the percentage cost of equity. the growth rate may be a long-term average of this figure. Thus.(K)X = (1-X)E / (1-X)K = E / K. if we assume that ROE is equal to the cost of equity. “E” for earnings. In fact. it is a better statement of the reinvestment rate. And – obviously – like any other dividend discount model. We have already been introduced to the Gordon Model. • 4) The Capital Asset Pricing Model (CAPM): This model is the preferred choice for capital structure analysis. However. Ultimately. the “risk-free” rate. What it does beyond dispute is to compare any security to a market index in a linear fashion. “required rate of return”. It then multiplies a derived comparative number by the difference between the index and what is termed. because it meshes well with the earnings and funding characteristics of a company. Why use the Gordon Model if it is only applicable to dividend paying companies where the dividend is growing steadily? The assumption of steady growth will not radically alter the viability of the calculation in the near term. comparative. and not to extrapolate the “fair value” of a stock. 2. it has many more uses besides determining a cost of equity. The major problem with using the Gordon Model is that it is so dependent on the fundamentals of the company. we use the cost of equity for near term comparisons between companies. . (D1 / P) + G = K. Despite years of academic and professional disparagement. P = D1 /( K – G). The individual security with the market. it can not be used with firms who pay no dividend. which becomes. we assume that the next dividend will be growing at the same rate as the present dividend. the CAPM seeks to be the essence of risk comparison but falls far short of any claims to precise accuracy. and summed with the growth rate will produce a cost of equity. the method comprises a “random walk”. In effect. a probability that avoids judgment about transitions. The simple adjustment that we make is to solve for “K”. and then adds that figure back to the risk-free rate. With a high volume of literature to both support and excoriate it. The market with what is thought to be a risk-free yield (usually the ten year treasury). “the expected rate of return” than an equity based. A modified dividend yield using the next expected dividend rather than the current. the concept continues to thrive like a Darwinian “missing link”.107 companies. Consequently. Without excessive extrapolation. it compares two sets of numbers: 1. each method of determining the cost of equity will tend to corroborate the other in the domain of change . it is rarely encountered in finance and its presumption can lead to poor decision making . No other method better incorporates the force of the market with a comparison between alternatives.108 It is this “required return” that forms the cost of equity. and therein lies its downside. The CAPM uses regression techniques (see the chapter on statistics).on a graph.e. ∆ Y / ∆ X. but each respective method may yield a different value. the individual security and the risk-free rate of return. the CAPM is basically a straight line in the form of Y = A + B(x) with some modification. but in the next week they might assume a polynomial relationship of the fifth degree.if the change in the cost of equity while using the “risk premium” method is 20 %. The key to this model is the comparison factor. When used correctly. that is . which is essentially. the relationship between assets and earnings might be linear. the CAPM is susceptible to periodic instability that will lead to incorrect decisions. Used in isolation. The CAPM needs to be perceived as a valuable but potentially misleading tool. and the proverbial “rate of return on companies with similar risk”. As much information about leverage. the familiar “rise over run” . there will be a corresponding increase using the CAPM method. and the emphasis must be placed on consistency. and derives a comparison factor named “Beta”.witness the implied chaos of wage and price controls in the early 1970s to observe just one example. i.. for example. The cost of equity is a measure of comparative risk and not an accounting standard.albeit at a different absolute level. In fact. the business cycle (whether the Fed has raised or lowered rates) and other methods must support the use of the CAPM. Ultimately. While true linearity breeds certainty. Most financial variables have an ongoing multinomial relationship that constantly changes and forms a curve. As long as the same methodology is used in the comparison. performance in one year will be compared to performance in the previous year. In one week. however. the forces . the dynamics of change define their utility. Similarly. accuracy and precision are a premium. The E / P method. While many companies use sophisticated Monte Carlo analysis to forecast a potential cost of equity. forms a more accurate measure of the cost of equity. It is when two different methodologies are used that trouble arises. • 5) Consensus Methods: The logic behind combining methods into a “fusion” cost of equity stems from the opposition of fundamental forces. and in some specific situations. For example. because the components of the measurements will be so different. The judicious use of these combinations gives the analyst an advantage when gauging changes in risk.partly because the political necessity of “telling the truth” about a project can actually undermine it. Thus. each points to optimality. the absolute values of these costs may be totally different. The most frequent use of the cost of equity is not in investment analysis. but when used properly in the context of capital structure. The coefficients of the regression will be the weights of each method in the . One such method is to regress each of the methods against the performance of the stock over a five year period in one multiple regression.109 acting on the stock will be very similar and will be measured in similar ways. while the methods act as the sequence of X variables. The dependent “Y” variable is the five year performance of the stock. but in capital budgeting. historical data with consensus techniques may work for others. More mistakes are made from underestimating the need for capital inputs than any other . combining the return on equity (ROE) with a market based method like the “E / P”. for example will measure market price much more than the “risk premium” method which will measure the comparative risk on debt. When a corporation outlines the capital needs for future projects. utilizes the difference between book and market values which is an implicit factor in the cost of equity. combining the CAPM method with the Gordon model gives weight to both market comparisons and internally generated company fundamentals. . The reality of financial data is that it is actually “heteroskedastic” . or for capital budgeting. By its very simplicity and flexibility. That is: they describe a probability distribution with a central tendency around a mean and an area that is defined by a set number of standard deviations. for a brief moment in time. While this method is linear. however. is not so comprehensive. sometimes trading utility for structured precision.one that maximizes the mean return while minimizing the standard deviation thereof. for portfolio management. Indeed. given a narrow range of variables.almost a flat line. the student/investor is encouraged to study this model from the perspective of creating a working portfolio . then much of the return would be extracted from it. Any discipline that is as probabilistic as finance must adapt to existing conditions. The area defined by a handful of “big winners” and a lot of “losers” is heavily skewed. The distribution would be like a savings account at a bank . In fact only part of its potential is realized in deriving a cost of equity. if finance worked like physics which is mathematically precise and predictable (on a non-theoretical level).defined by extremes with heavier weights at the ends.110 consensus. However. many professionally accepted statistical techniques assume “normality” where none exists. THE CAPM AS A BUILDING BLOCK FOR CAPITAL STRUCTURE ANALYSIS The CAPM is integral to an area of finance that commands the utmost respect modern portfolio theory. the CAPM can be used in several different ways – as an investment tool. the combination of securities that will diversify away risk and maximize returns. Its use in capital structure analysis.highly predictable but with little return . While it is typical of an eclectic system like capital structuralism to use what is practical. it will allow the analyst to cite the most historically accurate method and to better understand the forces behind the stock’s changes. For example. the investor must realize that it describes only a “best estimate”. but nevertheless encouraging experimentation. a tradeoff. because of inflation and the nominal probability of . The financial executive makes this tradeoff when he or she decides to raise dividends and retain fewer earnings. The market .en masse will make this decision when it favors investment grade bonds over the equities market the classic flight to quality that occurs in a downturn. While no security is truly “risk-free”. but notice that both the return and standard deviation are less than they are at point A where the greatest return with the smallest standard deviation is achieved. Point B is termed “the risk-free rate of return”. The individual investor makes this decision when investing in a debt laden company rather than a high-flying cash generator that has just peaked. all of these tradeoffs have a single element in common with the CAPM: whenever more risk is engaged. a greater return is expected. At point B. there is almost zero risk (no standard deviation). Essentially. using the market rate of return and the standard deviation of the market as measures of risk. Figure 5-1 Market Return A B C Standard Deviation The green curve is a combination of stocks that encompasses the market. The capital market line displays this tradeoff graphically.111 If one concept can characterize capital structure theory it is that every decision requires a balance between competing alternatives . and so the capital markets line inclines. reality dictates that in many cases the returns will not substantiate the risk incurred. and a portfolio can languish for many years without moving toward the line. the slope of the line changes when the risk/return characteristics of the market change. and the regression line of an individual security against the market. and have upward pricing pressure on them. Speculators (and unscrupulous fund mangers) will choose point C because it has the highest return associated with it. However.S. Analogously. this point represents U. . and that the risk/return line would change in proportion to the change in the market . In bull markets. most professionals would eschew this point because the chance of loss is too high. the investor should realize that some optimal portfolio exists even during recessions. but it should not be confused with the CAPM itself. On the other hand.inclining during a bull market and flattening during a bear market. However.112 catastrophe. points below the line would have more risk than warranted by their returns. The capital asset pricing model is actually a compendium of the capital markets line. The capital markets line displays the theory of risk behind the CAPM. These points are only temporarily achievable and there will be forces that put downward pressure on their prices. In effect. In fact. treasuries and has the lowest amount of investment risk associated with it. In a bear market. and so the line flattens. it is typical to gain large returns with little risk. any amount of risk incurred seems to yield negative results. point A represents the “efficient frontier” which is where fund managers and investors want to be. That regression line is called the “characteristic line”. Points above the line have a higher level of return given any level of risk associated with them. and that extreme Y intercepts point to less association between an individual company and the market. The monthly return and number of data points makes the comparative numeric change. less susceptible to error through volatility. using monthly changes in the stock price and the market. The market data is usually derived from a widely encompassing index like the S & P 500. Portfolio managers are quick to point out the relevancy of alpha to the particular situation of the company. for a total of 60 data points. The “Y” intercept or “alpha” component has a litany of its own. but much debate has centered around which index if any.113 Figure 5-2 Company X % Stock Increase Market % Increase This line is usually developed over five years of returns. What the investor needs to know is that interpretation of alpha leans heavily on investment experience with the characteristic line. The proper method to calculate the CAPM is as follows: • 1. Calculate the characteristic line as stated. can truly interact with this model because each has discrepancies that bias its distribution. A comparatively large or small alpha can indicate everything from industry dominance and protectionism to a decaying company with a predilection for bankruptcy. The “Y” data are the percentage changes in stock price over sixty months for the individual company. (N + 1) entries of the type . “beta”. the difference serves as a good indicator of how strong the equities market is compared to the fixed income market.usually five years.1) will produce a decimal percentage. The three components are arranged as follows: Risk free rate +(Beta)(Market rate Risk free rate). However. What happens to the Y intercept. a large risk premium will peak and then shrink when interest rates are raised to combat inflation. Once a regression line is formed. “alpha”? While alpha is not entirely out of the purview of capital structure. the process will be swift and mechanical: data can be inputted and regressions processed in less than thirty seconds. . The average market return over the five year period is inputted 4. the business cycle is often an expression of the size and acceleration of the risk premium. Sixty-one of these entries will produce sixty working data points. by using a shorter span. The risk-free rate is the average yield for the ten year Treasury bond over the period of the regression . (Market rate . because the beta component should correspond to the same length of time as the other components. At the beginning of a recovery. in combination with a low correlation coefficient (R) would indicate that the association of the firm with other companies of similar risk is weak. and can be applied to other assets. we show how to create a “Federal Reserve bias”. A comparatively extreme “alpha”. it is used to gauge the relevancy of beta. Although the CAPM is not exclusive to the stock market. • 2. Once the student gets used to doing these regressions in a spreadsheet like Excel. In subsequent chapters. This difference is called the market risk premium and delineates the excess return of the specific market over the risk free rate. the proper method is to use the average over five years. The difference. In essence. • • 3.Risk free rate) is especially significant. In fact. the coefficient of X is the “beta” component. The “X” data are the market changes over the same period. If needed. the market risk premium is especially significant in that arena.114 ((March Price / February Price) . refer to the chapter on statistics. While earnings and the cost of capital work on a microeconomic level. may have led to some of its lack of acceptance in the academic community. the connection between portfolio diversification and capital structure (a portfolio of capital funding) was not so readily apparent. however. It can be used on any asset (commodities. the components of the risk premium work on a macro level. rates are raised and businesses have a harder time staying afloat. and the risk premium is lower as well. junk bonds). Although selling bonds will raise yields. but it also defines different types of risk.115 the risk premium begins to widen. The premium was in finding the combination of securities with the greatest return per unit of risk. The main purpose of the CAPM was to enumerate these risks more clearly and create an investment pattern that allowed the investor to take a broader look at the reasons why a stock rose or declined. This sensitivity to the interaction between credit and equity markets is one reason that the CAPM is a good tool for gauging the cost of equity: capital structure is dependent on the business cycle and the inherent changes of risk in various sources of capital. Harry Markowitz pioneered modern portfolio theory nearly fifty years ago (as of 2008). Analogous to the difference between the rates of change of earnings and the cost of capital. Eventually. At the peak of the equities market. demand for equity is at its highest point which raises the risk premium. the acceleration difference between the risk-free rate and the market rate determines performance in the capital markets. The equities market declines much faster than rates can be lowered. Part of the value of the CAPM is that it is so flexible. which is an odd mathematical combination: it needs to maintain a straight . Although the CAPM developed into a necessary adjunct. the “Y” intercept in the regression is called the “alpha” component and may be combined with other assets to diversify away risk. real estate. the stock market accelerates at a much faster pace. inflation creeps in. The flexibility that the CAPM engenders. It combined flexibility with linear constraints. However. The beta component is called systematic risk and is non diversifiable. fund managers sell bonds and buy stocks. return tradeoff. Jensen and. Since the validity of the CAPM is contingent on a risk .116 line.a downward sloping curve was encountered that described less return with more risk. there were several periods where the relationship was inverted: that is . Scholes study of thirty-five years of stock performance. but it does so by reconciling variables that would not ordinarily have a linear relationship.1965) . Figure 5-3 Black. They found that although the actual relationship between beta and average monthly returns matched the theoretical model. Wall Street quickly rejected the concept and became more enamored with technical analysis. Jensen and Scholes Study Actual Theoretical Return Beta (35 Years) Figure 5-4 Return Theoretical Actual Beta (1957 . In fact some professional rejection occurred after the famous Black. such a comprehensive and imperfect model has stimulated many attempts at improvement. like the Miller/Modigliani model which also has stringent rules. However. 6. AN ESTIMATE OF RISK AND NOT PREDICTION Part of the inconsistency of the CAPM has stemmed from its use by fund managers who expect it to be a predictive tool. Investors are risk averse. Treynor and Mossin. it is tempting to invent some way around it. which are termed “market’ models more conducive to displaying the actual behavior of a security. Lintner. Whenever theory straitjackets behavior. The original model also had a bevy of rules that made it difficult to apply in a realistic situation. all of whom tried to elaborate on it through other techniques like the “single index model” or “multiple index model”. experimentation is encouraged because the rules are impossible not to break. The original model can be attributed to a unified effort by Sharpe. 5. Firms can lend and borrow at the same rate. variation is confused with performance: . No taxes or transaction costs exist.117 Naturally. What are the effects of high taxes on the model? What if the dollar is slipping and I want to apply the model in Euros? Does the model work better for a stock with a rock steady beta? These are essentially investment questions and outside the purview of capital structure. These empirical constraints were as follows: • • • • • • 1. The market portfolio (index) chosen is the appropriate one. Betas are stable Every one of these rules will be overturned in an investment scenario. 3. Investors are more concerned about domestic currency returns than exchange rates. In this case. 2. 4. While adherence to these rules removes some of the conceptual liability the model would have if a stock did not behave as expected. it also opens up several more avenues of theoretical approach. and different combinations will be either in place or suspended at any given time. They prefer more return and less risk on a constant basis. Since earnings and the cost of equity are highly correlated. In other words. then a beta of 1 will do the same.118 variation with the market is expected to present a forecast of expected behavior. If we disassociate risk from absolute performance (market returns). In fact. The return on the market. When using the CAPM as a proxy for the cost of equity. However. a firm can substantially increase earnings. the analyst must stay well diversified in his or her measurement activities resorting to other cost of equity measurements as well as relating it to the proportional increases in debt. Obviously such a prospect can be embarrassing and financially disastrous. Since we are measuring risk and not combining portfolios. and the required rate is not used as a gauge of absolute performance. a skeptic may ask.. if the proportion of debt to equity rises. one period of equity risk is compared to another. examination of the “alpha” component in the original regression may hold the key to the mystery. if the market goes up 17 %. we know that something may be awry. Therefore. and each stock has a unique “alpha” component that is not so easily quantifiable. An alpha that increases in a rising market.e. For example. even if using some other method of analysis. and yet our CAPM measurement of the cost of equity diminishes. will improve stock performance but reduce market risk. the equivalence of variation may not lead to the “cause and effect” relationship required by prediction. because more leverage is supposed to increase that cost. and beta interact in ways that make the market more comprehensible. the analyst would wrongfully interpret an increase in risk as a decrease in risk. even while beta shrinks. it is to the investor’s advantage to inter-relate the three components of the CAPM. “What if the risk-return line suddenly begins inverting for one comparative year and not the other?”. but decrease the cost of equity simultaneously. At times. i. we can extract the components of the cost of equity and add them to capital structure analysis for a more comprehensive perspective. a large alpha signifies a stock that will rise even as the market is stagnant. the myriad components of alpha have as much to do with a stock’s actual performance as beta. For this reason. . the riskfree rate. it was stated that using book values to determine the total cost of equity (percentage cost multiplied by equity . most proportional rises in debt will occur in subsequent years because there is a lag time before the investment begins to pay off. who eyes “a hot stock”. ANOMALIES PERTAINING TO THE USE OF THE COST OF EQUITY IN CAPITAL STRUCTURE ANALYSIS In a previous section. we combine market risk with changes in equity that are “organic” .either . and then multiplies it by the book value of stockholders’ equity.119 Besides the sometimes-periodic dysfunction of the CAPM. it is almost prohibitively expensive to make major changes in capital structure and then try to “undo” them. data lags the performance of the given sector.the fact that data is only available subsequent to company performance. One of the main tools in capital structure analysis is the production of a “hybrid cost of equity” which uses the CAPM derived required rate of return. In effect. even quarter to quarter observation by analysts will fail because there is no guarantee of continuation. When capital structure analysis focuses on proportional capital changes. When an economy undergoes a fundamental shift into a different phase of the business cycle. For example. any anticipatory changes in earnings are derived from capital shifts rather than observed momentum from quarter to quarter. Unfortunately. the earnings data may describe a totally different company than the one that is currently functioning. it is because they are more indicative of a trend than past earnings. In other words. Thus. the analyst must contend with what is termed. As covered in the chapter on the cost of debt. makes the mistake of buying a company whose earnings will do much worse than currently expected. few companies will merge in consecutive years or issue large amounts of stock on top of each other because of fear of dilution. Market values reflect the true total cost of equity although some textbooks use book values for the sake of simplicity. “look ahead bias” . The investor. Momentum is a function of being in a favorable leverage state that is trending.either market or book values) was considered improper. there is no reason to invest in stocks except for speculative purposes. an “adaptive expectations hypothesis” would give more weight to recent events when making a decision. capital structuralists will harbor an assumption that has historical credibility. Researchers have estimated the average risk premium to be between four and one half to six percent and it will be this number that will be inputted into the model when the actual premium dips below it. we use that figure to emphasize increasing risk. but may not be currently operable: the market risk premium is given a “floor” of at least five percent. RATIONAL EXPECTATIONS Previous sections encouraged the student/investor to do a sixty month regression on both an individual stock and the market while using the average market returns and ten year yields as inputs into the CAPM. we are imposing an artificial equilibrium factor onto the market to make the cost of equity a workable percentage figure. For . Thus. the cost of equity might even be measured as negative! To calculate a usable risk premium. The “rational expectations hypothesis” which professes that investors take all information into consideration when making a decision. In certain markets. the risk premium will narrow enough so that there is absolutely no reason to be in the market other than speculation: the risk. when derived from the CAPM. ADAPTIVE EXPECTATIONS VS. The more random and speculative market values are filtered out. Another tactic is to modify the risk premium. any rational decision would consider the possibility of reversion to the mean. On the other hand. Under these circumstances. would encompass such a long regression. Essentially.free return is actually larger than the return from the market index. the cost of equity has an upward bias because it is meaningless as a gauge of risk at extremely low levels of the risk premium. However. A stock market with very low risk premiums is simply not sustainable for a long period. If an investor can get a larger and more stable return from investing in a certificate of deposit which has zero volatility.120 increases in retained earnings or new stock issues. when the market risk premium goes well above five percent. modifying beta in a one year regression. depending on the time frame of the forecast. If the investor chooses to input the actual yearly figures for market return and ten year bond yield.a function of change in the derived cost. our objective is to determine which year has the lesser or greater amount of equity risk . an analyst who chooses five-year averages may be closer to determining the actual cost of equity. would be counterproductive. How about beta? Should we modify it? While betas have proven to be unstable. any market surge or interest rate hike will be a determining factor in this “adaptive” cost of equity. any acceleration of interest rates in the near-term would make us believe that inflation was becoming an obstacle. the result will be a much greater or less required rate or return. we can get functional results as long as the method is applied equally to each comparative year. Even if we choose the E / P “rule of thumb” method. the modification of beta would lead to redundancy and volatility. Again. capital structure is constantly changing because the optimum target is in a state of continuous flux. For example. will violate this uniformity. However. we are doing periodic comparisons of the cost of equity: as long as we do uniform applications for each year. The long-term regression with long-term averages will be less volatile but less an accurate reflection of the current situation. than one who projects several current scenarios. Even in the Black. Jensen and Scholes . for example. the results will be comparable. When capital budgeting for a long project. Adapting the cost of equity to reflect current risk-free yields and market returns while maintaining the long-term beta is a combination that will bias the measurement in favor of the current market. allowing the first year to be a fiveyear average while using current market returns and interest rates in the second year. While each method may yield a different absolute percentage.121 example. if we have already used the current market rate for that year. A short-term beta would simply reflect the external risk of reaction to the market for that year only. That is: we can never mix one method with another or especially use different averages when comparing different years. Both of these hypotheses have merit. when we obtain beta from the characteristic line. any large or small correlation becomes an implicit part of risk. “R”. In a stock beta. THE DECOMPOSITION OF BETA (The reader is referred to the chapter on statistics and mathematics if the concept of covariance is confusing) Any beta coefficient can be broken down into the ratio COV(y. the x coefficient in a regression line which we term.e. In fact. and the degree of risk depends on all three factors. by the product of the standard deviations of x and y. While such observations must be taken in context. σ(x) and σ(y). Essentially.at least for this measurement. adaptive expectations can not only gauge the risk of a stock. as a measure of performance and a wholly short-term CAPM will yield volatile results that exhibit the increase in stock price for that year. reversion to the mean is typical for beta which seems to exonerate the rational expectations hypothesis . i. The expression is simplified as Rσ(x)σ(y). but they can define our position in the business cycle as well. The real difference in this area is confusion between the concepts of performance and risk. while “y” is equal to the distribution of an individual security’s changes. we can compare this figure to the figure that uses the proper five year (long-term) averages and obtain a measure of risk. any declining risk premium with a larger risk-free rate would indicate a market top.x) / Variance of x. A fund manager may use the required rate of return. When we do a long-term regression and modify the CAPM with current market inputs. R. sometimes markets are incoherent and lack even random logic. Thus. Any covariance can be obtained from a regression by multiplying the correlation coefficient. .122 study that observed the instability of beta over a long period. ex-post. we can look for a rise in the required return and vice-versa. the theoretical thirty-five year rate was closely correlated with actual returns. “beta” is made up of the covariance of x and y divided by the variance of x. Thus. If we are currently below the long-term figure. “x” is equal to the distribution of market return changes. but any stock chart can tell us the same thing. there is often a discernible pattern.. there is a net income factor in each stock price change. Moreover. The same expression reduces to: (R σ(x)σ(y). Return on the market) / Variance of the Return on the Market = ROE x COV(% ∆ Net Income. m) The most simple way to envision beta is thus. When a constant is multiplied by one element in a covariance. %∆ Net Income .COV(c x . the change in net income. COV(ROE x % ∆ Net Income.m). R market) / Variance of R market. this change in net income was derived from using total leverage as a predictive tool. If we substitute DFL and DOL for the degrees of financial and operating leverages respectively. he or she will remember that total leverage multiplied by the percentage change in sales yielded the percent change in net income. Return on a market index) / Variance of the return on the market index. This is equal to COV(x.m) + COV(y. there is some return on equity that is also implicit in stock price changes and that this constant could be factored out and then multiplied by the series of earnings changes.example . (COV(x+y. ) / σ(x) . The covariance of a sum COV(x+y) with another element. • 2. 2 . COV(Return on the stock of a company. m) = c COV(x. the research team of Mandelker and Rhee did much to provide the financial logic behind the decomposition of beta. “R market” is merely the return on the market and is not a reference to the correlation coefficient in the regression. They theorized that since stock prices are the present value of future earnings. In the mid 1980s. According to rule number two above. m)) is the sum of the covariance of each element of the original sum (x+y) with the other element. is equal to (DFL) x (DOL) x ( % ∆ Sales). it can be factored out and multiplied by the entire covariance .123 Two simple rules to apply to covariance are: • 1. If the reader will review the chapter on leverage. “m”. The change in sales and the change in operating income (EBIT) cancel out and yield the change in net income: (% ∆ Sales) x (%∆ Net Income / % ∆ EBIT) x (% ∆ EBIT / % ∆ Sales) = (%∆ Net Income) In a stable company with controllable leverage. Net Income(p) / Equity(p). The first assumption. the first and foremost being that the return on a stock can be decomposed into a return on equity and a series of net income changes. In this new model. ROE = Net Income(p) / Equity(p). R market) / Variance of R market. it is equal to [(Net Income(p)) x (Sales(c))] / [(Equity(p)) x (Sales(p))]. We eliminated decimal percentages by using the ratio of current year / previous year and then subtracting one (“1”). For the second time. and when multiplied by the ROE constant. was that the percentage change in net income could be derived by multiplying total leverage by the percent change in sales. To obtain this expression. Thus. while re-multiplying ROE back into the equation and eliminating decimal percentages. R market] / Variance of R market. That is a mathematical fact attributable to the nature of the components in the equation. which is backed up by leverage theory. We factor the leverage constants out of the equation. we reasoned that % ∆ Sales is equal to (Sales(c)) / (Sales(p)) . we again factor out ROE. is more controversial because . p = previous year and c=current year.1. we have two new subscripts.124 We can substitute this expression for the change in net income in the original covariance to yield: ROE x COV((DFL) x (DOL) x ( % ∆ Sales). This final expression is: (DFL(c)) x (DOL(c)) x (ROE(p)) x COV(Sales(c) / Sales(p). Mandelker and Rhee used theoretical assumptions. Notice that sales is no longer a percent change but the return on the market and variance of the market remain decimal percent changes. This expression is then divided by the variance of the market index. and again multiplied by the covariance of the periodic ratio of sales with the market index. however. The entire expression is now: (DOL)(DFL)COV[(Net Income(p)) x (Sales(c)) / (Equity(p)) x (Sales(p)). it is the current total leverage multiplied by last period’s ROE. We next do an exchange of constants. The second major assumption. R market / Variance of R market In effect. which becomes the first term in the covariance. Both DOL and DFL are constants as is ROE. Note that the ROE was the previous period’s and that the leverage is current. we can use these identified variables to gauge risk. the decomposition identifies the key elements of risk. the use of both total leverage and the growth rate of sales. By using the degree of financial leverage (DFL). Although the argument for a precise decomposition remains elusive. Moreover. From a capital structure perspective. Mandelker and Rhee made equity an implicit variable in the decomposition. on a pragmatic basis. While it may seem pretentious to attribute the return on a stock to wholly known components. both ROE and the increase in net income have greater long-run correlation with stock increases than most any other element with the exception of earnings per share. thus limiting the amount of equity issued. because the equation defines several important relationships among the components. some element will always be “out of sync” with the actual results of a regression.without being discounted at a specific cost. We need not submit them to actual beta equations because they exist in the domain of probability. the return on equity 3) ∆ Sales and its standard deviation 4) The degree of operating leverage 5) The degree of financial leverage . Producing a valid beta measurement then. Steady sales can allow more financial leverage to be used. Mandelker and Rhee gave us a series of significant variables and connected them in a logical manner. Therefore. The variables are: • • • • • 1) ∆ Net Income and its standard deviation 2) ROE. which is the cost of equity. the assumption of implied changes in equity is clear. might just require a long-term regression that contains the relationship of ROE and net income increases as prime components. Sales(c) / Sales (p) also contribute to the overall induction. That is a disposable circular argument .125 it assumes that a stock is worth its intrinsic value in net income relative to the amount of equity . Much of the practical application of capital structure theory comes from the conflict between these two elements . an implicit economic cost derived from several risk factors that incorporate market variance with an individual company’s performance. with the theoretical risk factors of the cost of equity.we develop those calculations in other chapters. adds a different dimension to that figure.126 • • 6) ∆ Return on the market and its standard deviation 7) The variance of the return on the market The reader should observe that the cost of equity is related to the changes in net income and existing equity. but is neither the return on equity (ROE) nor the percentage amount of growth in net income. On the other hand. net income is a derived deduction of accounting costs like interest and taxes from operating income. their relationship points to the concept of economic profit. It combines market forces with internal corporate dynamics to produce a risk factor entirely separate from income fundamentals. When the objective is to increase periodic economic profit. Equity risk is a composite of these other risks and when compared to actual net income. The smaller the standard deviation. These relationships are ex-post indicators of performance but ex-ante estimates of the amount of equity financing a firm can viably do. balancing the correct amount of equity with the risk of any changes in required 3 The acronym “EVA” and the term “economic value added are registered trademarks of Stern Stewart. Together. Inc. The essential observation is: the subtle connection between a sustainable return and the cost of equity. financial statement-profitability of net income. the cost of equity is an opportunity cost. and the development of variants such as “Economic Value Added ®3. Within our decomposition of beta were several types of standard deviations which stand as the generic statistical measurements of risk. the more sustainable is a flow of income. The key word is “sustainable”.the concrete. .that is: the percentage cost multiplied by common stockholders’ equity . While the decision to add or subtract equity is crucial to forming a total cost of equity . and “the capital dynamic”. AND THE CAPM One misconception that both students and financial professionals harbor is that the optimal capital structure is a function of cost rather than risk. the cost of equity would be the cost of diluting earnings per share by a specific amount. BALANCING LEVERAGE. then the risk of owning the stock should go up or down. a total of $500 in capital would be raised. RISK. These three elements are always in conflict: to create more income tends to require more equity which causes more risk when the cost of that equity gets too high. the cost of equity is (2% x 500) + 25 =35. In effect. because of flotation costs of -let’s say 2 % . However. If the company needed to expand and issued 25 more shares for a total of 125 shares outstanding. consider a company with earnings of $100 and 100 shares outstanding. Thus. there is a relationship between the cost of .an optimal level of risk may not be the least risk or the greatest risk but a level that is perceived by the investing public to be the best for the firm. The price of the stock is $20 per share and EPS would be $1. When measurable risk is at an optimum. But . To illustrate this concept. Pure logic would dictate that if a company changes its default probability. This myopic approach to pricing equity turns it into an accounting cost. besides flotation costs. Retained earnings would be similarly priced by the administrative cost differences between keeping it in an interest bearing account and managing cash. The subjective connotation of “best” is objectified by measurements from credit agencies. A firm that finances with “junk” bonds should have a very high cost of equity as well. and further substantiated by the present and projected cash-flow of the company.and a dilution value of ($1 -(100/125)) x (125 shares) = 25.127 return is counterpoised with the amount of net income. Cost is a function of risk. Indeed. they believe in “shopping” around for the lowest priced capital and then combining it with the right weights of equity and debt to achieve an optimum. While equity bears its own costs and risks. the cost of capital is minimized. those firms who choose to finance with debt have the added ability to affect the cost of equity by increasing or decreasing its risk. the CAPM could be used to draw the same conclusions as Miller/Modigliani. The effect of leverage on the cost of equity was now being comparably quantified. and managing default risk would determine the amount of leverage and have an effect on the cost of equity. the development of the CAPM allowed the freedom to measure an estimated amount of change in the risk of equity due to changes in leverage. His seminal article in the May 1972 edition of the Journal of Finance was entitled “The Effect of the Firm’s Capital Structure on the Systematic Risk of Common Stocks”. It essentially stated that the required rate of return on equity was a function of the amount of leverage.128 both debt and equity that we treat as deterministic for practical reasons. These extreme corner solutions eliminated the concept that capital costs were determined outside of the confines of the corporation. inflation and interest rates: leveraged companies were at the . Hamada told the reader by how much and by what mechanism. While the Miller/Modigliani propositions were postulated in the domain of many constraints. Therefore. the optimum proportion was one hundred percent debt. BETA AND LEVERAGE Robert Hamada took the relationship between beta and capital structure to new heights. The Miller/Modigliani proposition II stated that more leverage increases the risk and cost of equity. Essentially. the interest rate and the corporate tax rate. they were not dictated by the macro economy. which was correctly attributed to the link between exchange rates. or bankruptcy costs. Hamada developed the standard on the theory: not only was the principle set forth that more debt raised equity risk. capital structure was immaterial. a common observation was that more leverage in a publicly traded company led to greater volatility in the stock. only affected by it. which further substantiated their research: in a world without taxes. World events appeared to affect debt laden companies much more than unlevered ones. In fact. a combination of tax policy. but when taxes were added. For many years. A SIMPLE EQUALIZER To make Hamada’s analysis workable. Essentially. as it was on prevailing interest rates. Through a simple transformation of beta. a mathematical proof could now be set up that indicated how leverage increased beta. Given the restrictions of the CAPM. only four inputs were necessary. and secondly. Moreover. and even observe the risk in a stock that was attributable to operations (business risk) alone. the foundation is linear because the best estimate of extreme variation in either direction is a straight line. Hamada gave analysts an imprecise “ball park” figure for changes in equity risk due to leverage changes. A “ball park” figure was certainly better than none at all. and the increased number of variables that affected the flow of earnings made the stock more sensitive to changes. this new view of leverage emboldened the field of risk management. and now analysts could experiment with combinations of different capital sources and levels that purportedly minimized risk.from banks to foreign suppliers. it became apparent that the optimal capital structure was founded as much upon the effect of leverage on the cost of equity. beta had two states: an unlevered state in which risk would flow mostly from operations because the company had no debt. Although beta is affected by myriad other variables and a precise determinant is elusive. They were: • • • • 1) The current beta of the stock which was available through regression 2) The market value of debt 3) The market value of equity 4) The current effective corporate tax rate These variables were combined in the following expression: (1 + [(1-tax rate) x Market value of debt / Market value of equity]) Thus. Again. the analyst could observe how much risk was added to the stock after a certain amount of debt was increased. the unlevered beta became a function of the current beta divided by this expression: . a leveraged state that reflected changes in the level of debt.129 mercy of a “creditor chain” . as well as how assets are structured in the firm. One would expect inordinately high or low betas to foreshadow the inability to make payments on interest. The steady cash-flow that is implicit in such coverage is an outgrowth of operating risk and the amount of debt already incurred. beta is not a measure of performance but of volatility. Banks charge interest based on credit worthiness which stems from the ability to cover a loan. While more debt may lead to higher interest payments. a firm that is near bankruptcy when the market surges may have a stock that languishes. Since more debt increases the probability of default. does it follow that the cost of debt determines capital structure? Only so far that interest rates are a “perfect” transmission mechanism for both government policy and the probability of default. the “real” cost of debt is the cost of bankruptcy which is attributable to the interface between operating income and the amount of interest expense. The relationship between beta and the probability of default has not been well developed. As in Mandelker and Rhee’s . higher interest rates alone will not lead to more equity risk. The inability to make payments may stem more from a lack of income generation than it does from too much leverage. it raises the cost of equity. While the cost of debt is a contributing factor through the effect of interest rate changes in the macro economy.130 Unlevered beta = Current beta / (1 + [(1-tax rate) x Market value of debt / Market value of equity]) The levered beta became: (Unlevered beta) x (1 + [(1-tax rate) x Market value of debt / Market value of equity]) A change in the beta due to more leverage became: (Original beta) / (1 + [(1-tax rate) x Old Market value of debt / Old market value of equity]) x (1 +[(1-tax rate) x New Market value of debt / New market value of equity]) If the risk of equity is dependent upon the amount of debt. While more debt may increase both beta and the probability of default simultaneously. However. many high debt companies that have steady incomes. sales that are languid could counterbalance the effect of higher leverage. the “alpha” variable in the regression would takeover and become very low. Analogously. these firms will have large differences between “expected” and “required” rates of returns. of course. Even when beta is up. attracting more equity capital into the company. a stock with too low a beta whose firm is not out producing their respective industry may under perform the market because too little risk is being taken.finding out how much equity risk it has without any debt. On the other hand. many firms that have more leverage than warranted by their betas will have naturally riskier stocks and become the fodder for speculative betting on Wall Street. can increase beta at the same time that the market is picking up. The happy medium. these companies see little rise in the probability of default even when they increase debt. Thus. making beta non-reactive. By the time the next downturn occurs. They will also offer the most return for the least amount of risk. The strategic implications of Hamada’s research (and many others in that same period) should be apparent. cashflows improve. is the firm with a low beta stock who is indeed out producing their sector. In that case. “R”. will have low unlevered betas. would be low as well.131 decomposition of beta. diminishing some of the higher cost of equity that would have occurred because of higher interest rates and a surging market. the stock would still under perform the market and the correlation. the proportion of debt to equity will have decreased. like many tech stocks. can afford little debt before the stock becomes prohibitively risky. CONCLUDING COMMENTS . Any firm that is wealthy enough to afford the lower interest debt financing that is offered at the beginning of a recovery. like utilities. mostly because they are heavily regulated. When the investment in debt begins to pay off. The effect of leverage on beta is most observable when we unlever a firm’s beta that is . Firms that have a comparatively high-unlevered beta. 5) It establishes the risk of a stock in comparison to the overall market. 6) It establishes a cost of equity for the company. Without understanding and seeking out alternative analyses. 2) It can measure operating risk and any additional risk derived from the combining of assets. That is a risky proposition. The watchwords in any analysis system are always corroboration and balance. (Back to Table of Contents) . the investor is “putting all of his or her eggs in one basket”. • • • • • 3) It establishes the required rate of return on a firm’s common stock. • 8) It can be used as a tool in portfolio management to diversify away risk .132 Any model that purports to be as comprehensive as the capital asset pricing model (CAPM) needs to be used with judicious caution. 4) It can be compared to the “expected’ rate of return and used as a “buy/sell” signal. by any standard. 7) It can be used to gauge the business cycle as well as the differences between risks in the equity and fixed income markets. it can resolve several issues: • • 1) It can measure the additional risk caused by incurring more financial leverage. When it is accurate. the net operating income (NOI) will be the same for each situation. Another assumption is that beta is stated in terms of ROE rather than the stock price. the optimal structure is one of all debt. and conclude that they are derived from the amount of leverage. Secondly. we use these constructs to exhibit the difference in the cases of both taxes and no taxes. First.133 APPENDIX: THE MATHEMATICAL RELATIONSHIP BETWEEN LEVERED AND UNLEVERED COMPANIES IN TERMS OF THE CAPM The capital asset pricing model corroborates the Miller/Modigliani propositions. making beta a deterministic function of net income and equity will display a cause and effect relationship. Thus. . However. when taxes are applied. capital structure is impertinent. In this section. The certainty that we impose helps to exhibit possible trends. 1) EQUATING THE LEVERED FIRM WITH THE UNLEVERED FIRM Students tend to interpret this proof as referring to two different firms. firms that fund with equity have an equal amount of risk as firms who fund with debt.with debt or without it. Since the stock price will tend to mirror ROE over the long-run. In the case of no bankruptcy and no taxes. but in the unlevered state. but the return on equity (ROE) will be different because in the levered state. It is more enlightening to think of the comparison as the same firm in two different situations . tax deductions on interest give a distinct advantage to leveraged firms as long as there is no probability of bankruptcy. it does not. although we fully realize that beta and stock price have a more volatile and probabilistic relationship. we establish the differences between beta in both leveraged and unlevered companies. the firm has to pay interest. we examine the Miller/Modigliani arguments from the standpoint of the CAPM. Therefore. This is the same as COV(ROE(U). STEP 4) Determine the beta of the levered firm.rB(L) / S(L) Table 5-3 UNLEVERED FIRM (U) ROE(U) = Return on Equity for the Unlevered Firm NOI = Net Operating Income S(U) = The Value of The Stock of the Unlevered Firm. We merely broke the expression into two parts. This is more complex because we now express the levered firm in terms of the unlevered firm (steps one and two above). Remember that R market is the return on an appropriate market index. R market) / Variance of R market. Next. First. This can be substituted into the NOI of the levered version of the company so that unlevered and levered can be equated. we eliminate the constants because they have a value of zero in the covariance. STEP 3) Determine the beta of the unlevered firm. rB(L) is Interest Expense. we multiply the variables within the covariance by the term S(U) / S(L) to equate .134 Table 5-2 LEVERED FIRM (L) ROE(L) = Return on Equity for the Levered Firm NOI = Net Operating Income B(L) = Amount of Debt for the Levered firm S(L) = Value of Stock of the Levered Firm r = Interest Rate on Debt. we use the additive law for covariance to combine the entire expression into one covariance. ROE(L) = NOI . S(U) = V(U) = Value of the Unlevered Company ROE(U) = NOI / S(U) = NOI / V(U) STEP 1) Note that NOI = (S(U))(ROE(U)). R market) / Variance of R market. Thirdly. We are using the unlevered ROE rather than the stock price and so: beta is COV(NOI / S(U). Remember that beta is a covariance divided by the market variance.[r (B(L) /S(L)]. giving it the common denominator of S(L). STEP 2) The levered firm is now [(S(U))(ROE(U)) / S(L)] . R market))(1 / Variance R market) Notice that in part d. The reader will remember that for the unleveled firm.rB(L)) / S(L)) . R market / Variance R market = (1 / Variance R market) (COV((NOI / S(L). (S(L)). If the levered beta is a given. Both betas are equal by a constant (S(U) / S(L)) that reflect the difference in financial leverage. S(U) = V(U). R market)) (1 / Variance R market) is the beta of the unlevered firm that we stated in step 3. Now BETA(L) = (1 / Variance R market) (COV(NOI /S(L). this smaller amount of stock will be the determining factor. Since the levered firm issues debt instead of stock to raise the same amount of capital. we factor out this same term (S(U) / S(L)) from the covariance. the leveraged beta is the same as the unlevered beta if we multiple the unlevered beta by a factor of S(U) / S(L). R market). We eliminate COV((rB(L) S(L). the term NOI / S(L) will remain the same. making the levered firm’s beta always greater than the unlevered firms . This figure will take on some integer value greater than 1. b) Eliminate one of the covariances because it is a constant. We have broken the expression into two separate covariances. Thus. That is: the value of the unlevered firm was totally dependent on the amount of its stock. The expression (COV (NOI / S(U). a) Using the additive property: BETA(L) = COV(((NOI . merely by multiplying the unlevered beta by a factor of the ratios of market values (S(U) / S(L)). we can factor out S(U) / S(L) so that now BETA(L) = S(U) / S(L) (COV(NOI / S(U).135 the unlevered beta with the levered beta. R market) d) By the law of covariance factoring. . R market). we divide it by the same term to yield an unlevered beta. we can determine the beta of the levered company.by a factor of the amount of leverage. and so we are comparing a whole value (S(U)) to a partial value. Fourthly. R market))) (COV(rB(L) / S(L)). R market) c) If we divide NOI by S(U) and then multiply the term by S(U) / S(L). That is BETA(L) = (1 /Variance R market) (COV((NOI / S(U) x (S(U) / SL)). In this instance “E” is not multiplied by ROE but is the expected return thereof.r) / RETA(U) = [E(ROE(U) . We further establish two premises. We obtain: (E(ROE(U) . “E” is the expected return. divided by the firm’s beta.r) / BETA(L). • Premise 1: (E (ROE of a firm) .136 2. while (S(U) / S(L))(BETA(U)) is substituted for BETA(L). • Premise 2: (E(ROE(U)) . In this expression. on the right side of the equation by algebraically multiplying by its inverse. S(L) / S(U). S(U) / S(L).rB(L) / S(U) .recall that we have established two substitutions: • • A) BETA(L) = S(U) / S(L) (BETA(U)) B) Since NOI = (ROE(U)) (S(U)). while “r” is equal to the risk-free rate. PROPOSITION I In proposition I.r. Again (ROE(U))(S(U)) is substituted for NOI.r) / BETA(U) = [(E(ROE(U))(S(U)) / S(L) .r) / Beta of a firm = E(ROE market) . is equal to the expected value of the return on the market. This assumes that the expected return of the return on equity of any firm. Miller/Modigliani basically stated that capital structure did not matter in a world without taxes or bankruptcy.r) / BETA(U) = (E(ROE(L)) .rB(L) / S(L) . (E(ROE(U) .rS(L)/ S(U)] / BETA(U) . then ROE(L) = [(ROE(U)(S(U)) / S(L)] -[rB(L) / S(L)] which essentially states that the difference in the two ROEs is because interest expense (rB(L)) is subtracted from one and not the other. This states that the expected value of any security divided by its beta is equal to the expected value of any other security divided by its beta. equating the market returns with the individual company returns through the mechanism of beta.r] / (S(U) / S(L))(BETA(U)) STEP 2: We eliminate the term. STEP 1: We substitute for ROE(L) and BETA(L) in premise 2. USING THE CAPM TO CORROBORATE THE MILLER/MODIGLIANI PROPOSITIONS To review the last section . tax rate).137 STEP 3: We eliminate terms and multiply it through: .r = [(.r S(L) / S(U) STEP 4: We multiply by -1 and then factor: r((B(L) / S(U)) +( S(L) /S(U)) = r or alternatively. Either a higher tax rate. In the tax case. or more debt.r = . then V(U) = V(L). PROPOSITION II In proposition II.tax rate) B(L) = S(L) or alternatively. S(U) = S(L) + B(L). Capital structure makes no difference in the valuation. or both.[rS(L) / S(U)] STEP 3.tax rate) rB(L) )/ S(U) ] .r B(L) / S(U) . STEP 1: ROE(L) = [(ROE(U))(S(U)(1 . the optimal capital structure would be made up of one hundred percent debt. (Back to Table of Contents) . we go through the same set of equations except that NOI (and the substituted variables) are multiplied by a factor of (1 . S(U) + (tax rate)B(L) = B(L) + S(L) STEP 4: Applying logic: Since S(U) = V(U) and since V(L) = B(L) + S(L) then V(L) = V(U) + (tax rate)B(L) The value of a leveraged firm is equal to the value of an unlevered firm plus the product of the tax rate and the amount of bonds.(1 . Reducing the expression and eliminating “r”: S(U) = (1 .tax rate)r B(L) / S(L)] STEP 2: Multiplying the equations through yields: . Since the value of the unlevered firm is equal to its stock (V(U) = S(U)) and the value of the levered firm is equal to its stock and bonds (S(L) + B(L)) = V(L).[(1 .tax rate))/ S(L)] . and so the value is maximized at one hundred percent debt. Miller/Modigliani argued that in a world of taxes. will increase the value of the firm. At first. THE PERCENTAGE TRAP Although it is more mathematically sound to use logarithms rather than percentages as the measurement of an increase. However. the upside return occurs “early in the game” when the risk is greatest. and the firm is very close to its objective. several difficult “hurdles” must be negotiated. Wall Street rewards those companies that have successfully overcome obstacles at a point where the greater investment community is unaware of their occurrence . the risks of not achieving it are greater than when the achievement is “a done deal”. Finally. If this statement seems paradoxical. earnings stock price. creating short-term hyperbole rather than scientific exactitude.before the knowledge of profitability becomes commonplace. In fact. their raison detreَ stems from creating distance between net income and the cost of equity. they are somewhat akin to an American using the English concept of “foot” rather than the metric system. expectations are increased by analysts or a target would not have been originally set. Fortunately. etc. Percentage gains are the standard in business primarily because they are understandable in a comparative sense. the financial community prizes the transitional progress toward a goal more than its actual achievement. their own dependence on reference allows their bases to shift. while it appears that there is substantial downside risk from not meeting objectives. . and not from the time-dependent requirements of maximizing a function.sales. when a company is a great distance from the target. this text submits to the percentage imperative. at this point.138 6 THE CAPITAL DYNAMIC AND OTHER TOOLS This chapter will provide the student/investor with some practical tools for detecting movement toward an optimal capital structure. Next. While mathematical optimization remains controversial and inexact. these measurements exploit correlation and probability. consider the risk factors involved when a company is a long way from reaching any goal . we do not sell investments or tout earnings increases. forms the backbone of capital structure measurement.both new issues and paid in capital. for example. a nickel increase on a nickel investment. and common stock . it is only a 20 percent decrease. preferred stock. simply because the reference base has shifted. WACC changes can tell as much about the direction of the company as earnings. percentage gains are a simple method of enumerating change. Indeed. In capital structure analysis. Those component costs may include long-term loans. In fact. it is a 25 percent increase.139 Consider as an example. when an investment goes up from 80 to 100. while ten million dollars on a billion dollar investment is “only” one percent.become the building blocks for changes in more prominent measurements. THE WEIGHTED AVERAGE COST OF CAPITAL Introduced briefly in the chapter on capital structure. each capital item needs to be specified as a separate cost. In fact. retained earnings. bonds. peaks and plateaus as well as any economic barometer. percentage changes in these “background fundamentals” – opportunity costs . but if it goes from 100 to 80. For generic estimation. however. The proportional component cost of each source of capital in aggregate makes up the WACC. will have much less meaning to the average businessperson that a twenty percent gain in the cost of goods sold. without a need for reference. a twenty percent gain in the cost of equity. we have grouped long-term debt and stockholders’ equity together to form our definition of “capital”: an investor who needs to compare several . On a technical level. the weighted average cost of capital or WACC. percentage gains are relevant because absolute size matters less than the movement itself. For that very reason. Although it does not always adhere perfectly to its theoretical underpinnings. the WACC can predict slowdowns. and it is essential that this cost be minimized. such as would occur in capital budgeting. Dissociated from emotional content. capital leases. when used in tandem with earnings measurements and information about the economy. such an increase can be touted as a “one hundred percent gain!”. The liberties that are taken become a practical extension of turning a theoretical “opportunity cost” into a practical decision-making tool: the cost of equity is removed from the contingencies of accounting. for example. . and so the risks have to be defined by model interpretation. Together. at times when the Federal Reserve raises rates. Most companies are risk averse enough to perform at a level substantially below that threshold. a corporation who precisely itemizes its cost of capital needs to ensure implementation and return. To resolve such problems of interpretation. For example. the minimum cost of capital is also theoretical. the major concern for the investor is to observe the changes in the WACC and not worry about a theoretical absolute. but more of a need for comparing alternatives. Some of the risk will always remain subjective. is more concerned about the risk of installing them: there is less need for detailed knowledge. we relate the WACC to other capital structure variables. not just risk and return. A bank. and compare it to what companies of similar risk would yield. each entity uses the weighted average cost of capital to form the comparison. Corporations will derive an internal rate of return and compare it with the discounted cash-flow of a project. however. Once the differences in methodology are discarded.140 companies in rapid succession needs “best estimates” rather than technical precision. Therefore. although we try to eliminate it by defining as many variables as possible. since a competitor’s WACC rises as well. Thus. the WACC may rise although the firm has minimized its capital costs. the implementation of any project has massive logistics problems on a relative level. The investor. the increase is entirely relative. Investors will observe the return from a company. On a pragmatic level. and almost completely dispense with the observation of its individual movement. The tax effects of interest deductibility mandate the use of debt up to the point when a company’s risk manifestly begins to undermine its stock price. must do a complex itemization of cost effectiveness when it installs automatic tellers. the decision vehicle will be similar. At small levels of debt. and greater proportion of the debt curve. Once the WACC hits a minimum. debt becomes more expensive. yield curves become inverted and short-term interest rates are higher than longterm rates. Thus.141 THE THEORETICAL SHAPE OF THE WEIGHTED AVERAGE COST OF CAPITAL CURVE Combining the linearity of the cost of equity with the upward trending of the cost of debt forms a convex curve. A realistic rendition will observe the sometimes absurd divergence between bond and equity markets: at times. and it begins to curve upwards to reflect the steeper slope. The risk premium between the markets may evaporate for long periods and equities will have great volatility with only a minimum of return. At other times. Figure 6-1 Cost of Equity Cost of Capital WACC Cost of Debt Optimum D/E Debt / Equity The shape of the curve is wholly dependent on the spread between the respective costs and risks of equity and debt at each level. the WACC curve slopes downward to reflect the flatter slope and greater proportion of the equity curve. Capital structure optimizes at the inflection point and this is also where the stock price theoretically maximizes. the risk premium moves in the opposite direction as investors sell bonds and buy stocks. more debt actually decreases the cost of capital. over . ACCELERATION RATES Are higher earnings and a downward trend in the WACC a signal to invest? While such a cross current may lead to an optimized capital structure. A firm with a low probability of default can most benefit. However. When market conditions change. The change in this comparison cost will confirm or deny movement in the cost of capital. the WACC may rise when it is supposed to fall and vice versa. the rate of increase in earnings far outpaces the rate of the WACC and stocks consequently rise. the WACC curve shifts up or down. To corroborate its movement. in the rare occurrence of Federal Reserve rate cuts. In the meantime. firms who can still afford a lot of debt have the opportunity to decrease both the costs of debt and equity simultaneously. To understand the need for a combined capital structure measurement. consider the theoretical relationship between debt and equity. that investment decisions can be made. and the Federal Reserve raises rates to contain inflation. the cost of capital mirrors the inherent risks of each component source and forms a hypothetical compendium. a comparison cost ifs formed. Once the market is back in equilibrium.142 the long term. It is this disparity that creates opportunity in the equity markets. When the WACC is multiplied by the amount of capital and then subtracted from earnings. and it is in this “vectoring” context. the information is too spotty to make a decisive judgment. several relational variables allow the examination of a directional flow. the analyst uses both earnings and changes in the proportions of capital sources. the cost of capital becomes higher or lower at each level of debt to equity. when more debt is incurred. corporations . changing the optimum mix. and the hypothetical relationship is readily restored. for example. both earnings and the WACC may rise together because the market gets better. During a recovery. but will lag the rate of earnings in either direction. the cost of equity rises. While short-term volatility prevents the analyst from pinpointing an absolute minimum cost of capital. In effect. What is of primary importance is the acceleration of earnings in comparison to the WACC. the cost of equity rises as more debt is incurred. the firm needs to balance the tax advantages of debt with the probability of default. a company must consider the probability of default first and foremost. When debt is especially inexpensive. there is usually a compensatory downside that creates forces in the opposite direction. common equity and retained earnings. the type of assets. It may show positive or negative trends. a firm may load up on a large amount of zero coupon bonds. Only when a firm balances the stability and amount of earnings. As long as the cost of debt is lower than other sources of capital. the WACC will optimize at a capital structure of all debt. THE MECHANICS OF THE WACC: RISK ADJUSTMENT When the component proportions of all sources of capital are multiplied by their respective costs. Thus. While some see opportunity in the various economic “bubbles” that arise. the WACC cannot be minimized on the basis of its own parameters. a weighted average cost of capital is formed. and the WACC would diminish. and the ability to make prompt payments. the WACC tells the analyst to load up on it because it is the least expensive. but is not the final arbiter of corporate action. the tendency to “go overboard” is justified by a lower WACC. After calculating individual costs. however. will an optimal capital structure arise. the crux of the equation revolves around the proportions and the definition of “capital”. The smaller .143 can shift their mixes of debt to equity to take advantage of the distance between earnings and the WACC. Since interest rates change frequently. not factoring in the potential distress of payment in the future. the WACC is an indicator of risk. For example. they will include both preferred stock and short-term notes in the capital denominator as well as long-term debt. Without the context of potential bankruptcy. For that decision. THE WACC AND RISK The market does not always price risk efficiently. To properly gauge risk. simply because there are no upward bounds on it. Corporations and large investors need a complete itemization of each source of funding. but it is not definitive. and will rarely be above the cost of equity for a solvent company. and will lead to gross errors if used in the context of capital budgeting. and less with implementation of a project (or a controlling stake) expediting the WACC will create utility. may want to make quick comparisons and use just the latter three components. AMOUNT 1000000 700000 70000 140000 490000 300000 250000 50000 0 PERCENTAGE 100 70 7 14 49 30 25 5 8% 10% 12% 7% 4% 0 COST CORPORATE METHOD For the sake of illustration. However. and both long-term debt. and common equity will be a larger proportion.144 investor. Consequently. most firms diversify away risk by funding from as many different sources as possible to avoid this exact type of “corner” solution. it would be proper to exclude them from the capital denominator because they . however. The reader is referred to previous chapters for the methodology of determining individual cost. Had they been enumerated. Table 6-1 TYPE OF CAPITAL ASSETS STOCKHOLDERS' EQUITY Preferred Stock Retained Earnings Common Stock DEBT Long-term Current Liabilities Notes Other Current Liab. observe the following balance sheet where each component is itemized by a pre-tax cost. To contrast the two methods. this “risk adjusted” WACC is not the true WACC. other current liabilities are assumed to be zero. the denominator will be smaller. Since the smaller investor is more concerned with risk. long-term debt and equity are the major risk factors and the contraction of the denominator will recognize this. 25(0.2 %. common equity and long-term debt.68 28. then the proportion of common equity would be (490000 / 975000) or 50. RISK ADJUSTED METHOD We change the capital base to the sum of retained earnings.1026 or 10.05 (0.9 55. debt is calculated as (interest rate) (1-tax rate). The other costs are multiplied by their proportions and then summed (costs are in parentheses) . had other current liabilities been 25000.14(0. the risk adjusted WACC is substantially higher. and notes 25000.284(0.1) + 0.5568(0.4 100 COST 10% 12% 7% The calculation is as follows: 0. Assuming a tax rate of 30 %.04) = 0. all other sources of funding would rise in proportion.7 x 0.09205 or 9.26 percent. The component proportions of these elements will rise: Table 6-2 TYPE Retained Earnings Common Equity Long-term Debt Total Capital AMOUNT 140000 490000 225000 880000 PERCENTAGE 15.08) + 0.7 x 0.07(. In that case. however.25 % and not 49 %. For example.56 % as a component proportion.12) + 0. Such a bias will inflate the importance of long-term debt and common equity in the capital structure which is its intended purpose. The full capital budgeting type of analysis yields a 9.145 are a source of internal and not external funding. In .159(0.07) = 0. Not only will changes in the economy also change the target mix of debt to equity. THE MARGINAL COST OF CAPITAL Few modern corporations can attain an optimal capital structure and remain there for any significant length of time. Notes. would go down to (25000/975000) or 2.49(0.1) + 0. but the firm can be penalized for taking too little risk.12) + 0. Thus.2 % WACC.07) + 0. However. Moreover. most investors will demand equity when they see sales and profits growing. and then plan to find a source of funding if needed. For example.146 effect.that is-each additional dollar of capital raised will be more expensive. The challenge at this point is perhaps one of the most difficult in business: to begin generating a profit when the infrastructure is new and untested. when combined with the amount of earnings. if the target structure calls for 35 % debt and 65 % equity. Knowledge of this limit can help a CFO gauge the feasibility of capital projects. may take a longer time to integrate. it is cost effective from an administrative perspective to raise capital in large increments because such inflows are often purposeful and act to focus management on an objective. Moving off the target will raise the marginal cost of capital . but will provide greater potential return if the risk is negotiated well. and capital budgets. creating stagnation and uncertainty. such rapidity implies a return for the extra risk that the firm incurred. . It is at this juncture when short-term performance is sacrificed for long-term gains that careers are made or broken. then 500 / . In the mean time. large capital infusions. the company will move away from its optimal structure from time to time by taking on inordinate amounts of risk. But . the CFO can set dividend policy. some of which will supply capital through retained earnings. management has to act in a temporary capacity not to maximize the price of the stock. The process can take up to three years or more but will come to fruition with consistent performance.35 = 1429 can be raised in total capital. The solution to this problem is to begin movement back to the target capital structure which can be taken in increments.it is much easier to move off target and then back again than to be constantly constrained by the amount of earnings and capital. and then ideally. and the firm has 500 million in debt. In effect. whether in debt or new stock. move back toward the target as rapidly as possible. management needs to be aware of its target capital structure and that a higher WACC will be implied when the firm moves away from it. the investor needs to have faith in future projects enough to warrant remaining a shareholder. 5) New issues can entail an ongoing obligation of dividend payments. and then subtracting that figure from net income: (300 . The first question to ask is “How much funding can be done on the basis of retained earnings alone?” Had the retention rate not been given. 4) New issues can dilute control from existing shareholders. It is at this point when new stock is issued. 3) New issues can raise the cost of equity through flotation costs. and the investor must deal with at least five diminishing characteristics: • • • • • 1) New issues can dilute EPS.147 For investors. 2) New issues can dilute market price.10 10% 68.18 (Million) 300 ( Million) 75% This example brings several issues to the forefront: it shows the relationship between the variables as well as the advantage of knowing the target capital structure. DECISION MAKING AND THE MARGINAL COST OF CAPITAL Consider a firm with the following limitations: Table 6-3 FUNDAMENTAL Optimal Target Next Dividend Growth Rate Shares Outstanding Net Income Retention AMOUNT 40 / 60 Debt / Equity $1. it could be derived by multiplying the shares outstanding by the dividend. Realizing that excess capital is being raised (or that retained earnings are inadequate to meet the target). the most important breakpoint comes when retained earnings are exhausted and debt is also high. and new tax advantages will be needed to balance it. Rather than meet stringent requirements on capital funding. This scenario will lead to eventual earnings disasters and should be avoided. To determine the total amount of additional funding that can be done.(68. In fact. In terms of bankruptcy costs. At this juncture. This amount of retained earnings is in addition to the amount already retained in stockholders’ equity. ration capital and under fund some projects. Issue stock. Fund the shortfall with debt. The options are as follows: • 1. which will imply a new level of optimal target. To observe the potential dilemma of being constrained by the marginal cost of capital. Moving past the target in any given year will require a counter movement in the opposite direction. Thus. • 4. However.148 . and may reduce the share price of the stock. • 3.6. the cost of capital goes up because flotation costs have been incurred. (375 . 0. By cutting dividends. the extra capital could have been raised through retained earnings. the total amount of additional funding is 225 / . • 2.06 = 375. consider a scenario where 390 and not 375 million was required. but more earnings are retained. This strategy will also lead to a diminished stock price. This strategy will move the firm off its target. This relatively small amount of 9 million will keep the company near its target level. we divide the 225 by the target proportion of equity. the student/investor should notice that had net income been greater.18 x 1. the amount of loss rises. most firms can create strategic movement toward the optimal target on a perpetual basis.10) = 225. Cut costs. and the firm still moves off its target structure. more debt would have been raised .225 ) or 150 million would be in new debt. Out of this amount. not only are expectations lowered. Such a strategy allows flexibility in capital funding because opportunities will arise that may require inflows greater than the amount of financing that is condoned by the optimal target. while raising WACC. Only 60 % x 15 million has to be raised in newly issued stock. Cut dividend growth to 7 or 8 %. is subtracted from the earnings derived from the . Similarly. Fortunately. and a firm may need a compensatory amount of capital from one source to move back toward the target level. conditions in the industry may change that will abruptly change the optimal mix as well. small issues of balancing equity can both preserve stock price and fund capital shortfalls. Thus. it will need much more equity than debt to make up the shortfall. Why issue stock instead of raising debt? Obviously. there is an analysis system that can both detect changes in the cost of capital as well as movement toward the optimum . and less people want to own a stock unless they are sure of a payoff.acquisitions. both strategies will move the company away from its target. economic profits are configured differently. Mergers are uncertain.the theory of economic profit. Such a large issue has strategic value . but raising a small amount of stock is less disruptive. However. those constraints are hardly realistic. ECONOMIC PROFIT. The market will reward a company for issuing large amounts of debt because generating profits will consistently move the company back toward its target. firms would be at their optimal targets and continually fund at that level. EVA® AND THE CAPITAL DYNAMIC: UTILIZING THE OPPORTUNITY COST While accounting profits are made by subtracting costs from revenues. This balancing act is much more difficult than it appears: not only will the firm need to look at alternative sources. small issues of debt are looked on as “moves of desperation”. the tax advantages will be less than the increase in bankruptcy costs. a large stock issue can undermine the market price through dilution. to keep the company solvent. large projects. In a perfect world. even leveraged buyouts. For example. the amount given up by pursuing one course of action over another.149 as well because the increase in net income would have lowered the probability of default. An opportunity cost. However. if the optimal target calls for a 50-50 mix of debt to equity and the firm is at 60 percent debt. 150 original course of action. Its unique approach enabled many major corporations like GE. . its economic profit is growing because the opportunity cost (what is given up) is so small. calling it “EVA®” or economic value added4. because that is the figure that would be made on alternative investments. several other deductions will be made at this time as well. Therefore. if one firm out performs those of similar risk. and the measurement never competed with the more simplistic “P/E” as a favorite of mutual funds and individual investors. EVA calculations can be very complex . the tax advantages of debt become implicit in the WACC. requiring knowledge of tax law and accounting skills. ELEMENTS IN AN EVA CALCULATION Users need to derive a figure called “NOPAT”. and corporate tax law. we take EBIT (earnings before interest and taxes) and deduct just taxes from it. and we end up comparing those figures after multiplying WACC by the amount of capital. and deriving a coherent NOPAT will require knowledge of itemized deductions. EVA becomes a serious management tool when used properly. to not only build capital but to compensate employees based on improvement in the measurement. Thus WACC becomes similar to 4 EVA is the registered trademark of Stern Stewart. In its simplest form. Inc. However. The firm. economic profit theory occupies a hypothetical “middle ground” between cost / benefits analysis and risk / return analysis and uses elements of each. pioneered the concept of economic profit in a practical “hands on” accounting environment. If I am a stockbroker and I give up a career in medicine. The comparative opportunity cost is the product of WACC and capital. A T & T and Coca-Cola. which is an acronym for “net operating profit after taxes. Thus. leaving interest untouched. it is a comparison cost. Stern Stewart. In most companies. In the case of economic profit. The second variable in EVA calculations is the WACC (corporate version). I need to subtract the cost and income of being a doctor from my own. Without interest deductions in NOPAT. This step assumes no other tax deductions except interest. The tax . Derive WACC = [(Interest Rate x (1-Tax Rate)] x (Component Percentage of Debt) + [(Cost of Equity) x (Component Percentage of Equity)] = [(.[(0.[(Tax Rate)(Operating Income)] = 145 .6897)(0.9645 THE CAPITAL DYNAMIC Without being versed in manageable deductions.2857)] + [(0. In effect.151 ROC (return on capital) except that it is applied to companies of similar risk .03 % 8% 10% 700 200 500 STEP 1. Derive NOPAT = (Operating Income) . The concept of economic profit remains relevant. STEP 3.87193 or 8.[(WACC) x (Capital)]= 100 . the cost of equity.the cost of equity.(0.0355 = 38.872) x (700)]= 100.3103)(145) = 100 STEP 2. This next example gives a step by step rendition of a simple EVA calculation: Table 6-4 FUNDAMENTAL Operating Income (EBIT) Tax Rate Interest Rate on Debt Cost of Equity Capital Debt Stockholders' Equity AMOUNT 145 31. and the value of stockholders’ equity. the simpler framework allows less room for error.= NOPAT . the equation becomes: Net Income . an investor can form this same result using just three variables: net income. This adaptation allows the investor to quickly extract these figures from financial statements and then determine only one cost .1)(0.61.72 %. Calculate EVA. and the investor can focus on obtaining an accurate cost of equity.08)(0.[(Cost of Equity) x (Stockholder’s Equity)].through the cost of equity component.71249)] =0. 6897) = 88. neither does the investor need to be privy to the latest negotiation over interest rates (the risk premium that is attached to the risk-free rate).97.6897 is a figure for (1 . 10%.Tax Rate). which is the same figure as for EVA. and the investor will not have to itemize the different interest rates with corresponding debt maturities.(16)] x (0. Finally we subtract 50 from 88. Next we derive a product of the cost of equity. ECONOMIC PROFIT AND CORRELATION . the capital dynamic becomes an investor friendly version of economic profit. and stockholders’ equity. Interest expense of 16 is calculated as 8 % of 200. In the above example. 500: (0. because deductions are not itemized as in NOPAT. However. The investor forces the book value of debt to equal its market value which gives the capital dynamic less resilience as a predictor than EVA.97. While EVA more accurately reflects the current cost of capital and the WACC. net income is derived as ((145) . The capital dynamic offers the investor a legitimate comparison between two main sources of risk: net income growth and the size and volatility of equity. However. THE RELATIONSHIP BETWEEN EVA AND THE CAPITAL DYNAMIC EVA and the capital dynamic will be the same figure when the interest rate on current debt matches the actual interest expense that is paid out. Analogously.1)(500) = 50.152 advantages of deducted interest expense become implicit in net income. the capital dynamic better reflects the current investment outlook because income growth is inherent in the calculation. managerial control is less apparent in this simpler structure. nor does he or she need to be mired in intricate calculations. the typical investor is much less concerned with accurate capital budgeting which would be a major concern of the corporate EVA practitioner. The 0.97 and derive 38. or by examining the growth rates of each component. and especially. However.that is . the result is “MVA”. the cost of equity. and not just because the earnings component accounts for so much of the variation. In effect. Thus it is not too far reaching to make a connection between EVA and stock price. higher equity diminishes beta. MVA is the difference between market values of capital and their book values. market to book value is highly correlated with stock price. In the capital dynamic. The WACC captures the interface between corporate risk and the state of the economy. any economic profit formula will not optimize capital structure mathematically . the effectiveness of the function is derived by observing one or more variables declining while the other rises. Stern Stewart trademarked another concept called MVA® or “market value added. EVA analysis is not as popular as some other systems such as the “PEG” ratio (price-earnings growth). higher interest rates that are implicit in the cost of equity may diminish net income. it is highly correlated with stock performance.in a deterministic fashion. the similarities between stock price and EVA can be observed. the correlation value is so great that improvements in the measurement can be read as movement toward the target mix. each variable is affected by the other: net income increases equity through retained earnings. there is high correlation between all three variables. As researched by the team of Fama and French. Nevertheless. To display the correlation between EVA and stock price. When all future EVAs are discounted into the present at the cost of capital. Inc.153 Despite the prodigious efforts of Stern Stewart to educate investors. equity risk. While such extrapolation may be debatable (as it is in any valuation model). 5 MVA® is the registered trademark of Stern Stewart. while the capital component measures proportion and implicitly encompasses market to book value Since default probability is not an explicit variable in the function.”5 Essentially. . which decreases the cost of equity. The tax deductibility of debt brings potential income but the possibility of default. These firms depend on equity management and changes in operating risk to keep a high economic profit. the priorities are similar but the focus is different. these firms must perform many of the same actions as debt laden firms.two characteristics that are often in conflict. On the other hand. . and these firms must manage credit adeptly. for a company who funds only with equity.154 MANAGEMENT OF ECONOMIC PROFIT Implicit in the capital dynamic is the specter of debt. it is never outwardly acknowledged. but emphasize sales and earnings growth and stability . and yet it has the greatest effect on all three variables. Buy back stock on the open market. 4. 3.Work to lower operating risk by focusing on fixed costs. 5. . Fund capital shortfalls in the short-term credit market. 2. Focus on retained earnings and pay special dividends if needed. 1.155 Table 6-5 ALL EQUITY FIRMS NET INCOME 1 Match marketing strategies with demand trends in the business cycle. 2. Tax strategies 3. COST OF EQUITY STOCKHOLDERS' EQUITY 1. Be wary of expansion at the top of a business cycle . Diversify among products. Beta is regulated by operating risk Work to keep sales as stable as possible through diversification and focus on fixed costs. customers. acquisitions 4. Manage the issue of outstanding shares through small increments. and the urgency of meeting . these debt laden firms must work for a high mean and low standard deviation in sales and income. An indebted firm must balance the amount of interest. Beta is decreased by increasing the proportion of equity in the capital mix. the price extracted is the greater risk of default. 3. Such an effort will lower beta risk. when the firm becomes a takeover target. COST OF EQUITY 1. Both long-term and short-term debt will work to limit the amount of shares outstanding. Share buybacks can be implemented with debt Leveraged firms have greater financial flexibility and the ability to maneuver through less profitable phases in the business cycle.156 Table 6-6 DEBT CARRYING FIRMS NET INCOME 1. With less dependence on equity funding. They can grant a steady dividend and “brace” the company through leverage. Debt laden firms must concentrate on the same income generating strategies as all equity firms. 4. STOCKHOLDERS' EQUITY 1. but the firm must monitor the probability of default. 2. leveraged firms must pay strict attention to retained earnings and new issues. Like all equity companies. 2. However. these companies can pay a steadier dividend. 2. and tax deduction with the increased variability of net income. Like all equity companies. however is often false in the short-term. many of these companies must expand into unrelated territories like finance (GM Capital) or even auto repair (Wal-Mart. the Federal Reserve will raise rates when the market is accelerating to combat inflation. the investor must realize that equity growth is actually permissible when interest rates are high. And yet. will the performance of the component parts be comprehensible. both net income and the cost of equity will often rise together. The assumption of coherency. each component has several countervailing effects that may occur when the market is in equilibrium. . K-Mart. but a lot of it is related to the performance of the business cycle. leveraged companies have both a commitment to shareholders . The greater effect is to increase both the level and the size of the risk premium (difference between the market and risk-free rates). Some of that positive correlation has to do with increasing sales just as the market is rising. Sears). While efforts may be directed at keeping equity growth at a minimum. While investors do not expect all-equity companies to be “big earners” during some phases of the business cycle.. Only when the measurement is taken as a cohesive whole. COMPONENT MOVEMENTS OF THE CAPITAL DYNAMIC The component movements of the capital dynamic follow a sequential chain of logic on which the investor will focus i.and . They are put into the position of “grow or fail” quite frequently because of the added necessity of meeting both interest and dividend payments. This double commitment creates the phenomenon of growing beyond the confines of “normal” growth of the industry. and lowering beta.157 earnings targets. As an example.e.creditors. The result is a larger cost of equity. Thus. the rate of change for earnings must be greater than for stockholders’ equity or the capital dynamic will fall. any random variable can put pressure on a component (especially the cost of equity) that makes it behave eccentrically. and so the context of each change is most significant. consider the correlation between net income growth and the cost of equity: net income should actually decrease the cost of equity by stepping up the proportion of retained earnings. 158 The following scenarios have the capital dynamic laid out as component driven changes. The arrows indicate whether the component is rising or falling. In reductionist terms, the stock market is simply an aggregation of these three component changes. • 1) Earnings are adequate, and the company begins to pay off its debt. The scenario for the capital dynamic would be: (Net Income ↑ ) - [(Cost of Equity ↓) (Stockholders’ Equity ↑)]. Net income increases retained earnings which decreases the proportion of debt to equity and reduces beta in the cost of equity. • 2) The firm funds capital requirements with more leverage. The scenario would be: (Net Income ↓ ) - [(Cost of Equity ↑ ) (Stockholders’ Equity ↓ ). In this case, net income may be decreased by higher interest payments, while stockholders’ equity may have less retained earnings and benefit from less of a need to issue shares. The cost of equity would rise because the greater proportion of debt to equity increases beta. • 3) Market forces are taking over and inflation is being curbed by interest rate hikes. The scenario might be: (Net Income ↑) - [(Cost of Equity ↑ ) (Stockholders’ Equity ↑)]. The firm may be at the top of the market, and performance will depend on the rate of acceleration of the three factors. • 4) A market downturn arises as investors flee to high quality bonds. Such a scenario puts negative force on all three components: (Net Income ↓ ) - [(Cost of Equity ↓) (Stockholders’ Equity ↓ )] There is little demand for any company’s equity while sales and earnings are falling. The cost of equity falls because the level and size of the risk premium declines; the market is descending and the Fed has lowered rates. • 5) The initial phase of a recovery: (Net Income ↑) - [(Cost of Equity ↓) (Stockholders’ Equity ↓ )] This may be the best time to invest because profits are recovering but pricing pressure on the stock is so low. • 6) A stock buyback purchased with leverage. That fortunate scenario looks like this::(Net Income ↑) - [(Cost of Equity ↑ ) (Stockholders’ Equity ↓ )]. The cost of equity rises because leverage forces beta to ascend. Net income rises because the amount of 159 debt is not substantial enough to raise the financial leverage ratio (EBIT / (EBIT Interest). Stockholders’ equity is targeted for decline. The movement in stock price is almost always concurrent with the change in the capital dynamic, although sometimes market inefficiencies arise and it follows it. Changes in the components are a matter of degree and the acceleration of each is as significant as its absolute level and its direction. THE COMPARATIVE CAPITAL DYNAMIC Although some performance can be gauged by increases in the actual size of the capital dynamic or EVA, it is an absolute measurement that is a function of corporate size. To place companies on a more comparative basis, it will be necessary to create a ratio between net income and the total cost of equity that equalizes qualitative gains. The comparative capital dynamic is merely Net Income / [(% Cost of Equity) (Stockholders’ Equity)] and is a very applicable measurement for companies in the same industry. It can be used with more caution for companies in different sectors as long as the analyst recognizes that stock performance is related to eclipsing the gains within an industry. For example, if a comparative capital dynamic of 2.5 is high for a specific industry, that valuation will carry more weight than if 2.5 were typical. Some industries will naturally do less equity financing, and they need to be compared with similar companies. 160 APPENDIX: THE EFFICIENCY OF EVA VERSUS ROE Many CFOs will concentrate on the return on equity (ROE) measurement as a gauge of corporate performance. The basic measurement is the product of profit margin (Sales / Net Income), asset turnover (Assets / Sales), and the equity multiplier (Assets / Stockholders’ Equity). In effect, these indicators do not measure progress toward an optimal capital structure as efficiently as EVA. With both ROE and its distant cousin ROC (return on capital), debt can be directly substituted for profitability and the ratios will still rise. Observe the following table: Table 6-7 Profit Margin Year 1 Year 2 0.07 0.06 Asset Turn. 1.2 1.1 Equity Multiplier 1.8 2.4 ROE 0.1512 or 15.12% 0.1584 or 15.84% Year 2 has a higher ROE with more debt but a lower profit margin and asset turnover. Technically, the firm can go through a downturn but “redeem” itself with more leverage. By substituting debt for equity in the correct proportion, ROE was able to rise despite the lower profitability ratios. On the other hand, had the firm used EVA as its measurement of progress, a smaller increase in net income may have reflected the lower profitability ratios and balanced the decrease in equity. More leverage would have increased the percentage cost of equity as well. (Back to Table of Contents) 161 APPENDIX: THE REAL COST OF CAPITAL AND WHAT THE INVESTOR NEEDS TO KNOW The cost of capital is defined as the return a business could make if it chose an alternative investment with similar risk. All of the component costs of capital – debt, common equity and preferred stock- are considered opportunity costs and are determined in the realm of market values when configuring the weighted average cost of capital (WACC). However, the investor is left in the dark when determining some of these values. While the cost of capital may be the most important fundamental in determining a company’s direction, the process of calculating one of its component costs, the cost of debt, is privileged information. Just as attorneys and clients have a private relationship, so too do creditors and debtors. Your next-door neighbor need not know your mortgage rate, and neither do competitors in any business need to know other competitors’ borrowing rates until those rates become public knowledge. The real cost of debt is the next interest rate that a corporation can incur after analyzing its risk criteria (leverage ratios) and interfacing this analysis with the market (a potential creditor): a default premium is derived and is added to the risk-free rate (an appropriate Treasury yield). When multiplied by the tax rate reciprocal (1 – Tax Rate) and then applied to the market value of a firm’s debt, a cost of debt is formed. If the corporation decides to incur debt at this rate, interest expense will be tallied and transcribed to the next financial statement. However, no matter how high or low past interest rates have been, the cost of debt is configured at the new rate. If interest rates have been 10 % and for some cataclysmic reasons go down to 5 % in one month, the new cost of debt is configured at 5 %. What happens to past interest rates and their collective effect on corporate debt? The effect of ongoing interest payments is reflected in the market value of 162 the debt. Just as with any bond, the market price of debt tends to rise when interest rates are falling and vice-versa. Where does this leave the investor? If the “nominal” cost of debt is derived by averaging each debt maturity with each corresponding interest rate, the amount of interest payments will be the one that makes book and market values equal. A much less accurate cost of debt would be formed by applying interest expense to the book value of a firm’s long-term debt. These accounting versions of the cost of debt, despite their defiance of the definition of true “opportunity costs”, may better serve the investor in gauging risk. In the following example, a company decides not to incur new debt at a lower interest rate because it considers itself overleveraged. The “new” percentage cost of debt is much lower than the interest rate it actually pays because the risk-free rate has decreased. While the average investor can copy a model that simulates the market value of debt, he or she has no access to the negotiations that determine the real interest rate, nor would such “transparency” be cost effective. As shown in the chapter on leverage, Chapter Two, the financial leverage ratio has predictive value based on past interest expense which is an explicit component of both the capital dynamic (EVA) and the risk-adjusted WACC. DETERMINING A MARKET BASED WACC: AN EXAMPLE The XYZ Company is negotiating with its underwriter to configure a new interest rate for a prospective bond issue. Although a lower Federal Funds rate collectively decreased interest rates by approximately 2 %, XYZ is beyond its optimal target proportion and anticipates a downturn. Thus, they will eschew lower-rate debt and are unable to refinance because its bonds have a no “call” provision. The main criteria that the underwriters use is XYZ’s interest coverage ratio which determines their default spread. 163 XYZ has earnings before taxes (EBT) of 500 (million) and interest payments of 250, giving them a coverage ratio of 2. On the following chart, that coverage ratio yields a default spread of 2 % over the Treasury yield. Interest Coverage > than Interest Coverage < than Default Spread Percentage 0.8 1.25 1.5 1.75 2 2.5 3 1.249999 1.499999 1.749999 1.999999 2.499999 2.999999 4.249999 5% 4.25 % 3.25 % 2.5 % 2% 1.5 % 1.25 % To configure a market-derived WACC, we need to apply the most up-to-date interest rate to the market value of a firm’s debt. We will consider XYZ’s position over two years. The first year (Year 1), the market value of the debt is the same as the book value. In the second year (Year 2), we apply a model to determine the market value of XYZ’s debt. Besides this market derived difference in capital proportions, the only other differences between the years are the risk-free rate and consequent new cost of equity, and the new interest rate. The model for estimating market value is as follows: Market Value of Debt = Interest Expense [(1 – (1 / (1 + New Rate) Average Debt Maturity)) / New Rate] + Book Value of Debt / (1 + New Rate) Average Debt Maturity. 164 The following charts delineate the changes in the cost of capital over two subsequent years. The change in the cost of equity is a consequence of the lower Treasury rate, while the market value of changes in capital is derived from the effect of the new interest rate on the price of the firm’s debt. XYZ YEAR 1 Risk-Free Rate CAPM Cost of Equity Interest Rate Market Value of Equity (Shares x Price) and Percentage in Capital Structure Market Value of Debt and Percentage in Capital Structure Book Value of Equity and Percentage in Capital Structure Book Value of Debt and Percentage in Capital Structure Tax Rate Cost of Debt (Book) Cost of Debt (Market) Average Maturity of Debt WACC (Book) WACC (Market) VALUE 6% 10 % 8% 7200 = 0.6973 =69.73 % = 7200 / (7200 + 3125) = 69.73 % 3125 = 0.3027 = 30.27 % =3125 /(7200 + 3125) = 30.27 % 4000 = 0.5615 = 56.15% = 4000 / (4000 + 3125) = 56.15 % 3125 = 0.4385 = 43.85 % =3125 / (4000 + 3125) = 43.85 % 0.3 = 30 % (1-0.3)(0.08) = 0.056 = 5.6 % (1-0.3)(0.08) = 0.056 = 5.6 % 10 Years (0.4385)(0.056) + (0.5615)(0.1) = 8.07 % (0.3027)(0.056) + (0.6973)(0.1) = 8.668 % 165 Because interest payments are equal to the product of the interest rate and the book value of debt, the market value of debt is equal to its book value: 250 [(1 – (1 / (1 + 0.08)10) / 0.08] + [3125 / (1 + 0.08)10] = 3125 For the next year (Year 2), we must take the new interest rate and apply it toward the market value of debt. Since the new Treasury (risk-free) yield is 4 %, the new interest rate (as determined by the default spread of 2 %) is 6 %. The market value of debt is: 250 [(1 – (1 / (1 + 0.06)10) / 0.06] + [3125 / (1 + 0.06)10] = 3585. The increase in market value occurs because XYZ’s price is bid up by investors who want the higher coupon rate on its debt rather than the new lower rate. Notice also that interest expense remains at 250 Million because no new debt has been incurred – just a change in market interest rates. XYZ YEAR 2 Risk-Free Rate CAPM Cost of Equity Interest Rate Market Value of Equity (Shares x Price) and Percentage in Capital Structure Market Value of Debt and Percentage in Capital Structure Book Value of Equity and Percentage in Capital Structure Book Value of Debt and Percentage in VALUE 4% 8% 6% 7200 = 0.6676 =66.76 % = 7200 / (7200 + 3585) = 66.76 % 3585 = 0.3324 = 33.24 % =3585 /(7200 + 3585) = 33.24 % 4000 = 0.5615 = 56.15% = 4000 / (4000 + 3125) = 56.15 % 3125 = 0.4385 = 43.85 % 166 Capital Structure Tax Rate Cost of Debt (Book) Cost of Debt (Market) Average Maturity of Debt WACC (Book) WACC (Market) =3125 / (4000 + 3125) = 43.85 % 0.3 = 30 % (1-0.3)(0.08) = 0.056 = 5.6 % (1-0.3)(0.06) = 0.042 = 4.2 % 10 Years (0.4385)(0.056) + (0.5615)(0.08) = 6.94 % (0.3324)(0.042) + (0.6676)(0.08) = 6.74 % The accounting version of the cost of debt did not recognize the new interest rate and maintained its relationship between interest and principle (0.08)(3125) = 250. On the other hand, the market value version of the cost of debt created a more exaggerated change in the WACC. Academic tradition acknowledges the market cost of debt as the only true opportunity cost and creates consistency with the cost of equity. However, the computational rigors of the measurement bring it away from the purview of the individual investor and into the corporate boardroom. Indeed, the reader can assess the complicated effect of changing market values of equity on the cost of capital and can view the temporal character of the WACC. Since the investor needs to make rapid comparisons between companies, it is recommended that he or she use the risk adjusted version of the WACC; the relationship between interest and principle when combined with earnings information, is just as forward looking and perhaps more stable than the more volatile market version of the WACC. DISTORTION AND ACTIONABILITY 167 For the financial professional, a market-based cost of capital can only be minimized by adhering to the risk-oriented criteria that lowers the interest rate proffered by creditors, and by matching the life of corporate assets with the maturity of liabilities. The market value of equity may fluctuate spasmodically which directs strategic attention to the only controllable variables in the cost of equity – operating risk and the proportion of debt to equity which both affect beta. For the investor, a market-based cost of capital is simply not actionable and he or she must be content with substituting the juxtaposition of interest expense with the book value of long-term debt. However, interest expense and long-term debt are both major risk factors when making investment decisions. In fact, applying the real cost of debt to book values will create a distortion because the cost of debt will not have prior interest payment obligations as an explicit variable. For example, in the above scenario, using book values, XYZ’s cost of debt would have decreased to 4.2 % from 5.6 %, understating the ongoing coupon rate of 8 %. While not wholly integrating opportunity costs, the capital dynamic better quantifies risk for the investor than does an EVA that uses the market-based cost of debt. Assuming that prior interest expense quantifies the interest rate being paid is not theoretically correct in terms of a weighted average cost of capital. Such an assumption, however, allows the investor to gauge risk from financial statements without the need to be privy to creditor negotiations and may even be more profitable than discerning a marketbased WACC that fluctuates with volatile stock prices. (Back to Table of Contents) 168 7 FUNDAMENTALS AND CAPITAL STRUCTURE “Fundamental” is a word with diverse meanings. In a religious context, it usually connotes a literal interpretation of a sacred text. In finance, it refers to concrete performance measurements - sales, net income, and assets. In the context of capital structure, we are aware of the literal, concrete aspect of balance sheet items, and then we turn them upside down: only in the domain of change do these figures have any great significance for the investor. For over forty years, there has been some equivocation in academia about the proper teaching of fundamentals. Since earnings forecasts are based on fundamentals, one would certainly see the importance of teaching fundamentals in any business curriculum. Indeed most business students need to take at least a couple of accounting courses in order to graduate. However, among academics, there is a righteous adherence to what is termed the “semi-strong” form of the efficient markets hypothesis. In brief, that doctrine proclaims that no publicly disclosed information can correctly forecast the future path of a stock; the market is so “efficient” at pricing a security that the price reacts to information before it is announced. This divergence between the great amount of financial data available, and the inherent inability to utilize it, has frustrated many a student and professional alike. To the capital structuralist, the amount and risk of capital inflows determines the fundamentals themselves, and therefore becomes the building block of earnings forecasts. While most executives and analysts are optimistic about sales and earnings increases, few will consider their source. The nebulous world of capital proportions and allocation entails risk and interpretation, and not just the straightforward pronouncement of a rise in sales. However, capital structure analysis can be classified as a more sophisticated form of fundamental analysis -albeit one that anticipates fundamentals rather than reacting to 169 them. While a typical fundamental analyst might look at industry margins to anticipate revenues, the structuralist looks for more indirect numbers that would signal an environment conducive to earnings increases - a smaller interest expense, lower federal funds rates or an increase in long-term debt, for example. Similarly, a fundamental analyst would be concerned with a stock’s intrinsic value, the present value of discounted future earnings as compared to the actual market value. Alternatively, a capital structuralist will be as much concerned whether the number of shares outstanding is the proper amount; the structuralist sees earnings in the domain of immediate changes in risk. To quantify this risk, he or she steps out of the realm of relational fundamentals and uses statistical techniques to compare the mean, skew, and standard deviations of a distribution. Only when risk and return are optimized through actionable changes in the proportion of debt to equity, will he or she be satisfied. If fundamentals by themselves lack predictive value, they may be the most significant tool in the education of a financial executive. While the efficient market may anticipate fundamentals and discount their value, it is the executives who will implement the changes that affect those ratios for better or worse; these are the changes that portend a rise or fall in stock price - a new marketing strategy, a lower cost of debt, or a broader customer base. Thus, in the most ultra-efficient market, the stock will rise concurrently with sales and earnings, and not after the fact. Although momentum in these figures can spur even more capital infusion in later periods, that type of a rise is a “hit or miss” proposition, based as much on continued sector domination as on immediate earnings history. While greater earnings can help make the cost of capital less expensive, future prospects are geared to funding projects with a high net present value (see the chapter entitled, “Capital Structure”) and not on the generation of past earnings. In essence, an ideal combination of low cost capital sources, some of which may be retained earnings, must be coordinated with the capital budgeting process. a priori. fundamentals are like a “yard stick”: they enable executives to compare and contrast the variables that need changing. it created the underpinnings for examining the probabilities and changes behind each input. When comparing companies. It is in this revelatory role that a structuralist uses fundamentals .performance. • 1. For example. ROE is an accurate. As a measured result. Ultimately. the Du Pont system set the foundation for further analysis. no accounting system can project an operating income.170 Notwithstanding their value as a predictive tool. • 3. general indicator of current . and whether they need to be changed. The component parts of ROE point toward a company’s capital structure. The components of ROE will indicate what risks need to be addressed. how they are balanced. The decomposition of a firm’s return on equity (ROE) is a multipurpose exercise with four distinct rewards for the capital structuralist. albeit deterministic system that linked leverage with sales. the difference between finance and accounting is established by the balance between choosing the action with the best probabilistic outcome and then measuring its result. profits and even the effective corporate tax rate. and delineate a firms operating and financial leverages. . that will lead to this result. the measurement EBIT / Assets must exceed the interest rate if leverage is to be justified. However. and which elements have potential for improvement. DU PONT ANALYSIS The Du Pont Company developed an extensive.to gauge capital inflows and determine whether they are being channeled efficiently. • 2. especially asset turnover. The resulting five component parts are then multiplied together. which this author believes negates the role of debt in reducing stockholders’ equity.171 • 4. Some instructors will refer to ROE as: earnings before interest and taxes (EBIT) / Stockholders’ Equity. Table 7 -1 COMPONENT FUNDAMENTAL 1) Earnings Before Interest and Taxes (EBIT) 2) Sales 3) Assets 4) Stockholders' Equity 5) Earnings before Taxes 6) Net Income PARTS OF ROE LOCATION Income Statement RATIO 1) EBIT / Sales DESCRIPTION Operating Margin Income Statement Balance Sheet Balance Sheet Income Statement 2 )Sales / Assets 3) Assets / Equity 4) EBT / EBIT Asset Turnover Equity Multiplier Cost of Debt (Financial Leverage) Tax Retention Income Statement 5) Net Income / EBT The shortened form of this equation is: (Net Income / Sales) x (Sales / Assets) x (Assets / Equity) = Return on Equity. Each fundamental is extracted from a financial statement and then divided by another fundamental to form a component part of ROE. but the full array of components better describes the dynamics of the equation. Since capital structure analysis is founded upon this crucial difference. Each component of ROE is affected differently by each phase of the business cycle and any imbalance in the components will affect performance. we submit to the Du Pont Analysis system and end up with ROE = Net Income / Stockholders’ Equity. A prime observation that will be readily apparent when one works with these ratios is their tendency to be consistent within certain boundaries . . 3 27872 . the world renowned consumer retailer.8 become a company with a ratio of “2”. Fixed costs in a department store chain. if it is in fact. and the companies within the industry will share similar levels. or working on their respective weaknesses. are totally different from those required in a semi-conductor company. a developer of risk management software with high research and development costs. by either emphasizing what they do best. Table 7 -2 1999 COMPANY BARRA WAL-MART NET INCOME 23. but neither will Microsoft match Wal-Marts’ large asset turnover or asset to equity ratio. Thus. profit margins and equity multipliers. The two companies can obtain the same ROE.172 profit margin and assets to equity . and Wal-Mart.the three components of the shortened form of ROE. Again. the amount of fixed assets and technology within the industry will determine the boundaries of capital inflows. the same company. and each will have respectively different asset turnovers. we will contrast two companies: Barra. To display the difference in ratios that an industry can impose. the size of the components of Du Pont analysis conveys the risks of working in a particular industry and can be increased or decreased within the limitations proscribed by the industry. But . The characteristics of each industry dictate limitations on the components. for example. as with leverage.never will a company with a typical asset turnover of 0. however. Wal-Mart might struggle to obtain a 4 % profit margin and never match Microsoft’s 20 %.4 5575 SALES 187 165013 ASSETS 169 70349 EQUITY 101. For example. more established industries that churn . On the other hand. Hence. more cash-flow diminishes the probability of default.asset / equity. Without the necessary ”track record” of steady earnings and loan history.09 % 20. The reason? Higher profit margins sometimes entail more variation in sales and/or the use of retained earnings to raise capital.4/187=0. creditors are less likely to loan at favorable rates. they are exposed to competition that can under-price them. They are often from older.high profit companies usually have pricing power but lower asset turnover and diminished use of credit. While Barra has no credit risk.524 ROE 23.1 5575 / 165013 = 165013 / 0. In retail. pricing power implies the creation of unique “niche” products that cannot be easily duplicated.3=1. high asset to equity type companies will almost always have higher asset turnovers because a large turnover diminishes risk and provide greater security for their credit. the advantage is more qualitative and stems from the higher prices such products may garner.35 Asset / Equity 169/101.1251 187/169=1. an analyst can learn much about a firm.173 Table 7 -3 1999 ROE COMPANY BARRA WAL-MART Profit Margin Asset Turnover 23. the advantage comes from asset turnover and the safe use of credit . Each company has different inherent risks. In specialty software development. For example . On the other hand. Moreover.0337 70349= 2. and will consequently have a lower asset turnover. Wal-Mart has a large credit risk but cushions it with a large asset turnover. higher profit margin companies tend to be incapable of “churning out” product at a moments notice because of technical or marketing limitations. The student/investor will find different patterns among the industries that balance each other in the context of ROE.00 % By merely observing these three ratios. The return on equity is similar but the way each company arrived there is totally different.67 70349 / 27872 = 2. it is doubtful whether investing on the basis of improved ratios can beat the market for any length of time because gains in share price happen concurrently with the ratio improvement. Since stock price is highly correlated with improvements in ROE. The one big advantage that company insiders enjoy is that they know which ratios are improving and by how much. In deference to the efficient markets hypothesis. without “synergy”. the reckless improvement in one variable to the detriment of another can lead to “shocks” in the system that need to be reconciled in future years. because they strategically set out to change them. However. the need for balance is even more imperative because decreasing one variable has a tendency to increase the other. Analogously.whether they are tires or bakery bread. The three largest competitors will usually supply enough data to establish a valid average. even a one percentage point gain in profit margin would have earned the company approximately 1.6 Billion . it is rewarded by a higher stock price. it will do so by creating implicit benchmarks in these areas. In fact. Thus. .174 out simple but very necessary products . However. these firms can concentrate wholeheartedly on this single lack. and these companies may have a difficult time increasing profit margin. In Wal-Mart’s case. both of these companies could have shifted component parts to find the factors that produced an optimum. the most helpful tool is to research the industry averages over a five year period. Thus. For the capital structure analyst. the observed confidence that management has in its own stock is one of the few ways that the average investor can indirectly profit from fundamentals. and some websites will have done the work already. their own productivity interferes with the price that they can charge. increasing the profit margin by even two percentage points on a consistent basis would be considered an amount that would dominate the industry. if the market will price stocks “efficiently”. Barra needed to both limit its amount of equity in proportion to its assets and increase sales. Buying opportunities carry more weight than sales because many insiders will sell stock for tax purposes. In fact. in many industries. and warrant a soaring stock price. once a firm improves on a deficiency. To establish a benchmark component ratio. For example.. Finally. but they need to balance any change with reinforcement from other components. i. such a move in the short-run might artificially pump up asset turnover.175 the primary mission is to coordinate a “map” of the territory . keeping one component very stable allows the other components to vary (in a three component ROE). If a firm has large sales in terms of assets. it probably is not growing. However. a firm can plan inventory levels and sales districts around asset turnover. Another potential advantage of ROE analysis is that a firm can use its most stable ROE component as a planning tool. some “insiders” are simply fervent optimists and we need to corroborate their enthusiasm. In this regard. if a firm disposes of assets (net). there will be one component that gives the firm the majority of its ROE strength. This component will represent its competitive advantage and may or may not be the same as the “stability” component. apparently. That is CSX’s competitive advantage. The exhibit of ROE components . if the asset to equity ratio is stable. Most of these firms are in different industries which are reflected in both the size and stability of the components. Similarly. but low profit margins. Moreover. the analyst wants to find a situation where earnings will accelerate much faster than the cost of capital. the challenge to any firm is to improve this weak link without damaging the performance of other components. They may lower this ratio and still remain competitive. Among those other two components. cash-flow is stable enough to fund its heavy need for capital with debt. a one percent rise is profit margin may be very substantial. a company like CSX (see charts) has a huge advantage in its asset to equity ratio. The objective is to exceed the industry standard for that component.deciding whether the investment environment is conducive to earnings and observing how the sector is meshing with the business cycle.e. For example. COMPARING ROE COMPONENTS The following charts display the three basic ROE components for ten different companies. one will represent the “weakest link” to a healthy ROE. a firm may want to perform near its target capital structure at all times and not take the risk of brief directional movements away from it. 99 2.9 1.88 Asset Turnover Profit Margin Asset/Equity . Table 7 -4 DILLARDS (DDS) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 3.15 2.95 1999 1.9 1.1 2.176 displays the component number over five years and follows by ranking each company for which component is stable.8 1.65 1.1 1. weak and/or competitive.19 1998 1.7 0.19 2.74 2001 0.8 2000 1. 7 0.63 2001 12.177 Table 7 -5 ST.6 0.49 0.8 0.7 1.72 1.4 0.07 2001 29.56 1.83 3.68 1.48 Asset Turnover Asset Turnover Profit Margin 1998 12.95 1999 31 1.37 2000 28.3 1.46 2000 11 0.83 1.73 1.5 0.72 1999 2.89 Profit Margin Asset Turnover Profit Margin 1998 32 1.77 1.94 6. JUDE MED (STJ) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 5.38 Table 7 -6 US TOBACCO (UST) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 31.47 .2 0. 7 0.77 1.32 2.75 1999 3.7 0.2 1.93 2.93 Asset Turnover Asset Turnover Asset/Equity .77 2000 9.178 Table 7 -7 ECOLAB (ECL) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 8.56 2.59 2001 9 0.31 1.14 1999 8.2 1. RECTIFIER (IRF) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 0 0.7 2000 9.5 1.57 Profit Margin Asset Turnover Asset / Equity 1998 8.73 2.71 1998 3 0.26 2001 8 0.25 2.2 1.62 1.87 Table 7 -8 INT.16 2.07 1. 8 0.32 Asset/ Equity Asset Turnover Profit Margin 1998 11. (NATR) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 7.27 2000 10 0.3 1999 10.2 2.2 2.2 2.4 0.4 Asset/Equity Asset/Equity Asset Turnover .35 2001 5.25 Table 7 -10 NATURE'S SUN.85 1999 6.67 1.2 0.07 1.44 1.9 1.99 1.99 1.93 1998 7.6 1.38 2000 5.9 2.37 1.4 2.32 2001 8.94 1.43 1.179 Table 7 -9 MOLEX (MOLX) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 10.7 1. 4 3.46 Asset Turnover Asset Turnover Asset / Equity 1998 5.9 0.17 1999 2 1.49 1999 0.5 0.36 2001 1.6 2000 6.4 0.48 4.13 Asset/Equity Profit Margin Asset/Equity .39 2.96 2.5 0.4 2.22 1998 2.26 1.9 1.88 2.41 2001 3.5 1.94 Table 7 -12 ARCH.7 1.02 2. (ADM) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 2.6 0.3 0.53 3. DAN. MID.180 Table 7 -11 CSX (CSX) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 7.52 3.25 2000 2. 29 5.81 1999 6.181 Table 7 -13 ARGOSY GAMING (AGY) YEAR / COMPONENT Profit Margin Asset Turnover Asset / Equity Stable Component Weak Component Competitive Component 1997 0 0.37 * The Asset / Equity ratio is assets divided by common stockholders’ equity and not the full stockholders’ equity that may include preferred stock. Often. more business activity. while “short-term debt” is synonymous with current liabilities (for the sake of this analysis). 2) it may indicate more trade credit which is often treated as an interest free loan.7 1. 3) if used judiciously. MODIFYING AND ENHANCING DU PONT ANALYSIS The capital structuralist needs a detailed view of a firm’s debt structure.598 7.05 9.1 1. he or she will further decompose the assets to equity ratio (commonly called the equity multiplier) and also use the full array of components(five) in the model.9 13.3 0. and additionally.16 2001 8. The asset / equity ratio is decomposed as follows: ASSET / EQUITY = (LTD / Equity) x (Asset / Short-term debt) x (Short-term debt / LTD).61 17.74 2000 6. it can reduce exposure to interest rate risk.13 NONE Asset Turnover Asset/Equity 1998 1. “LTD” is an acronym for long-term debt.4 0. This last item needs qualification: . funding with more short-term debt and less long-term debt will accomplish one of three objectives: 1) it may reduce overall interest expense as short-term rates are lower than long-term rates in a normal market. the student/investor can observe the behavior of the complete asset to equity ratio in a profitable year: (LTD ↓ / Equity ↑) x (Asset ↑ / Short-term debt ↑ ) x (Short-term debt ↑ / LTD ↓) = Asset / Equity. By following the changes in each component.182 while funding long-term fixed assets with short-term debt can lead to instability. while Net Income / EBT determines the effect of taxes on shareholder earnings. the result will be an increase in stock price. Of the four remaining ratios in the full model. The student/investor should . Although an increase in current liabilities may affect short-term solvency and the probability of default. Table 7 -14 COMPONENT LTD ASSETS EQUITY SHORT-TERM DEBT DIRECTION OF CHANGE DOWN UP UP UP REASON Less interest expense. Our objective is to cite the mechanisms that make that happen. it is the size of the long-term debt to equity ratio that is significant. and not whether long-term debt is actually decreased. funding with short-term credit when interest rates are anticipated to drop may reduce exposure to long-term commitments . and is an important tool in managing capital structure. risk Growth in sales and earnings More retained earnings More business activity If the reader has been following the text. he or she will realize that in certain situations the opposite changes can affect the ROE measurement in a positive manner. Since many companies will further raise long-term debt in very profitable years. two are imperative to capital structure analysis: the ratio. This is simply an example and not meant to convey one set positive pattern: when tax benefits soar while the probability of default does not. it is both a source of internal financing and free cash-flow.as long as such a move is temporary. EBT / EBIT distinguishes the cost of debt. Interest Expense. . Net Income / EBT increases as interest increases. but the cost of debt is a reciprocal that actually decreases as interest rises. this figure will increase as interest expense increases. For example. This more “understandable” form is termed the “financial leverage ratio” or alternately. To put these figures on common ground. The entire five component model is configured as follows: (EBIT / Sales) x (Sales / Assets) x (EBT / EBIT) x (Net Income / EBT) x (Assets / Stockholders’ Equity) The reader can notice that the first four components reduce to a return on assets (ROA) because all but the two fundamentals. each decision made about any one ratio will affect all of the others and financial management must pursue a proper balance. are no less important than the aforementioned components. later in the chapter we will reverse the cost of debt ratio and turn it into EBIT / EBT. but they are less directly manageable. Thus. it would not be “stepping out of bounds” to say that the goal of ROE strategy is to optimize the risk/return characteristics of these two ratios by controlling the other three. and that interest is a tax deductible expense. operating margin (EBIT / Sales) and asset turnover (Sales/ Assets). and since the target company has a large asset turnover. understanding the repercussions of each action. and competitive pressure. It is EBIT / EBIT . Indeed. age of the industry. and that issue would have to be addressed. The two remaining ratios. (Net Income / Equity). cancel each other out. net income and assets. When we multiply (Net Income / Assets) by (Assets / Equity). In that case. financing an acquisition with leverage may raise the asset to equity ratio. the deciding point would be the fear of potentially decreasing the profit margin. Thus. we get the full ROE effect. They are more sensitive to business cycle fluctuations.183 be aware that earnings before interest and taxes (EBIT) minus interest expense will equal earnings before taxes (EBT). the level of technology. the degree of financial leverage” and is the same ratio we have used in previous chapters. operating risk would be diminished. the “ideal” ROE has been observed many times when a dominant company in a dominant sector comes to fruition. While debt is being paid off from enhanced profits. in the vast majority of capital transitions. To the capital structuralist. and Asset / Equity to keep the two profit ratios (EBIT / Sales. Eventually.184 To corporate management. it is the changes in each ratio that are significant. take on new projects. In fact. . the capital structure oriented ratios can temporarily make up the difference. this is where talent in financial management is realized. a thorough sensitivity analysis involving all the components is imperative. and the proportion of equity increases. enhancing an ROE with an increased equity multiplier even while operating margins and asset turnover are depleted. However. is a risky strategy with a large payoff if the outlook is favorable. the cycle will begin again. the firm will load up on low interest loans. profits will be extracted from less use of long-term debt in comparison with equity. as high and stable as possible. More earnings are retained. and sometimes move past the point where the proportion of debt is optimal . the company faces asset cutbacks and/or downsizing. In reality. there may be an interim when operating margin and asset turnover do not increase enough to pare down debt. profitable projects must flow into” the pipeline” to keep the company competitive. However. more long-term debt is needed. A less risky strategy is to use the three leverage ratios (EBT / EBIT. While financial executives often attempt to extract the largest ROE possible. Although new. Sales / Assets). Net Income / EBT. In a handful of cases. if sales are not generated from new projects. When the economy “heats” up and the cost of retaining earnings becomes expensive. and a corporation will make several adjustments to the other ratios to keep ROE from diminishing. and the tax benefits will enhance greater sales and earnings. the investor must gauge the risk of interactive change toward an optimal capital structure. interest expense is better covered by operating income and the probability of default decreases. Would it be expedient to lower assets to equity to a more permanently tolerable level? That question . consider a scenario in which a firm under performs the industry in asset turnover. EBT / EBIT. The tax retention ratio.yet. While the absolute size of the ratios is important to the investor. Net Income / EBT.185 Table 7 -15 THE IDEAL ROE COMPONENT Operating margin Asset Turnover Cost of Debt Tax Retention LTD / Equity Asset / Short-term debt Short-term debt / LTD RATIO EBIT /Sales Sales / Assets EBT / EBIT Net Income / EBT LTD / Equity Asset / Short-term debt Short -term debt / LTD DIRECTION OF CHANGE UP UP UP (Indicates decreased costs) UP (Indicates reduced taxes) DOWN (More retained earnings) NEUTRAL (Increase both) UP (More business activity) At high levels of earnings. more concern is placed with the interactive dynamics between them . (Assets / Equity). By concentrating on the cost of debt ratio. Assets / Short-term debt. there is a penalty for retaining earnings and not distributing them as dividends. Short-term debt / LTD). we implicitly affect tax retention. Our Du Pont model is not sophisticated enough to deal with these concerns . we have narrowed our perspective to seven ratios. interest expense is a common factor to both earnings before taxes and the cost of debt.which components are increasing or decreasing and how will they affect the other components. are a decomposed version of the equity multiplier. three of which (LTD /Equity. As an example. and also a penalty for issuing too much equity in the form of newly issued shares of stock. So far. is subject to influences from sales and operating income for which there is little tactical control. in which return on equity changes occur. an under covered. increases in equity do not automatically translate into a company moving toward a more optimal capital structure. the dynamics will be mathematically changed. ROE components represent an objective. By replacing the fundamental. Wall Street has never played a game of absolutes when it comes to either fundamentals or near-term stock price increases. In the ratio. The rewards are more derived from beating high-risk odds and doing the unexpected. What the ROE components lack as tools of managerial finesse. long-term debt to capital (LTD / Capital). and without a strategic overlay of project analysis and low cost funding. and a better ratio can unify the three elements of Asset / Equity into an understandable whole. with “capital”. Obviously. retained earnings can be excessive and new stock can be issued at an inopportune time. however. goes down. the cost of debt (EBT / EBIT) goes up when the actual cost of debt. Therefore. Thirdly. In particular. The challenge of going from 0. 30 per share to 0. the micromanagement of these ratios can lead financial executives into a trap. And it is that time frame . the asset to short-term debt ratio is ambiguous. They represent performance goals but do not clarify the distinction between risk and return enough to act as tools with which to manage a firm. but would be a difficult choice otherwise. We must. but not a means to an end.70 per share in EPS is often rewarded more than a high performing stock that goes from an EPS of three to four dollars per share. Capital is defined as the sum of long-term debt and stockholders’ equity. small cap firm can outpace a large-cap Dow component. THE RETURN ON CAPITAL RATIO Through a simple modification.the “near-term” or short-run . they more than make up as signals for analysts and investors. financial engineering of a desirable outcome would be impossible.in the short-run. Secondly. adding . as measured by interest expense. In essence. than from consistent performance . modify them to encompass capital structure changes.186 would be self evident if the equity multiplier consistently exceeds the industry standard. the capital structuralist creates a “return on capital” or ROC ratio. “Equity”. LTD / Capital. there may be some other element that buffers the risk of an increase. Increases in this ratio will usually signify that the default probability for the firm has risen. observe the following: ASSETS / CAPITAL = (LTD / Capital) x (Assets / Short-term debt) x (Short-term debt / LTD). In fact.the business cycle. which will show increases of interest expense in relation to operating income. increasing the ratio. Thus. However. It is when these elements interface with the profit components. that risk is reduced in the domain of larger returns. whereas the former long-term debt to equity ratio might not. EBIT / EBT and Asset / Capital. On the other hand. the amount of absolute debt. EBT / EBIT. we have the elements of risk reduction. Assets to Capital becomes more of a risk management tool because it can be applied to several leverage situations where short term debt increases in proportion to long-term debt. With these three components. is inverted to EBIT / (EBIT . This is the much referenced “financial leverage ratio” which is very much like the interest coverage ratio in bond ratings. the other leverage factors. Any increase in assets to capital implies a relative increase in short-term debt by default. will always contribute to the return . Moreover. and become integrated with the dynamics of the greater economy. it is dependent on other factors for interpretation . Reducing all three components will lower risk. the ratio asset / capital is a very neutral element in terms of absolute risk because it can be used as a capital substitute when the cost of debt is very high. since default probabilities have countervailing components. To see how this fits into the new expression. However. Assets to Capital. creating a less volatile and more statistically significant measurement. long-term debt to capital will show significant changes.187 long-term debt increases both the numerator and denominator. Assets / Capital.Interest Expense). operating margin and asset turnover. The cost of debt ratio. the ratio. becomes more volatile and sensitive to change because we are making it smaller than Asset / Equity and not changing both numerator and denominator in the expression. but not necessarily increase return. We bring this back to basics with: (Net Income / Sales) x (Sales / Assets) x (Assets / Capital) = ROC (Return on Capital). a “trifecta”. it will increase both risk and return. as long as raising it does not tacitly contribute to a depleted operating margin or asset turnover. (Back to Table of Contents) . Back in the 1990s. the completed return on capital equation is: (EBIT / Sales) x ((Sales / Assets) x (EBT / EBIT) x (Net Income / EBT) x (LTD / Capital) x (Assets / Short-term debt) x (Short-term debt / LTD).188 on capital on a superficial level. Although stocks were inflated and momentum was rampant in that decade. When all three of these components were raised simultaneously in any given year. it is prudent to look at the shortened form because many comparisons need to be done to find the “right “ company. The normal success rate for a stock increase was about 73 % according to our data. an “after the fact” increase of that magnitude was phenomenal. typical for a “bull” market. From an investor’s perspective. adherence to the principle of increasing all three components yielded a situation we affectionately called. In full form. the success rate hovered at around 90 %. Huge earnings can be made. he failed to “foresee” shifts in the economy which successful managers will anticipate. A personal story .189 8 CAPITAL STRUCTURE AND THE BUSINESS CYCLE The genius of management is to place itself “in the right place at the right time”. even a mediocre manager can prosper if he or she is prescient enough to anticipate the dynamics of an industry which is “suddenly” favored by the economy. the economy follows a discernible pattern. but if they are garnered in an inflationary period.My nephew Randy was a whiz at computers and helped design many Internet sites early in the game. When stocks tumbled at the end of 2000. when coupled with even a slipshod leadership ability. He day-traded stocks at a time when such a passion was unique. they have isolated common characteristics of most markets: . and although each market is different from the last. Despite the protestations of “more talented” underlings. for example. However. he got hit. distinct similarities emerge. His famous line? “But the fundamentals were great on that company!” Corporate fundamentals are always secondary to the actions of the entire economy which the market anticipates. the market will totally discount them. COMMON ELEMENTS OF BUSINESS CYCLES At times.other potential managers) will be the one who is rewarded. Therefore. While economists cannot predict the peaks and troughs of this pattern with precision. so to speak. the business cycle is the ultimate source of all stock gains and “structures” capital structure. While a talented manager can take the helm in a recession and slowly guide a company into recovery. the manager who can out foresee the competition (in this case . That foresight. may be enough to garner huge bonuses at the end of the year. each market is different in some risk-taking aspect . Recovery 3. Expansion 4. tax-wise. Phases rarely make smooth transitions from one to the next.190 • 1) Four phases appear valid (some economists would argue for more or less) and smaller sub-cycles are sometimes prevalent. After about three to five years into a recovery/expansion. Recession/contraction 2. In fact. They are: 1. The danger of business cycle analysis comes from expectations. Plateau • 2) A similar pattern of interest rate changes occurs over the cycle as the Federal Reserve responds to both the need for investment and the potential blight of inflation. Fortunately. in the amounts of inflation or the rates of foreign exchange. THE YIELD CURVE AND INTEREST RATE BEHAVIOR Students tend to think of the yield to maturity curve as a supply and demand curve which is not a correct assumption. The reason for this confusion is that the demand (and price) for a debt issue moves in the opposite direction from changes in the interest rate. prospects may again seem dim and the Federal Reserve may engineer what is termed a “soft landing” or “growth recession” . • 3) Companies prosper at different times over the entire cycle. Sometimes entire sectors will be left out of a recovery and expansion because conditions in that industry have changed since the last cycle. However. contrary to the best pundits’ predictions. separating themselves into industrial sectors that have similar cash-flow patterns and borrowing habits. He or she can concentrate on the capital structure decisions of the best sectors and seek out patterns of leverage that will replicate throughout the cycle.a period of low GDP growth which allows a “bull” market to continue for a few more years. we naturally assume that past patterns will be duplicated and begin to extrapolate into the future.legislatively. and leading indicators such as M2 or the stock market can even be lagging in some cycles. A logical question might be: “Why would . it is much more a gauge of investor expectations about the direction of future rates than a demand graph for loanable funds. the student/investor does not need to make accurate predictions in order to make money. As we shall see.191 anyone pay more for an investment with a lower rate?” The answer is: “They don’t. As the economy heats up. As interest rates are raised by the Federal Reserve. Analogously. With the yield curve. based on the maturity of a loan. Banks. because investors need to be compensated for the risk of holding an investment that is more exposed to changes in interest rates and inflation. and the ascending yield curve is universally accepted by economists as “normal”. when the issue is sold before maturity. giving it parity with new issues – when held to maturity. the price will be below par. Investors can make more on an issue with higher rates and sell off any bonds that have the older and lower rates. Companies with large amounts of fixed assets demand the use of long-term funds. the price of existing issues with lower rates goes down. when interest rates spike above the debt issue with lower rates. During the business cycle. financial institutions make profits by borrowing at lower short-term rates and lending at long-term rates. have many short-term liabilities and they will match those maturities by investing in short-term securities. the Federal Reserve begins raising rates at the .”. the lower interest rate now yields more return. as the Federal Reserve raises or lowers rates. This theory implies that the supply and demand for loanable funds is derived from the cash-flow patterns of a business. while companies with more current assets will borrow short-term. will minimize capital costs. just the opposite is often true because rates reflect investor expectations as much as the price of debt. mostly treasury bills. the yield which relates price to interest rates is a function of the time to maturity. The process is dynamic and not static. However. Thus. for example. At a lower price. Financial professionals analyze yield curve behavior through what is termed “the segmented markets hypothesis”. and interest rates have risen. the curve changes shape to reflect investor sentiment and the prevailing rate for each level of maturity. long-term rates will be above short-term rates. A second incorrect assumption is that a firm’s choosing the lowest interest rate. If investors expect a normal spate of inflation to occur. there is less demand for the old issue and the price goes down. “the expectations hypothesis”. it is because the economy is in bad enough shape for the Federal Reserve to begin lowering rates. these theories are not mutually exclusive. generally analyze yield curve behavior through what is termed. In essence. the uncertainty of the level of default rises. . demand for loanable funds is a function of both the financial structure of a company and expectations about the direction of interest rates. The segmented markets hypothesis proclaims that short and longterm debt cannot be substituted for each other which is backed up by empirical evidence. the opposite does not occur because firms do not want to pay variable interest rates. flooding the market and pushing up the yields on them. If longterm rates are temporarily lower than short-term rates (the dreaded inverted yield curve). and both appear to explain yield curve behavior. it went bankrupt. on the other hand. While a large company cannot delay all of its funding until interest rates are lower. Although many companies will convert short-term debt into long-term once it reaches a specified level. Academics. Therefore. Braniff Airlines in the 1980s is a case in point: when this company began using short-term debt as a substitute for long-term debt. banks will sell these short-term securities. Therefore. it may ration capital by limiting outlays to current projects and forego any new projects until a downturn transpires and the Fed lowers rates. these graphs serve to unite both the segmented markets and expectations hypotheses. The volatility of short-term rates began “eating” the company’s profits. GRAPHS THAT UNITE THE TWO THEORIES To match corporate behavior with the shape of the yield curve in each phase of the cycle. To meet both required reserve ratios and higher loan demand. the shape of the curve is a reflection of future expectations which will determine basic supply and demand.192 same time that loan demand is also high. Long-term rates are viewed as the sum of shortterm rates combined into a longer maturity. and the price of long term debt is declining. Any company with steady demand will generally out perform the economy. As stated above. The lack of capital expenditures is a leading indicator of “”trouble” down the road. even if profit margins are squeezed. Consumer spending is now declining. With higher debt levels. 3) Segmented Markets Hypothesis: Banks have raised short-term rates by selling securities to meet loan demand in prior periods. firms delay major purchases of plant and equipment because they are uncertain about the direction of their respective sectors. Thus.193 Figure 8-1 YIELD TIME 1) Phase: Contraction/Recession 2) Expectations Theory: Short-term rates are above long-term rates and long-term rates are expected to decline. this is “self-fulfilling prophecy”. Companies do not want to lock in a loan at a higher rate and wait for interest rates to decline. . and become worse when firms delay building their inventories. healthcare maintenance firms. However. and companies with steady demand and short product cycles will do well. because recessions begin when investment and loan demand declines. industrial companies may not be demanding loans to maintain fixed assets. 4) Company Behavior: Most companies retrench and attempt to deplete existing inventories. “sin” companies like alcohol and tobacco. Long-term debt is now in demand. and the slightly higher price they charge for their services. but rates are now lower and companies begin to borrow at these lower rates for large capital expenditures.194 and consumer staples like bathroom tissue etc. Figure 8-2 YIELD TIME 1. Housing. 3. These sectors are deemed “interest sensitive”. Lower risk companies have the greatest acceleration of EPS beyond the cost of capital. Segmented Markets Hypothesis: Companies with more fixed assets and steady operating leverage can best benefit from lower interest rates on long-term debt. . Expectations Theory: The yield curve tends to be flat (it may be slightly ascending or descending). 4. The expectation is that the Fed is through with cutting rates and that now would be a good time to “lock” them in. will at least maintain the value of most portfolios. Company Behavior: Firms will extend inexpensive loans among themselves and to consumers. Phase: Recovery 2. Refinancing loans is again common for financial institutions. banks and utilities will benefit from lower cost borrowing. These firms tend to use a lot of debt and have large periodic expenditures which would lower capital costs if bought with lower cost loans. before interest rates rise and inflation takes hold. . perhaps several times before inflation appears. Phase: Expansion 2.so called “big ticket” items. Higher GDP growth and higher consumer spending offsets the higher risk. Intermediate industrial goods (used to produce consumer end products) and transportation companies begin to do well. Segmented Markets Hypothesis: Long-term debt is still in high demand. this scenario represents a collective race for immediate consumption. The economic imperative is to borrow and buy “now”.195 Figure 8-3 YIELD TIME 1. 4. Expectations Theory: Companies expect interest rates to rise as the economy “heats” up The Fed does indeed raise rates. At this point. Company Behavior: Consumer spending has been bolstered by a great employment market and loans are extended for “consumer durables” . 3. the equity markets are in full swing which represents the main source of financing for many of these firms.autos. boats. Companies with higher operating leverage begin to do well. appliances. Essentially. only to face the consequences of a high beta during the latter part of the phase. In this market. The financial sector begins to succumb to interest rate worries and lobbies for a cut. Expectations Theory: “Mixed signals” are given off by an economy that is unsure of which direction it is going. . Company Behavior: Both personal income and interest rates are relatively high. while others see diminishing returns and hope for a rate cut. Segmented Markets Hypothesis: At the market peak.196 Figure 8-4 YIELD TIME 1. it is “winner take all” and some sectors are heavily favored over others. 3. In essence. Some think the Fed will raise rates to curb inflation. 4. there may be more volatility in the market. Phase: Plateau 2. and some of these stocks will “sky-rocket”. the flat yield curve mimics an early recovery and yet interest rates are too high for another expansion. causing volatility as investors are unsure of where to put their money. companies with high operating leverage and little debt do the best. but most companies take a “wait and see” approach as merger and acquisition activity slows down. However. Consumers are still spending even though there is less industrial activity. The tech sector can benefit at this time because their financing is independent (superficially) of high interest loans. Capital goods have benefited from replacement needs. Inevitably. STRATEGIC CONSIDERATIONS • 1) The market precedes actual business activity by approximately six months. increasingly cherish “noninterest income” that diversifies them away from dependence on the Fed. Borrowing at low rates in Japan. is doomed to failure. Those rates affect the equity markets and the type of assets that generate income. But . for example.and also determine how those cash-flows are funded. and in the period 2003-2008. By the time an investment is made. the level they are at. The length of time that they maintain a pattern. in order to invest in China. One reason that insiders profit is their ability to anticipate increased activity in their respective businesses. the distance between short and long. especially the behavior of foreign markets. The real utility behind the shape of the yield curve is that it graphically represents the business cycle at each point in time. we would be very close to the truth. as well as their inherent volatility. that curve would reflect economic conditions throughout the developed world. can turn them plausible.if we combined our own yield curve with that of other countries and weighted it by GDP. we witnessed another anomaly: gold and stocks were perfectly correlated.197 The student/investor should realize that the characteristic of a phase is not derived from its predicted patterns. the increased complexity of the economy. there may come a time when the stock market no longer reflects risk in the United States alone. If we were to conclude that capital structure decisions are an outgrowth of the relationship between long and short-term interest rates. the yield curve shifts and begins to favor another sector. Although it fluctuates and shifts to reflect immediate conditions. The one highly recommended . This is also the reason why “chasing profits” by investing in the most profitable companies. Banks. no other measurement so precisely captures investor behavior and expectations. makes companies less sensitive to the yield curve. Can a market go from recovery to plateau and miss an expansion? Can banks do better when interest rates are high? Can stocks do well in the first part of a recession? As improbable as the scenarios behind these questions sound. all determine the amount and timing of cash flows . but defined by its anomalies. The span can be from six months to two years before a stock becomes “overbought”. If earnings decelerate compared to the cost of capital. such a move would signify that the Fed considers inflation to be at an acceptable level and would penalize creditors with deflated dollars. Since debt levels . At this point. but that does not mean that the company is not a good long-term prospect. The greatest gains in the market occur before the Federal Reserve raises the interest rate three times in succession.198 strategy is to invest in a general market index about six month’s after a recession is officially announced. the Fed will try to increase the money supply with open market operations and other lending facilities. although it does not mean a company will make less than optimal capital structure decisions or lose value. No one except market professionals will be watching the index. Unfortunately. If after three to four years of recovery. it usually signifies that the sector is about to wash. 2. At this point. responding to individual cashflow/capital cost circumstances. the confluence of ideal income and capital cost conditions will only be temporary. prudent investors will curb trading because the potential risk outweighs the returns. History dictates that a market begins to recover at this point. as interest rates rise and the yield curve shifts. the market becomes more sector oriented. • 3) Two other aphorisms have historical merit: 1. While enormous speculative opportunities exist. • 2) A favorable leverage state will move a firm toward an optimal capital structure and a higher stock price.”a soft landing”. it does have the potential to flatten out and begin ascending to continue the bull market. and that the greatest gain will be at the beginning of a recovery. it may be time to shift money out of the market. When an investor has seen a company double its stock price. If the inverted yield curve is at low enough a level. investors begin looking to other sectors. and the market seems to respond to this level of anonymity. trying to engineer what is termed . and shortterm rates exceed long-term. However. Secondly. the Dow Industrials follow the utilities. pushing them down. follows the Dow industrials. Interest rates get too high for the “interest sensitive” utilities. Since the public equates economic behavior with the prevailing presidential administration. • 5) Major trends are confirmed by sequential movement in the same direction of all three Dow component indices. the greater proportion of beta is derived from the relationship between sales and market return. In fact. utilities. Historically. Unless a trend occurs in that specific order. the Federal Reserve tries not to raise rates in the last two years of an incumbent’s reign. If inflation gets out of hand. up or down. and may lead to a less predictable business cycle and a chaotic market. stocks generally do twice as well in the last two years of a President’s term than in the first two years. electric utilities lead the way. the Dow Theory is completely rational. forcing the Fed to take even more drastic measures. Moreover. Fewer orders push down the industrials because companies do not want to commit to projects at higher rates. it is an assumption of capital structure theory that operating and financial leverage balance each other. Lastly. THE BUSINESS CYCLE AND THE COST OF EQUITY The first misconception that investors have about beta is that companies with greater financial leverage have higher betas. the transports suffer because they are the main service unit for the industrials. losses would be high unless inflation was actually held in check. • 4) Political considerations may trump economic discretion. The average investor best not tempt fate. Lastly.199 are higher. it is not considered “confirmed”. Beta = (DOL)(DFL)(ROE) [(COV % Sales. At first. while leverage increases beta. the postponement of a rate hike may make it worse. a company with more financial leverage will have less operating leverage and vice- . the transportation index. Sales is a prominent part of both the degree of operating leverage and the covariance component of the equation. From a sector rotation standpoint. % Market)/Variance % Market)]. To reiterate the Mandelker and Rhee equation. industrials and transports. the firm continues to diversify with acquisitions.200 versa. after a downturn. In the second phase of its strategy. From the perspective of capital structure. the firm actually begins to lower beta in response to higher interest rates. Concurrently. Although some sectors employ more of both types. attracting equity through its higher EPS. It began to broaden its customer base by investing in risk lowering acquisitions that would diversify its operations. Playing the cycle. the company begins to retain more earnings and lowers its long-term debt to capital ratio. lowering operating risk. within any sector. it would take advantage of low interest rates until the market picked up. First. . the mixture of leverage will be similar and balanced. Simultaneously. 2. the firm begins to pay off some of its old debt. Lastly. Theoretically. the firm uses leverage to accelerate the change in EPS well past the rate of change in the cost of equity. while maintaining high demand for its products. its investment in assets would begin to pay off. Secondly. it nullifies the risk of leverage at higher rates. and rates were increased. As the Federal Reserve raises interest rates to stave off inflation. and trying to broaden its customer base for the next profit cycle. the “ideal” company would have a low beta because it would be well diversified and be able to increase its beta with more debt and/or acquisitions. it took advantage of lower interest rates and began taking on debt and raising its beta just as the economy was improving. During the stagnant market. it lowered its debt ratio and began restructuring its capital towards an equity base. as interest rates were continually raised. Thus. the cost of equity was at a cyclical low because the market had declined and the Federal Reserve had lowered interest rates. and the company would raise its return on equity (ROE). the company prepared for a downturn in two different ways: 1. EPS was rising and an “over heated” economy ensured that demand for its products was stable. decreasing beta just as the business cycle transitions to a plateau. At this point. the “ideal” company engineered a strategy that took advantage of three cyclical characteristics. By paring down its proportion of debt to equity. It jettisoned its high interest debt and positioned itself for greater solvency. the rate of earnings increases will slow and be suddenly eclipsed by the cost of equity. we often see firms with twenty percent earnings increases .e. Performance was spotty at best. The market simply factors in diminished future prospects. beta was used as a tool for predicting stock prices. earnings will usually outpace the decrease in the risk-free rate. However. many firms hire economists to guide them through the various pitfalls and missteps.201 On the other hand. The banking industry in particular is exposed to cyclical risk. beta encompasses cumulative volatility and is not stable in the short term. institutional investors’ attempt to time the market was futile. which carries over to all those who are influenced by the prime rate . Such anomalies can be further punctuated by . Preposterous you say? A fantasy? While even the best run corporations cannot go through every cycle with such machine like precision. it will find itself “out of sync” with changes in the cost of capital. and so it attracts equity funding with its higher EPS and substitutes it for debt. Beta rarely performs as expected in the short-run. At this point.twenty percent in stock price. the firm is both prepared for a downturn and yet “cautiously optimistic” about future prospects. It is always possible (although not likely) for a high beta stock to only react violently when the market is declining and to make meager gains during the expansion phase of a typical cycle.which encompasses at least some aspect of nearly every sector in the economy.. right after the theory of the CAPM was proposed. Again. it can do nothing about the systemic risk of an overheated market. which will begin to accelerate. and not to gauge comparative risk. but even if earnings are stable.lose and not gain . Wall Street immediately jumped on the bandwagon. the premium is placed on foresight and not hindsight because if a company adopts these strategies as a reactionary response. Investing in high beta stocks during an upswing and then in low beta stocks during a downturn. a higher beta will create an over reaction to a market downturn. the real culprit was methodology. In fact. i. On the downside. In essence. The third phase of the strategy lowers its beta in response to a stagnate market but also prepares for the next business cycle by diversifying away some of its operating risk. then that industry may prosper with the effect of a greater “alpha”. the risk-free rate is manipulated outside the system by the Federal Reserve. it is no longer as dependent on the market. the singular relationship between beta and the risk premium will determine the coherence of the model In the following example.the part of a regression that depends on factors outside of market influence. The industry may have a collective beta of perhaps “1”.05 HIGH 0. Periodically.098 0.6 % 9. For example. but hardly reacts at all during a market downturn. is the primary economic factor affecting the cost of equity. 0. steel. known as the market risk premium. but it changes yields (not coupon rates) based on demand for treasuries. The more volatile stock market index can vary from negative twenty percent to positive twenty percent over the course of a year..048 CAPM 8. Table 8-1 EQUILIBRIUM Risk Free % LOW 0. Ultimately. and less beta.048 0. and the difference between that figure and the risk-free rate.048 0.e. and high. we will observe the effect of a one percent change in interest rates on the CAPM.25 Market 0.202 a low correlation coefficient and a high but volatile alpha . THE CAPITAL ASSET PRICING MODEL AND SENSITIVITY ANALYSIS The true worth of beta is gauged in relation to the other components of the CAPM. The CAPM is a dynamic model that changes daily as the market changes. i.098 Risk Prem.098 0.8 % 11% . defining three levels of beta: low medium. if tariff policy suddenly favors a particular industry.05 MEDIUM 0.75 1 1.05 Beta 0. 038 0. and when either earnings or the market declines.25 Market 0. 0.203 Table 8-2 INCREASE 1% LOW MEDIUM HIGH Risk Free % 0. This next example will show sensitivity to market changes.098 Risk Prem.05 0.75 1 1. high beta stocks fall precipitously. This is a systemic advantage that encompasses firms who must finance with more equity. and temporarily out perform the market. as the market keeps rising.75 1 1.6 % 9.048 0.85 % 9.098 0.038 0. Companies that have high betas may have a difficult time turning to the credit markets for necessary financing. but then we must remember that earnings are highly correlated with the market. the cost of capital goes up for those firms who use leverage.06 Beta 0. increasing the risk-free rate actually decreases the cost of equity (from 11 % to 10. However.048 CAPM 8. Table 8-3 EQUILIBRIUM LOW MEDIUM HIGH Risk Free % 0.06 0. On the other hand. these companies have a much higher cost of capital.06 0.098 0.75 %).05 Beta 0.098 0.05 0. a market decline actually reduces the cost of equity. When interest rates rise.8 % 11% . It also shows the importance of always examining the cost of equity in the context of earnings.038 CAPM 8. a decline in market value implies that earnings may be decreasing by an even greater amount. those who use equity have a competitive advantage.25 Market 0. The higher beta allows a firm to both escalate earnings with lower capital costs.098 0. especially if the market ignores the higher rates and keeps ascending.048 0. 0.8 % 10. If we examine the CAPM in isolation.75 % Notice that with the high beta stock (1.25).098 Risk Prem. 068 Risk Prem. it is doubtful whether a high beta stock could maintain its price.2 or 20 % . One of the keys to understanding the cost of equity is the recognition of how market dependent it is. the cost of equity dropped more than that of low beta stocks. Table 8-5 The Market goes from 20 % to 15 % Return .05 0.018 0. must have higher operating volatility by definition. without a market response.2-.05 0. The next table should convince readers that market reaction is paramount: we cut the interest rate in half.018 CAPM 6. Although interest rate changes can move the market.25 % Even with the introduction of extraordinary cash flow during this period (beating the average company in earnings gains).75 1 1.068 0.05) = 0.204 This scenario calls for a three percent drop in market return to 6.05 + (1)(0. Certainly. Any firm that finances with equity.8 %. 0. they will represent only a small fraction of the cost if the market does not respond.35 % 6.15-.Riskfree).15 or 15 % .25 Market 0.8 % 7.068 0. and has a higher beta.usually higher operating risk attributable to volatile sales and earnings. but the student/investor should consider the source of the higher beta . DECREASE LOW MEDIUM HIGH Risk Free % 0.05) = 0.05 Beta 0.05 + (1)(. The equation is the CAPM with a beta of one: Risk-free rate + (Beta)(Market rate . Table 8-4 MKT.018 0. 2 .025 +(1)(0. Figure 8-5 Interest Rate Market Return The most constructive advice that an investor or corporate professional can receive is to keep beta at a minimum. the market creates inflation. i. In statistical terms. which increases the market which causes inflation.risk-free rate).05) = . the relationship is in a state of perpetual change which diminishes forecasting ability. which lowers rates. Low beta stocks are rewarded out of proportion to their risk profiles. However.. the danger of . a downward bias exists for high beta stocks.2 or 20 % Moreover. Expressed as a logarithmic curve.2 or 20 % . while high beta stocks are penalized. the relationship is a function of yield curve behavior.205 Table 8-6 Interest Rates are Cut in Half (the risk-free goes from 5 % to 2.. The correlation between rates and market is not explicit and cannot be permanently quantified because volatility in both credit and equity markets changes the relationship.. inflation and aggregate demand among many factors.2 .e... Ultimately.05 + (1)(0. the empirical evidence is unchallenged. which raises rates which lowers the market.025) = . there is no factor in the model that expresses the relationship between interest and market return except for the comparative risk premium (market return .5 %) . asset growth was leveraged through the stability of sales. the small beta. No one ever bragged about how Fisher was beating Wall Street estimates and the stock was not heavily traded. Fisher had upside potential but was protected from downside risk One famous study was conducted by the formidable research team of Black. While some high beta stocks offer an exception and achieve some degree of stability by their very size alone (Intel. is with Chapter One’s introductory illustration. but the low profit margins put the company in the “high turnover” class. it took off like a guided missile. The company was well diversified with a low operating leverage.higher beta stocks produced lower returns . 1965. the imperative is to seek out firms who can raise beta and take advantage of upswings. With a small profit margin of 3 to 4 percent. waiting for word of the next growth opportunity. for example). Mathematically. these large jumps would not be interpreted as volatility because they were isolated to specific short periods. The “mystery” behind Fisher was that the stock simply did not move on the basis of sales or profits. They found that the entire period from April 1957 to December. Nevertheless. sales and cash-flow were comparatively large. but not have the type of operating risk that would damage the stock during a downturn. Eventually. but had a debt to capital ratio that sometimes approached seventy percent because of numerous leveraged buyouts. Jensen and Scholes. thus. In fact. the company maintained a beta of about 0. How was that possible? Besides balancing operating and financial leverage. Fisher was capable of increasing sales even in a downturn.6 and was able to increase their share price sevenfold in about eight years. was characterized by a skewed risk-return tradeoff .206 making this generalization is to lose the upside potential that so many higher beta stocks provide during the expansion phase of the business cycle. One of the best examples of how beta can be deceptive. but once the word “acquisition” was mentioned. it would again settle in for another soporific interim. Fisher Scientific (now Thermo-Fisher Scientific). this firm rarely reacted as much as the market. In other words. The reason for the lower “price” of short-term debt is that the lender incurs less . In fact. Very reputable financial professionals thought that a “new era” had dawned. a high beta should not condemn a stock’s investment potential. low beta stocks can out perform the higher betas because of a tendency to hold their gains.207 than low beta stocks. the anomaly can be explained by the extremity of the movement. However. Its beta is no longer exceptionally high and it has diversified into voice generated programming and video games. Therefore. Microsoft has become more investor worthy as it gives out special dividends and trades between a much narrower range than it once did. while it may take a year or more for it to achieve a new high during an expansion. could be bought for ten dollars and change. Again. characterized by permanently high price to earnings ratios. these findings are often pointed out as a condemnation of the entire CAPM. On the other hand. even during an upswing. Another example of being “burned by beta” was during the tech stock speculation of the 1990s. many high beta stocks that commanded a hundred dollars a share just a few years earlier. because it is over exposed to a downturn. and volatile stocks that would “go back up”. If income streams are certain over a long period. By 2001. Unfortunately.given a “normal” yield curve. Even a well run company like Google is not an investment grade stock as of this writing (2008). a stock with a high beta can triple the markets downward path in only a few short days. That jack rabbit-volatility almost always represents a return to relative book value and away from speculative risk. This period offered rewards to high beta investors only if they were adept enough to part with their investments before they “fell through”. the volatility of certain tech stocks has led to more speculation and less investment. the cost of borrowing short-term would be significantly less than borrowing long-term . CIRCUMVENTING THE OPTIMAL CAPITAL STRUCTURE The relationship between short-term interest rates and long-term rates defines the business cycle. a well managed high beta company can lower its beta as it takes advantage of lower cost equity. Most businesses need to plan far enough ahead to be competitive and frequent rate changes on top of frequent payments will increase default probability. as would default probability. a firm may temporarily substitute the increasingly expensive short-term loans for long-term debt. However. Lower beta companies. Interest expense would rise in the interim. but this movement away from an optimal capital structure would position the firm for greater gains once a recovery commenced. there is a tendency for short-term rates to rise. who have less operating risk can make a strategic play to lower capital costs in the long-run by accepting these immediate risks: they do not face the prospect of great market volatility which would damage a high beta firm that accepted the same risks. in the hopes of avoiding being “locked in” at a higher rate. there is always the temptation to use less expensive debt when the firm is in a temporary “holding” pattern at the top of the business cycle: it may wish to delay major projects because interest rates are expected to decline. THE GAME OF CAPITAL STRUCTURE “GOTCHA” . In essence. the borrower incurs much greater default risk . acting like an insurance premium. Thus. This disparity between the price of debt and the risk of debt skews the absolute cost of capital in the direction of greater risk. the company is forced to choose higher cost debt because it is less risky. the higher price paid for the short-term loans buffers the firm against uncertainty. Any pricing anomaly in the borrower’s favor is short-lived because the loan will be frequently renewed at the current rate. and counterbalancing the risk of default because the action is only implemented temporarily. This trade-off of more immediate risk and expense for the prospect of lower rates in the future is a luxury that high beta companies cannot afford. Essentially. As banks sell off short-term securities to meet loan demand. Since a well run firm matches its cash-flow with its funding needs.loans might have to be paid when cash-flow is negative .there is greater risk incurred because of the potential for interest rate fluctuation. indicating the prospect of a contraction.and.208 risk. Unless earnings are maintained and even accelerated. company officials sometimes feel indestructible. because cash-flow is not stable enough to warrant it. expecting to “dump” the shares before getting . pushed up by a market that has surged through a few years of expansion. When a high beta stock is at its peak. it is kept until sold. retained earnings provide sufficient equity given the lower economic prospects that occur during the plateau phase of a business cycle. In fact. Thus. the cost of equity is relatively high at this stage. Once stock is issued. the investor must enter the deal like a gambler. At the same time. taking on more risk than is warranted by the economic outlook. the cost of equity will be rising and eclipsing the rate of earnings.209 High beta companies who peak near the top of a business cycle primarily finance with equity. The unbridled optimists are matched with the unsophisticated investors. the high beta firm is minimizing capital costs by raising equity when the stock price is high: more funds will be raised with fewer shares. the loan does not expire like short-term debt. sometimes vectoring off into a different direction altogether. an investor needs to view equity issued late in the business cycle just as a company views short-term debt. Naturally. one entity demands plenty of equity and the other supplies it.in this game. However. Therefore. Their unbridled optimism is punctuated by huge bonuses and a stock price that is soaring.except . the” perfect storm” occurs in capital structure. The higher risk needs to be counterbalanced with the potential for appreciation . Long-term debt is both prohibitively expensive and risky. With the prospect of both lower capital and consumer spending. an expanding market creates many unsophisticated investors who have never seen a downturn. What each player in this scenario does not realize is that they have upped the “ante” at the wrong time. Opportunities for growth are sighted and the firm may begin raising capital to fund large projects. the forecast for a stable stock price is dim. the investment begins to implode as soon as any of the major industry participants misses earnings expectations. The market seems like a Lotto ticket that always pays off. Under normal circumstances. but offers the opportunity to observe the workings of other forces. IDEALIZED TRENDS Business cycle behavior defies expectations more than it confirms them. Thus. the investor can therefore be alerted to the danger of an exception . both parties end up like poker players in a Las Vegas casino.210 “burned”. Temporarily. Speculative excess. . the cost of equity may actually decrease during an upswing. By recognizing the rationality behind the “ideal” business cycle. a “bubble” created by government action or inaction. The market may go into a tailspin just as the expansion phase is expected. In such a scenario. Exceptions to the rule do not negate its logic. and the result will be a sector (housing or Reits) that remains profitable beyond its capacity to generate income. banks and brokers causing a severe credit debacle in 2007. An example of such a bubble was the housing speculation that occurred in the early millennium which was purported to be an outgrowth of interest rates being too low for too long. the laws of supply and demand are circumvented by creating an artificial scarcity for example. Often. High beta stocks may recover earlier than low beta stocks. This imbalance can only end up affecting other sectors. can occur. In essence. even with the best of intentions.whether it be a speculative bubble or a more complicated supply and demand issue such as occurred with the price of oil. the investing public will put too much capital in one area to the detriment of others. a sector whose internal dynamics would normally let it prosper for one phase only. the political motives of a lobby can be exhibited in some piece of legislation that seems inconsequential but has repercussions in the market. and indeed we saw a fallout with mortgage companies. ended up being favored by the economy for three. A Time B C VARIABLE Interest Rates Market Index Cost of Equity Relative Beta Relative Op. Financial Leverage 2. The reader should observe that economic risk is less in phases A and B and it is at this juncture that corporate risk can be increased. and attempt to lower operating risk through diversification. Equity is less Expensive than Debt (relatively) The managerial imperative is to find its own “efficient frontier” within the parameters of its industry. Market Spending Peaks 2. Leverage EPS Acceleration PHASE A LOW LOW LOW LOW LOW INCREASING PHASE B MEDIUM MEDIUM MEDIUM MEDIUM MEDIUM HIGH PHASE C HIGH HIGH HIGH HIGH HIGH DECREASING PHASE A LOW BETAS PROFIT 1. Once a firm enters phase C. Flexible Mixture of Debt/Eqty 2.it should strive always to achieve the greatest return with a given level of risk. that is . Stability of Demand PHASE B BETAS NEAR “1” PROFIT 1.211 Figure 8-6 Market Cap. the Internet and even telecom equipment. Without the low beta of a . A very obvious example of an industry changing its risk profile is land line telecom. More Market Spending PHASE C BETAS > “1” PROFIT 1. reduce debt. corporate risk is no longer an option: it should decrease beta. These formerly regulated companies have increased their collective risk by branching out into wireless. but will still be buying bread and going to visit the .will be favored as well. During a contraction. when credit risk is still high. these firms peak much later in the business cycle. AT&T with a beta of around “1”. Consequently. banks. Extreme amounts of capital poured into these firms right at the market peak of the late 90s. telecom suffered more than most. When the market imploded. In fact.. more applicants qualify for credit and any “big ticket” item requiring a loan will be favored. stocks that do well during a plateau should be chosen on the basis of lower risk simply because the higher risk stocks that do well during this phase will be the first to fall.autos. In addition. is much like chasing earnings: the investor might gain from momentum. etc. The real value in sector rotation is to recognize that a diversified portfolio can be achieved by concentrating on firms that benefit from the recovery and expansion but should also have some defensive stocks in case of a contraction. For example. SECTOR ROTATION To flawlessly predict which sector will be the next to profit is a pipe dream. which was the wrong time to incur such risk. but could have avoided some of the damage by expanding in the earlier phases of the cycle. The typical utility has more financial leverage but not nearly the greater operating risk that AT&T has incurred over the years. Nor will it be possible to predict how long a sector will profit although six months of accelerated earnings will be an indicated minimum to be considered as “outperforming” the economy.212 regulated utility. Most sectors will be segmented by performance and all firms will not prosper at once. when interest rates are low. consumers will not be taking out loans for purchases. any industry that depends on interest rates . but just as likely will lose money because the large investors have already spotted “the next big thing” and moved on. housing. “Chasing” profitable sectors. however. SECTOR LOGIC The concept behind sector rotation is that the cycle favors specific industries within the bounds of interest rates and demand. only had a long term debt to capital ratio ranging from 15 % to 33 % throughout the 1990s. As a general guideline. electric utilities) Recovery II 1. Paradoxically. most economists will debate which phase the economy has entered as well as how many phases or sub phases actually exist. homes.. heat expanders etc. Food and Beverage Recovery I 1. Consumer Discretionary 3. The consumers with the best credit will be purchasing the homes and autos that stimulate the economy in a recovery. small motors. It is at this juncture that unused capacity disappears and manufacturers begin to expand by purchasing “capital goods” . Consumer Staples 3. “big ticket” items) 2.the tools and machinery that constitute the final product of the intermediate goods. Consumer Durables (autos. Transportation One pattern that is worth noting: both consumers and businesses follow similar borrowing patterns. Many sectors will “bleed” over into a phase whose interest/demand characteristic does not fit the industry. consumer staples and the healthcare industry are favored. all of this freight must be transported. as interest . often called intermediate goods. Likewise. shipping. will be in demand and set the stage for the late expansion.railroads. and the various transportation stocks begin to do well .213 doctor if needed. Capital Goods Plateau 1. When the recovery and expansion phases hit. the largest manufacturers with the least operating risk will expand with the greatest financial leverage. While the astute reader will notice that sector rotation is merely a detailed elaboration of the “Dow theory”. Consumer Discretionary 3. early plateau. Healthcare 2. the following table offers a rough estimate of a typical business cycle: Table 8-7 PHASE GOODS Contraction 1. trucking. Capital Goods 2. appliances. Additionally. Interest Sensitive (banks.sub assemblies. goods that are used in the production of other goods . Thus. Intermediate Goods Expansion 1 Intermediate Goods 2. the premium for economists is to observe how it is different. Naturally. at the top of the market there will be investment in junk bonds. IPOs (initial public offering) and even in firms without any earnings. This situation presents another anomaly: when debt is in demand by investors. albeit at higher rates. the company must pay a higher price. makes debt less expensive in the capital structure. As the market expands. the difference between stocks and the risk-free rate. the axiom that the most credit worthy customers need the least amount of credit is true. there is a “flight to quality”. This higher demand for debt instruments in the initial phases of the recovery. thus qualifying customers for loans. The greater certainty in the bond market attracts capital away from stocks. In fact. wealthy businesses and people use debt as a moneymaking tool rather than as a necessity. high quality debt instruments like treasuries and AAA bonds. This recipe for default costs financial institutions billions but no alternative system seems practical. credit is even more obtainable and less credit worthy customers will be picking up loans later in the business cycle. The successful firms who finance with equity do so not because they cannot qualify for loans. it is relatively inexpensive for the issuing company. At the prospect of a contraction. However. Individuals and companies who are the least credit worthy “somehow” end up paying more interest for loans because they did not possess the collateral early enough in the “game” when rates were low. a general movement into low risk. but when equity is demanded. but as risk premiums rise. The reader should notice that the risk premium. but because it is the most cost effective method. On the other hand. The basic reason for this anomaly is that personal and business incomes do not rise until later in the cycle. investors take on more risk. and the cost . firms who have a volatile cash flow will sometimes take on debt when credit terms ease and will suffer the consequences late in the plateau phase. the rising market raises the risk premium. helping to maximize stock price. firms can garner more funds from an equity issue despite its higher cost because stocks are in demand. and stocks are again favored.214 rates rise. also mirrors sector rotation. inflating the risk premium. This “survival of the fittest” scenario can be combated with knowledge of the business cycle and applied judgment. INDUSTRY RESPONSE TO THE BUSINESS CYCLE The following industries respond (not always positively) to the respective phase: Table 8-8 1) CONTRACTION Utilities Consumer Staples Tobacco Food. both beta and the cost of bankruptcy proceed to drop . Autos. Forestry Chemicals Steel Household Furnishings. Beverages Publishing Drugs Healthcare Apparel Table 8-9 2) RECOVERY Electric Power Paper Products. Appliances Crude Oil Banks Small Machine Tools Intermediate parts Defense Electronics Pollution Control Waste Management Table 8-10 3 ) EXPANSION Capital Goods Machine Tools Gold Mining Tobacco Beverages Drugs Cosmetics Oil Equipment Computer Systems Financial Services Table 8-11 4) PLATEAU All Types of Mining Oil Refineries Telephone Systems Communications Equipment Specialty Chemicals Transportation Aviation Aerospace .215 of equity When such an equity issue is supported by higher earnings.beta by the decrease in debt to equity and bankruptcy costs by an earnings generated decrease in default probability. 216 Table 8-12 CROSSOVER SECTORS INDUSTRY Publishing Beverages Mining Oil Electric Utilities Tobacco Drugs Defense Electronics PHASES Contraction. For example. Plateau. Expansion. Recovery Expansion. it is much better to take advantage of low interest rates and speculate during the recovery phase. why would any investor buy stocks on margin when interest rates are high during a plateau phase? If the investor has a speculative bent. it is almost too late to move into a sector because all the money is made through early anticipation. The decisions to invest in a certain sector must anticipate its economic milieu. Plateau ECONOMIC SIGNALS Economic indicators are so mixed that it is very difficult to achieve consensus among leading economists. Plateau Contraction. Charts and tables set up the illusion that anticipation of an economic phase is effortless. and much money is made in making a correct forecast. No one needs to be a fortune teller to almost guarantee that interest rates will drop during a contraction. However. Recovery Expansion. Recovery Contraction. Expansion Recovery. once the majority of economists agree on the state of the economy. The following tables give a brief outline of what signals to expect in each phase: . Contraction Contraction. The average investor can benefit from knowledge of economic signals if only as an instructional tool for “what not to do”. Plateau Recovery. but these tables do not function as “tea leaves”. 5. The percentage change in M2 turns from Negative to Positive. Interest Rates Increase. Non Farm Payrolls Increase. GDP Percentage Change turns Positive. GDP Declines as a Percentage Change (2 % instead of 4 % for example) 3. 6 Month Moving Average of the Rate of Change for the Real Monetary Base turns Positive . PLATEAU 1. Non Farm Payrolls Increase 4. GDP Declines as a Negative Percentage 2..217 Table 8-13 CONTRACTION 1. 4. Initial Unemployment Claims Decrease. 2. 4. 3. 12 Month Moving Average of the Rate of Change in the Federal Funds Rate turns Positive. The Moving Average for the Rate of Change of Industrial Production turns Negative. Table 8-14 EXPANSION 1. 12 Month Average Percentage Change in Federal Funds Rate turns Negative. Unemployment Claims Increase 2. 2. 12 Month Average of Industrial Production Change becomes Positive 3. Interest Rates Decline. 6 Month Moving Average in the CPI (Inflation) turns Positive 3. RECOVERY 1. but investment on the basis of such indicators is in itself a risky proposition. In some economies there will be inflation and recession at the same time and so the acceleration in the rate of change will certainly not be an expansion phase indicator. (Back to Table of Contents) . the premium is placed on anticipation of the phase. Common sense and coordination of several factors (some not enumerated) will be the watch words. Again.218 The arbitrariness of some of these signals is apparent if one views an indicator like the CPI in isolation. The delineation of such costs is crucial. data was obtained from old newspapers and microfiche. Nevertheless. We can now obtain detailed information of both high quantity and (some would argue) quality. If academic theory was readily observable in the marketplace at all. A correlation between R and D (research and development) expenditures and near-term sales. would be considered time-consuming research. sometimes losing track of overall objectives in pursuit of esoteric data. the greater number of sources has enabled us to compare data and ground ourselves more firmly in the truth. the FASB (Federal Accounting Standards Board) and the SEC has not made such disclosure mandatory although it would immeasurably aid the investor in gauging risk. Before the ubiquity of the PC. Although it is standard practice to itemize costs between fixed and variable. The one area. . As quaint as that may seem. because the greatest amount of corporate activity rests entirely on the concept of operating risk and its concurrent measurement. in which there is little more transparency than there was in 1965. The SEC (Securities and Exchange Commission) does not require corporations to break down their costs into fixed and variable categories in required financial statements. is in the accounting for fixed and variable costs. it was restricted to such general truisms as “higher earnings lead to higher stock prices”. not typically initiated by the average investor. Although more communication has also created the exchange of information with dubious quality that most find obfuscating and confusing. operating leverage. confusion exists when asset categories overlap and the two costs become indistinguishable.219 9 OPERATING RISK One major advantage of the information age is that we can observe the transformation of academic theory into practical application. investors would plot charts out of stock guides. for example. Technology has transformed data. for example. and variable costs can be differentiated into (Quantity x (Variable Cost/Unit)). operating leverage provides the foundation for structuring a business. and automation . Thus. this . it is most affected by the business cycle and the timing of demand for that company’s product. Since fixed costs reflect the level of automation. even the average level of employee education is dependent on operating leverage. there is little discretion. Profits = Sales . etc. will compete on the basis of a much higher level of technology than a department store. but the need for long-term commitment of capital is also an imperative. Almost without exception. marketing. finance.many observers consider them implicit in the type of product one is selling. would reducing them change the nature of the business? Is such reduction possible? The answer to both questions is affirmative. Sales can be further differentiated into (Quantity x Price).Fixed Costs. Fixed costs in such a merger would also remain stable.220 Since fixed costs are often associated with the technological inputs of a business that is machinery. However. and all competitors have one. Increasing profits is a simple matter of either increasing sales or decreasing costs. but the quantity of units would increase. For example. the basic production process still requires the same inputs and remains unchanged. and almost beyond management’s control. and thus reduce their cost per unit by increasing capacity. it will be deemed “capital intensive” and have a higher level of fixed costs. This operating risk is sometimes referred to as “economic risk”. the type of business is the greatest determinant of fixed costs. FIXED COSTS AND ECONOMICS In the classic breakeven point equation. A semiconductor firm. I must buy the emitter and pay the fixed costs for maintenance. In fact. One look at the breakeven equation dictates that variable costs would increase as well. if my business requires a neutron particle emitter. The objective of many mergers is to “share” fixed costs. test equipment.Variable Costs . Consequently. and especially the amounts and consistency of its component parts. the level of technology characterizes the amount of fixed costs in a business. These companies have almost invariably high operating leverage. Such competition becomes a rather expensive proposition for shareholders who tend to invest in these companies at specific points in the business cycle and then sell their shares when the cost of capital climbs and further expansion is unwarranted. While prices are determined by the market. it is its failure to correlate higher fixed costs with higher prices.such participation connotes a higher profit margin and high unit prices. higher fixed costs raise the breakeven point for sales. that only a few enterprises are willing to take the risk and participate. revenues increase . and the old technology becomes obsolete. Another. Eventually.221 proportional rise in variable costs to fixed costs lowers the overall risk to the merged company. and the sources of supply increase.usually . The result is a lower price. because it can be more easily duplicated. Thus. If such technology offers a competitive advantage. If there is one major flaw to the breakeven point equation. However. They are a fundamental sign that a process requires so much capital. but the price charged for the product decreases as well. the supply of this technology will begin to eclipse demand for it because “new and improved” processes replace it. not only do fixed costs proportionately decrease. the smaller will be its supply. more individuals jump into a profitable market. Industries become “capital intensive” because the competitive “edge” goes to the businesses with the latest technology. its relative complexity decreases. profits are driven by churning out more of a product. i. it will be in demand.e. and command a higher price merely because of its limited source. which must be purchased at “any cost”. When a process is gradually improved over time. And .. The greater the novelty and relative complexity of technology. when technology is cheapened to the point where its supply is vast and unending. at least one of those variables needs to change. High fixed costs are a barrier to entering a market. much more insidious reduction in fixed costs occurs when production processes quietly become obsolete. Since sales are a function of both price and quantity. This need for competitive strategies has been thoroughly analyzed and documented by Michael Porter. A “partnership” arises between the two corporations. “high tech”. His outstanding research produced three generic strategies that most companies emphasize and share: 1) Cost leadership 2) Product differentiation and 3) Focus on niche markets. Eventually. almost at will. but by this time.222 because unit quantities increase and not because prices rise. more of a good can be produced with less investment. a time in the business cycle which favors its sector. the very ease of production has allowed numerous competitors into the market. if demand is stable and not increasing. The chain of cost reduction continues: cutting a firm’s fixed asset costs will allow more participants to enter that firm’s product market albeit at a lower profit than was previously acceptable. the market will be glutted and prices will have deflated. An example would be a department store that rents unused space to another business like McDonald’s or Starbucks.e. cost leadership. It is the prime reason that companies need to remain flexible and willing to change. and the department store is spreading out the same fixed costs over more revenue. i. No company can afford to wait for “its time in the sun”. a company can charge lower prices because it cuts variable costs and spreads fixed costs over greater production. which will make a firm’s inputs less expensive. the firm can stamp out a product that was previously deemed. In fact. In the first.. At this point. Not only is the company spreading fixed costs . This “negative economic synergy” is one of the bailiwicks of technological change. vendors cut prices to meet competition. forcing companies to compete with more “asset turnover” and less profit margin. Each strategy uses fixed cost manipulation to gain market share. or encumbered by a limited number of products and processes. If a company is fixated on a specific method of “doing things”. and tests it with the same devices it uses for chips in PCs. The second strategy would be exemplified by a semiconductor company who creates a new chip for a medical prosthesis. As more participants enter a market. it will be doomed by a lack of diversification. However. Increasing demand for quality computers (ones that did not lose information when you turned on the vacuum cleaner). THE CASE OF COMPAQ COMPUTER During the 1990s the computer business became so hotly competitive that many companies fell by the wayside. the research and development in pharmaceutical companies. creating a less risky cash flow and changing the character of its operating leverage. created a commercial environment for the Internet that stimulated shopping. Fighting to keep their heads above water. While computers are certainly high-tech. such accessibility came with a price of its own: many of these companies were well managed and yet saw their profit margins shrink. but one does not have to be an “electronics whiz” to connect several of these modules together and form a computer.. with greater “mean time between failures”. the niche status can deteriorate over time if technology becomes easily duplicable. Because of frequent “price wars”. there was less . computers became very accessible to the average person. battled to remain viable. a PC manufacturer who swallowed up the floundering enterprise system maker.e. and social activities as well. and many participants enter the same market. Compaq. As in the case of “negative synergy”. The modules themselves are relatively sophisticated.223 over more units. only later to see itself swallowed up. meant that machinery no longer had to be replaced because it “wore out” like tennis shoes. User friendly operating systems like Windows 95 from Microsoft. consolidation became a survival mechanism rather than an opportunity. Flexible companies must ordinarily adopt all three strategies. and naturally require higher prices to cover those costs. game playing. they can be easily assembled from a series of modules. it is diversifying its product line. The computer industry was changing – in what seemed like a nanosecond. Niche markets are sometimes specialized enough to require higher fixed costs i. In fact as storage space became great enough to eclipse our ability to use it. sometimes emphasizing one to the exclusion of the others and sometimes creating a hybrid of all three. Digital Equipment Corporation. Gateway. Custom built computers by the local electronics store could compete with huge PC makers who had the added burden of high fixed assets and costs. and parts were not standardized enough to be interchangeable. measured by Operating Income / Sales. Frequent changes in the production process. Manufacturers like Dell stopped selling computers in stores. computer speed needed to keep pace with ever more sophisticated programming. they were a requirement. THE NATURE OF COSTS AND MARGINS The tell tale sign of a deteriorating economic predicament is simply falling operating margins. The breakeven point equation best establishes this relationship: . Forecasting demand became a hit and miss art. In other words. and unsold inventory forced Compaq to adopt a new distribution model called. Note that operating income in this instance is before depreciation and amortization has been deducted to derive EBIT (earnings before interest and taxes). computers marched right out of the scientific laboratory and into the living room. “the distribution alliance program”. ended up closing the majority of its distribution centers. When operating systems becoming “user friendly” and programs became compatible. Computers were not only accessible. For years. coupled with an increase in market participants led to an inflexibility within the largest manufacturers. putting one module on top of another was like putting Chevy parts into a Ford: It might get you to the corner store and back. computers would be made to order and not to forecast. almost like a service company. computer makers would compete on this “custom style” basis. and computer giant. but not much farther.224 of a need to replace computers with an “upgrade”. Inevitably. 66 % rise in variable costs totally dilutes the effect of a one hundred percent increase in sales: Operating margin declines back to 20 %.225 Table 9-1 FIXED COSTS (FC) = 100 SALES (S) = 500 QUANTITY (Q) = 100 PRICE (P) = 5 VARIABLE COST PER UNIT (VC/UNIT) = 3 (P x Q) .(200 x 3) . Table 9-2 FIXED COSTS (FC) = 100 SALES (S) = 1000 QUANTITY (Q) = 200 PRICE (P) = 5 VARIABLE COST PER UNIT (VC/UNIT) = 3 (P x Q) .FC (100 x5) .FC (200 x5) . variable costs that rise as a percentage of sales.Q(VC/U) . a mere 16. In the next table. . As long as more units are being sold.100 = 100 OPERATING MARGIN = 100 / 500 = 20 % As long as fixed and variable costs are stable (that is fixed costs remain the same and variable costs are a stable percentage of sales) operating margins will increase as sales increase.Q(VC/U) .(100 x 3) .100 = 300 OPERATING MARGIN = 300 / 1000 = 30 % Notice that quantity is the only variable that changed. do not adversely affect margins. Taken to the extreme.100 = 200 OPERATING MARGIN = 200/1000 = 20 % Over time. The real cost differences are in such things as TV advertising. it . regardless of the mixture between fixed and variable costs. endorsements etc. Their collective investment in fixed assets is large.FC (5 x 200) . “No-Brand” corn flakes never have a prize inside in order to keep costs low. breadth of distribution. The concept is obvious when one walks into any supermarket and prices generic corn flakes vs. that push up the cost to sell the product. In some industries. (seats) is limited. variable costs rise as a percent of total costs and even accelerate as fixed costs decline: increased competition “raises the bar” and each sale involves more distribution. Fixed costs of the process are similar. although some “quality” premiums are built into the name brand that would make it slightly more expensive. The two cereals are nearly identical in production costs.226 Table 9-3 FC = 100 S = 1000 Q = 200 P=5 VC/U = 3. a name brand. When a plane (a fixed asset) carries its capacity of 300 people. Such “parity” would not be destructive of margins if only variable costs did not rise to compensate for the reduction.5 (P x Q) . when the increase in total costs is greater than the increase in sales. but their productive capacity. The airline industry.Q(VC/U) . operating margins suffer. operating margins rise when the percent change in sales exceeds the percent change in total cost. advertising. fixed costs and not variable costs are the culprit. packaging and marketing expenses.5) . The tendency in most businesses is for fixed costs to become a smaller percentage of total costs over the life span of a product. In fact.(200 x 3. has never been profitable for an extended period of time. for example. they showed the capacity of Compaq management to adapt to a situation that was rapidly getting out of hand. The ideal situation for Compaq would have been to effortlessly increase its capacity and incur the cost increases that were normally contingent on expansion . which it found later in Hewlett Packard. While these investments were immediately extraneous to the production of PCs. the main problem with Compaq’s costs were not a rise in fixed costs. However.by selling more units. Thus. The “higher ups” could see the writing on the wall . in an effort to make up in variable costs what it must spend in fixed costs. In Compaq’s case. and also the reason why seats are created to conform to the dimensions of a runway model. Such a strategy would have raised variable costs in proportion to fixed and margins would have stabilized at a very low level. Ultimately. However. The only way that Compaq could have maintained its margins would be to either expand production and sell at drastically reduced prices. Most airlines have responded by cutting frills such as pillows. Compaq was plagued by an overcapacity and resorted to the aforementioned change in its distribution model. or cut its variable cost per unit enough to bring up operating income. albeit at a lower price. which acted so long as a barrier to entry and were now diminished. they also entail more risk because raising them raises the breakeven point for sales.227 cannot add capacity or spread the fixed costs of a plane among a greater number of units. but a decline in . both fixed and variable costs were rising faster than salable production. and meals. FIXED COSTS AND THE BREAKEVEN POINT FOR SALES While higher fixed costs act as a barrier to entry.they needed to diversify and do so rapidly. because competition had lowered prices and flooded the market with PCs. peanuts. No company can cut its operating margin in half and continue to operate unscathed. it had finished a deal to buy Digital Equipment and invested heavily in the Internet (CMGI and Alta Vista). Neither path was feasible and it was natural for Compaq to look for a merger partner. Compaq needed to increase margins and could not rely on sales alone. That is the nature of the technology. a company would have to raise its unit quantity when fixed costs rise. a utility company who has higher fixed costs (turbines etc.228 pricing power that was derived from an acceleration in variable costs that replaced fixed costs. and yet the result is the same when fixed costs are raised. The risk of raising fixed costs (or even buying the stock of a company with high fixed costs) stems from the relationship between sales and operating income. It is far easier to raise a price than to gear up production processes and churn out more units.) will also have a high degree of sales stability because their product (electricity) has a demand during downturns. A company who can raise prices without customer complaint is in a far superior position than those who must do what the market dictates. Table 9-4 FC = 100 VC/U = 3 Q = 50 PROFIT = 0 PRICE = 5 FC = 200 VC/U = 3 Q = 100 PROFIT =0 PRICE = 5 Note that without pricing power. In fact the volatility of stock prices is somewhat dependent on the relationship between the amount of fixed assets and the stability of demand. a small change in sales will cause a large change in operating income. sales must rise to meet the increase. Those companies faced with both higher fixed costs and unstable demand. and an unstable . For example. While this scenario may sound like a terrific profit making venture. which has an almost negative correlation with very high operating leverage. a semiconductor company who may have the same level of fixed costs faces a much more daunting task because the demand for its product is much more elastic: both businesses and consumers tend to do less capital spending to upgrade equipment (or life style) during recessions. When higher fixed costs are incurred. will tend to have higher betas. such profits are only contingent upon the stability of sales. On the other hand. albeit in a more direct manner. which measures the implied amount of capital a firm needs. it must be reiterated that the investor must decipher the amount of fixed assets in a company as such information is not directly proffered. It is merely. But when prices fall due to competition. . these companies tend to do enormously well when their respective sectors are favored and then lag behind the market the rest of the time. To understand the gravity of the situation. it can offer a good comparative snapshot. In what was termed the largest computer company acquisition in history. and like the first. They will also destabilize operating profit because higher capital expenditures are needed to maintain a high level of fixed assets which will also require a higher level of sales. Total Assets / Total Revenues and discerning students will notice that it is the inverse of the “asset turnover” ratio. The company could not compensate for market price decreases with additional sales volume. COMPAQ COMPUTER: THE REST OF THE STORY By 1997. The first is called the “capital intensity ratio”. Tandem Computer. Compaq saw them as a stepping stone to regaining operating margins. it tends to reveal the amount of fixed assets and consequent fixed costs in a company. it acquired UNIX operating Digital Equipment Corporation to gain a foothold in the enterprise systems market. in newer industries.229 stock price as well.Current assets) / Total Assets. (Total Assets . Since it is rare to have such ideal conditions occur simultaneously. they act as a barrier to entry and pump up prices because the supply of technology is low. two measurements need to be introduced. The second measurement is the non current asset ratio which is merely. rising variable costs deplete any profit. Compaq added it to its previous purchase. Enterprise systems were “big ticket” items in comparison to the commoditized PC industry. When it is compared to both year to year change and an industry average. Desperately seeking to increase sales and profits. Compaq began to see margins decline. measures the amount of fixed assets. Thus. fixed costs are a double edged sword. Although this ratio can be obfuscated by large amounts of intangible assets like “goodwill”. Compaq had a monumental task to integrate them efficiently. Had the acquisition been consummated four years earlier in 1994. By 2001.58125 0.huge enterprise systems that could run a major city.systems that were both complex and massive.5951 0. The Windows based PCs were rapidly being replaced by the proverbial 800 pound gorilla . Tech stocks were completely slammed.1787 Year of Digital Equip. Not all computer systems are created equal.230 Table 9-5 COMPAQ YEAR CAPITAL INTENSITY NON CURRENT ASSET RATIO 1996 0.3412 The Digital Equipment purchase was paid for with over nine billion dollars in cash and stock for what was.12908 1997 0. Since the breakeven point for sales was rising at the same time that . increased to fifty and even seventy percent for many stocks in the tech-heavy NASDAQ. Purchase 1998 0. many stocks that had sold for eighty or ninety dollars a share could be purchased for as little as ten dollars. The “higher ups” knew they were taking a major risk. In essence. What was anticipated to be a “normal” ten to fifteen percent decline. such a “shot in the arm” would reformulate the company away from high quantity. Compaq’s valiant attempt to revive its profit margins was doomed to failure. the largest acquisition in computer history. at the time. With the purchase of large amounts of fixed assets.7395 0. Compaq may have had a chance of riding out what was about to occur: one of the worst market corrections in modern history. low leverage PCs and toward a niche product . the higher fixed costs of the Digital Equipment merger entailed more potential profit but more operating risk. but rightfully believed that a path to higher margins included more operating risk with higher fixed costs. fixed costs determine the magnitude of the ratio. and fixed costs remain truly “fixed” (they do not change) then the ratio of S . i.FC or S . The lower leverage from PC manufacturing could not help weather the storm because PCs were a discretionary item.FC).231 the economy was in decline. one must realize that the ratio has multiple uses.VC .e..VC /EBIT at the end of this year. because the other variables occupy both the numerator and the denominator with the same relationship. The paradox of seeing margins decrease because of technological change. The total ratio is thus: S . and then divided by EBIT (earnings before interest and taxes) in the denominator. OPERATING LEVERAGE AND PREDICTION To understand operating leverage and the concept of a small change in sales having a magnified effect on operating income. Since the ratio is a compendium of perhaps thousands of . The great “mystery” is how these concrete numbers actually transform into percentage change variables when we compare one year to another. the leverage was not low enough to make consumers buy these products when “money was tight”. often called “the contribution”.VC / S . will mirror % ∆ Operating Income / % ∆ Sales at the end of the next year. The basic ratio is formed by sales minus variable costs in the numerator. It not only displays the risk of higher fixed costs. % ∆ Operating Income / % ∆ Sales? The explanation is: if variable costs remain the same percentage of sales. and thus have a predictive relationship with those variables. Simply stated. What happened to Compaq was “the perfect storm” in operating risk.VC . It should also be observed that the components of EBIT are sales minus variable costs minus fixed costs (S . and then an attempt at higher leverage fail because of changes in the business cycle.VC /EBIT. but can predict the relationship between sales and operating income. How can S . should not be lost. the higher operating leverage jeopardized Compaq’s position: Compaq needed to sell high profit enterprise systems just as the economy was gong into a tailspin. a merger with Hewlett Packard became Compaq’s saving grace.VC / EBIT turn into the beta ratio. Ultimately. but more like “cost of goods sold” or “administrative and selling costs”. Therefore. we compare the stability of actual percentages to this ideal and try to coordinate the leverage trend with the types of products a firm is selling and its place in the business cycle. just as financial leverage does. which they can presumably do if operating income is stable. operating income. Creditors favor those companies with stable revenues because the risk of default is lower. what we observe as investors will rarely conform to the academic ideal. By observing the mean and variance of operating margins. just by discontinuing one business segment and adding another.232 internal breakeven points. over a five year period. It is not impossible. The discerning reader will notice that the concrete form of the ratio can never drop below “1” and yet many companies have processes where operating income percent changes are always less than sales percent changes. a company can radically change its percentages. more shares issued actually diminish the risk of unstable cash flow. if a company’s operating leverage is high it will have a tendency to lower financial leverage by financing with equity. will charge the “prime rate” to the customers who are most capable of paying off and renewing loans - . When we find such information. Conversely. companies with low operating leverage have a chance of raising EPS simply by financing with more debt. sales. banks for example. CHARACTERISTICS OF OPERATING LEVERAGE While higher operating leverage ultimately increases the variability of EPS. the investor needs to do a clever “end around” and infer operating risk from statistical relationships. Since costs are never published in the format of “fixed” and “variable”. it is debatable whether operating leverage for an entire company can be precisely measured. Consequently. and capital intensity. Some research reports may contain specific operating leverage information for an industry or a company. most companies try to minimize this variability by balancing the two types of leverage. we can better understand the risks of any firm’s cash flow. In fact. Thus. but a handful of “astute” investors saw the Compaq debacle coming and acted with foresight. These variables are difficult to predict. and the near inelasticity of bread holds. Consequently. Thus operating leverage is the major source of most other types of risk including: financial leverage. i. the combination seemed contrived and awkward. people would buy less bread. as they would in a famine. On the other hand.233 firms with large and stable operating characteristics. if prices quadrupled. during the 1950s and 60s. inflation risk and even random event risk. Diversification can smooth out this risk. as processes overlapped. plasma televisions and high-tech exercise equipment all have high elasticity and will see large shifts in demand when prices change. The cycle continues because these customers can use more debt in their respective capital structures when interest rates are lower. Analogously. Agricultural products. Elasticity is not calculated for non normal periods. political risk. it would not bode well for a maker of expensive golf clubs to start selling batteries. usually have low operating leverage and must be produced in quantity to make a profit. In the above allusion to batteries and golf clubs. The elasticity of demand determines the percent change in demand based on a percent change in price. most people would buy the same amount of bread if prices were to increase by ten percent. However. Stocks that pertain to these products will do well in the brief time their respective sectors are favored and then lag behind other sectors in the interim. The type of product often determines the scope of risk.e. foreign risk. a higher fixed cost per unit will be implicit in their development. On the one hand. Another characteristic of operating leverage is the ability to combine different operating leverages and change a risk profile. Vacation resorts. demand is unresponsive to price and these products are almost inelastic. Since price is determined by a competitive market of many buyers and sellers. such combos were frequently tried and “conglomerates” were formed for the expressed purpose of diversifying away risk. Corporations . for example. products with high elasticity imply a higher operating leverage although the relationship is certainly not cause and effect. but each firm must maintain a core competency. inefficiencies occurred that diminished sales and profit for each product. Unfortunately.. we will walk into Wal-Mart and see a kiosk set up to serve McDonald’s hamburgers. is because the payoff is uncertain. the payoffs need to be more rapid than in a low operating leverage scenario. who can assume the risk of a huge fixed cost operation. and too many other profitable investment opportunities abound for that to occur. Instead of playing tennis with the sales manager. Private investors expect a return. Consider that fixed assets often involve processes with high technology that can rapidly become “new and improved” and replace existing technology. Such “reengineering”. Moreover. Finally. When a venture capitalist takes the risk of investing in a high-tech operation. The reason that a major project like the interstate highway system was completed is simply because it was undertaken by the government. and instead of having a golf club company owning a battery company. although firms still attempt to diversify sales within a narrower framework. Without a rapid payoff. This behavior is readily observable in which stocks get . the investor risks being “eaten alive” by the competition who may already have the new technology in place. a sales rep may play video games with a programmer in India. massive projects require the organization of many investors. and with higher fixed costs. One reason that America has failed to invest in “alternative energy sources” (as of 2008). No investor wants to tie up capital for five years in the hopes of doubling his or her money “somewhere down the road”. A “good” product from an average company was discarded in favor of a “great” product from a good company. While tax breaks and incentives can stimulate some action.234 were unsure of what business they were in. often required numerous dislocations and downsizings but increased the overall flexibility of the company. “Outsourcing” and “partnerships” became the buzz words. a “niche”. capital budgeting time frames are an important characteristic of operating leverage. The diversity in cash flows have been offset by diversifying the processes. and many firms began to refocus during the 1980s and 90s. that return needs to be greater with at least some modicum of certainty. investment in capital intensive fixed assets involves the risk of obsolescence. Similarly. THE QUALITY OF AN OPERATING MARGIN For long-term investment. it will be a short time before it is duplicated with a different. it is tempting to formulate trend rules such as never investing in a company that has a three year decline in margins or looking for a reversal after three years. we know that such trends stop when more competition enters the market. We can observe increasing risk. Traders want to get in and then leave the stock before the technology becomes obsolete or duplicated. OPERATING TRENDS AND REVERSALS The following sections are going to discuss evaluating the size and stability of operating components. If a pharmaceutical firm. “designer” formula by some other company. one of the hardest decisions to make is to find a company with outstanding size and stability of operating components. the size and stability of an operating margin is one of the determining factors. one has to enter the market before a trend even starts. isolating trends in operating margin will be self defeating because the investor will be “chasing earnings”. In fact. We apply both the mean variance method and the coefficient of variation to five years of operating margin data. While the mantra of “buy and hold” serves a well thought out portfolio. In a need for certainty. and then view a period of declining margins as an arbiter of investment. trend spotting requires coordination between company. Sometimes. products. When two investment options differ in . industry and the overall economy. many traders want to “get in and get out”. owns the patent on a miracle cure. if we see increasing margins over a long period. they can not predict the future. and no analyst can predict when such momentum ends by observing operating characteristics by themselves. for example. Such rules can blind an investor to big opportunities. The trend in operating components must be seen in the context of an overall movement toward a target capital structure.235 “day traded” and in what periods. but we can never say that a company is about to “implode”. The reason is simple. While statistics give us insight into the quality of sales and operating income at any moment in time. 5 593 11.09 5. 1997 107 599 17. The following tables present the operating margins of two companies along with a mean variance and a coefficient of variation for each: Table 9-6 INTUIT YEAR Operating Inc.86 % 1998 70.28 % 2001 265 1261 21. Indeed. we usually choose the one with the lowest coefficient of variation. and its response to changes in the business cycle. Sales Operating Mar. Standard Dev.02 % Table 9-7 Statistic Mean Standard Deviation Mean Variance Coefficient of Variation Methodology Average of Operating Margin Square root of variance Mean minus Standard Dev.26 11.236 this measurement. This ratio defines the company .89 % 1999 254 848 29.95 % 2000 200 1094 18.6819 .its market. The reason that this is such a telltale ratio is that most firms will go out of their collective ways to ensure the preservation of this figure.17 0. its funding. divided by Mean Result 16. we saw how Compaq even made a nine billion dollar acquisition to do so. Our five year time frames contain a limited number of data points. A RISKY PROPOSITION: CONFIDENCE INTERVALS Any decision making based on the extrapolations of data is very risky.J. For a much more minute examination. there is no better analysis than doing a coefficient of variation on operating margin data. more data points translates into more accuracy. and future trends rarely mirror past performance.43 3. Sales Operating Mar. For a quick and decisive snap shot of operating history.129 12.71 % 1998 1834 9209 19. divided by Mean Result 15. As investors.74 % 2001 1282 9430 13.59 % Table 9-9 Statistic Mean Standard Deviation Mean Variance Coefficient of Variation Methodology Average of Operating Margin Square root of Variance Mean minus Standard Dev. our information is much more limited than that of professional analysts. So why make trend predictions? Why take the chance and have to “bite our lips” when we are wrong? Statistically deriving a prediction interval is not the same as making a prediction. Heinz YEAR Operating Inc.92 % 1999 1412 9300 15.2028 Heinz’s’ margins are as steady and slow as its catsup. there is a . that from our very limited data.18 % 2000 1575 9408 16. an acquisition specialist could use 60 months of data. 1997 1096 9357 11. Standard Dev.237 Table 9-8 H.14 0. which is why the company has become an institution. We are merely implying. and we need to trust a small sample. but we should not make it the primary criteria for decision making. This figure is then added to (or subtracted from) the mean . would most likely revert to the mean. To form confidence intervals that express volatility. we need six years of concrete figures. In our quest for a “ball park” figure. we try to use the 99th percentile as a rule because it will be so extreme. We then form confidence intervals by multiplying the sample standard deviation by a T score (see Statistics Primer) and then dividing by the square root of the sample size (5). we consider small sample predictions to be the least reliable. we know there is a specific percentage chance that it will move back toward the mean. which is % ∆ Operating Income / % ∆ Sales. we take the five year mean and sample standard deviations of operating momentum. Since data is limited. in that case we may use an unreliable indicator only to confirm a reliable one Regression. as we are determining five years of percentage changes. is properly applied to data that has a “normal distribution”. We add this information to many other indicators before making a decision. A figure that would violate that extreme a constraint. and so we will have to weight each measurement on the basis of reliability. and collectively. we are looking for as many diverse indicators as possible. We may use it to confirm inferences about distributions that are anything but “normal”. for example.238 certain percentage chance of an event occurring. Additionally. If our current figure lies outside the interval. thus forming the interval.but even more importantly. The following example is very typical of a small sample size: . some data is so skewed that no measurement is meaningful. To achieve this amount of data. it suggests an upheaval in the production process and gives a signal for further investigation. Some of these will be contradictory. Such reliability is established by the number of data points and the proper methodology. it covers a percentage area of a curve by multiplying a specific number by the standard deviation.64 2002 17 23 0. We divide the sample standard deviation by the square .776 2. Like the sample standard deviation.53904 respectively. and the following T scores correspond to the given percentage: Table 9-11 PERCENTAGE N=5. Each T score also corresponds to the size of the sample. it automatically adjusts for the expected volatility found in smaller samples. with a smaller sample size indicating a higher T score.79 The student’s T score is very much like the Z score adapted to a sample size under 30. FREEDOM 99% 95% 90% 80% 50% T SCORE 4. we need to determine the mean and sample standard deviation which is 1.294 and 0.239 Table 9-10 VARIABLE YEAR Operating Income % Sales % Operating Momentum 2000 20 15 1.132 1.533 0. Moreover. 4 DEG.741 For the sample data. Therefore. like the Z score. the mean plus or minus a number of standard deviations covers a percentage area of the curve.33 2001 29 18 1.604 2.739 2003 14 7 2 2004 19 24 0. Our own sample size is five (for the number of years). We then "exponentiate" this figure with the inverse of the number of periods between years (4). this interval would be: 1. and then create an operating momentum out of the quotient.294 ± (4. that if the next operating momentum indicator is not within these numbers.49 % growth rate. The inverse of 4 is 1/4 or 0. so there is a heavy reversion to the mean. it will revert back to the mean. We can not say with certainty. We multiply this number by the percentage confidence level and corresponding T score. The methodology is as follows: Table 9-12 GlaxoSmithKline Year 1999 Operating 2702 Income Sales 8490 2000 5190 18079 2001 5508 20489 2002 8498 21312 2003 7365 21441 To determine an estimated growth rate.604)(0. At 99 % confidence.294 ± 1.25 or 1.184 and 2. we may want to do some extensive examination if we are serious about investing in this company.25 or 1.53904 / √5. or 0.241) = 1. sales and income. The relationship between the operating components. 5. 1. the growth rate would be: (21441 / 8490)^0. the figure is (7365 / 2702)^0. or subtract it from the mean. we determine the percentage growth rate of 26. we form a ratio of the near term figure in the numerator and the far term figure in the denominator.06 % per year. determines the character of the company.404.2606 is the growth rate and if we subtract “1”. but since it will violate the constraints (0.404). and then add it to.240 root of the sample size. Dividing the two. but a characteristic mean that can be used to compare industries or individual firms.2849 for a 28. divide them. Large permanent increases in these components are rare.2606.11 or 0. This mean is not a true mean.25. Thus for sales. we determine the characteristic operating momentum: . For operating income. Another method we can apply is to determine the individual growth rates for sales and operating income.184 ≤ X ≤ 2. the probability that the operating momentum will be around “1” is far greater than the probability of sales increasing by exactly 26.241 (28.0932. each with its own response to economic and financial risk. has a different “beta” than another division. even a “ball park” risk measurement for operations opens up innumerable other options because sales and profit can be better related to the cost of equity. the model is flexible enough to comprise a wide variety of financial assets and will relate them to interactions between the risk free rate (government). one division of an oil company. A corporation can be viewed as a portfolio of assets. The collective sum of betas from each division. Nevertheless.49 / 26. Can we filter out the financial risk and find a beta that pertains to operations alone? This theoretical “unlevering” of the company was pursued by both the team of Miller/Modigliani and Hamada well before 1975. will make up the overall corporate beta. Such a measurement can be used for mergers and acquisitions. Thus. They found that they could extract the financial risk from beta if they factored in the debt to equity ratio as well as the tax rate. THE UNLEVERED BETA EQUATION The consensus equation is simple and is just: . and even for the evaluation of firms that are not in the market and which need an estimated beta for comparison. performance evaluation. others would declare that this volatility reflects the reality of constant financial change. The mathematical ease of doing so depended on the linearity of the function.06) = 1. OPERATING BETA The capital asset pricing model (CAPM) is very adaptable. the market (many buyers and sellers). and the individual asset (beta).06 percent in the following year. and the CAPM is conveniently a straight line. Although some would argue that the model is inherently unstable and reveals only a fleeting glimpse of financial truth. Notice that we did not proclaim the predicted growth rates of either component. exploration for example. but used the results to form a more meaningful number. While not precisely accurate. weighted by asset value. 35)(0. all we need to do is to weight each beta by its associated market value to derive a combined total. any deviation from the ideal would be considered unsound.696 TENNIS RAQUET COMPANY: 1.204 The Ardco Barbell Company takes on two acquisitions: • 1) A chain of gyms with a beta of 1.4 and the company uses little debt. we have made up a scenario where book and market values are the same for the ease of computation.6 and a debt to equity of 40/20 or 2. Both tax rates are 35 %.4 / (1 + [(1-.35)(0.5. about 10/40. There is limited demand for old fashioned weights although the company keeps its leverage high by offering niche products.35 OR 35 % 1. With a tax rate of 35 %.25 debt to equity. MV is “market value”.6/(1 + [(1-. we can determine the operating beta of each unit. The logic behind betas is that if they are added linearly. a market value of 10 Million. The market value is the same as book equity in this case – 40 million. what would be its operating beta? Table 9-13 VARIABLE BETA DEBT / EQUITY TAX RATE UNLEVERED BETA 1. The current beta is 1.1 / (1 + [(1-. In professional risk management.25)]) = 1. Step 1: Find the operating betas of the new acquisitions: GYM CHAIN: 1. while T = tax rate.25 .1. The market value of this enterprise is 20 million.8301 . or 0.35)(2)]) = 0. and a debt to equity of 5/10 or 0.5)]) = 0. Using book values in place of market values is theoretically improper.4 0.242 Unlevered Beta = Current Beta / (1 + [(1-T)(MV Debt / MV Equity)]). THE ARDCO BARBELL COMPANY: AN EXAMPLE OF UNLEVERED BETA The Ardco Barbell Company needs to expand. In our example. • 2) A tennis racquet company with a beta of 1. 243 Step 2: Multiply each operating beta by its market value weight. The combined value is 40 + 20 + 10 =70 (40/70) (1.204) + (20/70) (0.696) +(10/70)(0.8301) = 1.0037. This is the operating beta of the new company. Step 3: We “lever” the beta back up to reflect the new debt position. Notice that the combined company has 10 + 40 +5 = 55 in debt. For illustrative purposes, we assume these companies were purchased with cash, but Ardco could have incurred more debt to buy these companies and we could have adjusted that proposition into our analysis. The equation is: (Unlevered Beta)(1 + [(1-tax rate)(MV Debt / MV Equity)]) = New Combined Leveraged Beta (1.0037)(1 +[(1-.35)(55/70) = 1.5163 1.5163 is the newly combined beta. It is higher because Ardco had to assume the gym chain’s debt. The benefit of diversification can be observed in Ardco’s much lower operating beta (1.0037 vs. 1.204), which they hope will contribute to paying off the larger debt obligation. The “top down “ approach to beta would multiply each separate leveraged beta by weighted market value to determine a final beta. In this case, that beta would be (40/70)(1.4) + (20/70)(1.6) + (10/70)(1.1) = 1.412. Which approach is more valid? While the “top down” approach is more conventional, it assumes that separate entities are correlated by market value and beta response. We know from previous chapters that the coefficient of determination, R , is usually a value between 0.2 and 0.4, and that the “alpha” component sometimes supersedes the beta component in importance; we often assume that the correlation factor is stronger than it really is. Since we use the required rate of return to determine a cost of equity and not for predictions, the effect of a low correlation gets buffered in the analysis. However, the “top down” approach assumes that a “single index” market correlation is valid enough to capture the diversification effects of a combined 2 244 company. On the other hand, the levered beta technique assumes that beta is made up of a mathematical sum of operating risk and financial risk without additional outliers. This is an oversimplification, but is valid for the purpose of analysis. In fact, beta is made up of some random fluctuations in the economy, statistical “noise”, and other undetermined factors. Since beta is unstable, a consensus average between “top down” and levered approaches is sufficient. “MOM AND POP” STORE BETAS: COMPANIES WHO ARE NOT ON THE MARKET Most transactions are not enacted within a corporate environment. An example would be a local building contractor who wants to add a roofing company to his portfolio. These business deals are not glamorous but are very typical, and need to be evaluated for risk. To examine them, analysts use one of two approaches: 1) An accounting beta can be determined. Sales or income data from the business is regressed against the S & P 500 (or appropriate index) as percentage gains and losses. After the accounting beta is determined, it is unlevered for debt, and an operating beta is derived; book values must be used instead of market values and we assume they are the same. 2) The analyst uses the “pure play” technique. The investor takes the betas of the three largest market-traded businesses in the same industry, levers them down using the average debt to equity, and derives an unlevered beta. An example: The Ardco Barbell Company now has a beta of 1.52, a market value of 70 ,and a debt to equity of 55/70. It wants to buy a chain of tanning salons with no actively traded market. The salons have a debt to equity of 15/20 and a price of 40 Million. The three largest actively traded competitors are: Table 9-14 COMPETITOR BETA DEBT TO EQUITY MARKET VALUE SMITH 1.8 2 60 BRONZE 1.9 3 60 FRYE 2.1 3.5 40 245 The combined “top down” beta is (60/160)(1.8) + (60/160)(1.9) + (40/160)(2.1) = 1.913 The unlevered beta for this hypothetical combination is 1.913/ (1 + [(1-.35)(2.8333)]) = 0.673. Note that the 2.8333 figure was the average debt to equity for the three firms. Also note that financial leverage can have a significant effect on beta. Without debt, the three firms have an operating beta of only 0.673 which seemed so “safe” that they piled on the financial leverage to improve EPS. Thus, the operating beta of the new firm is only 0.673. The firm is very attractive to Ardco because they have so little debt in an industry that seems to thrive on it (hypothetically). To form a completely levered beta, we multiply the unlevered beta by the firm’s debt to equity multiple: That is (Unlevered Beta)(1 + (1-.tax rate)(D/E)) The completed beta is 0.673 x (1 +(.65)(.75)) = 1.00111. The acquisition is probably a “young” firm in the industry, and has not established itself enough to incur the massive amount of debt of its peers. This is an opportunity for Ardco, and they immediately snatch it up. OPERATIONS RESEARCH FOR THE INVESTOR Most investors do not have access to detailed corporate data that itemizes the types of assets (prepaid insurance for example) or the types of projects that need capital budgeting attention. There is a bond of mutual trust between top management and shareholders when acquisitions are made; synergy has been evaluated and the risks of a purchase have been evoked. However, good management is fearless. They welcome questions from the least of employees and the smallest of investors because there is an objective principle that substantiates their position. Part of risk management is to find balance and common ground between two sets of data. When acquisitions are made, there may be a history of operating income and sales for both companies, and the investor can use this quarterly data to measure the covariance 246 between percent changes. If the acquiring company is high risk (operating and financial leverages are high), a low covariance will indicate a strategy of risk reduction. Analogously, a low risk company may attempt to step up its profit margin by targeting a company with a high covariance. Other strategies include: sector timing. A high risk company may want to target another high risk company if that sector is anticipated to grow as soon as assets are efficiently integrated. Since few combinations can be this strategically facile, the investor needs to “look twice”, when this scenario occurs. Similarly, when a low risk company buys a company with low covariance, it may not need the implied diversification as much as it needs to seek out high profit, “riskier”, projects. The natural inclination is to trust operations to those who know it best; they often have large stakes in the success of a merger. Nevertheless, asking questions is the prerogative of the shareholder, and doing so will help management as well. The following example continues with a table of cash flow data, comparing Ardco Barbell with the proposed acquisition of yet another tennis racquet company, and also a vitamin company. The means, standard deviations and covariances are at the bottom of the chart. Table 9-15 QUARTER (Period) 1 2 3 4 5 6 7 8 9 10 Mean Standard Deviation R (with Ardco) COV ARDCO BARBELL 22 17 19 31 14 15 22 26 15 11 19.2 6.106 TENNIS RAQUET 14 10 9 11 21 24 20 30 31 32 20.2 8.94 -0.5882 -32.108 VITAMINS 20 16 18 25 12 13 14 19 17 4 15.8 5.61 0.77294 26.477 247 *ALL DATA ARE PERCENT CHANGES In this shortened example (at least 20 quarters should be used), it is obvious that the vitamin company is more risky from a diversification standpoint. On the other hand, the tennis racquet company sells tennis racquets in the spring and summer, while Ardco sells fitness equipment in the winter months as people “get in shape” for summer. Also, the chain of distributors would be similar for Ardco and the tennis racquet company, but there would be a need for negotiations with a health food chain if the vitamin company were purchased. Additionally, any fixed assets from the vitamin company would be unique to that manufacturing process and would need to be separated from Ardco. Similarly, trucks and warehouses that carry fitness equipment have different priorities that may not include the more delicate storage of vitamins. Therefore, Ardco makes the decision to purchase the tennis racquet company - and- tries to put its label on generic vitamins with another company in charge of distribution. TWO MASTERS: FISHER AND BUFFETT Two of the greatest analysts of operating risk displayed an almost intuitive quantitative sense. They were well known for strategic investing, but shared a mathematician’s perception for minute detail. They are Philip Fisher and Warren Buffett. Philip Fisher was a financial analyst during the depths of the depression - almost a prerequisite for tenacity - and ascended from that morass with dedication and vision. His book, Common Stocks and Uncommon Profits, became a “must read” in the financial community. His best known tenet was to evaluate a company’s long-term prospects by studying growth and sales potential, and then favoring those corporations who were dedicated to producing at the lowest cost. This qualitative approach had as its foundation, a thorough knowledge of management skills and manufacturing processes which Fisher would exhaustively research. Fisher rejected what he termed, “marginal companies” those that showed high profits during economic upswings, only to lag behind in the rest of the business cycle. 248 Fisher also excelled in detecting operating synergy - areas of growth in which products or processes complemented each other. For example, while the average investor might notice an aging population and invest in nursing homes, Fisher would have invested in the products and services needed to operate those homes and that were used in collaboration - perhaps disposable needles and “sharps” containers. Two of the best examples of this acumen were investments in Du Pont and Alcoa. Du Pont was a gunpowder manufacturer who standardized processes and went on to produce unique synthetic materials like cellophane and Lucite. In addition, their management team was responsible for developing the preliminary analysis used to evaluate capital structure, the aforementioned “Du Pont equation”. Alcoa was an aluminum company who capitalized on the airplane manufacturing that was at the forefront of technology in the 1930s and 40s. Both companies remained flexible and adapted their operations to new product demand and technologies. Warren Buffett is in a class by himself. Few investors have matched his dedication to gauging risk and positioning for long-term gains. Curiously, he does not fret at all about economic fluctuations or Federal Reserve actions, and appears to have unbridled faith in the notion of a “good idea”. However, he does possess an economist’s insight about most macro topics, including inflation. His natural inclination is to avoid companies with large amounts of fixed assets, because initial rises in asset turnover are depleted when capital expenditures finally need to be made. Inflation acts like a smoke screen blinding those firms with a high percentage of fixed assets, because sales will initially outpace the need for investment. A second major principle of Buffett’s is to invest in firms with consistent operating history. He reasons that those companies who are in the process of a major change in operations are not good investment prospects because of higher risks associated with cost and revenue inconsistencies. 249 Thirdly, and most importantly, Buffett seeks out companies with what he terms, “economic good will”. These are companies that produce above average long-term profits because they possess “franchise value” - the ability to raise prices when needed. They tend to produce unregulated products for which there is no close substitute, and often induce a “brand loyalty” among their customers. Appropriately, Buffett shuns what he calls “commodity type” businesses - low profit businesses that churn out undifferentiated products. He reasons that these businesses can only compete on the basis of price and can only prosper during the rare event of a short supply. While most analysts treat the future as an extension of the past, both Buffett and Fisher created a different type of mental calculus. They appeared to compare the rates of change among several variables which coalesced into an investment strategy that produced long-term gains among many commonly known brands - Coke, The Washington Post, ABC, Etc. Buffets desire for both “pricing power” and operating consistency, seems to seek a company with a high but consistent operating leverage that will not deflate at the prospect of a downturn. High technology firms would be eliminated by that constraint, but “niche” companies who have created a brand consciousness among consumers would not. A high “mean” and a low “standard deviation” among both operating momentum and operating margins would meet that objective. However, a small capital intensity ratio (one without a lot of fixed assets) would also fit Buffett’s objective, except that these often designate the very “commodity type” industries that he avoids. Naturally, a productive balance is implemented when these several constraints are satisfied, but such a strategic combination is difficult to find. A BRIEF OPERATING ANALYSIS OF FED-EX AND STAPLES FOR THE YEAR 2000 Most analysts examine companies in the same industries, because large disparities may be more indicative of the industry rather than the company; there is sample bias when different companies in different industries are compared. The need for precision requires focus, but when one company is subject to forces (demand, legislation, trade barriers) that 250 the other is not, the analyst can not form valid, actionable conclusions. We close this chapter by summarizing our indicators through analysis of two separate companies in two separate industries. While such comparison is statistically unconventional, it is realistic from an investor’s perspective. We offer a five year operating history of both Fed-Ex, the next day transit specialist, and Staples, the office supply fixture. Sales and operating growth is determined, and then changes in fixed assets and capital intensity are derived. Finally, we do mean-variance statistics on the operating margins of each company. The theme that runs through the analysis is the choice between growth and stability. Are the risks of slowing growth greater than the effect of stagnation, for example? Does a period of long growth preclude investment, i.e., not getting in soon enough? When the statistics become contradictory, we add some “qualitative” analysis and try to perceive the situation as Fisher or Buffett would. STAPLES AND FED-EX OPERATING HISTORIES 1994 - 1999 Table 9-16 FED-EX Year Sales Op. Income Current Assets Total Assets Table 9-17 Fed-Ex Beta 1994 9392 1244 1995 10274 1344 1728 6699 1996 11520 1477 2133 7625 1997 15873 2047 2880 9686 1998 16773 2198 3141 10648 1999 18257 2376 3285 11527 BETA (2000) EQUITY (2000) DEBT (2000) 0.98 8191 1899 251 Table 9-18 STAPLES Year 1994 Sales 2000 Op. 110 Income Current Assets Total Assets Table 9-19 Staples Beta 1995 3098 191 926 1403 1996 3968 260 1151 1788 1997 5181 355 1666 2455 1998 7123 514 2064 3179 1999 8937 708 2192 3814 BETA (2000) EQIUITY (2000) DEBT (2000) 1.03 1837 2152 ANALYSIS AND STATISTICS All betas are not created equal. Without unlevering the betas, we can readily observe that Staples’ operating beta is lower because they have much more debt in their capital structure, and yet both companies’ betas are nearly the same. Since Fed-Ex’s operations are oriented around transportation, which requires a heavy investment in fixed assets, Fed-Ex has more economic risk. Additionally, they are exposed to the risk of higher prices in the oil market which are totally out of the firm’s control. These are inherent risks that come with the nature of the business and can only be minimally diminished by diversification. However, as we will observe, Fed-Ex compensates for this detriment with a high level of efficiency and stability. 252 Table 9-20 FED-EX Percent Change YEAR Operating Income % Sales % 1995 8.04 9.39 1996 9.9 12.13 1997 38.59 37.79 1998 7.38 5.67 1999 8.1 8.85 Table 9-21 Staples Percent Change YEAR Operating Income % Sales % 1995 73.64 53.4 1996 36.13 29.34 1997 36.54 30.57 1998 44.79 37.48 1999 37.74 25.47 The following table tabulates the extraordinary growth of Staples’ sales and operating income and compares it with Fed-Ex’s rather ordinary growth. However, what the investor should observe is sustainability. No company grows at twenty percent forever. Although the growth occurred in the context of a long bull market, five to six years can be the length of an entire business cycle. Staples’ growth was derived from a structural change in the way America did business: the Internet spawned numerous home businesses that needed small amounts of office supplies. Such a transition would not take nearly as long as IBM’s, for example, which computerized entire industries. Competition would crop up easily in office supplies, but not in industries which required huge investments in fixed assets. Thus, while Staples’ growth rate looked like a tempting investment, it would have been far riskier to put capital into it in 2000 than in 1996 or 1997. 253 Table 9-22 FED-EX SALES OP. INCOME MEANVARIANCE 1.697 0.898 STAPLES SALES OP. INCOME MEANVARIANCE 24.22 29.797 *We measure mean-variance as (mean - standard deviation) which is a typical adaptation. Table 9-23 Average Operating Momentum 0.99 1.265 Last year (1999) operating Momentum 0.915 1.4817 FED-EX STAPLES Staples’ growth is both high and steady, and it would be easy for an investor to be deceived by these operating statistics. In capital structuralism, our desire is always to anticipate growth and never to “chase” it. Capitalizing on an overly long growth cycle is reserved for the “lucky” few; a handful of investors can make money “without knowing any better”, because they happen to be in the right place at the right time. While there is money in high risk momentum trading, it is not recommended for even the most skilled investors, because it is such a gamble: the odds are better playing blackjack in Las Vegas. However, there may come a time when the investor observes three to four years of growth, and may have information about movement toward an optimal capital structure. At this point, both equity and the cost of equity may be low enough to justify investment. The investor appears to be capitalizing on momentum, but is not investing on the criteria of prior growth: he or she has correctly anticipated a profitable change in capital structure which should be actualized by a jump in economic profit. THE CONFIDENCE INTERVAL TOOL 254 When we sight a near term number that is uncharacteristic of others in the sequence, it is cause for concern. New developments (changes in costs or demand for example), may affect our investment. Therefore, we recommend using the 99 % confidence interval to gauge any oddities we may encounter. With small sample sizes, it is suspect as a measurement, but we merely use it as a check to see whether further investigation needs to be done. The following capital intensity and non current asset ratios apply to the companies: Table 9-24 FED-EX YEAR Capital Intensity Non Current Asset Ratio 1995 0.652 0.7421 1996 0.662 0.7201 1997 0.61 0.7027 1998 0.6348 0.705 1999 0.6314 0.715 Table 9-25 STAPLES YEAR Capital Intensity Non Current Asset Ratio 1995 0.457 0.3399 1996 0.4506 0.3562 1997 0.4796 0.3214 1998 0.4463 0.3492 1999 0.4267 0.4252 All of these ratios look like they are in an orderly sequence, except for the 1999 value of non current assets for Staples. At 0.4252, it is larger than the others, and would be the most likely to affect our potential year 2000 investment. We decide to submit it to the student’s T test and derive confidence limits at 99 %. Table 9-26 one ratio can act to confirm the other. decisions made on the basis of operating characteristics should be long-term. Since 0. However. However. In fact. Also. the simple expenditure of a large amount of cash for an acquisition can deplete current assets and inflate the non-current asset ratio. distribution changes. the probability of default is lowered when interest coverage ratios rise.23606 4.2765 on the low side and 0. the large change in non-current assets is not confirmed by a jump in capital intensity.4394 on the high side. If both measurements rose.(4. we will not investigate the measurement any further. OPERATING MARGIN The size and stability of operating margins is paramount.604 0.4252 is between that interval. As growth is often stimulated by a sector’s position in the business cycle. The following tables exhibit the operating margin histories for both companies: .03956) / 2.604)(0.358396 0. in the absence of other information about financial leverage and capital structure. So far our analysis has encompassed the amount and stability of growth.358 +/. etc. it tends to waver and attract investors in the shortterm.255 STAPLES NON CURRENT ASSET RATIO Mean Sample Standard Deviation Sample Size Degrees of Freedom Square Root of Sample Size Student's T Confidence Interval 0. the eccentricity is even more evident because this last number is part of the sample itself. Remember that financial leverage will respond to the size and stability of sales and income and not necessarily its growth. we would investigate their acquisitions.236 The limits of the interval are 0.03956 5 4 2. which is the reason for obtaining corroborating evidence. Standard Deviation FED-EX 12.8623 % STAPLES 6. In this respect.1197 12.1301 0. They are a low debt. They were making a stable profit in a business that had much more inherent economic risk.0722 1999 0.256 Table 9-27 OPERATING MARGINS (Decimal) YEAR 1995 FED-EX 0.6523 6. A mean-variance analysis reveals the difference: Table 9-28 OPERATING MARGIN % Mean Sample Standard Deviation Mean . while growing sales leads to a larger operating income. On the other hand.954 % 0.0655 1997 0.131 0. their inherent business risk might be too high for an investor of . Choosing to finance with retained earnings lowers the risk of interest rate changes and buffers the effect of having a high level of fixed assets. the cost structure of a business can be optimized but not changed. instead of credit. Fed-Ex had an operating margin that was nearly twice the size of Staples’. when all efficiencies were realized. that can be used for purchases.0685 1998 0. high equity company.1308 STAPLES 0.982 % 0. 2) More dividends flowing to investors and 3) Retained earnings. Thus. and yet their earnings are large and stable enough to finance with much more debt if they needed to.3017 % The higher operating margin gives Fed-Ex much more financial flexibility.0792 A large and stable operating margin can lead to three positive scenarios: 1) A larger cash flow.1282 0. Staples had growing operating margins that would potentially “top out”.129 0.0623 1996 0. or even long and short-term moving averages.if we add financial leverage and market information to the mix. but they do have “pricing power” as one of the few players in a limited field. (Back to Table of Contents) . and operating characteristics will not tell the entire story.257 Buffett’s caliber. a higher operating margin is an arbiter of investment. company hype. Enough data must be coordinated to form a cohesive picture. we try to coalesce as much information as possible. much of what we need to know will be in the margins. neither depending on earning forecasts. Such “franchise value” and “economic goodwill” would be right up Fisher’s and Buffett’s alley Rather than the proverbial “coin toss”. over night shipments to Greenland or Africa ensure that Fed-Ex will be known as a “brand”. what if Fed-Ex’s margins decline as Staples’ rise? For that very reason. And yet . However. Since the real operating leverage may be only 1. At certain times. consider the premise that we define operating momentum as the percentage change in operating income divided by the percentage change in sales. these outlying measurements create volatility and tend to make wild swings in companies that are undergoing major changes.5. and the long-term average of momentum may be only 1. We smooth this volatility by taking separate growth measurements for both operating income and sales. large. This inherent volatility causes it to rise and fall ten fold in some periods. or even composed of negative numbers. and -11/-11 are all the same number. it can be alternately. Thus. The need for data creates a link between years that allows this ratio to be more comprehensive and forward-looking than it would first appear. any change in cost. creating an uncharacteristic ratio of fifteen. While it mirrors operating leverage when costs are stable. price or quantity will affect operating momentum. In any given year. . 15/15. its very imprecision and relativity forces us to obtain as much information as possible. especially when sales and earnings are out of “sync”. Since it is a measure of velocity rather than magnitude. An increase of this ratio requires three years of data. instead of averaging individual ratios for each year. As an example. its own variance is so volatile that it defies rational use. a one percent increase in sales can lead to a fifteen percent increase in operating income.2.258 10 OPERATING MOMENTUM Operating momentum has never been a standard measurement of risk. small. While operating leverage usually changes in response to new technology that changes the proportional requirement of fixed costs. 50/50. we need to obtain long-term averages (at least five years) to use the measurement in a meaningful way. leading to the conclusion that an application of this ratio is an exercise in futility .and not utility! Most statisticians would argue that it is an irrelevantly imprecise measurement. In fact. while the other involves a decrease of the same magnitude.259 REASONS FOR STUDY • Operating momentum is relatively unexplored territory among academics. • Analysts often misunderstand operating momentum. Once we can relate operating momentum to other ratios and the earnings potential of a company in the near-term. Consider a small manufacturing company with the following operating income and sales distributions: Table 10-1 Variables Sales Operating Income Total Cost Year One 100 10 90 Year Two Probability = 1/2 50 5 45 Year Two Probability = 1/2 150 15 135 In the hypothetical absence of fixed costs. the behavior of the combined ratio has not been examined thoroughly. Although both the numerator and denominator contain “essential” measurements (sales and operating income respectively). Statistical methods will often neglect its importance because it does not fit the parameters of “normality”.although one scenario involves an increase of fifty percent in costs and sales. If fixed costs became ten percent of total . confusing it with “operating leverage”. operating momentum and operating leverage for both scenarios would be equal to “1”. and abandoned businesses will all extract cash from operations even if they are accounted for separately. it is placed in a more meaningful context. which it mirrors in theory. Neither measurement would fully reflect risk. operating momentum reflects the real cost over-runs and volatility of that process. restructuring charges. • Operating momentum has cash value for the investor. While operating leverage reflects the stability of a cost structure in a production process. Litigation. discontinued business. While those fixed costs are an explicit part of operating leverage. and an over-dependence on a small base of either vendors or customers. and foreign currency translation. Operating leverage. the closer operating momentum is to operating leverage. causing firms to either attempt to reduce fixed costs with even newer technology. which ultimately spreads fixed costs over a greater quantity. it would seem that the decreasing capacity scenario would be more risky merely because unused capacity is more difficult to manage than the prospect of full capacity. fixed costs still need to be paid although revenues have declined. for an example.260 costs. Comparing industry-average operating leverage to company-specific operating momentum will reveal production risks not apparent in even the standard deviations of sales and income. each of these outlying factors drains cash from collective production processes. From the perspective of common sense. reacting to extraneous factors like law suits. but operating momentum would remain at 1.9. an operating momentum that is actually lower than industry operating leverage can signify that the company will under perform the industry when that sector is favored. operating momentum responds to the vagaries of the market. Thus. or to leave the business entirely. the true operating leverage would increase to 1. may not be reported as part of normal operations. While such variables would not be reflected in basic operating leverage. Ultimately. on the other hand. they remain implicit in operating momentum: we measure sensitivity to total cost with the latter ratio and infer that fixed costs are increasing when operating income increases at a faster rate than sales. the more stable the production environment. prices will decrease. While diversification can lower risk. even when they are counted as one time separate charges. Reasons for divergence include: changing the basic nature of the firm through acquisitions. When a state of over-production is reached. but will affect production for years to come. they are an inherent part of operating momentum. Since the objective of diversification is to lower . A multi-billion dollar settlement against a tobacco company. is dependent on the industry and changes gradually over time as new technology is added. 261 overall risk in the long run. Therefore. For example. If all three measurements are higher than average. The small winery must depend on a favorable growing season. and makes large profits when conditions are “right”. is not diversified. possessing what Warren Buffett would call “franchise value”. the investment requires extensive analysis. Its operating momentum may be measured to be less than the industry average of “operating leverage”. because earnings will tend to increase at a slow and steady pace. These elite companies are usually good investments.66 Operating Momentum = 1 (Given as a Hypothetical) . To illustrate the effect of changes in price on both operating leverage and operating momentum. it is important to combine information about this divergence with the standard deviation of both income and sales. compare Gallo wines to a small winery in France. but it trades immediate high profits for long-term viability. companies with lower operating risk are at a temporary disadvantage when the relative sector surges ahead in the economy. Gallo is a diversified winery much less dependent on a favorable market. and can sell wine far into the future. or might warrant shelving altogether. consider the following scenarios: Table 10-2 HIGH OPERATING LEVERAGE Variable Costs = 700 Operating Income = 300 Sales = 1200 Price = 4 Quantity = 300 Fixed Cost = 200 Operating Leverage = (1200-700)/300 = 1. In a few industries there is the power to change prices without affecting demand. Those “sacrificial” profits are gained back in long-term viability. 5 % PRICE INCREASE FROM 4 TO 4.5 % PRICE CUT FROM 4 TO 3.66 OPERATING MOMENTUM .5. Table 10-4 HIGH OPERATING LEVERAGE SCENARIO (1050-700)/150 = 2. operating income increases to 450 from 300. increasing four fold.33 LOW OPERATING LEVERAGE SCENARIO (1050-800)/150 = 1. Sales become 1350 from 1200.50 / .5 = 4 Notice that both ratios displayed the greater risk of a price cut by increasing.5 Sales become 1050 from 1200.33 Operating Momentum = 1 Given as a Hypothetical) THE EFFECTS OF A 12.262 Table 10-3 LOW OPERATING LEVERAGE Variable Costs = 800 Operating Income = 300 Sales = 1200 Price = 4 Quantity = 300 Fixed Costs = 100 Operating Leverage = (1200-800)/300 = 1.12. Other variables remain unchanged. while operating income declines to 150 from 300. . but that operating momentum reacted more violently. Other variables remain unchanged. THE EFFECTS OF A 12. Since variable costs increase when quantity increases. This movement has predictive value.263 Table 10-5 HIGH OPERATING LEVERAGE SCENARIO (1350-700)/450 = 1. When price can increase at a faster rate than quantity. which in this case we assumed to be “1”. Even a price shift that has positive acceptance from the customer must be met with a countervailing force because more risk has been incurred.22 OPERATING MOMENTUM 50 / 12. Whenever it moves substantially away from the mean. Thus. or price to redirect momentum back to its mean. .5 = 4 The essence of pricing power is that it significantly reduces risk (operating leverage) while raising operating income. that the firm will be forced to control cost.44 LOW OPERATING LEVERAGE SCENARIO (1350-800)/450 = 1. quantity. we can expect it to compensate by reverting toward the average in the following year. after a quick rise to “4”. the higher costs that accompany quantity driven changes in sales are avoided. investors will actually demand it because it means production is stabilizing. However. the firm may have created a niche that competes on the basis of quality-driven product differentiation. analysts would be looking for the conditions that would lower operating momentum . discontinuance of some operations or even restructuring charges. such risks could entail increased salary demands. Although operating momentum does not differentiate between specific risks as well as operating leverage. it should be noted that operating momentum increases whether prices are increased or decreased.greater quantities and more variable costs . any large shifts will so upset corporate equilibrium. While the reversion is by no means a “hard and fast” rule.hopefully with increased earnings. On the other hand. they may have to do so all at one time to keep operations running smoothly. If the company fails to counter the “positive” risk of increased prices. only to be followed by a sell off because the situation is temporary and the inherent risks return to normal.. risk to these companies actually increases because not only do they have to make the necessary capital expenditures. because demand is essentially inelastic -it will not decrease as prices increase.264 A good example of operating momentum-shifts occurs in industries that sell commodities. variable costs do not increase. sacrificing longterm viability for short-term gains is thus a risky strategy. The consequent rise in operating income relative to sales occurs because sales rise without a quantity-driven change in costs i. the entire industry can raise prices. Another example of counteracting the risk of higher growth occurs with a price announcement. OPERATING MOMENTUM SENSITIVITY A comparison between operating momentum and leverage reveals the implicit volatility of operating momentum. which will increase operating momentum. companies with more fixed assets may actually benefit. Another topic that affects both leverage and momentum is the impact of inflation. . When supply of a commodity is particularly low.e. Marketing strategies are geared toward the risk of operating momentum: Can the company cut quantity and still make a profit if customers find a less expensive substitute? Can the company counter this move in the following period with a quantity driven product line that appeals to a more cost-conscious customer? These are the types of questions that fuel diversification and are the reason that more diversified companies have a lower operating momentum and inherently less risk than “one product” companies. The variables in the breakeven equation are: price. it must suffer the consequences of another shift in operating momentum. The stock price will rise for a brief period. The student will observe that in times of inflation. The distinct lag time between initial sales and the need to upgrade equipment will temporarily boost income above those companies that must deal with more immediate rises in variable costs. Table 10-6 BREAKEVEN POINT VARIABLES . and keep the other breakeven variables constant. we can observe the effects on both leverage ratios. variable cost per unit and fixed costs. If we change each breakeven variable by a factor of 5 % and then 25 %. The changes in variable and fixed costs assumes a one percent change in sales We establish a hypothetical base and then display the effects of both five and twenty-five percent increases and decreases:.BASE Quantity = 100 Price = 4 Sales = 400 Variable Costs = 2 per unit Fixed Costs = 100 Operating Income = 100 .265 quantity. 91 1.91 2. = 2 Increase 1. Lev.25 2.9 .5 -23 Decrease 4 2 52.11 2.83 1.5 29 Table 10-9 OPERATING LEVERAGE Variable Change = 5 % Price Quantity Variable Costs Fixed Costs Base Op.105 Decrease 2.11 1.266 Table 10-7 OPERATING MOMENTUM Variable Change = 5 % Price Quantity Variable Costs Fixed Costs Increase 4 2 -6 -1 Decrease 4 2 14 9 Table 10-8 OPERATING MOMENTUM Variable Change = 25 % Price Quantity Variable Costs Fixed Costs Increase 4 2 -48. and as variable costs rise. but hyper-sensitive to changes in variable costs. Reality dictates that the breakeven variables interact. that product becomes a commodity. isolated changes occur only in the laboratory. eventually eclipsing any operating profit. The much narrower range of values makes operating leverage a better gauge of risk than momentum. except in the last year when fixed costs increase to 400. periods of stability can be readily contrasted with disturbances.5 1.1700 and 1800 respectively. Each year. However. The Bee Good Honey Corporation has three sales increases and sales for each year of 1000. As quantities increase. When a company can only compete by churning out more of an undifferentiated product. or both. variable costs become more prevalent.66 1. which assumes statistical “normality”. momentum captures the volatility of year to year changes. Naturally. fixed costs are 300 and variable costs are 60 % of sales. there are other techniques from “extreme value statistics” that may better characterize this fleeting measurement. In four years. raise prices. If we assume “stable” production processes (a concept that many operations managers would scoff at). Besides reversion to the mean.6 Operating momentum is insensitive to price and quantity changes. the company may cut production. because of its very instability. the next example will show the mirror relationship between the two measurements.267 Table 10-10 OPERATING LEVERAGE Variable Change = 25 % Price Quantity Variable Costs Fixed Costs Base Op. .1500.67 3 2. operating momentum may be the better forecasting tool. While true operating leverage may change over the long run through changes in the production process. Lev.66 Decrease NA 3 1. = 2 Increase 1. Since there is equivocation about the definition of “fixed” and “variable” in real accounting costs.33 % = 2 . Once fixed costs change as they do in Year 4.78 (1800-1080)/320 = 2. especially the barriers to entering the field 3. = %OP/%SALES NA 200 % / 50 % = 4 26. The predictive ability of operating momentum is predicated on the probability of increasing operating margin.268 Table 10-11 YEAR Year 1 Year 2 Year 3 Year 4 OP. and it is an astute management that will lower momentum with a large sales increase. The type of business and industry 2. LEV.6 Observe how operating momentum will perfectly mirror operating leverage when costs are stable. management can maximize potential margins. higher margins will not continue indefinitely. On the other hand. = (S-VC)/(SVC-FX) (1000-600)/100 = 4 (1500-900)/300 = 2 (1700-1020)/380 = 1. that mirror diminishes and the predictive power of operating leverage is lost.79 % / 5.15.e. .9 % = -2. but must counter its limitations with additional strategies. a “real” measurement like operating momentum is less reflective of risk and harder to interpret. The following table illustrates the probabilities of operating margin and momentum increases in a data set of about 180 different sample points. the introduction of a quantity-driven product that lowers operating risk.666 % / 18. The economic outlook and business cycle Within that framework.25 OP MOM. i. Operating margin is affected by: • • • 1.. The level of competition. operating leverage is relegated to a hypothetical. we are measuring how correlated income is with sales. any operating income decrease will lead to a negative momentum number if sales increase at the same time. and designating the amount of risk by the covariance.14 70. no valid decision rule exists. Without reference to both size and variability together.e. . The reader should notice that the sample of 182 data points displayed a distribution of exactly one half (91) momentum increases and decreases – an indication of random variation. In essence. while percentage sales increases are the independent “X” variable.96 Therefore. Analogously. comparative risk is vague.33 43. and whose major risk is economic i. The percentage increases of periodic operating income are the “Y” variable. REGRESSION When linear regression is applied to operating momentum.269 Table 10-12 INCREASING OPERATING MARGIN STATE Total Operating Margin Increases With Increasing Operating Momentum With Decreasing Operating Momentum NUMBER 104/182 64/91 40/91 PERCENTAGE 57.. the probability of an operating margin increase is greater when operating momentum increases. By default. simultaneous decreases of both the components will lead to a positive number and a potential increase. but without recourse to the changes in fixed and variable costs. operating risk. The following tables are comparisons between three companies that have little debt. .66 14.881 0. there is little that regression can tell us.41 14.70 38.36 5.62 29.35 1998 5.05 21. The best indicators of risk remain the standard deviations of sales and income respectively. even when operating risk is considered so great that no long-term debt is incurred at all.58 Table 10-14 COMPANY MOLEX BIOMET FAIR.29 11.49 16.69 30.177 R 0.129 The problem with regression is in interpretation.39 1999 -4.50 12.63 5.60 8. and their various adaptations like the coefficient of variation. and the mean-variance rule.24 25.44 1997 16.37 11.61 12.62 15.15 7.223 4.270 Table 10-13 1996 MOLEX OP INCOME SALES BIOMET OP INCOME SALES FAIR.842 0.133 0. All numbers in the data table are percentage increases.56 11.48 2000 29. ISAAC Y INTERCEPT (ALPHA) -4.69 15.9684 1.28 23.871 COEFFICIENT 1. As the student will observe. ISAAC OP INCOME SALES 7.359 -2.94 37.16 12.27 33.66 8.92 23. 92 12. For Fair. Dividing the two yields 1.66 % 13.1619. but one look at their respective distributions makes us choose the latter. Without the ability to inter-relate sales and income through specific costs. the coefficient of variation for operating income is 1.5731.68 % 22. the coefficient of variation for operating income is 0. THE GENERAL ELECTRIC SOLUTION Even to this author. while that for sales is 0.26 14. Isaac is on the upside. while that for sales is 0.259.23 % 11. Isaac.271 Table 10-15 COMPANY MOLEX SALES OPERATING INCOME FAIR.59 Notice that Molex and Fair Isaac have almost identical regression profiles.7385. Thus. our only recourse would be to evaluate operating momentum by creating a ratio of coefficients of variation for both operating income and sales. General Electric. Dividing the two yields 1.78 % 21.656. a company with enough personnel to populate a . The ratio of standard deviations alone would misstate the risk for percentage values that were particularly large.5211.43 % 10. Molex has a riskier stream of income and we would use these figures to evaluate operating momentum. While it enumerates the separate risks of both sales and income. small values would have a smaller standard deviation and appear less risky. ISAAC SALES OPERATING INCOME MEAN STANDARD DEVIATION 9. it seems to create independence between the two fundamentals that does not exist in reality. Most of the risk for Fair. the coefficient of variation seems like an incomplete solution. For Molex. we know that sales and income are interdependent and the risk that we need to measure must flow from their working together. The following two tables show the relationship between operating margin and momentum.86 -9.16 -3.% Total Cost solution has great practical value. 10 18 22 20 35 40 Op. The fine line between cost.57 % Rev. operating momentum must be greater than one. risk. pushed by the changes in momentum. Table 10-16 “COMPANY A” Sales 100 120 132 140 165 150 Total Cost 90 102 110 120 130 110 Op Inc.84 9. ROC/ROE.1 0.266 Op. the specificity of true operating leverage can not be superseded by any other measurement. Marg.38 Distance NA 6. Like the previous coefficient of variation. While operating margins can increase over time.09 % TC NA 13.167 0. defined as operating income divided by sales. Performance is measured through the mean-variance method.199 -1.83 -15. and the “GE version” of operating leverage. Consequently. and higher margins translates to a better return on equity and higher stock prices.47 .15 0..53 24.5 4.but will not explicitly divulge the risk of fixed assets. they define it as the percent change in revenue minus the percent change in total cost. In their 2006 prospectus. and the %Revenue .33 7. they usually move cyclically. operating margin. NA 4 2.06 17. any type of performance evaluation can reveal the return and stability of both margin and momentum . Again.67 2. NA 20 10 6. or % ∆ Revenue .03 9. which is a utilitarian concept. It will yield a number that is highly correlated with operating margin.272 major city. when that condition arises.143 0.% Total Cost formula is termed “Distance”. must rise anytime the percent change in revenue is greater than the percent change in total cost. 0. has their own definition of operating leverage. and stock price.212 0.% ∆ Total Cost.09 8. Mom. and both of those variables are positive.22 -1. The % Revenue . 42. i. Although margins increase when we implement our method. σ Distance = 10.85 Distance NA -2. the operating momentum was less than “1” and operating margin still increased.e. Thus. the mean-variance for σ Company A is . there is still the chance of more variability because risk is not explicit. any company can view the necessary requirements.948 %. the mean-variance for σ Company B is -9. Inc. By coordinating operating momentum and “Distance”.2.28 % TC NA 7. µ -σ = -9.7 14. there are no indications of changes in fixed costs which can have long-term effects on production.99 -3. µ -σ = -2.15625 Op Mom NA -2 0 5. 10 9 9 21 10 25 Op. 0.0909 0.43 µ Distance = 0.96. The following chart is a four year comparison of changes in fixed costs and various operating risk measurements: .14 11.592. sales and income changes. The second condition of the relationship was not met. but only operating leverage can specify the amount.14 -2. The operating margin is growing more vigorously for Company A. Thus.0714 0. When operating momentum is growing over time.14 20 11.34 10.107 0.22 15.86 22. Table 10-17 “CONPANY B” Sales 80 84 99 121 140 160 Total Cost 70 75 90 100 130 135 Op.11 30 3. Mar. The student should observe that in the last data point. NA 5 17..11 -14.948.125 0.3 10.1725 0. σ Distance = 10.273 µ Distance = 7. the Distance parameter requires that the percent change in revenue is positive.42 %.38. a higher level of capital expenditures and fixed assets may be needed. for increasing operating margin.5 % Rev.54. 2941 1.33 2 0.as long as that “Rock of Gibraltar” type company “ups the ante” by balancing stable income with greater leverage.67 3.24 Of the three operating risk measurements. Leverage 0. inflation and market activity. NA 4 2 -2. Picking growth over stability is fine as long as the growth is Xerox in the 1970s or Google in the new millennium. at first declining and then going negative. and investing in growth potential. the answer is more dependent on the situation. If companies specified fixed and variable costs on financial statements.2272 0. deriving a generic measurement is a monumental task. than a formula. the investor who sticks with stability can do well . but their measured response is more volatile.25 Op. The market will reward risk-seeking behavior when financial leverage has a “counterweight”.78 0.6 % Rev. While it is futile to time markets. it is never too difficult to gauge the interaction between interest rates.%TC NA 16. steady interest coverage and a low default probability. . Since some assets have dual use and have both fixed and variable costs. Like the equivocation between large cap stocks and small caps. Mom. The two other measurements react to the increases through the interface of total costs. Since such choices are rare.33 0. everyone is well aware of when the last “downturn” occurred. and decide whether one is in the early or late stages of a recovery.5 4 0. the investor would not worry about momentum because all of the information to increase margins in the domain of risk is contained in the operating leverage measurement. Other dilemmas occur between investing in the size and stability of income.25 0. i.2702 Fixed/Variable Op.3703 2..33 -6. only operating leverage fully reflects the increases in fixed costs.274 Table 10-18 Year Year 1 Year 2 Year 3 Year 4 Fixed/Total 0.e. “ Distance”? The answer can easily be deciphered from the numbers: .and not through the fault of analysts. and that concentrating on growth factors like the “PEG” ratio is a risky proposition at best. As an example of this marginal analysis. While the growth of operating margins is constrained by the production characteristics of most industries. one will hear complaints that profit maximization theory is not “realistic”. when the argument is phrased as “risk versus return” rather than “value versus growth”. There are numerous other variables that can be enhanced. particularly in commodity industries. estimates made from fundamentals were off by as much as twenty-four percent . The company decides to “ramp up” production. The next year is very fertile and revenues go up by fifty percent. CLASSICAL MICROECONOMICS AND OPERATING MOMENTUM Students often take introductory microeconomics as a prerequisite to upper-level courses.275 It is when the analyst attempts to time the peaks and shifts between cyclical stages that predictions go awry. new avenues of approach are encountered because interdependence is recognized. the methods by which return is maximized are not. that “marginal revenue never equals marginal cost”. momentum. Frequently. and sees a rise in total costs of fifty percent. For many years. Most investors find that growth estimates are as much “art” as “science”. Professional analysts get direct “guidance” from companies because accurate forecasts of rising earnings will attract capital. are essentially ways of reducing fixed costs and minimizing risk Thus. Such classic arguments like “leasing versus owning” or modern methods like “outsourcing”. except that it is directed toward the flow of income. consider an agri-business whose revenue in one year is 200 with 40 in operating profit. such a scenario is valid. In fact. and the GE invention. The debate between “value” and “growth” investing that has been ongoing since the days of Philip Fisher and Benjamin Graham is still raging today. What happens to operating income. even as margins temporarily decline. and 160 in total costs. There were simply too many uncontrollable variables to make an accurate forecast with the consistency needed for investing. .5 While the provision that operating momentum equals one and “distance “ equals zero. a negative change in distance combined with a movement of operating momentum toward zero indicates that quantity is becoming counter productive. operating momentum will decrease and “distance” will decline past zero. Table 10-20 VARIABLE Sales Operating Income Total Cost Operating Momentum Distance (% Revenue %TC) YEAR ONE 200 40 160 NA NA YEAR TWO 300 40 260 0 -12. does not imply that profits are being maximized. when costs are rising faster than revenues.276 Table 10-19 VARIABLE Sales Operating Income Total Cost Operating Momentum Distance (% Revenue %TC) YEAR ONE 200 40 160 NA NA YEAR TWO 300 60 240 1 0 In an alternative scenario. 277 Figure 10-1 TOTAL REVENUE (TR) $ TOTAL COST (TC) Quantity Figure 10-2 Operating Profit $ Quantity . (Back to Table of Contents) . In fact. there is pressure to produce less. there is some optimum of percentage changes in sales and operating income that allows this “hypothetical ideal” to occur. as is operating profit. Thus. and the distance between the concrete values is maximized. the percentage change in operating income will approach zero. because the increase in total costs exactly offsets the increase in revenue. When ∆ Total Revenue < ∆ Total Cost. producing more of an item becomes counter productive – a nightmarish scenario for any firm. when operating momentum approaches zero. When ∆ Total Revenue > ∆ Total Cost.278 The top graph shows total revenue and total cost. the slopes are the same. When marginal revenue equals marginal cost. there is pressure to increase the quantity produced. Although the equality is consummated with absolutes and not percentage changes. while the bottom graph shows operating profit. At the outset. and that entails “taking one step back to move two steps forward”. and profits. “ex-post” analysis can be both humbling and beguiling.e..from either investors’ or managements’ perspectives. We present several simplified methods of comparing an idealized calculation of a company’s capital requirements to its actual additions. we can extract a comparative logic behind setting short-term benchmarks for capital requirements. Short-term manipulation of corporate fundamentals can damage the long-term viability of a company. i. assets. a cumulative three year gain of fifty percent is better than a twenty percent gain over two years. While the material presented in this chapter entails methods of raising capital to meet the objectives of the investor. Often. if we observe some of the realities behind “funding”. the investor needs to realize several concepts before we proceed. • 2. Determining capital requirements can be a complicated mathematical exercise given the complexity of most large firms. an after-the fact. The ideal model assumes linearity while reality imposes a jagged curve. Economies of scale and different utilization rates of fixed assets create an exponential relationship between sales. THE REALITIES OF FUNDING • 1) The best companies often move away from their target capital structures for a year or two. They function merely as guidelines to educate investors as to why a firm behaved in a certain manner. Nevertheless. • 3. Certain large risks must be incurred to ensure large returns. it does not advocate them.279 11 STRATEGIC CAPITAL REQUIREMENTS Taken out of context. individual assets are itemized and the requirements of smaller units are aggregated. Moreover. and are not meant to be final arbiters for decision-making . the calculations are models that are imperfect in their simplicity: they leave out many variables and contain many assumptions. . • 7. a restrictive covenant in an existing bond contract may preclude the accumulation of debt above a specific level. once funding occurs there are practical constraints that prohibit efficient combinations of debt and equity. On the same note. optimal combinations of funding are not always possible unless “ideal” internal conditions exist for the firm. Capital structure optimization in the short-term may not take into account the exigencies of the competitive environment. To maintain control. it may make sense to take on more debt than is deemed optimal. and buy back shares of stock. the capital dynamic may rise to greater heights in future years with fewer shares on the market. Even if net income is lower in the immediate year. Although most executives will concentrate on minimizing the WACC. For example. Acquisitions require much more capital than would be warranted by normal increases in sales. For example. Short-term interest rates may be so high that normal one year loans are discarded in favor of less expensive long-term debt Thus. • 6.280 • 4. because investors expect large returns from these capital infusions. over capitalized companies often do better than under capitalized firms. It is a mathematical exercise that carries little . and still improve a firm’s EVA. It is also possible to have an over abundance of either equity or debt. the need for funds escalates. or the outlook for the economy. • 5. Using the EVA/Capital Dynamic to determine capital requirements is premised on the accuracy of the CAPM . THE PROPER AMOUNT OF CAPITAL The greatest impediment to capital structure optimization is the addition of too much capital. Since most publicly traded companies grow as much through acquisitions as through internal “organic” growth.which has been shown to be randomly unstable and not always correlated. shareholders may want to restrict stock issues even if they are warranted. Then why is the amount of capital so important? The EVA/capital dynamic is a universal filter with a few limited variables. if interest rates are historically low. but the “value” of the investment may not pay off until earnings are actualized. This is another danger of using EVA expectations to make capital requirements decisions: it may impose artificial restraints on capital that normally would be raised to meet competitive needs. Conversely. Also in the chapters concerning the Spearman rank correlations. because a company may be perceived as more risky if income has yet to come to fruition. Even large equity increases when coupled with large earnings gains usually lead to higher stock prices. When compared to previous years’ performances.a diminution of market value because corporations “over spent” in the prior year. a higher income attracts capital into the stock. and retain earnings. which occurs because both capital and net income are heavily correlated. implied by transitivity. greater equity increases will accompany higher net income because there is a tendency to both pay off loans. While the situation was hypothetical. most of the market gains will be speculative when a company incurs more debt: tax benefits are immediate. On the other hand. it represents the true danger of over capitalizing. If the student examines the Spearman rank correlations in the section about probability. In essence. he or she will find that more capital is correlated with stock price increases. other profitable projects may be discarded in order to concentrate on the capital infusion. and will be mirrored by a decreased EVA/capital dynamic. and the lower the amount and cost of equity. the potential for efficient capital utilization increases. and EVA will increase accordingly. we observed the effects of capital rationing . large debt increases are rarely consistent with high net income increases because greater interest payments cut into operating income. If prospects are truly favorable. Thus. An excess of additional capital will be too costly in relation to the income it is required to produce. in addition to the inclination for “troubled” companies to take on more debt.281 restraint: the higher the net income. the higher the function. it creates a chasm between investors who want immediate results and financial executives who are . it must work within specific limitations in order to improve. the imminent disaster may have been prevented. Moreover. Thus the capital dynamic / EVA does not have great predictive power because it focuses on immediate earnings . Sun Microsystems and Lucent Technologies were the darlings of the investment set. the large betas for these stocks would have been a tip-off. “Buy and hold” strategies did not work in an era of merger and acquisition.282 concerned about long-term viability. it may lead to a “worst case scenario”: the manipulation of balance sheet fundamentals to appease the largest shareholders. Ultimately.the vision of a transformed economy that would always out perform “the past”.even though historical earnings fluctuations and a lack of outstanding credit might have deemed otherwise. because they indicated a level of risk that only a few investors could have safely taken. Although the market would not have “sky rocketed”. THE DEBT / EQUITY TRADEOFF AND EVA In the very speculative year of 1998. The “safe. When the market for tech stocks finally collapsed. the lack of long-term debt would have been a vague signal that the income stream was risky. From a capital structure perspective. their unbridled earnings potential seemed to warrant a massive influx of equity capital . companies like PMC Sierra.5. it might have stabilized at a slightly above average level which would have attracted capital far into the new millennium.5 instead of 1. a thorough analysis of trends in both the cost of equity and the rate of change in equity. little company” that Grandpa put into his retirement account five years ago. How could this have been prevented? While net income was climbing rapidly. These were the companies that were going to change America and bring us boldly into the twenty-first century. millions of investors were left with a large amount of shares worth only a few dollars each. Technology provided investors with an elixir . would have revealed otherwise. was now a high tech behemoth that was about to “bet the ranch”. The rest of the story is well known. Even as a higher EVA may have indicated movement toward the optimal (through abnormally high earnings). Had most of the tech stocks carried betas of 0. By definition. Since it is imperative to protect EPS and dividends. Given a projected EPS. and the risk of compromising dividend income with more interest expense. indicative of a recapitalization such as a leveraged buyout. “opportunity cost” which rarely dilutes market value. it occurs when the difference between net income and the absolute cost of equity is large. Since equity is proportionately more expensive than debt. and not a leading one. Manipulations can and do occur so it is important to know the context of these shifts. the firm is presented a tradeoff between two forms of risk to shareholder value: the risk of diluting market value with more shares. a company will increase debt but decrease equity by an even greater amount. EVA will increase: the net effect of increases in income more than offset the corresponding increases in stockholders’ equity. Less debt will reduce interest payments creating a consequent rise in net income. In fact. EVA/capital dynamic analysis will be impervious to techniques like shifting expenses from income statements to capital accounts simply to “pump up” immediate earnings. and will not forecast trends over the course of the business cycle.given the default ratings of agencies like Moody’s. while the latter has only a comparative. it is a concurrent indicator.283 rather than potential capital utilization. these are usually isolated occurrences. a knowledge of limits on equity will produce a corresponding calculation of allowable debt . If the earnings that are retained from this income rise faster than the cost of equity. On rare occasions. the investor should certainly differentiate between equity issues and retained earnings. we can use this data to set a rough “guesstimate” for capital requirements that would be ideal from an . we can set limits on the amount of equity capital that can be raised without diminishing movement toward the optimal structure. This is the leverage mechanism that is so correlated with movement toward an optimal capital structure. However. Within capital constraints. causing a shift upwards in the EVA/capital dynamic. Together. because the former is often used as “currency” for acquisitions. 235. 115 . The following example will suffice: COMPANY XYZ Table 11-1 YEAR Net Income Equity CAPM Percentage Cost of Equity EVA/Capital dynamic ONE 100 500 7% 35 65 YEAR Net Income (estimate) Equity CAPM Percentage Cost of Equity EVA/Capital dynamic TWO 115 ? 8. it does not necessarily mean that it is time to sell the stock.235. and so the value of this analysis . and 3) Last year’s data on net income and stockholders’ equity . That cost of 50 is then divided by the estimated CAPM percentage to obtain a value for equity. if XYZ comes in at 600 for an equity figure.X = 65. sector and market . but it will signal the need to examine the context of the decline in EVA . In the scenario above. 2) An estimated required rate of return from the CAPM. Thus given the correct estimates. X = 50.085 = 588. 50/. or even one’s own research is treading on shaky ground. Since the investors’ immediate goal is to at least preserve the difference between the cost of equity and net income. the maximum that stockholders equity could increase and still maintain the size of EVA would be 88.284 investor’s immediate perspective -and may or may not be conducive to a firm’s long-term growth. Reliance on estimates.more closely.5 % ? Goal = 65 or more The cost of equity is derived by subtraction.to calculate an absolute limit on this year’s equity. but professional opinion is usually available for the component parts of this rate. whether they be from analysts. the market.sales. and a firm’s beta. it is a simple matter to take: 1) An analyst’s projected net income. The estimation of the CAPM percentage can be difficult because one is correlating the relationship between interest rates. ex post.285 comes after the fact .such as accepting those projects with a positive net present value. we would need to know the default limits set on debt by ratings agencies like Standard and Poor’s . the greater the potential for movement toward an optimal capital structure. BBB. Since we know the limits on equity. capital funding would have to be provided through long-term debt which increases the variability of EPS. EVA / CAPITAL DYNAMIC BASED IMPROVEMENT In previous chapters. rather than on sound financial principles .and establish the threshold limits of the company’s rating class (AAA. it is . The lower the level of equity in comparison to net income.6 % increase in equity was enough to absorb the 15 % increase in net income. which usually acts as a buffer against risk. we suggested in this chapter that an optimization based on EVA increases might set the company up for a later fall because capital funding was not contingent on the net present value of projects. To use this material properly. can have the opposite effect if the level is high enough to warrant more debt financing. But what if equity were already high? What if it could only move up five or six percent to absorb the increase in net income? In that situation. high equity. AA. it is important to know how much leeway is available for an equity increase. Moreover. and the risk premium of the capital asset pricing model. Thus. The XYZ example above suggested that a 17. we suggested that the EVA / capital dynamic improvement is based upon the relationship between interest rates on debt. If the EVA/capital dynamic indicator is primarily a concurrent gauge of stock performance. why is it important to know what it is after the fact? In order to analyze the potential for an increase in the EVA and by association. etc. the entire company suffers through higher capital costs and a depressed stock price. A disconnect occurs any time that capital budgeting becomes based on the short-term goals of the investor.). such as in a high interest rate environment. When that scenario occurs at inopportune times. the market value of the stock. which make them invalid inputs for a working corporate model. we will obtain an interest expense that can be used to calculate the new financial leverage ratio.286 imperative to have a legitimate standard for debt/equity . the interest rate does not rise in correspondence to the amount of debt. Such an equation is a “win . The “ball park” figure is a hypothetical ideal meant to gauge the potential for shifting capital proportions. In essence. The amount of dividends acts as a stress test for net income and additional shares can be compared with the necessary dividends that would justify them. retained earnings are added to stockholders’ equity once dividends are subtracted from net income. 2) beta does not rise when more debt is incurred. the maximum amount of equity is derived. Indeed. When we multiply this figure by the current interest rate. EVA defined requirements coalesce several other forecasts involving variables such as taxes. We set up a simple algebraic equation and solve for retained earnings under the constraint that no additional shares are issued. As in the previous section. we determine what level of retained earnings must be achieved to improve EVA. dividends and net income – each of which can be fundamentally wrong. as well as determining any level of additional equity (issues) that would be warranted by the difference between the optimal and proposed amounts. company five year averages. Additionally. From the given level of total capital. While corporate capital requirements are contingent on sales expectations. any use of an EVA determined capital model must harbor some assumptions based on forecasts and would be prone to error.win” for investors because it essentially displays where a break in the marginal cost of capital will occur. Three other assumptions are relevant to this analysis: 1) in a given year. we determine the change in debt by subtracting the new level of equity (previous equity + new retained earnings). we are working backwards to find the level of debt and equity that would increase EVA. and 3) operating income shifts to accommodate either more or less interest expense. Those are unrealistic assumptions.which can be industry averages. . or even a study of cyclical peaks of the stock price. given a specific level of income. (Net Income (yr1)) + (Net Income (yr2))) / % Cost of Equity (yr2) This figure is the maximum amount of equity that can be raised in Year 2 for an improvement in EVA. Year 2 = (yr2) Year 2 Stockholders’ Equity = ((Equity (yr1) x % Cost of Equity (yr1) ) . Year 1 refers to the current year while Year 2 is considered the next year. allocating capital on the basis of immediate economic profit. Table 11-2 VARIABLES Year 1 Net Income Year 1 Percentage Cost of Equity as determined by the CAPM Year 1 Stockholders' Equity Year 2 Net Income (projected or estimated) Year 2 Percentage Cost of Equity (estimated from changes in components of the CAPM) Year 2 Stockholders' Equity ? Determined by an algebraic solution Year 1 = (yr1). and its relation to net income.((new cost %)(X) = EVA. i. If indeed net income rises and the calculated amount of capital proves inadequate. THE IRRATIONALITY OF RATIONING CAPITAL Without reference to the internal dynamics of capital budgeting. An easier method is to set up an EVA equation and solve for equity: Net Income . In the example.. there is .e. This is termed the threshold EVA. That process entails funding projects based on positive net present values and having an internal rate of return (IRR) that is greater than the cost of capital. is doomed to fail. The EVA/capital dynamic makes improvements based on changes in the short-term cost of capital. The EVA should be as least as great as the previous period’s.287 INCREMENTAL EQUITY IMPROVEMENT The definition of the incremental equity improvement is the amount of equity to raise. but a company’s long-term viability is centered on developing products that meet market demand over the entire business cycle. that would maintain or improve a firm’s EVA. EVA. There is no implication that this figure represents the ideal level of debt. DIVIDENDS AND RETAINED EARNINGS When we solve for net income. INCREMENTAL DEBT The maximum amount of equity to raise presents an extreme corner solution to capital funding if no debt were raised. that standard can be relaxed. this assumption is significant because it creates a level of funding where the marginal cost of capital will not be raised. For sensitivity analysis in corporations. On the other hand. the change in equity implies a change in debt. it is the only realistic choice if EVA is going to improve. The dividend amount that we subtract from net income must project the next dividend but use the last period’s shares outstanding. By setting the equation to the previous period’s EVA. Raising debt would dilute net income with interest payments but also decrease the required amount of equity and thus its cost.288 a danger of projections not being met. This optimum can be subtracted from actual capital to . However. within the limitations set by the model. given a level of EVA. because the amount of raised capital may be incorrect.which may be a cost that is actually below the cost that was calculated when some of these projects were developed. This rationing of capital may favor projects that only return an amount above the current cost of capital . we also add retained earnings to the previous year’s equity. For investors. we have already identified the maximum amount of equity. CAPITAL FUNDING FROM EVA: TWO METHODS METHOD 1: SOLVING FOR THE OPTIMAL EQUITY Using the actual net income from a company. we want to develop a net income without reference to an equity issue. Subtracting an idealized equity from a given amount of capital will yield a residual debt figure that we multiply by the interest rate to determine interest expense. the return on both equity and capital may be greater in the short run since only projects with the highest net present value will be accepted. and a company can find a level of additional shares that would least dilute stock value. While this assumption is pure speculation. The derived optimal equity value is then subtracted from this quotient of proposed net income and previous ROC. proposed dividends. Once debt is determined. . it has some validity because we are improving EVA as well. and realistic equity amounts are used. The student/investor should note that any amount of hypothetical EVA can be set. Actual or proposed net income is then divided by (1-tax rate) to produce a new EBT. There is no need to calculate dividends in this method because retained earnings are implicit in the equation. it is multiplied by an estimated or current interest rate to derive interest expense. a “target” operating income is determined. Dividends are determined by multiplying last period’s number of shares outstanding by the next proposed dividend. and retained earnings. METHOD 2: SOLVING FOR THE OPTIMAL NET INCOME We look for a minimum level of funding that would improve EVA. and so we set the new equation to the previous EVA and simultaneously solve for net income. as long as actual shares. Once interest expense is added to EBT. If we use analysts’ forecasts for earnings.289 determine the residual amount of debt. our only alternative for determining capital is to assume that the return on capital (ROC) will be at least the same as last period’s. Dividends Equals Retained Earnings + Stk.(Yr 2 Dividends)) +(Yr 1 Equity)) = Yr 1 EVA Algebraically. Yr. Table 11-4 DETERMINING CAPITAL PROPORTIONS AND REQUIREMENTS VARIABLES NET INCOME TAX RATE YEAR 2 EBIT EBT IMPLIED INTEREST EXPENSE CURRENT INTEREST RATE LONG-TERM DEBT EXPLANATION Derived from Previous Calculation Current Effective Rate Derived from ( EBT + Interest Expense) Equals Net Income / (1-tax rate) (Interest Rate) x ( long-term Debt) The Effective Rate Equals Actual Interest Expense / Actual long-term Debt Equals Capital .Equity Once net income is derived. we simultaneously calculate the new amount of equity. multiplied by the projected dividend. The Year 2 Dividends variable is the product of the last period’s outstanding shares. 1 The full equation is Net Income .290 Table 11-3 VARIABLES NET INCOME PROPOSED NEXT DIVIDEND OUTSTANDING SHARES (Year 1) STOCKHOLDERS' EQUITY (Year 1) CAPM % (Required Rate of Return) RETAINED EARNINGS NEW EQUITY EXPLANATION The" X" Factor Projected or Known Last Period's Number of Shares Outstanding Last Period's Stockholders' Equity Either based on past regressions or estimated Equals Net Income . Again. In order to solve for the new level of operating income. We . we solve for “net income” as all other variables are known. we need the current effective tax rate.. because we have obtained the amount of retained earnings that solves the equation. Hlds' Eqty.((Yr 2 % Cost of Equity) x (Net Income . there is room for adjustment and sensitivity analysis on a full spreadsheet. FINANCIAL ENGINEERING: SETTING CAPITAL REQUIREMENTS FROM EVA If we assume that operating income is constant. We then add the implied interest expense to EBT to determine the amount of operating income needed to achieve this EVA improvement PROJECTED ANALYSIS While the determination of optimal net income is used to evaluate the performance of companies in an ex. We start with an optimal net income. average .291 subtract this rate from “1” and divide this difference into our derived net income to yield a new EBT (earnings before taxes). and calculate the implied long-term debt. and by default. which is only achievable in a stable economy. then our optimal net income will imply a specific level of interest expense. In essence. we work deductively backwards. it is a small step to actually project EVA scenarios. we would most likely formulate a minimally acceptable level of capital funding by determining the quotient of last period’s ROC and analysts’ projected earnings. One would have to extrapolate both the cost of equity and interest rates. When we add the equity amount to the debt amount. 3) some determination of capital requirements.post manner. determine a corresponding equity. The three main variables needed would be: 1) a projection of the next dividend per share. CONOCOPHILLIPS: 2005 . ConocoPhillips had to raise enough capital to pay for the purchase. This method is particularly insidious because no analysis of capital budgeting or economic viability needs to be done. as with the optimal equity method. and assumed that company’s outstanding debt. As a percentage of sales. which moved it away from its optimal target structure. The method is: Net Income / (1-tax rate) = EBT. the oil company ConocoPhillips bought Burlington Resources with a large amount of stock. 2) an accurate forecast of the next EBIT. an amount of long-term debt that corresponds to it. We determine capital outlays from potential profits regardless of their productive source. we determine a derived amount of capital.2006 A REAL WORLD EXAMPLE After four years of record profits. 6 % to 1. (0.13529 +15550) / . which would make it equal to approximately .038 We observe that ConocoPhillips increased its financial leverage ratio only 1.4% (EBT.85 .(Yr1 Net Income) + (Yr2 Net Income)) / (Yr2 % Cost) = ((52731 x .0825)(X) = 9933.82 % 52731 10758 4.25 % 82646 23091 4.1 % 9933 497 179442 1388.(.98(millions) 1. The following data applies to 2005 2006.0825 = 68088. METHOD 1: CAPITAL PROPORTIONS FROM OPTIMAL EQUITY STEP 1: Determine the optimal equity ((Yr1 Equity x Yr1 % Cost) . X=68084.038 and that a larger increase would have possibly deleterious effects. Table 11-5 CONOCOPHILLIPS Variables Net Income CAPM % Stockholders' Equity Long-term Debt Interest Rate (effective) Tax Rate (effective) EVA/Capital Dynamic Interest Expense Sales (from operations only) Outstanding Shares Dividend per Share Dividends Paid Financial Leverage Ratio 2005 13529 6. A ten percent increase in the financial leverage ratio is within historical bounds.1 % 8732 1087 183650 1581.0825)(X) = 5617.71 % 45.44 2277 1.0682) . The goal is to create a level of funding that would preserve the previous 9933 in EVA. not EBIT).25 (millions) 1.18 1639 1. and the company would not normally descend below this level with more interest expense. given the prevailing higher interest rates in the economy and shrinking margins on oil.62 % 42.292 earning before taxes is 6.0212 2006 15550 8. A second way of achieving nearly the same results: 15550 .1.14. 0471) = 1773.78 FINANCIAL ENGINEERING METHOD .0825) x ((X-2000.549 = 28324.0825X + 4185.23 + 1773.5 ConocoPhillips would need 677. X= 15387.293 STEP 2: Subtract From capital to determine debt 105737 . Yr2 dividends are ** ** determined by multiplying the last period’s shares outstanding by the latest dividend or 1388. 105737 ..23.79. Xstk = 66118.X .33 STEP 4: Determine the amount of interest Expense (39618.In this example it is X.33)(0.13 STEP 2 Compute the amount of Equity.3) = 9933 or 0.13) + (52731))) =9933 which reduces to . METHOD 2: CAPITAL PROPORTIONS FROM OPTIMAL NET INCOME STEP 1: Determine the threshold level of net income. .76 + 1866.98 (1.67 STEP 3: Determine the amount of debt. (Xstk equals new equity) Net Income .3.27 = 30097.((Yr 2 % Cost of Equity) x (X-(Yr 2 Dividends)) +(Yr 1 Equity)) = Yr 1 EVA . EBT + Interest Expense = Operating Income.9175X = 14118.76.(0.79/.5 (million) more in operating income to achieve this EVA ideal.02 STEP 5: Determine the target operating income 15387.79.(.79 .68088 = 37649 STEP 3: Determine Interest Expense 37649 (.66118.44) = 2000.0471) = 1866.((.02 = 29894.27 STEP 4: Determine Operating Income Target EBT = 15550/(1-tax rate) = 15550/0.(Cost of Equity %)(Xstk) = Previous EVA = 15387.67 = 39618. Add interest expense 28324.549 = EBT = 28028. The equation is X .0825(Xstk) = 5454. 28028.0825)(Xstk) = 9933. Within those ratings. On a default probability basis. Actual Effective Tax Rate = 45. operating income must be high enough to absorb interest payments without exceeding industry averages for financial leverage. The capital that ConocoPhillips actually raised was 42248 which is the difference between the two years’ sums of long-term debt and equity. which further determines the amount of allowable debt.28028.294 Derived Optimal Net Income = 15387.79. Derived Equity =66118.83 %. Proposed ROE = 23. Operating Income . This balancing act continues with operating margin: While greater margins lead to more retained earnings. 1391.EBT = Interest Expense Interest Expense = 29420 .79/0. (1-tax rate) = 0. Actual Operating Income = 29420.549 = 28028.76. The lower cost of capital was produced by using slightly lower cost debt.88 %. lower margins imply that less equity financing will be done.67 + 29538 = 95656. 29420). operating income determines interest coverage.549 EBT = 15387. Bankruptcy risk can cripple a stock and so most increases in leverage are within the boundaries of the industry. the additional debt may be excessive.0471 = 29538 Implied Capital = Derived Equity + Implied Debt. these margins have to be within industry boundaries or investors will presume that the firm is more risky. 66118. and correspond to the highest obtainable credit rating. New LTD/CAP = 30.76 = 1391.82 %.67. A COMPARISON While both methods added debt.24 Implied Debt = Interest Expense/ Interest Rate. The two boundaries for operating income are operating margin on one side and the financial leverage ratio on the other. Old LTD/CAP = 21. the optimal net income scenario produced a lower cost of capital and implied an operating income that was only slightly more than the original (29894.1 %. and yet low enough so that the tax advantages of those payments exceed the advantages of extra income. This analysis raises the same amount of capital with far greater debt.67 Actual ROE = 18.78 Vs.24/0. To obtain the optimal level of debt. .27 %. 6 % 30097. EVA would have been maintained.55 % 9933 Optimal Equity 37649 68088 105737 1773.67 1391. taking into consideration.84 % 29420 18. debt and interest payments.79 1.0496 30.84 % 14. Without a thorough examination of ConocoPhillips leverage.5 22. undermining the exigencies of capital budgeting. and this conformance may have been crucial to future EVAs. they extrapolate from projected sales. While we do not know whether ConocoPhillips’ projects were properly funded.31 % 9933 2006 23091 82646 105737 1087 15550 1. debt becomes implicit in most net income projections. Their proposed capital requirements were most likely proper. we can not second guess their management. Capital requirements would have met the 105737 standard. and so a fifteen percent increase in operating income will imply fifteen percent increases in both interest expense and net income respectively.82 % 14.9 % 29894.24 15387.67 95656. 39018.67 105737 1866. 29538 66118.0626 35.0666 36. with a major increase in debt. margins. Therefore.0212 16.78 23.086 % 9933 Optimal Net Inc.27 % 14.27 15550 1. Effective tax rates remain fairly stable for most companies. we do know that capital rationing can make things “look good on paper”.66 % 21.27 % 16. considering .95 % 24044 25.71 % 8732 Financial Eng. Table 11-6 COMPARISON 2005 Long-term Debt Equity Capital Interest Expense Net Income Financial Leverage Ratio LTD/CAP Operating Income ROE ROC EVA/Capital Dynamic 10758 52731 63489 497 13529 1.79 1.02 15387.295 When analysts forecast yearly earnings.33 66118.71 % 9933 Therefore. and equity would not have been issued.038 21.88 % 29420 23. Its basic premise is that “nothing happens without a sale” . THE ADDITIONAL FUNDS NEEDED EQUATION The standard methodology for assessing capital requirements is the percentage of sales method and its close relative. and then uses the sales forecast to project the estimated size of each item. short hand version of the percentage of sales method. but investors can use it as a screen to identify those companies and industries that have an absolute advantage over their peers. Only those items that increase spontaneously with sales are included. Those companies with the highest average sales increase. these assets are compared with the internally generated funding available from short-term credit and retained earnings. Corporate analysts use the AFN equation to project needed funding. or AFN. In fact.externally generated capital like bond issues. which itemizes each balance sheet item as a percentage of existing sales. the additional funds needed equation.296 that large acquisitions were being made.a principle often forgotten by contemporary hedge funds who are more enamored by risk management than the mundane “ups and downs” of the business world. Within the component parts of the equation lies a wonderful financial logic that is elegant in its simplicity. and seem almost passé given Wall Street’s modern day penchant for mergers and acquisitions: the equation rarely matches capital outlays because asset accretion through “organic” sales growth has been superseded by the outright purchase of other companies. their capital costs are low. and the deficit is made up from the “additional funds needed” . coupled with the least need for additional funds are more destined to move toward an optimal capital structure. These techniques have been mainstays of financial textbooks for over fifty years. the AFN is almost a comprehensive. To illustrate: . and that funding went well beyond existing operations. i.297 THE VARIABLES Table 11-7 VARIABLE A / S (assets / sales) EXPLANATION Assets that Increase with Sales ∆S Change in Sales from Year 1 to Year 2 (Absolute and Concrete not a Percentage) L/S Liabilities that Spontaneously Increase with Sales (not Notes Payable. for example) S1 Absolute and Concrete Figure for Projected Sales M Profit Margin (net income/sales) This is a Projection (1-D) D The Retention Ratio once Dividends are Paid Dividend Payout Ratio (Dividends/Net Income Since precise accuracy with this equation requires an accountant’s knowledge of the company. .e.. familiarity will breed flexibility. one can derive a ball park figure by assuming that L / S implies current liabilities and that profit margin is a five year average. THE ADDITIONAL FUNDS NEEDED EQUATION (A / S (∆ S) ) – (L / S (∆ S)) . Over capitalization usually requires a movement away from the optimal capital structure. then over capitalization is less risky. Sensitivity analysis is also easily performed with a data table or a one variable solver like “Goal Seek”. and the return on capital. consider a company who seemingly “over funds”. Just like EVA and the capital requirements model. it may be just the “stroke of genius” that the firm needs to buffer them from a down turn.the type of funding acquired by ConocoPhillips in our previous example. and lower fixed costs. the AFN lends itself to spreadsheet optimization through a linear programming module using an equation solver.D)) ∆ ∆ Essentially. and a consequent lowering of the return on capital. the AFN is mechanistic and oblivious to context. the demand for that product may be constant. These same companies will over capitalize for the sake of acquiring other companies . minus the internally generated funding available from increased profits. sales/capital. In essence. they can easily be projected through the geometric growth method (see statistics chapter). a low profit project much to the consternation of various corporate “quants”. and other business activity. and require little outside funding. As an example. even if the individual cost of the components is low. If that funding occurs at the end of a business cycle. . Thus. its capital costs will be too high.and offer the investor a great opportunity. if it is high. simply because the funds are not used for the immediate generation of sales. since the firm has more capital than needed. wages payable. and that scenario can cause a drop in both EVA/capital dynamic. the equation is composed of assets that increased with sales minus the internally generated funding available from trade credit. Like EVA however.(M(S 1)(1 . A negative or low figure is usually correlated with lower sales as measured on a year to year basis.298 If dividends are consistent. Evaluation is based on the company’s historical capital turnover rate. But .on occasion there are companies who consistently generate internal funds above their immediate needs. The real value of the AFN is to use it as an algorithm to gauge capital efficiency. holding a modern corporation to the AFN standard is myopic at best. Thirdly. Any large distribution chain knows the benefits of consolidating inventory into a few key locations. The significance of these three components is as follows: The A / S figure is a capital intensity ratio that gauges the amount of fixed assets in a business. While it is a useful tool to determine some levels of funding. The AFN gives us a comparative “ball park” figure to add to our decision making and can be a “red flag“ to check more closely. the larger the amount . The larger the A / S. followed by a steady increase. While sales do not increase exponentially. By dividing the last figure. as more units are distributed over the same amount of fixed costs. The equation reduces to: (A / S ) ÷ (L / S) ÷ (Retained Earnings / Sales). machinery will be idle that will lower expected sales. they do increase by a much greater linear factor. and not duplicating processes. Secondly. they will be under utilized. retained earnings. we find it to be a preferred measure of risk. by sales and then dividing. we can obtain a measure of capital self sufficiency. changes in the business cycle can offset additional funds because excess inventories build up during unforeseen downturns. This inconsistency causes a large drop in the assets to sales curve. At first. Most industries possess economies of scale in which more sales are increasingly generated with less assets. but so too will the unused capacity of fixed assets. generating less sales and then gradually producing at capacity. Not only will these excesses be unaccounted for in the AFN equation. the cost and nature of fixed assets require that they be bought and implemented in large “blocks”. When we compare firms within industries. The synergies of most mergers involve the cost savings of reduced inputs. THE MODIFIED ADDITIONAL FUNDS NEEDED FUNCTION The deficiencies of the AFN equation are readily apparent. one figure into another. It also gives a “rough estimate” to the amount of operating leverage.299 there are many extraneous variables that determine the level of capital funding. not subtracting. and financial executives have learned to exploit disparities in the yield curve that offer arbitrage advantages in capital funding. The greater the liabilities that rise spontaneously with sales. earnings may be more variable. the more funds that will be internally generated by business activity. One can observe the mechanism if one considers the capital intensity ratio. they should be distributed as dividends. Short-term credit is normally less expensive than long-term debt. When this ratio is especially high. along with volume discounts and the ability to extend loans. most dependable source of funding. retained earnings can reduce share issues without incurring accounting costs. with the implication that a business is more “capital intensive”. The gist of the modified AFN function is to have the lowest ratio within a specific industry. but for a company with both a high capital intensity and variable income. create a “capital sub structure” that is never adequately measured by analysts. especially at the end of a business cycle when the cost of equity is high. especially when compared to similar companies. Few companies have to worry about “too much” retained earnings.300 of fixed costs. they offer the least expensive. in the Gordon model. The reason that some companies quickly turn long-term debt into profit is that they create liabilities that significantly lower their cost of borrowing. and obtain a large number. . The L / S ratio is vital. A / S. it may mean that retained earnings were too small to provide adequate financing. as a hypothetical measure of operating risk. The retention ratio is a significant component of growth. If we divide the retained earnings ratio into a high A / S. we multiplied it by ROE to obtain a growth factor for the cost of equity. Companies with high operating risk can afford less debt and must generate more internal funds. We have also observed that too great an amount of retained earnings can raise the cost of capital. Vendor relationships that develop significant trade credit. and we would look for less financial leverage to be employed. The retained earnings to sales ratio will encompass those funds that are generated by more profit. When the company’s return on investment is lower than the cost of equity. even in the case of excess. 98 1974 4207 3505 1050 249. a decrease is relative to the prior year and the industry that contains it.53 56. Since each entity is capable of improvement.25 1978 5578 5902 1339 315. In its simplest form. and its absolute size depends on the type of industry. This equation contains the same types of components but gives them the common reference point of sales and then divides them. the modified AFN separates them by the amount of outside funding required. commodity industries have large modified AFNs. not positively correlated.85 76. the Du Pont equation multiplies profit margin by asset turnover and the equity multiplier.39 . Like EVA. and associated regression. The student/investor is encouraged to look for trends. except that it is negatively. it runs concurrent with stock prices. or: (Net Income / Sales) x (Sales / Assets) x (Assets / Equity). and should be used for intra industry comparisons. except on a different scale. much like the Du Pont equation for ROE. The following example will illustrate the scope of the numbers. Improvement is achieved when the ratio decreases. The elegant simplicity of the measurement is that it can be calculated in thirty seconds without reference to the cost of capital. while higher profit businesses fall on the lower end of the scale. Table 11-8 XEROX YEARS Assets SALES Current Liabilities Retained Earnings A/S / L/S / R/S 1969 1516 1357 419 113. ROE.301 The modified AFN function is similar to EVA/capital dynamic. While the measurement. does not distinguish between the types of industry.07 42. 04 120.6 13.82 Table 11-12 MICROSOFT YEARS Assets SALES Current Liabilities Retained Earnings A/S / L/S / R/S 1993 3805 3753 563 953 26.4 45.67 1996 84.302 Table 11-9 MANPOWER YEARS Assets SALES Current Liabilities Retained Earnings A/S / L/S / R/S Table 11-10 BARRA YEARS Assets SALES Current Liabilities Retained Earnings A/S / L/S / R/S 1994 1204 4296 668 75.3 16.49 2000 3042 10843 1522 155.47 Table 11-11 WALMART YEARS Assets SALES Current Liabilities Retained Earnings A/S / L/S / R/S 1992 20565 55484 6754 1775.97 1996 39604 104859 10957 2567.11 2001 52150 19747 9755 7785 18.51 1997 14387 11358 3610 3439 13.04 149.6 39.2 105 32.3 3.55 95.24 98.89 1993 38.5 20.51 103.14 2001 83451 217799 27282 5603.23 1997 2047 7259 1005 149.6 12.61 138.06 2000 226 224 68 45.47 . The advantage comes when one either decreases the value.as long as existing capital needs were met. like Barra and Microsoft. the firm may have a profit margin slightly above others in the same sector. The market will set a premium on the one quality that a specific company does not possess. there is no absolute advantage in having a low Degree of AFN. For Microsoft. Moreover. but anticipates better future prospects.303 Notice that companies who eschew outside funding. In the former case. Mathematically. funding with debt. In the latter case. the expression is (Assets x Sales: / (Current Liabilities x Retained . or possesses the lowest Degree of AFN within a sector. and then paying off the loan with the proceeds of an equity issue when earnings increase. we derive a quotient between cost components and funding components: (Assets / Sales) x (Sales / Current Liabilities) x (Sales /Retained Earnings). a higher degree of AFN will give a company more flexibility as to strategic capital allocation. if we multiply by the inverses of the equation. which would be likely if the firm were an industry leader. the challenge is to lower capital intensity. fall into a much narrower range than high turnover companies like Manpower and Wal-Mart. Each company is capable of improvement within their respective degrees of AFN. For Wal-Mart. DEGREE OF AFN LOGIC The logic behind the degree of AFN is straight forward. a decrease would most likely signify that a company was profitable enough to pay off debt . Thus. They are dependent on the vicissitudes of operating income to meet capital goals. Simplified. they have the option of turning to the credit markets. the premium is on profit margin and more retained earnings. With no long-term debt. if Manpower has a poor year. On the other hand. a company like Barra must sell stock if retained earnings are insufficient to meet anticipated needs. Higher asset turnover (smaller capital intensity) allows a company to take on more debt with less risk. but cuts into an already narrow profit margin. a more diverse mix of capital funding is available when a company uses external financing. . a more accurate five year CAPM regression would have changed the size of EVA.2000. a higher ROE can raise the cost of capital.which has a good probability of raising the cost of capital. an absolute increase in the ratio indicates more financing is derived from external sources . more retention. While it does not measure major changes in capital structure like ROE. the degree of AFN is dependent on the relationship of sales to the other variables.304 Earnings). such an indicator can be misleading. Since the respective levels of retained earnings and current liabilities are dependent on sales and assets. it may be a more accurate indicator of both risk and return. While we can never be certain that lowering the AFN will also lower the cost of capital. To illustrate how close these measurements can be. If a firm becomes obligated to paying high interest rates far into the future. but not the overall relationship between net income and equity. Without the optimization of proportional debt to equity. this is a better concurrent indicator of performance. On the other hand. and increased business activity are all correlated with minimizing capital costs. the ROE can improve despite massive loads of debt because leverage (Assets /Equity) is implicit in the magnitude of the calculation. On a purely utilitarian level. a higher profit margin. both of these measurements are imperfect. The cost of equity is skewed upwards because it was calculated from a one year rendition of the Gordon Model. How well the degree of AFN stands up to conventional indicators like ROE is the subject of continued research. consider the following analysis of Manpower in 1999 . 61 138.94 % 1.305 Table 11-13 MANPOWER . because they depend on internal. as well as long-term debt to capital.7 in 1999. While some companies avoid short-term debt when the yield curve is inverted as it is before a recession. rather than market dynamics.3 Notice that the degree of AFN mirrored the ROE as a performance indicator.12 % 1.59 would not have been as forward looking as a drop in current . Most investors would have avoided this stock because there are no clear indications of the direction of the company. However. However. The slight decline in asset turnover to 3.59 4.56 4. and indeed Standard and Poor’s gave this stock an “avoid” rating in 2001.18 2719 9770 1418 135 138. because of concerns about the employment market during a concurrent recession.56 from 3. use of the Gordon model makes EVA calculations more exacting but less accurate. EVA decreased ever so slightly.1999 – 2000 VARIABLE ROE Profit Margin Asset Turnover Equity Multiplier Assets Sales Current liabilities Retained Earnings Degree of AFN Net income Equity Cost of equity EVA 1999 22.53 % 3. Manpower’s’ ratio was probably more indicative of lower business activity.04 % 15.98 150 651 20.47 in 2000 from 2. the asset / capital ratio dropped to 2.56 2000 23.65 % 15.11 3042 10843 1522 155.58 % 3.74 171 740 21. A more thorough analysis indicates that the company increased both operating momentum and financial leverage. Even with a low beta of 0. Theoretically. if I know that the industry standard is 30 % long-term debt to capital. but it is not a substitute for analytical common sense. Additionally. numerous financial institutions undergo a mass dislocation. most major companies have some optimizing software that attempts to put constraints on the amount of capital spending within the domain of better performance. Although they would receive some of the overflow from businesses that did not want to fully commit their resources to hiring employees. By changing the levels of constraints to meet new economic standards. while “booms” would loosen credit and allow lower TIE (times interest earned) ratios. Thus. This data is then turned into a logit type probability function that chooses between default and solvency.306 liabilities as an indication of business activity. unless these credit standards are flexible. analysts can construct a collective probability distribution for similar companies consisting of interest coverage ratios and the frequency of default. it provided the logic for qualitative assessment: a series of tradeoffs and stalemates leading to uncertainty. distressed economic times would create a higher standard and “tighten credit”. there will be some penalty in violating . THE NEED FOR QUALITATIVE ASSESMENT Although numerical analysis was quite neutral. THE PROBLEM WITH OPTIMALITY In any given time period. even if the fundamentals tell you that it is. At the highest levels. the company creates its own default probability model. they could not compensate for cutbacks in temporary labor. They were in the employment business in the middle of a recession. they will be unable to mirror the changes in demographics and asset structure that comprise the larger economy.8. capital structure analysis possesses some powerful computational tools. However. the company was taking on debt as interest rates were peaking which may have raised the cost of capital. Manpower was “in the wrong place at the wrong time”. When disparity occurs. For example. with consequent fallout to smaller businesses and the requisite Federal Reserve involvement. Selling Christmas trees in July may not be such a good idea. . given the constraints of default. and arbitrary than the former. has the same risks that any non diversified business plan has: it may work well for a period of time. we get neural networks that primarily get their input from past performance and are unable to “foresee” new developments. capital spending optimizes at a structure of all debt. we must use the probability of default in a unique way: we need to solve for the variable that would most improve the algorithm earnings -and relate it the amount of debt. we again enter the territory of “over dependence” -an algorithm that can send us in the wrong direction. Thus. However. The solution of course. which encourages over dependence. the more random variation affects decision making. unique. which is even more undefined. without these self imposed constraints. but in how they are used. there is both a lag time and inadequacy in their breadth of coverage.must counteract it. Secondly we need to juxtapose immediate tax benefits to long-term benefits. and then collapses completely. Lastly. To circumvent this dependence. realizing the higher net present value of the immediate benefit. Since bankruptcy costs imply the use of default probabilities.307 that constraint and that some compensating factor . etc. the company faces a similar dilemma as most financial institutions: the greater the complexity of the economy. and earnings are high. we equivocate between adding debt when it is less expensive (usually when interest rates are low). The crux of problems with default probability models lies not with the human frailties of analysts or in mathematics. more diversified operations. and adding equity when it is attractive. we use the stock price as a barometer of value: any market value that was .. Basing consumer credit on a probability model that is derived from one company. Eventually. simply because interest is tax deductible. When we base capital structure optimization on a shifting parameter of minimizing capital costs.increased market share. Even when variables are computer generated. in one time period. Probability models will sometimes fail to work simply because they are unable to encompass new information. because it fails to adapt to changing circumstances. is to not so much concentrate on the cost of capital as the cost of bankruptcy. to our advantage . the indicator becomes more speculative. it is still counted in the “book value” of assets over liabilities. In fact. and only then on the level of default. stock price and tangible assets that can be used as collateral. the probability of default is as much a part of the solution.as a measuring stick that balances several other variables income. assets. rather than investors’ choices to fund a stock. it is because it has become prominent for two reasons: 1) hedge funds and private equity firms have used mergers and acquisitions to create the illusion of growth. although mathematical precision would be lacking. The difference is crucial: when companies pay a premium over fair asset value. MERGER MANIA If tangible book value per share is one of the linchpins of our optimization algorithm. the level of debt will depend on the interaction of profitability and tax advantages. 2) accounting policies have changed over from the “pooling” method to the “purchase” method of accounting. is that both . with the hoopla of CNBC crowing about “Merger Mondays!”. the scope of today’s mergers. we can at least obtain a ball park figure that would enable us to compare an idealized capital structure with our own. which makes any multiple of market price to book price seem less speculative. The legitimate purpose of mergers is to increase economies of scale and achieve synergistic earnings that would not otherwise be possible in a smaller enterprise. because it is based upon management’s judgment of the merger’s potential. Any generic probability of default can offer input on what the best level of earnings would be. Diversification reduces risk and any firm that decreases operating leverage can ultimately increase the value of the firm by adding financial leverage. However. as it is a constraint. is viewed as a transitory negative. the excess is termed “goodwill”. Nevertheless.308 created without lower default probabilities. Such an ideal will lack the market adaptability mentioned previously. but will be less dependent on creating an absolute constraint on default. we use our biggest liabilities. In essence. Thus. and becomes an intangible asset. However. or at least without more tax benefits from debt. taxes and debt. The investors believe that more assets will equal a higher stock price. “the illusion of growth”. there will be economies of scale and some synergy that help reduce costs and increase profit. What investors look for in a company is often an “emotional ideal” that is differentiated from the competition in some discernible way. while management has an unshakable faith in the efficiency that comes from pooling resources. even if no synergy or economies exist. What they don’t know is that a worldwide shortage of the company’s main input will eventually escalate costs. many mergers are based on what Eugene Brigham termed. If fixed assets can be shared by the two companies. crippling any chance of “organic” . a cash flow that is timed differently from the acquiring company’s will help reduce risk. not necessarily to the exclusion of the qualitative . In fact. Their P/E is 20 and their EPS is $ 1 / share. In fact. Sales have stagnated and they need to increase earnings per share or risk analysts’ downgrading the stock.but at least emphasizing it. What would be the products? How could investors choose sides? In essence greater growth is not always a cash cow for investors. or Coke and Pepsi. only to realize later that greater market share did not translate into greater profit per share because the market was declining. Here is how it works: Company X is not growing internally anymore. Defenders of mergers have three valid arguments. However. greater size will allow quantity purchases that will reduce variable costs. especially if greater demand is not stimulated. Thirdly. Secondly. Consider a legal merger between Fed-Ex and UPS.309 investors and management get conned by the asymmetry of information: many of these deals are simply methods of making a large “commission” for a few participants. both investors and management can suffer the consequences of a merger that looked good on paper. it is almost always lucrative for those who broker the deals. What these arguments have in common is that each depends on quantitative growth. For years they have been growing at 20 % and investors expect that rate to continue in the future. each earning $1 per share. their earnings. P/E ratios are in line with earnings. the ratio is 3. and the number of shares. Company X would adjust for the deficit like this: Table 11-14 NET INCREASE IN PLANT AND EQUIPMENT $50 MILLION NET INCREASE IN GOODWILL $50 MILLION NET INCREASE IN EQUITY $100 MILLION If company X originally had a 4/3 (1. In fact. each also earning $1 per share. rising from 3/4 (0. But if we go by tangible book value per share. If company X has 20 million shares outstanding. and at a P/E of 10. Company X makes a play for company Y. book value increases relative to market. On the other hand. Company Y has 10 million shares outstanding. Company Y has patents about to expire and so the value of its assets is much less than the $100 Million (5 million shares x $20/share) in stock that company X paid for it. for a total of 10 million dollars in earnings. which is less than the original value. After discussing ways to consummate the merger. The new earnings per share of the combined company are: $20 Million + $10 Million / 20 million shares + 5 million shares = 30 / 25 = $1. .8). 100 Million). looks like a bargain. Suddenly. If we reverse the ratio and go by book to market value.25).75) to 4/5 (0. and tangible book value would increase by only one half the asset value (50 vs. suppose company Y is only worth $50 Million.310 growth. company X has grown 20 % in EPS simply by combining two pieces of paper! Now what if company Y is a highly profitable bioresearch firm and company X is a tool and die maker. the companies decide that the most equitable method would be to set up an exchange ratio of 10/20 or 0.5/5 (0.7).33) market to book value. of course. are 20 million dollars.5 shares of company X stock for each share of Y. price. The merger seems to go unquestioned.20/share. it would decline to 5/4 (1. company Y is distressed. company X would have to take on increasingly larger mergers or suffer the consequences of less growth and a larger cost of equity. The main concern of an optimal target strategy is to ensure that the tax benefits of any leverage used to buy another company are equal to the additional distress costs. but large institutional investors have to be. Nevertheless. while company Y has a 1/10 or 10% cost of equity. and would have increased these distress costs without any additional interest payments or rate changes. To continue growing without productive synergy. Company X has a 1/20 or 5% cost of equity. Company X had a (20 . we have increased EPS by 20 % without any corresponding increase in productivity. If there is no . those marginal bankruptcy costs can offset any gain in tax benefits. we can apply EVA to both companies.!))(400)) or 30 . a decreased tangible book value per share was coupled with greater market value.(((. Assuming that the “rule of thumb” inverse of the PE gives us an accurate cost of equity. “goodwill”. The average shareholder can not be aware of all the purchases made by a firm.05) + (. and created $100 Million in market value In our simplified model of bankruptcy costs.25 = $5 Million EVA Ultimately. the final merger engendered a higher cost of equity for company X with no real gain in economic profit.05)(300)) or $5 Million economic profit. Since the deal was paid for in stock and not debt.1)(100)) or zero economic profit. and the brokers of the deal a lot wealthier MERGER GROWTH ILLUSION AND EVA Like earnings there is no tell tale outlier that tells us EVA is being manipulated.75)(. and attributing the deficit to another asset class. leaving a handful of short-term investors happier. Company Y had a (10 . If the gap between market value and tangible book value increases. and paying for its market value in stock.25)(.(. Together the combination produced (30 .311 Two significant points: 1) By purchasing a company with a lower PE.(. we have taken $50 Million in assets. there were no concurrent tax advantages that would have absorbed the higher default costs. 2) By purchasing a company in excess of its true market value. earnings would have increased by 20 % in the near-term. (Back to Table of Contents) . it may be time to look for another investment. but earnings are touted as “growing”.312 real gain in economic profit. 313 . 314 . Without such access. The second section will be a set up for optimization of the actual figures. The first section is for input of actual values from financial statements for sensitivity analysis. and will allow entry of only one variable. There is no special skill required. and row 2 is all calculations. either in computer science or mathematics . the user will find that a working model is instructive when he or she tries to follow the logic of the material. . The user should input the word “ACTUAL” in bold print in A1. SECTION ONE Section one has 14 columns and 4 rows. It will require two sections which are near mirror duplicates of each other. The columns are labeled “A” to “N”. The other columns and rows are set up as a pattern of one row of titles and another row of calculations below it. Row 1 is all titles. and row 4 is calculations. SETUP The program is set up for both sensitivity analysis and optimization. Row 3 is all titles again. the student/investor can still follow the conclusions of the experiments. and may want to review the entire chapter before proceeding. but column A is just a title column. However.except for remedial knowledge.315 SECTION II: BUILDING CAPITAL STRUCTURE MODELS 12 THE ECONOMIC PROFIT LABORATORY: COMPUTER APPLICATIONS This chapter will require the user to create an economic profit model in a spreadsheet program like Microsoft’s Excel. except for the entry of “Equity” in J7. Its titles form a perfect mirror image of those in section one . Risk-free rate BETA Market Rate Column/Row Title B3-B4 Interest Expense (B2-B4)-(C4 C3-C4 Tax Rate x (B2-B4)) ENTER D3-D4 Interest Rate C2 / D2 E3-E4 Long-term debt J2 / D2 F3-F4 Cost of debt ENTER G3-G4 Debt/Equity ENTER ENTER H3-H4 I3-I4 Total Capital WACC Calculation ENTER Calculation D4 x E4 ENTER ENTER H4 . with a calculation of (=L7 . section two changes BETA each time a new proportion of debt to equity is .except for two cells: in N6 the title is “EVA Difference” and not “ROC”. Technically.L2) placed below it in N7.(k2 x J2) E4 / H4 C2 / H4 J3-J4 K3-K4 L3-L4 M3-M4 N3-N4 Cost of Capital Total Cost of Equity Stock ROE EVA Stock Value SECTION TWO Section two is for optimization. The calculations in the optimization section are determined by the inputs in section one.J2 D4 (1-C4) E4 / J2 ENTER ((F4 x E4)/H4)+(( K2 x J2)/H4) I4 x H4 J2 x K2 E2 / K2 C2 / J2 L2 / I4 J1-J2 K1-K2 L1-L2 M1-M2 N1-N2 Equity Cost of Equity EVA LTD/CAP ROC ENTER G2+(H2 x (I2-G2)) C2 .316 SECTION ONE Table 12-1 Column/Row Title B1-B2 EBIT C1-C2 D1-D2 E1-E2 F1-F2 G1-G2 H1-H2 I1-I2 Net Income Number of Shares EPS Book / sh. (K7 x J7) E9 / H9 L7 – L2 .G7)) C7 . By holding operating income. a new net income and cost of equity is calculated which changes EVA. the model is based on the capital dynamic and not EVA. By changing just one cell (“Equity” in J7). Table 12-2 Column/Row B6-B7 C6-C7 D6-D7 E6-E7 F6-F7 G6-G7 H6-H7 I6-I7 J6-J7 K6-K7 L6-L7 M6-M7 N6-N7 Title EBIT Net Income Number of Shares EPS Book / sh. Properly speaking. The beta calculations are merely restatements of the Hamada research that was covered in the chapter on the cost of equity. but we use these concepts interchangeably when interest paid on debt mirrors the interest rate. and the interest rate constant.317 inputted. all changes in EVA are driven by changes in equity.B9) .(C9 x (B7 . Risk-free Rate BETA Market Rate Equity Cost of Equity EVA LTD/CAP EVA Difference Calculation B2 (B7 .B9) D2 C7 / D7 J7 / D7 G2 H2 / [(1+(1-C4) x G4) x (1+(1-C9) x G9)] I2 ENTER G7 + (H7 x (I7 . capital. They are listed here to make sure that you are “on track”. a number of “derived” variables will be calculated. SAMPLE DATA To employ both optimization and sensitivity analysis. A8 and A9 will be left blank. Another option that the user may find helpful is to copy the entire “ACTUAL” section. After inputting the variables.E9 C4 D4 H9 .318 Table 12-3 Column/Row B8-B9 C8-C9 D8-D9 E8-E9 F8-F9 G8-G9 H8-H9 I8-I9 J8-J9 K9-K0 L8-L9 M8-M9 N8-N9 Title Interest Expense Tax Rate Interest Rate Long-term Debt Cost of Debt Debt / Equity Total Capital WACC Cost of Capital Total Cost of Equity Stock ROE EVA Stock Value Calculation D9 . Cells A7. input the word “OPTIMIZATION” in bold letters. and paste it a few rows below the optimization section.J7 B9 x (1-C9) E9 / J7 H4 (F9 x (E9 / H9)) + (K7 x (J7 / H9)) I9 x H9 K7 x J7 E7 / K7 C7 / J7 L7 / I9 In cell A6. . the student/investor can make comparisons between the same firm or different firms since the module is “stand alone” and not dependent on an outside source for calculations. we will use some sample data to enter into the “ACTUAL” section. section one. In this manner. Therefore.75 % 43. There is no predictive model that can encompass either of these variables when massive changes in capital structure are made. optimization is more theoretical because we need to keep both interest rates and the number of shares constant.75 % Use the derived figures to check your calculations. Thus. .5 25 17. optimization assumes that a firm can raise as much debt as required at the given interest rate and that the amount of shares will not change.319 Table 12-4 INPUTED VARIABLES Equity Tax Rate Interest Rate Total Capital Shares Operating Income Risk-free Rate BETA Market Rate AMOUNT 500 30% 5% 1000 20 150 4% 1 10% DERIVED VARIABLES Long-term debt EPS Net Income Book Value / sh. On the other hand.13 50% 6. The number of shares to issue is sometimes a political judgment with a little mathematics thrown in. while interest rates change on a periodic basis. GE’s EVA is possible as well as an indication of the necessary inputs to improve it. EVA Interest Expense ROE ROC Stock Value EVA Stock Value Long-term Debt / Cap. The user can configure the program as decimals or percents but should probably stick to decimals in the beginning to make sure it works correctly. these variables are derived from sources outside the model. It is quite easy to place an inappropriate sign.5 25 37.75 35.375 87.5 % 8. SENSITIVITY VERSUS OPTIMIZATION Sensitivity analysis is more realistic than optimization because each variable can be changed to meet the actual demands of financial statements. WACC AMOUNT 500 4. a realistic rendition of say. and the earnings figure will grow to its maximum. and some improvement in actual EVA is viable. By minimizing the amount of debt. it is amenable to changing its parameters and maximization. A third nominal constraint is the amount of capital. it can point the firm in the proper direction. which is located in Excel’s tool section. we can emphasize the “net income” side of the function because interest will not be deducted. they can not calculate the proper amount of capital. However. this model assumes that the proper amount of capital is being raised and works within those limitations. the firm has the option of moving slightly in the suggested direction. we will use the sample data from above. At this time. under or over funding prospective projects creates numerous “shocks” to the system. Since EVA is a linear function. go to the information section and read the instructions on Solver. Solver is an “Add In” for solving linear problems with constraints. it is necessary for the student/investor to be familiar with a linear programming module like “Solver”. then there must be some combination of cost elements that creates an indifference between debt and equity. then the interest deductions may be both small and tax deductible such that equity should be minimized. “Fixing” the overage or underage with a shift in capital proportion will only delay the pain of inefficiency until a compensating capital inflow actually neutralizes it. Since capital allocation is the most important decision that management can make. PROVING THE CAPITAL DYNAMIC / EVA HYPOTHESIS Economic profit is a linear construct. other spreadsheet programs have very similar . SETTING THE CONSTRAINTS To examine this hypothesis. While these programs can exhibit how those problems can be circumvented. without making changes in the number of shares issued or the current interest rate that is paid. Minimizing equity would detract from the “Stockholders’ Equity” side of the function and create a larger EVA by default. If unfamiliar with the procedure. if the cost of debt is inexpensive enough. If EVA does indeed maximize at extremes of the capital components. given those assumptions.320 However. EVA will optimize at an all equity value. we type in J7 STEP 5 We set two constraints. optimization will be in the other direction . 500 in this example. For example. STEP 4. Click on the cell with the EVA calculation in it. L7. any number between 1 and the amount of total capital. we will be able to change any blank input cell (without a calculation in it) that is part of the economic profit function in order to optimize it. This extreme corner solution may not be practical or obtainable but it shows in which direction capital components must move in order to improve economic profit . Step 2. We do this because the nature of beta is to divide debt by equity and we need to have a non zero function.a capital structure composed of all debt. We add that cell J7 is greater than or equal to (>=) the number. The procedure is as follows: STEP 1. Where the options give “minimize”. In the second constraint. TRIGGER POINTS In a linear function like the capital dynamic. In essence. In the box that states “by changing”. or “set equal to”. Go to the optimization section (section 2) and input in cell J7. Click “Solve” Unless the combination of capital cost components is very unique. click “maximize”. It is helpful for a firm to know what these “trigger points” are so that they can . but at some lower rate.321 modules. STEP 3. That is the function to be optimized. we set long-term debt. the function will maximize when equity is either “1” or at the capital limit (in this case 1000). if the interest rate is high enough. and it behooves the student/investor to at least have a working familiarity with these. STEP 6. “maximize”.with existing parameters. cell E9 to any number greater to or equal to zero (>=) because we do not want to optimize with negative numbers which would be impossible to obtain. there may be some combination of cost components that totally neutralizes the effect of changing capital proportions. “1”. Engage Solver in the tools section. Say. STEP 3. and the interest rate that a firm pays on its long-term debt. STEP 4. a market that lacks equilibrium can temporarily skew these cost variables and mislead a firm into making bad judgments. and enter a zero. This cell subtracts the actual EVA from the optimized version. the tax rate becomes significant not just for the deductibility of interest expense. In the “by changing cells” part. The following brief procedure takes us through a determination: STEP 1 As stated. Click on cell N7. Risk-free Rate. . attempting to change EVA by changing capital proportions may be productive at some times and not at others. “EVA Difference”. an herein lies its greatest liability. change them to that configuration now.322 set priorities. It is helpful to know what this rate is. Otherwise leave them as they are: J7 >=1. SETTING THE TRIGGER POINT MODULE In the optimization section. wealthier companies can incur debt and improve EVA through leverage alone. the student/investor will find that the relationship between the “risk premium” in the capital asset pricing model. and E9>=0. those elements are: Tax Rate. for example. In effect. Click the “set equal to” button. Without a risk of default. If the constraints are not set as they were in the optimization module. When interest rates are especially low. Market Rate BETA or Interest Rate STEP 6. We then engage Solver to set the optimized difference to zero by changing one of the cost of capital components. In order. is perhaps the most important determinant of EVA. enter a “1” or the capital limit of H4 in cell J7 STEP 2. STEP 5. Engage Solver in the tools section of Excel. J7. Moreover. The cell needs to be a blank entry (no calculations) and should be a cost of capital variable. pick any cell for which you wish to find a trigger point. we input either a “1” or the capital limit (in this case 1000) in the “Equity” cell. but because it affects beta calculations and the cost of equity. EVA maximization becomes an exercise in combining cost components. EVA will optimize at one hundred percent debt. the emphasis should be on income management and not capital proportions . through the management of retained earnings alone. THE EARNINGS SOLUTION While managing capital proportions can have a profound effect on EVA. if a firm is near this combination of cost elements. “Interest Rate” as your trigger point. a firm can gain a competitive advantage.23 %. an indifference curve forms and the cost of making changes based on capital proportion would be less than the benefits. Although the cost of retained earnings can escalate and diminish EVA.23%.at least in the domain of EVA improvement. More net income can also give impetus to a well-managed share issue which will usually move a company away from its target structure. In fact.22 %.323 STEP 7. use cell D4. If we change the interest rate to 7. Thus. Thus. flotation costs are incurred and the risk of owning the stock goes up. what might seem to be a healthy increase in earnings may not be adequate to eclipse the potential increase in equity. At this juncture. is usually wealthy enough to have a compensating dividend policy or a share buyback program. Click “SOLVE” As an example. a firm who is worrying about “too much retained earnings”. Whenever a firm expands through share issues rather than retained earnings. Keeping all other cost of capital components constant. and economic profit declines. More earnings can buffer the dilution effects (both in EPS and share price) of a stock issue while attracting . if we change the interest rate to 7. that cell will change to 7. With the suggested data still in place. it is not nearly as correlated to stock market increases as earnings. few firms will be materially hurt by funding in that manner. operating income fuels the profitability of equity financing. However. The culprit in a scenario of costly retained earnings is always the percentage increase in net income which needs to be high enough to justify retention. the weighted average cost of capital rises. EVA will optimize at one hundred percent equity. Enter that cell in the “By changing” section. Firms fund in a diversified manner to lower the risk of dependence. That type of decision . EVA can still increase despite the non-optimality of capital proportions. “How much of both?”. there is less demand for an equity issue when earnings are depleted. it is improbable that such an increase would be relatively large.represents a temporary risk that shifts a firm away from its target capital structure in order to gain more potential return in later periods. but the question of. the stockholders’ equity side of EVA may rise even as net income increases are minimal during a debt issue.. as we lowered one component. In effect. From the net income side. The net effect will be a smaller EVA and not a “managed substitute” of debt for equity. Thus. OPTIMIZATION AND CORRELATION In our optimization model. there was a perfect substitution of debt for equity. During the issue. and once a “trigger point’ was engaged. and since more debt raises the cost of equity. While it is certainly mathematically possible to increase both the proportion of debt and EVA simultaneously. issuing debt when earnings are diminished.324 buyers and creating demand at a high market price. just as they may engage several vendors for the same part. additionally. interest payments may be greater and most companies want a net tax advantage over taxes paid. net income was reduced faster than equity would decrease. we could increase EVA during a debt issue because the reduction in stockholders’ equity was greater than the reduction in net income. a firm has a choice between varieties of other sources: the choice to finance with new stock seems cost effective if the price is high enough because more capital is raised with less shares. the choice is rarely one source of capital over another.to raise a large amount of capital . In this case. Anytime retained earnings are inadequate to cover capital requirements. it will be more expensive as well. However. more capital would be raised than warranted which would increase its cost. RAISING CAPITAL EFFECTIVELY . This scenario occurred at lower interest rates. we raised the other. Equity will be increased during most debt issues. much of the damage occurred by raising too much capital at the wrong time . the entire cost of bankruptcy diminishes because asset prices get depleted during a downturn: when the prospects of covering interest payments improve. there are simply greater tax advantages at a lower probability of default. The promise of a “New America” through technology almost mimicked the “Better Living through Chemistry” axiom of the 1960s when plastics were revolutionizing the product industry. For those who can remember the “Tech Bubble” of the late 1990s.at the end of a business cycle. In fact. This is the scenario in which many firms find themselves early on in a recovery: “the Fed” lowers rates to the point where firms cannot afford to eschew debt. . but firms who had little stable earnings history were issuing many shares of equity to fund whatever Internet-related project they had dreamed up. Thus. we can lower the interest rate enough to raise both EVA and increase debt to equity proportionally. raising more capital is encouraged because it can be done inexpensively and will set the firm up for a competitive expansion with other firms. During a market top. the hallmark will be high market values accompanied by a low probability of default. Over production in fiber optics required huge amounts of fixed assets. However.325 In sensitivity analysis. Something had to give. and yet the world was subtly shifting to wireless access. raising too much capital implied the use of a non-optimal source like convertible bonds or a stock issue. the addition of debt may move a company toward a more optimal structure simply by restoring growth to assets. however. not only were the costs of both debt and equity accelerating upwards by late 1998. becomes much less of a problem. The outlook for startups and venture capital seemed positive with no end in sight. In an environment of low capital costs. it behooves any firm to keep the WACC to a minimum because more capital can be raised with such a combination. More capital can be raised at a lower cost which increases the optimality for those who fund with debt. The previous example of too much capital causing a shift away from the optimal target. if capital remains the same. . any increase in earnings will have a positive effect on both EVA and stock price. hypothetical underpinnings notwithstanding. in a microcosm. STOCK PRICE. then EVA will decrease. THE CONNECTION BEWTEEN. then the percentage formula for EVA evinces the connection between earnings and capital: (NOPAT / Capital . The reason was a simple neglect of the principles of the cost of capital. there is both a cell for “Stock Price” and “EVA Stock Price”. In the case of “Stock Price” it is earnings per share (EPS) divided by the cost of equity. the reason was not a lack of innovation or vision. Investors were pumping money into companies with a diminished EVA. expecting excessive risk to produce an excessive return where none was warranted. AND EVA In the computer program. and the cost of capital declines. Therefore. the similarities are great enough to equate the entities. while “EVA Stock Price” is EVA divided by the weighted average cost of capital (WACC). It was not because we were under producing compared to the rest of the world. While NOPAT / Capital is not quite the same as “return on capital” or “ROC” which is Net Income / Capital. CAPITAL. These values have some theoretical validity in a “no-growth” situation where all earnings are paid out as dividends. which are artificial constructs that divide an earnings component by some element of the cost of capital. they each show earnings acceleration compared to acceleration in the cost of capital.326 When the market collapsed in the early years of the new millennium. When earnings (NOPAT) declines and capital stays the same. both EPS / Cost of Equity and EVA / WACC will also be greater. However. In essence. It was not because a few speculators were pumping up the market with excessive optimism and a quick” pull-out”.WACC) x Capital = EVA. If the student/investor can observe the high correlation between EVA and stock price. these are the main determinants of ascending stock prices. 5 25 37.5 25 17.327 SENSITIVITY ANALYSIS: THE EFFECT OF CHANGES IN OPERATING INCOME AND CAPITAL Increased operating income is not mathematically linked to equity issues or increases in equity. However. but there is no direct link between operating income increases and a rise in equity. Net income has a specific relationship to equity because interest is deducted when debt is incurred. EVA. EVA Interest Expense ROE ROC Stock Value EVA Stock Value Long-term Debt / Cap.375 87.75 % For this experiment.75 % 43. a strong correlation exists between the two because the conditions that are conducive to their growth interact concurrently: when operating income is up.13 50% 6.75 35. ROE and hypothetical stock price.5 % 8. and we are going to both add and then subtract one unit to observe how the other variables react. WACC AMOUNT 500 4. A change in operating income changes those variables more than any others. Using the sample data. Table 12-5 INPUTED VARIABLES Equity Tax Rate Interest Rate Total Capital shares Operating Income Risk-free Rate BETA Market Rate AMOUNT 500 30% 5% 1000 20 150 4% 1 10% DERIVED VARIABLES Long-term debt EPS Net Income Book Value / sh. and earnings get retained. . debt tends to get paid off. we can readily observe the effect of one more dollar of operating income on EPS. operating income is at 150. 328 Table 12-6 CHANGE OP INC.63 % 44.083 Table 12-7 CHANGE OP INC. The reason? There are no forward. . both debt and equity are 501.065 17. in the economic profit laboratory it is not.looking measures of prospective success and only current performance is gauged.4 While the amount of change in capital is also positively correlated with stock price changes in the real world. the economic profit laboratory sees it as “dead weight”. both are 499. TO 149 Variable EPS EVA ROE Hypothetical Stock Price Direction of Change DOWN DOWN DOWN DOWN Amount 4. the proportion of debt to equity is preserved by adding and then subtracting two units of capital.36 % 43. and in the decrease. TO 151 Variable EPS EVA ROE Hypothetical Stock Price Direction of Change UP UP UP UP Amount 4. In the following tables. neither producing more income nor minimizing the cost of capital.34 36.8 17. If a capital inflow does not immediately translate to more profit.4083 38. Thus in the additional capital example. earnings anomalies. While such efficiency evokes the process of optimization. and is correlated with a higher stock price. . and an absence of foresight can skew results. and that gains in EVA are heavily correlated with movement toward an optimal target.3768 37. the additional unit of equity eclipsed the tax benefits of deductible interest forcing EVA downward. However.3733 37. the parameters that govern EVA are very wide. The additional unit of debt created more interest expense and less net income. false readings of EVA are the exception rather than the rule.46 % 43.54 % 43. a movement toward non-optimality was enacted. it is not too radical to state that economic profit and stock price go hand and hand.635 17.329 Table 12-8 INCREASE CAPITAL BY 2 Variable EPS EVA ROE Hypothetical Stock price Direction of Change DOWN DOWN DOWN DOWN Amount 4.733 Table 12-9 DECREASE CAPITAL BY 2 Variable EPS EVA ROE Hypothetical Stock Price Direction of Change UP UP UP UP Amount 4. Although beta remains the same. earnings per share (EPS) decreased without any change in either the number of shares or operating income. lack of a default probability. Moreover. Creating greater returns through the efficient use of capital is not optimizing capital structure.768 During the increase in capital.365 17. we can not make predictions for next year off of this year’s EVA. Since the WACC is implicit in the capital dynamic. using a default probability like Altman’s Z Score in combination with EVA analysis would help corroborate it. But . and the WACC will rise when all financial logic calls for it to fall. risk management will adjust the rates of change in each variable by dampening its correlation value to make net income rise faster than either stockholders’ equity or the cost of equity. . When we mathematically oppose highly correlated variables we resolve the conflict by giving more weight to one than the other. but the risk premium will rise to counter that action: when the market as a whole improves. the precision of the measurement depends on how well the user calculates the cost of equity which is partially subjective.the investor needs to always be wary because both of these measurements are coincident indicators. However. more firms use equity financing despite its higher cost. The correlation value of economic profit with an optimal capital structure is primarily derived from its ability to counterpoise earnings with capital. Analogously.330 For those who crave certainty. Thus. and so corroboration in this regard is more dangerous. When the economy is not in equilibrium (there are inefficiencies ) risk may not be priced correctly. By forming a constraint between variables that would normally be correlated (net income and stockholders’ equity). both net income and the cost of equity rise concurrently because more earnings lead to a better market rate and a possible increase in interest rates. given earnings outlooks and normal increases in both the risk premium and corporate equity. Additionally. A less definitive combination would be to use the capital dynamic form of economic profit with the WACC. small changes in the WACC can largely be ignored. For example. economic profit has all the ingredients of an optimizing function. we can tell when the EVA components are at a point where improvement would be quite easy to achieve. an error in one will cause an error in the other. In effect. the beta component in the cost of equity will decline with an increasing proportion of equity to debt. the investor needs to anticipate this favorable . The mathematical model in this text may confirm real world data. and the investor will not beat the market and may even under perform it. it is not forward looking and does not consider new opportunities. Firms that are closer to the target but make favorable use of capital are generally more stable over the long run. investing in ex-post EVA increases is a strategy that is doomed to fail. When a firm is increasing its proportion of equity. By observing the averages over five years for the three leading companies in a sector. beta will decrease putting downward pressure on the cost of equity. only a small percentage of companies will produce large returns The “miracle” for which investors are looking is the firm with sustainable earnings that accelerate faster than the cost of capital. The economic profit model works best if equity is low and net income is about to increase substantially. or by coordinating stock-price peaks with capital structure for an individual firm. While some success may be generated from momentum. This may seem easier than presumed if the analyst is willing to do some research. • 2) Know the leverage position. but the student/investor should realize that it optimizes debt and equity in the domain of net income. • 1) Attempt to determine at least a “ball park” figure of what the optimal capital structure target should be. However.331 ESTABLISHING GUIDELINES Unless one is in the throes of a bull market. Therefore the first guideline is to have some estimate of what the optimal target structure should be. If the characteristics of the sector and firm are favorable. In that framework. EVA can increase for two to three years without the “hazard” of a buildup in stockholders’ equity. ultimately there will be as many transitions downward as upward. firms that travel the longest distance to reach an optimal target and yet are moving rapidly in that direction. the investor can achieve a workable “guesstimate”. at any given time. A bull market will produce about a twenty percent threshold of firms that will out perform their peers over a one to three year time frame. tend to have the highest risks and returns. • 3) Be aware of the business cycle. On the other hand. (Back to Table of Contents) .332 position by observing the leverage ratios and investing accordingly. will give the investor a perspective about the capacity for EVA improvement. Once the student/investor can spot a leverage position that might yield a favorable EVA. earnings estimates are probably available on a prospective investment target. and historic equity growth. Many insiders are required to take a stake in the company. check to see if company insiders are investing in it. • 5) Use executive trades as a signal. Firms with lower operating risk tend to do well in a downturn and early recovery. a firm who has built up debt is ready to start paying off its investment and attract equity interest. An entire chapter is devoted to this subject. Moreover. while riskier firms do well at a market top • 4) Use analyst’s forecasts to your advantage. Comparing those estimates with both existing equity. See the “Fundamentals” chapter for more information. Oftentimes. avoid judgments about individual firms. Simply observing the pattern of interest rate hikes or decreases and the market’s reaction to them will give some indication of where the economy is headed. however. Make the effort to research the leverage first. If there is a consensus about the future prospects for a sector. mistake blind optimism for an investment opportunity. There may be so many downgrades and upgrades on a firm that the evaluations become meaningless. and use the listed insiders’ trades as confirmation. by all means do some research on a few of the firms’ leverage positions. Do not. 0675 K(m) 10.00% # Shares 0 Interest Rate E EPS 4.778529412 EVA Stock 42.94118 Total Cap 1000 Beta Total Cap J Equity 500 Cost of Cap 67.75 EVA 45.00% # Shares 20 Interest Rate 5.75647059 Total Cost of Equity 24.0597 K(m) WACC #DIV/0! C Net Income 87.0875 EVA Stock 35.00% WACC 0.035 LtD/Cap #DIV/0! ROE #DIV/0! H Beta 1 Total Cap 1000 Beta 411.501 LTD 999 EPS Div0! LTD 0 F Book/Share 25 Cost of Debt 3.50% Book/Share 25 Cost of Debt 0.7214706 Equity Cost of Cap #DIV/0! K Cost of Equity (Rate) 0.75647059 Cost of Equity (Rate) 0 Total Cost of Equity 0 L EVA 37.50% Book/Share 25 Cost of Debt 3.5 Tax Rate 30.95 EBIT Interest Expense 0 I K(m) 10.00% M LtD/Cap 0.375 LTD 500 EPS 3.5 Equity 1 Cost of Cap 59.00% D/E 1 RF 4.035 Tax Rate 30.12880562 EVA Difference 7.333 APPENDIX EVA OPTIMIZATION WITH SAMPLE DATA A 1 ACTUAL 2 3 4 5 6 OPTIMIZE 7 8 9 10 11 COMPARE 12 13 14 G RF 4.1 Total Cost of Equity 50 Cost of Equity (Rate) 24.00% WACC 0.00% Net Income 70.72682084 ROC #DIV/0! EVA Stock #DIV/0! B EBIT 150 Interest Expense 25 EBIT 150 Interest Expense 49.5 ROE 0.724 EVA 0 Stock #DIV/0 ! .5 Stock 43.279 Stock 70.00% D/E 999 RF D/E #DIV/0 ! N ROC 0.00% Net Income 0 Tax Rate D # Shares 20 Interest Rate 5.175 LtD/Cap 0.999 ROE 70. 706 Stock 39.334 FINDING A TRIGGER POINT FOR THE INTEREST RATE ON DEBT A 1 ACTUAL 2 3 4 5 6 OPTIMIZE 7 8 9 10 11 COMPARE 12 13 14 G RF 4.706 Stock 34.853 EVA 29.23% # Shares 20 Interest Rate 7.23% # Shares 0 Interest Rate E EPS 3.294 Cost of Equity (Rate) 0 Equity Cost 0 EVA 29.1 F Book/Share 25 Cost of Debt 5.00% WACC 0.706 Tax Rate 30.0753 K(m) WACC #DIV/0! Tax Rate J Equity 500 Cost of Cap 75.626 ROC #DIV/0! EVA Stock #DIV/0! B EBIT 150 Interest Expense 36.294 Equity Cost of Cap #DIV/0! K Cost of Equity (Rate) Equity Cost 50 Cost of Equity (Rate) 0.5 ROE 0.079 EVA Stock 27.00% D/E 0 RF D/E #DIV/0! N ROC 0.294 Equity 1000 Cost of Cap 75.00% D/E 1 RF 4.1594 LtD/Cap 0 ROE 0.626 EVA Difference 0 EVA Stock 27.0753 K(m) 10.00% WACC 0.00% Net Income 105 Tax Rate 30.06% Book/Share 25 Cost of Debt 5.075294 Equity Cost 75.863 EVA 0 Stock #DIV/0! .134 EBIT 150 Interest Expense 0 EBIT C Net Income 79.9853 LTD 500 EPS 5.00% Net Income 0 D # Shares 20 Interest Rate 7.00% M LtD/Cap 0.105 LtD/Cap #DIV/0! ROE #DIV/0! Interest Expense 0 H Beta 1 Total Cap 1000 Beta 0.06% Book/Share 25 Cost of Debt 0.25 LTD 0 EPS #DIV/0 ! LTD 0 L 0.5882 Total Cap 1000 Beta Total Cap I K(m) 10. Many who hold power desire to create a system of asymmetrical information wherein specialized knowledge becomes a bridge to accomplishing a task. a third may merge with a competitor and become a completely different . the probability of default is the “missing link” between the research of the Miller/Modigliani team and the practical application of movement toward a target optimal structure. there may be many unique “costs of bankruptcy” that apply to each firm’s specific situation. all of the conclusions from their research were arrived at given the absence of bankruptcy. in an age that disseminates information very quickly. and compromising photos all harbor the correlation between specialized information and political power. The cost of bankruptcy is not a generic cost. power tends to rise and wane just as fast. nor is it a legally uniform term. In the framework of capital structure. “secret formulas”. the stakes in that occupation rose exponentially. In fact. However.335 13 THE MARGINAL BENEFITS EQUATION: AN EXPERIMENTAL MODEL One of the greatest conflicts in our time comes from the tension between obfuscation and transparency. Thus. and it was a simple “tweak” of their famous equation. When the statistician suddenly became infused with the power to decide who gets credit and who is refused. esoteric club memberships. it is used for political purposes. When the information is hoarded and isolated. patents. copyrights. is just one example of an “invisible hand” regulating most of our public affairs. V(L) = V(U) + TB that provided the impetus for optimization: we subtract the cost of bankruptcy. Therein lays the problem of optimization. because of the transformational quality of the information itself Default algorithms are powerful tools. The massive popularization of “credit scores” and the influence they entail. creating a certain position or status within a group. One company may sell all its assets and another may “reorganize”. Implicit in that interface is the utilization of assets: the rate and amount of generated income.the risk that the firm cannot make timely interest payments.the amount of loss is equal to: [(1 . because a maximum has been obtained Ideally. we create a functional model based on the Miller/Modigliani concept. assets. THE MODELED CONCEPT Under the premise that there exists some optimal proportion of debt to equity that maximizes the value of the firm. Tax advantages of debt are balanced with the product of the amount of loss and the probability of default. “the cost of bankruptcy”. When the change in bankruptcy costs begins to exceed the change in tax advantages. the amount of loss is constructed of the reciprocal of the ratio between tangible book value and market value multiplied by market value. we are tasked with creating a common cost of bankruptcy that attempts to reconcile two entities: the book value of the proportion of debt to equity. Furthermore. THE MARGINAL BENEFITS EQUATION . default performance. any additional value created by debt will be apparent when tax advantages exceed bankruptcy costs. and stock price. and the first derivative of the entire function will equal zero. we can interchangeably use per share values.336 company. For the ratio of tangible book value to market value. the change in tax advantages will equal the monetary change in the cost of bankruptcy.[( % Probability of Default)(Amount of Loss)]. and this is the limiting factor on additional debt.(Tangible Book Value / Market Value)) (Number of Shares Outstanding x Price per Share)]. In effect. an optimum is found. Consequently. and the market value of a firm’s stock. In capital structure analysis. which we have termed. we discover an interface that unites those values in the probability of default . the equation we want to maximize is: [(Tax Rate)(Long-term debt)] . Therefore. That is . In effect. this equation is telling us that the optimal amount of debt is based on a balance between tax advantages. and market value will be determined by the product in the last parentheses. Thus. because all unlevered firms would be valued less than their leveraged counterparts. bankruptcy costs represent a threshold amount that must be exceeded by possessing a higher earnings capability than a levered firm: what they lose in tax advantages.337 Once we accept the proposition that the incremental value of a firm is equal to the difference between bankruptcy costs and the product of the amount of bonds and the tax rate. is . the type of assets determines both the bankruptcy function and the amount of financial leverage available. By minimizing the number of common shares outstanding and yet generating a high level of sales. we assume that the unlevered firm has the same income generating capacity as the levered firm. In essence. sales may not be stable or assets may not offer good collateral value. it has some threshold amount and rises at a non-constant rate. Without such capacity. there would be no reason to be in any business that did not require leverage. When we subtract bankruptcy costs. they must make up in the potential for generating income. The incremental value of a firm who has no debt is less than zero. and are unique to that firm. because there will always be a minimum level of bankruptcy costs that would be derived from operational incapacity. these firms can keep bankruptcy costs to a minimum. we need to account for those firms who create value without debt. With the addition of bankruptcy costs. however. Those firms who produce a high output at a lower fixed cost will have a lower cost of bankruptcy and have the capability of incurring more debt. The cost of bankruptcy. however. the unlevered firm must generate a level of income that overcomes this functional disadvantage. and that the only difference between them is how they are funded. we assume that they rise with the level of debt. firms that have no debt will have a larger amount of potential loss to shareholders that will prohibit them from taking on leverage. Thus. Many companies lack the asset structure to support debt to the point where it yields a tax advantage. a plant asset has the same earnings potential whether financed by debt or equity. Hypothetically. the proposition changes because bankruptcy is not a linear function. To an unlevered firm. Therefore. especially concerning the fate of the common shareholder. not only will bankruptcy costs rise with the level of debt. Although a model can give an estimate of an optimal target. and the income generating potential thereof. In this context. The algorithm needs to be both prohibitive of excessive debt. has that capacity. a utility company may take on more debt at a higher default rate simply because the local government infuses it with cash. it works probabilistically. The relevancy of any credit algorithm changes periodically as the relationship between income and debt shifts in the greater economy. for viability. twice the debt will not have twice the risk. Banks and insurance companies. . Some may even be inoperable. and these are the inputs into the model DEFAULT PROBABILITY AND BANKRUPTCY Among the great number of both generic and commercial default algorithms available. have asset structures that are much different from a manufacturing company.338 more multidimensional. they will adjust for the type of asset. An adaptation of the Zmijewski algorithm. Not all companies will meet its constraints. a handful would fit the needs of capital structure theory. like sales or net income. for example. However. but flexible enough to input and solve for variables. defined by the components of the algorithm. but will take on any number that the bankruptcy algorithm gives it. “operating risk” is a confluence between tangible assets and the probability of default and operating leverage is implicitly defined. Likewise. These are anomalies that are outside the purview of the model and negate the possibility of inventing a “one size fits all” algorithm. It also needs to be checked against some other standard. any working capital structure model will always remain transient and experimental. there are commonalties among all bankruptcies that can be addressed. In this much more realistic. non-linear capacity. In the numerator. allowing for the optimization of both net income and debt when the net tax benefits of interest are divided by the net tax benefits of debt.[( % Probability of Default)(Amount of Loss)] / [(Tax Rate)(Long-term debt)] . Like most default algorithms. it is constructed from averages that will allow a company to be capitalized entirely by debt before it registers a “one hundred percent” chance of default. the algorithm is not robust enough throughout its entire range to optimize debt. Nevertheless.[( % Probability of Default)(Amount of Loss)]. it would not help meet the test of a stand alone model that maximizes when the first derivative is equal to zero. Given a specific level of capital. With the addition of the tax benefits of immediate interest payments. By itself. and attempt to maximize the function. * . we must modify the marginal benefits equation. without the addition of interest benefits. The final function looks like this: [(Tax Rate)(Interest Expense)] . it will yield an amount of debt that creates parity between tax advantages and the cost of bankruptcy. when the ratio Zmijewski’s original analysis of 840 bankrupt and solvent companies used probit analysis to form an algorithm.339 The Zmijewski algorithm* will allow for the solution of variables that help minimize it. In effect. a variable in the default algorithm. except that we replace long-term debt with interest expense in the first part of the expression. we input the full marginal benefits equation. allowing it to meet a comparative standard other than its own increasing value. it optimizes in the domain of net income. Thus. We then divide by the entire marginal benefits equation. when adapted for interest benefits. THE INTEREST BENEFITS MECHANISM To use the Zmijewski algorithm in capital structure optimization. Any increase in the ratio is interpreted as movement toward the optimal. and used in two counterpoised equations. This is a logiistic version of that research. Alternatively. we form a ratio between the marginal benefits of interest expense and the marginal benefits of long-term debt. The optimization is n the domain of net income. the equation then finds the level of debt that keeps default values low and income levels high. The tax advantages are greater when ROE is maximized. and the default rate is less when ROC is maximized. Together. Thus. With less interest payments for debt. when equity replaces debt. the two measurements are optimized by inputting the correct amount of long-term debt. when used in combination with interest benefits. the program will trade basis points between the two measurements because it attempts to create the greatest tax advantages in the domain of a lower default probability. In a year to year comparison. and B are the coefficients of the . Although the capital base remains stable. On the other hand. To achieve an optimum for debt. the optimum has been passed: the nature of the algorithm allows marginal benefits to increase past the optimum which is the reason it needs to be modified. X1 are the fundamental ratios. In the meantime. ROC declines when ROE is at a maximum. net income increases until it is maximized in a capital structure of all equity. However. greater interest payments deplete net income. When debt replaces equity. We algebraically eliminate the logarithm and solve for the probability.340 decreases while marginal benefits are increasing. ROC is decreased because the effect of more debt is to decrease net income. DEFAULT MECHANICS The Zmijewski model has a minimum number of variables (four). Ln [P1 / (1-P1)] = X1B where P1 is the probability of default. The product of parameters and fundamental ratios forms the logarithm of a probability of default. the opposite phenomenon occurs: ROC is maximized. and a “guesstimated” target can be obtained CHECKING RESULTS AGAINST A VIABLE STANDARD The gist of optimization occurs in the realm of balancing the return on equity (ROE) with the return on capital (ROC). equity decreases at a more rapid rate and the return on equity rises. the marginal benefits equation will be like the weighted average cost of capital (WACC): movement is a tentative indicator because other variables may interact with it and skew the results. However. the probability of default will be lowered in the domain of tax advantages. but the inherent flexibility of the model makes it effective. 00191 or 0. P1= 1 / 1 + [EXP((6.069 -1.069)(-1. CA / CL = 1. Shumway and Merton each explicitly inputted an assets variable.06)(-0. STRATEGIC IMPLICATIONS: FINANCIAL LEVERAGE .479 As an example.5. The following table contains a definition of the fundamental ratios and the coefficients of the algorithm.2577) = .5.341 algorithm. where the fundamental ratios are given negative signs. they are explicit risk indicators and especially work well in terms of measuring the magnitude of changes in overall financial position.479)(-1))] = 1 / 1 + EXP(6.07)) + ((-9.5)) + ((0. Table 13-1 Zmijewski Default NAME TL / TA CA / CL NI / TA Intercept FUNCTION Total Liabilities / Total Assets Current Assets / Current Liabilities Net Income / Total Assets NONE COEFFICIENT 6.5)) + ((-1. each of these is designed to detect major financial catastrophes. we can input typical ratios to show how the basic function works: TL / TA = 0. and now it is being multiplied by negative one (-1) to form a positive number. The intercept is also given a negative sign because it was originally multiplied by “1”.06 -9. a student/investor who researches generic algorithms will find that the Zmijewski probability shares some commonalties with others.07.19 % While we can question the accuracy of this determination. To obtain a probability.384)(-0. In fact.384 0. Ohlson. but fall short of predicting bankruptcy for a firm throughout its entire range of debt/asset combinations. an income variable and an existing debt variable into their distributions. NI / TA = 0. However. we turn the equation around and input P1 = 1 / 1 + EXP [-XB]. the tax benefits rise. the less effect on default. would raise the amount of loss. if more common shares are not issued. and may even increase the cost of capital but . there is less concern about generating income from assets because net income is implicit in equity growth from retained earnings. increasing the total incremental value of leverage. In this model. paying off debt and replacing it with equity is not conducive to asset growth. any company that takes on debt must monitor the default rate to see how the subtraction of interest payments affects net income. if market value grows without a “legitimate reason” . In effect. an increase in tangible assets will allow market value to grow and still maintain more marginal benefits. equity derived from retained earnings should decrease the probability of default.that is -without greater tax benefits or lower default probability. Secondly. . From the equity side. the quality of purchased assets becomes paramount because the infusion will initially cause an increase in the probability of default. affecting the amount of loss. Therefore.342 When debt is added to the capital structure. Moreover. the more rapidly a firm can turn a purchase into an income generating asset. and the better was the decision to use debt. Ultimately. the balance between market value and tangible assets may change. any stock issue would be registered as more shares outstanding which in isolation. a chain of events ensues that affects each component of the marginal benefits equation. and that is the primary signal to observe when the proportion of equity is increased. Naturally. However. Thirdly. it is observed to be a speculative run-up and is expected to fall precipitously.increasing the amount of both will often unite the twin objectives of growth and optimal proportion. Intangible assets are not necessarily less worthy than tangible assets but will be more difficult to collateralize and amortize. A second signal would be debt neutrality or even slight increases in long-term debt that would increase tax advantages. income may not be immediately generated from a purchase and the lag causes the probability of default to rise. and how the other ratios can be buffered to maintain the current level of solvency. First. there can be no tax advantages from debt.the same debacle that hit many NASDAQ stocks in the late 1990s. Consequently. Only by observing the effect of increased fixed costs on a firm’s breakeven point. What we have done is substituted asset class. making it an implicit factor. When operating leverage is both large and unstable. but they are experimental. attempting to . A lower default rate will imply the use of less debt and more retained earnings. and market value will be a function of more share issues and less price appreciation . unstable operating leverage can prosper at certain points in the business cycle. a capital structure model is incomplete. The problem with high operating leverage is that it wavers and may create an unstable default probability: no creditors want to take a chance on a company whose default rate is less than one percent in one year and ten percent in another. For example. Although firms with high. a large operating leverage will produce a larger increase in operating income. Intel and Microsoft have “softened “their approach by diversifying into related fields. the effect of operating leverage is paramount to income generation because it will determine the “base” from which financial leverage can operate. for the effect of operating leverage. The large “intercept” in the default algorithm accounts for the variance of input. Some of the bigger tech names. their stock will be speculative at best. Stable operating margins lead to the proper use of leverage. and default probability. However. can the proper amount of leverage be determined.343 STRATEGIC IMPLICATIONS: OPERATING RISK We assume that the parameters of this model are correct. which naturally diminishes the rate of default. without the inclusion of operating leverage. a firm who achieves this financial state with fewer shares outstanding will see their stock price grow. And especially . There may be better. A more precise bankruptcy cost can be developed. “stand alone” default algorithms that need no modification.a module that allows for the interaction of operating leverage can be inputted. Google. but also accounts for a large percentage of default probability. Cisco. In fact. While this model does not account for the “benefits of funding with retained earnings” in lieu of tax advantages. we can check through the model whether the decision is valid. Those firms who cannot take advantage of tax breaks must earn enough to cover a threshold amount of bankruptcy costs or be at a disadvantage to financially leveraged companies. we should look for a balancing increase in the amount of loss since it represents mostly market value. use the cost of bankruptcy side to measure performance by balancing the rate of default with the amount of loss. We can.344 maintain a high earnings capacity that is more stable. Figure 13-1 Default Probability Tax Benefit Amount of Loss Debt Equity Operating Leverage The student/investor must keep in mind that the model is a rendition of the incremental value of debt . SPREADSHEET CONSTANTS . however.and not the value of the firm itself. one of the provisions would be the generation of income above and beyond the tax benefits that are lost . but we cannot gauge the period to period performance of marginal benefits because they will be less than zero. If we see the rate of default go down.including the subtraction of interest payments and the advantages of a lower default rate. If a firm chooses not to fund with debt. Thirdly. Alternatively. i. “optimal” is only operative in relation to the constraints of the situation. the amounts that we input as capital are also actual amounts and so it is “realistic” to attempt to optimize at that level. The relationship between assets and capital is fixed. While all inputs can be changed. “Number of Shares Outstanding” can be changed but will not react as the other variables change because it is deemed to be outside the confines of the model. This ratio. some variables will remain constant simply because there is no realistic method of making them react. operating income responds to sales. For example. Most firms would desire to get more debt at the same rate. the term.e. we diminish year to year anomalies. Secondly. then we are optimizing debt at too large a level of capital and it will be incorrect. Similarly. but not realistic if a firm must go from a level of ten to thirty percent debt to equity. Since there are no production variables in the model. EBIT is a given. No additional amount of debt or equity should affect the quality and types of assets to purchase. carrying over to tangible assets and market capital as well. Additionally. variables that are determined to be realistically uncontrollable are made into constants. That is a realistic assumption if a firm is already near the optimal target. but the WACC is predicated on increasing rates for increasing risk. the inputted interest rate is the “effective rate” as applied to long-term debt. the amount of capital is and assumed to be the correct amount to raise. it is assumed that the interest rate will not change as new levels of debt are engaged. making calculations depend on an unusual amount of capital raised in any one year.345 The marginal benefits model is partly static and partly dynamic.. . By using long-term averages. Thus. if the amount that a firm raises is well beyond its requirements. the “Market Price” of the stock will not react to changes in the model because no known formula can determine the price change of a common stock. the variable. However. (Interest Expense / Long-term debt) is effective for the period and denies inconsistencies between interest expense and the actual rate. Analogously. which implies interaction with the greater economy. SPREADSHEET LOGIC . • 2. it does not change in response to changes in fundamentals that very well might dictate a new rate. • 4. Finally.The mid-range price during the period is inputted. While the average effective rate is used. It also implies that the amount of assets and the relationship between tangible and intangible assets is constant. Capital . The Number of Shares Outstanding . etc. divided by two.The interest rate becomes a function of the relationship between interest expense and long-term debt but remains unchanged in terms of the movement of other variables • 3.These variables depend on complex interactions outside of the model. The effective rate for a banner year (when the model increases the amount of net income) is the same for a dismal year of near negative income. The mid-range is merely the high price of the stock during the period. Tax Rate .The tax rate remains constant despite vicissitudes in net income.346 Fourthly. the tax rate is a constant. vendor credit. The interactions are too varied to input a model function that would be realistic.Capital is inputted as a given amount. Current Liabilities and Current Assets. much like the stock price. we make current liabilities and current assets constants because they react to myriad variables. The Market Price of the Stock . plus the low price.Inputted as the actual amount at the end of the measured period. The amount of capital acts as a base for the tradeoff between debt and equity. 7. • • 6. Interest Rate . There can be no deterministic function for either variable because they depend on such outliers as interest rate relationships. type of industry. • 5. Operating Income (EBIT) This value forms the base for net income before deductions of interest and taxes. The list of constants is: • 1. the benefits must be weighed against the potential for increasing bankruptcy costs. In the case of the one that is not (taxes paid). The variables that change when long-term debt is changed are called “dynamic” variables. the user can calculate the figure or create another cell that does that job. The change in debt causes a corresponding change in total liabilities and net income which affects the probability of default. we include WACC and beta . interest is deducted from operating income at an appropriate rate. While more debt will create more tax advantages. the program should register zero or negative interest benefits and indicate that a capital structure of all equity is most effective. we can observe how following that recommendation affects EVA and ROC. the default rate is not so stringent that interest benefits are denied. For example. one of the saving graces of the Zmijewski algorithm is that it seems to be “forgiving” enough to allow interest benefits in the first place. we see optimized debt in the context of more EVA and / or balancing ROE and ROC. The appropriate amount of debt will be the amount that maximizes the ratio of interest benefits to marginal benefits when both of those figures are greater than zero. The following are “dynamic” variables: • 1) Equity . However. we can view the recommendations in the domain of repercussions. Moreover. we have added many of the functions from EVA optimization that allows us to view the program in another context. a chain of events ensues that affects the other variables associated with marginal benefits.347 When long-term debt is increased. Most of these variables are explicitly set forth in the spreadsheet. For those firms with large intrinsic values well past the value of tangible assets. and can be realistically determined from either accounting functions or spreadsheet logic. because the program simply subtracts long-term debt from capital. DYNAMIC VARIABLES Once we input a greater or lesser amount of long-term debt. If the program recommends an all equity structure. In fact.It is possible to include preferred stock in this figure. in other words. MODEL SETUP The model is set up as a “stand alone” optimizer for the purposes of illustration. • 6) Tax Benefits on Debt . preferred stock will be an outlier and it is better not to include it. The only additional entries would be the three variables that comprise the CAPM. which are: the risk-free rate. The program is also set up for comparisons so that the module can be copied and configured for the input of a comparison firm. The list of .When interest is subtracted from operating income. the two models can be integrated if one makes the EVA variables dependent on the Zmijewski variables.The optimization substitutes debt for equity and adds to (subtracts from) total liabilities. An extra cell can be created for the calculation. EPS.The program multiplies the interest rate by long-term debt and then subtracts this amount from operating income. and forms part of the interest benefits mechanism.348 calculations to supplement the marginal benefits information. • 2) Interest Expense . • 3) Total Liabilities .This expression is interest expense multiplied by the tax rate. which are configured by the Zmijewski algorithm.As mentioned above. Together. • 4) Net Income . allowing the student / investor to observe the effect on the cost of capital. The base figure is current liabilities which remains constant. the provision for income taxes is implicit because the accounting total is not needed for marginal benefits. However. taxes are deducted and the expression becomes “net income”.This expression forms a major part of the marginal benefits equation and is tax rate multiplied by long-term debt. • 8) The Probability of Default . • 5) Taxes Paid . ROE etc. Such interdependence allows the user to test the optimum against EVA.This is an interaction between some of the other dynamic variables. • 7) Tax Benefits on Interest . the market rate and the beta of the company. the EVA sensitivity module in chapter twelve can give corroboration of the recommendations from marginal benefits. Assets 2. The effective tax rate. and the entire set-up of the spreadsheet are available in the appendix. The entry variables and decision variable are listed as follows: • A) Eleven variables located in vertical sequence in the “Zmijewski Variable” section. Current Liabilities 4. It is this cell which will yield the optimized value for debt and change the entered values. 10. Effective Interest Rate (Interest Expense / LTD) 6. EBIT (earnings before interest and taxes) 7. LTD (actual) 5. 11. thus allowing period to period input. eleven variables are entered in sequence in the “Zmijewski Variables” section. It is best to use a historical five year average in all of the entry variables. The amount of capital (with preferred stock deducted for better accuracy). 1. THE PROCESS: ENTRY VARIABLES There are eleven entry variables that are entered into the gray cells of the spreadsheet Once the entry variables are completed. Market Price for the stock. these variables will carry over to the “OPTIMIZATION” module and be transformed as different numbers are entered in the red “LTD” cell. the program can still give a fair “guesstimate” if operating income was typical for a particular year. Highlighted in red. Current Assets 3.349 formulas in the program. there is one (red highlighted) decision variable for long-term debt that is entered in the optimization section. which is an average of the range. Thus. The number of shares outstanding • B) The decision variable. . The Tangible Book value per share 9. below the top section . However. 8. there is one decision variable located in the blue optimization section. It proceeds by evaluating the probability of default with the existing amount of income and then makes a determination about tax benefits. In fact. the program does not uniformly try to lower the proportion of debt to equity in order to raise net income. marginal interest benefits / marginal debt benefits.350 THE PROCESS: OPTIMIZING WITH SOLVER Optimization occurs with a minimum of constraints. By changing cell B28. case three is the recommendation to use no debt in the capital structure because no marginal interest benefits are obtained. three separate cases arise: case one is where there is a recommendation for less debt. it is because there is no positive marginal interest benefit throughout the entire range. L16 which contains the optimized amount of marginal interest benefits: J16 which contains the optimized amount of marginal debt benefits. which naturally raises the return on capital and EPS while lowering beta. case two is where there is a recommendation for more debt which lowers EPS and ROC but raises beta and ROE. “Equity”. The key cells are: B28 which contains the decision variable “LTD” in red. 3. . THE RESULTS: THREE EXAMPLES While it may be next to impossible to prove and verify the recommendations as “optimal”. “no feasible solution” is registered in Solver. Maximize cell O16. The full ratio to be maximized. and either the probability of default is too great . is located in O16. If Solver has no feasible solution. In this last case. 2. B35 >= 1. We set the parameters as follows: 1. so that the change in beta can register properly (when used in conjunction with an EVA module). and B35. which must be set to a value greater than or equal to “1”. Therefore. or the amount of loss is great. the optimal amount of debt for that company is zero. Therefore. B28 >= 0. Many high beta stocks fall into that category. Subject to: L16 >= 0. the student/investor can detect changes in EPS or beta that reflect efficiency. or both. the limited return on capital of 10. Effective Tax Rate Capital Number of Shares a.1 % $1.18 0.89 In a model of this type.Equity % BETA Actual 2698 32.37 8192 764.73 10.69 % 1.99 Optimized 1774 21. The program registered a fairly high .37 7.99 12% Key Measurements Long-term Debt LTD / CAP Financial Leverage EVA ROC ROE EPS ROC .83 % 0.WACC % ROE .93 % 1.88 % 14. However.15 5. Beta c.5 % 0. / sh.27 percent (optimized) is much more sustainable at this level of debt. the “distance ratios” that measure the difference between ROC and the cost of capital for example.0541 1431 25. Table 13-2 LOWE'S 2000 (LOW) ENTRY VARIABLE Assets Current Assets Current Liabilities LTD Effective Interest Rate EBIT Market Price Tangible Book Val.27 % 13.74 % $1. On the other hand.10 0.351 CASE ONE: LESS DEBT RECOMMENDED * Note * in each case. while decreasing the return on equity (ROE).8 % 0.87 9. Risk-free Rate b.06 0.071 117. did not improve.75 % 2.65 % 1. decreasing debt will automatically increase EPS and the return on capital (ROC). the amount of capital is composed of long-term debt and common equity. Market Rate AMOUNT (Mil) 11376 4175 2929 2698 0.1136 153. 04 6217.14 28.08 27. Market Rate AMOUNT (Mil) 44007 12962 11544 4799 0.02 % 1.73 % 0.24 34. Effective Tax Rate Capital Number of Shares a.37 % 44. It raised beta and ROE at the expense of the return on capital (ROC) and earnings per share (EPS).93 % 45.47 In case two. Risk-free Rate b.02 % 0.72 % 1. the program saw that Merck had phenomenal earning power and a low probability of default.37 % $3.29 20849 2319. Beta c.0656 6174.2 % 0.33 76.42 Optimized 6823 32.WACC % ROE .352 probability of default at the thirty-two percent debt level and pared it down by limiting the amount of debt to approximately twenty-one percent.42 10% Key Measurements Long-term Debt LTD / CAP Financial Leverage EVA ROC ROE EPS ROC . The emphasis was obviously on increasing tax .93 % $3.025 3.1 4.0968 10721. / sh.77 0. A higher operating income would allow a higher debt level because the probability of default would be lower.Equity % BETA Actual 4799 23.26 % 50.95 34.23 % 38. CASE TWO: MORE DEBT RECOMMENDED Table 13-3 MERCK 2001 (MRK) ENTRY VARIABLE Assets Current Assets Current Liabilities LTD Effective Interest Rate EBIT Market Price Tangible Book Val. CASE THREE: THE RECOMMENDATION FOR NO DEBT Table 13-4 CITRIX 1999 (CTXS) ENTRY VARIABLE Assets Current Assets Current Liabilities LTD Effective Interest Rate EBIT Market Price Tangible Book Val. Effective Tax Rate Capital Number of Shares a.WACC % ROE . Throughout the nineties.Equity % BETA Actual 314 37. this company eschewed debt but decided to take on three hundred and fourteen million in long-term debt .8 14.4 % 1. Market Rate AMOUNT (Mil) 1038 570 137 314 0.9 % 1.11 Citrix Systems is a very representative case.92 0. / sh.42 13. However. We assume that this type of volatility is implicit in the relationship between earnings.52 16% Key Measurements Long-term Debt LTD / CAP Financial Leverage EVA ROC ROE EPS ROC .32 % $0.1 % 1.125 2.38 % 21.0407 196. Beta c.58 -1.32 % 14.353 benefits.22 % -0.38 847 195 5% 1.62 -2. Risk-free Rate b.27 % $0.9 % -2. but other intangible measurements may capture this risk better.52 Optimized 0 0% 1 -24. analysts know that drug companies can fall precipitously because of lawsuits and expired patents.6 59.0699 -2. assets and share price. the model often yields a realistic rendition. In optimization. however. Ultimately. an over heated market had pushed up the risk of all equity. (Back to Table of Contents) . and that a certain consistency in operating income must be arrived at.354 in 1999. This program does not find a feasible solution with Solver because bankruptcy costs are greater than the marginal tax benefits of interest. if optimization tells the user to retain more earnings and shift to a lower debt to equity ratio (Lowe’s) then it must be assumed that such changes cannot occur “over night”. EVA DISCREPANCIES The student/investor will notice that EVA plunged in each case of “optimization”. which is reflected by a high beta. or those same proportions during the timing of a stock peak during a business cycle. The recommendation of zero debt brings beta down to a far more reasonable 1. The addition of debt into its capital structure added to more equity risk which made it excessively difficult to achieve a positive EVA even in a banner year like 1999. It is fallacious to argue that this model will set standards of optimization for every company. For example. making them negative. By this time. Like many “tech” stocks its operating leverage is fundamentally high. long-term changes are combined with current EVA information that will be “out of sync” with the periodic improvements that are necessary. We know from previous chapters that EVA optimizes at extremes of equity or debt depending on the relationship between their respective costs. it balances key variables and offers management several ancillary targets. However. if the reader will observe the average proportions of debt to equity in an industry.11. no recommendation for more debt in the capital structure can be undertaken during a “credit crunch” when interest rates are especially high. Analogously. we assume that the weighted average cost of capital will reflect risk over the long-run. even in the cases that it does not optimize. Reconciling a long-term optimization with volatile changes in the cost of capital is tenuous. 479 -6. ROE #DIV/0! LTD/CAP #DIV/0! LEV #DIV/0! ROC #DIV/0! J 1 amount of loss 2 3 DIV/0! 4 5 6 7 8 9 Marginal Benefits 10 #DIV/0! L M O cost bank default intermediary exp final #DIV/0! #DIV/0! #DIV/0! #DIV/0! Interest Benefit #DIV/0! INT / Marg. 5 LTD 6 interest rate 7 interest expense 8 EBIT 9 EBT 10 net income 11 total liabilities 12 equity 13 mkt price 14 book value 15 market value 16 tangible price 17 tax rate 18 capital 19 # shares B D variable intercept TL/TA CA/CL NI/TA 0 0 0 0 0 #DIV/0! 0 Actual ROE #DIV/0! LTD/CAP #DIV/0! LEV.355 APPENDIX: SPREADSHEET FORMULAS AND ZMIJEWSKI OPTIMIZATION A 1 Zmijewski var. #DIV/0! ROC #DIV/0! E negative coefficient 9. 2 assets 3 current assets 4 current liab. #DIV/0! 15 Marginal Benefits 16 #DIV/0! Interest Benefit #DIV/0! MAX #DIV/0! .06 G variable LTD tax benefit H amount 0 0 OPTO.384 -0.069 1. 069 1.384 -0.356 A B D E F G H 23 OPTIMIZE 24 Zmijewski var. 28 LTD 29 interest rate 30 interest expense 31 EBIT 32 EBT 33 net income 34 total liabilities 35 equity 36 mkt price 37 book value 38 market value 39tangible price 40 tax rate 41 capital 42 # shares variable 0 0 0 0 0 0 0 0 0 0 0 #DIV/0! 0 0 0 0 0 Marginal intercept TL/TA CA/CL NI/TA negative coefficient 9.479 -6. 25 assets 26current assets 27current liab.06 variable amount LTD tax benefit 0 0 Tax Benefit 0 Bank Cost #DIV/0! EQUALIZE #DIV/0! J L M O 24amount of loss 25 26 #DIV/0! 27 28 29 default intermediary exp final #DIV/0! #DIV/0! #DIV/0! cost bank #DIV/0! . =B17*B5 M4. =1/(1+M27) B29. =M5*J3 H4. =B28 J26. =H4-O3 L10. =H27-O26 L16. =EXP(M26) M28. =B33/B41 J16. =B8 . =(B7*B17) . =B3 H26. =B4+B5 B12. =B13*B19 D16. =B28*B29 B31.357 FORMULAS FROM RANGE A1:O42 H3. =B10/B18 J10.O3 O10. =B5*B6 B9. =B28/B41 F16. =B5/B18 F10. =B12/B19 B15. =B5 J3. =L16/J16 B25. =E3+(E4*(B11/B2))+(E5*(B3/B4))+(E6*(B10/B2)) O3. =(B30*B40)-O26 O16. =B40*B28 M27. =B31/(B31-B30) G16. =B6 B30. =(1-(B16/B13))*B15 M3. =B8-B7 B10. =L10/J10 B11. =M28*J26 B27. =B33/B35 E16. =1/(1+M4) B7. =B9-(B17*B9) D10. =(1-(B39/B36))*B38 M26. =E26+(E27*(B34/B25))+(E28*(B26/B27))+(E29*(B33/B25)) O26. =B8/(B8-B7) G10. =B2 B26. =B10/B12 E10. =B18-B5 B14. =B4 H27. =EXP(M3) M5. B40. F42. B39. =B31-B30 =B32-(B40*B32) =B27+B28 =B41-B28 =B13 =B35/B42 =B13*B42 =B16 =B17 =B18 =B19 =H27-H4 =O26-O3 =E42-F42 (Back to Table of Contents) . B33. B35. B37. E42. B42. B41.358 B32. G42. B34. B38. B36. Outside of the amount of leverage. but uses the “expected rate of return” from a valuation model as its cost of equity. One merely compared the income that a firm was actually generating to what similarly risky firms were generating.it was decided that ROE was not a true opportunity .359 14 AN INTRODUCTION TO RESIDUAL ECONOMIC PROFIT THEORY: USING A CONSTANT DIVIDEND DISCOUNT MODEL AN INTRODUCTION TO RESIDUAL ECONOMIC PROFIT THEORY The opportunity cost of an action is the benefit lost by not choosing the best alternative. Secondly. few of the variables were controllable by the firm. analysts would equate the cost of equity with the return on equity (ROE) and use them interchangeably. It allows the user to control a level of current dividend. Thus. But . new share issues or the amount of earnings to retain. Years ago. given a level of net income. the “cost of equity” was very market dependent. the firm can optimize its dividend policy and attempt to improve its residual economic profit for the year. it can decide whether the growth rate for dividends is optimal or needs to be changed. new stock issues. has the same foundation and framework as EVA. One of its strategic advantages is to find a level of net income where a stock issue would not undermine capital structure. The more pragmatic “residual” economic profit. the opportunity cost was the link between the comparisons. and the amount of retained earnings in the realm of net income increases. and indeed. neither technique is sufficient for “hands on” determination of dividend policy. However. That standard concept from economics became the foundation for capital evaluation techniques like EVA and the capital dynamic. it can create an optimal amount of retention that will maximize economic profit. Thirdly. next dividend. the economic profit in that situation would be zero. The user must accept that the growth factor in the Gordon model is composed of the product of ROE and the retention ratio. dividends. • 4. • 2. and that it will not change in response to any changes made to the fundamentals. and dividend theory OPPORTUNITY COST . Although it would lack the performance accuracy of EVA or the capital dynamic. it was more amenable to corporate control. yielded an expected rate of return that could be compared with ROE to gauge a firm’s performance. • 3. although in reality. The assumptions in the model are numerous. • 1.360 cost because. The user must accept that the stock price is determined outside of the model. To be a true opportunity cost. It is a constant in the model. it may change daily. the only market driven variable was the price of the stock. can be used as a near-term proxy for the cost of equity. by the market. Since this “expected rate of return’ was driven by internal fundamentals and did not have as many market variables as the CAPM oriented “required rate of return” had. it could be used to fine tune capital structure. When a firm maximized this distance. and even dividend growth would be derived from the model. debt. Valuation models. the cost of equity had to be some level below the return on equity. which is only theoretically applicable to firms who pay a constantly growing dividend. The three pillars of residual economic profit theory are. Non dividend paying stocks cannot be inputted. As we have seen.[(ROE) x (Stockholders’ Equity)] = 0. especially the Gordon model. Net Income . opportunity cost. it was functioning above the level of its peers. Firm’s who retain one hundred percent of their earnings would have both a growth rate and a cost of equity equal to its ROE. The user must accept that a model like the Gordon model. The amount of retained earnings. dividend discount valuation. the residual economic profit cites the “opportunity” of either reinvesting net income. In effect. They add a time element that places a current-dollar figure on anticipated earnings. it lacks the far-reaching market orientation that the CAPM generated rate possesses. or paying it out in the form of dividends. simply because the outlook for new projects is so poor. At that point. While the required rate of return used companies of similar risk to compare generated income. the analyst must match these with a suitable cost of . valuation models can be very accurate. but on the other hand. When compared to the market value of a security. if the economy is near recession. companies of similar risk would have similar returns. Like most analytical tools. Fundamentally. residual economic profit theory dictates that the choices made in reinvestment or distribution are derived from balancing the needs of the company with the exigencies of the market. a firm might decide to distribute earnings as dividends rather than to retain them. this value will help the analyst determine whether the stock is under or over priced. their projections may be so off base that the method loses credibility. the optimal amount of funding is implicit in the model.361 The opportunity cost that is conveyed by the residual economic profit model is much different than the required rate of return. it is as much a reinvestment rate as it is an opportunity cost. For example. The result is a figure that is referred to as “fair” value or “intrinsic” value. The expected rate of return will approach the true cost of equity only when the market is in equilibrium. It has a small market interface when it uses the expected dividend yield because the current stock price is the denominator of that component. However. VALUATION MODELS Valuation models offer an opportunity to gauge market response to earnings fundamentals by balancing future growth with the cost of capital. While some firms have very predictable dividends. Thus. there would be no difference between “required” and “expected“ rates of return. the stock is over valued. but the growth rate of dividends will often mirror the growth rate of the company. almost like economic profit. most companies want to ensure payment. the word “anticipate” connotes the discounting process of evaluating future income in terms of present dollars which involves “exponentiating” the quotient of cash-flow (dividends) to the cost of capital . they should be paid out as dividends. it is the amount and growth of these dividends that determines how much an investor is willing to pay for a share of stock. shareholders will see the action as a sell signal. If he or she (hopefully) pays less. two additional factors need to be considered: since a stock is owned in perpetuity until it is sold. Not only are stocks evaluated by their potential streams of dividends. and any decrease sends an extremely negative signal to investors. a stock is rightfully valued by the production of income compared to the cost to produce it.for each year the stock remains active. is founded on theoretical absolutes without recourse to periodic disparities. Since predicting such rates for even one year is a gargantuan task. forecasting in this manner is open to numerous errors. the stock is under valued. cutting the dividend is perhaps the single most extreme symbol of managerial defeat outside of declaring chapter 11. and some analysts simply revert to capitalizing operating income to determine a fair value. The validity of valuation models. DIVIDEND THEORY Dividends are sacrosanct. Thus. the acceleration in dividend growth will be less than the acceleration in earnings. its only source of return during that interim is the dividend. Thus. Moreover.362 capital over a long interim. Even when warranted to save a company. For this reason. On the other . The theory behind valuation is sound. If the investor pays more than the anticipated flow of dividends. Earnings belong to the shareholders. In fact. Unless they are reinvested and anticipated to return a higher rate than shareholders can receive on investments of similar risk. it measures growth over long time spans. Secondly. and it can be sold at a profit at a future date. however. alternative theories have often crept into the academic literature. As long as the amount of equity is optimal. which leaves 15 million for dividend distribution. which might be devastating . any new stock issue represents a future obligation to pay dividends as well. fixed percentages of the profits. the firm cannot . However. While that concept opposes the rationality of an opportunity cost.363 hand.the residual theory of dividends. and the best companies have a solid track record of steadily increasing dividends Most companies peg dividend growth at a rate very close to earnings because the common stockholder needs to be compensated from the profits of the firm.depending on contractual demands and the historic pattern of payments. the firm could renege on dividend growth. then (65 / 80) or 81. dividend growth is cited as a symbol of managerial confidence. advocating different payment rates. if the capital budget were to grow suddenly (perhaps because of greater opportunities). However.25 % will be retained earnings. In effect. 4) Pay dividends only when retained earnings are not fully exhausted For example. and even a concept tailored to stock price optimization . funding with retained earnings will not incur flotation costs nor will it potentially dilute the market value with new issues. a firm pays out dividends even when the reinvestment rate may be higher on the issue of new stock. such a policy does indeed minimize the cost of capital. 3) Finance this amount of equity with retained earnings to the greatest extent possible. The residual theory of dividends states four principles behind creating a coherent dividend policy: • • • • 1) Determine the optimal capital budget 2) Determine the amount of equity needed to finance this budget. then 65 million of the 100 million will be in equity. it is a realistic affirmation of the psychological milieu. if Company X needs 100 million in funding (the capital budget is 100 million) and their target capital structure calls for 65 percent equity and 35 percent debt. If 80 million in earnings is available (net income is 80 million) to meet that budget. From a mathematical perspective. the loss in accuracy will not be as profound as the utility in determining proper proportions. In fact. With five year smoothing. while newer. it is far more volatile than the “required rate of return”. albeit one that is much more dependent on the fundamentals of the company than market averages. a more reliable figure emerges. is that it provides a short-term answer. net income will not always be sufficient to provide retained earnings. even when it is not applicable to the actual dividend distribution. it has the same framework as the other economic profit models. or will interest rates and credit availability be favorable enough to incur debt. Shareholders will continue to demand higher dividends despite an uncharacteristically bad year. In reality. the dividend distribution implication may even be realistic given the volatility in some markets. Older. Since the economic profit in the current term is being compared with adjacent years. dividend policy is an unsolved conundrum in the business world. without any smoothing from a five year moving average. but consistency is implicit in the model. The analyst can use this dependence on internal fundamentals to gauge the . RESIDUAL ECONOMIC PROFIT The residual economic profit model applies the expected rate of return to stockholders’ equity and then subtracts it from net income. The value of using the Gordon model. In this manner. no “one size fits all” policy exists since each entity is structured differently. “start ups” may establish a policy that is tailored to their earnings flow that might include only “special dividends” when they are periodically declared. the user can declare any dividend for this year or next year. a direct rendition of current growth (ROE multiplied by retention). will give the expected rate of return a current bias. established companies cater to pension funds and retirees and rarely cut their dividends. However.364 assume that shareholders desire to incur the risk of postponing immediate consumption for future potential returns. The residual economic profit model allows flexibility in this area. As most financial observers will note. The choice between special dividend and buyback is significant. it appears that any level of net income will cause enough earnings to be retained such that economic profit declines. Indeed. Either the company can pay a special dividend that depletes equity by an adequate amount. a firm will attempt to accelerate earnings at a faster pace than dividends. originally implemented at about the same time the tax law changed. Microsoft was retaining earnings and building up equity at too rapid a pace. the market value of the company is immediately depleted by the amount of the dividend pay out. Without a hugely active market for new issues. the total cost of equity becomes too expensive . Most companies choose to do the latter. The company has fallen into the “dividend trap”. From a model perspective. a firm will find that stockholders’ equity has built up to an untenable level: retained earnings keep growing in perpetuity. After several successful years of this policy. put the firm into an entirely different category. the only difference between ROE and the expected rate will be how a company retains earnings and distributes dividends. Some companies. Almost “out of no where”.not only because earnings have accelerated the percentage cost of equity. Both techniques will require massive amounts of cash. . When Microsoft paid a special dividend. such a policy also raises capital in the form of retained earnings. On the other hand. there are only two ways out of “the trap”. but because stockholders’ equity is too high. like the one in the following examples. growing dividend. a buyback will raise market value by taking shares off the market but creates a glut of treasury stock and makes it difficult to issue new shares when the time is advantageous. circa 1997. and buy backs have become standard practice. Most companies end up “taking it on the chin” and have a mediocre year. analysts had no idea that the decision to pay any dividend at all. or it can buy back shares of stock. while net income is tabulated on a yearly schedule.365 effects of dividend policy and share buybacks. THE DIVIDEND TRAP To ensure a stable. great earnings notwithstanding. In effect. while residual economic profit displays the internal dynamics of the company. ROE and retention. it too is a constant. While the market price of the stock is a very small part of the model. Nevertheless. MODEL OPTIMIZATION Mathematically. EVA is in tune with what the market demands. Therefore. this model will optimize when the pay out to the shareholder is one hundred percent and no earnings are retained. the absolute factors of growth. the “dividend trap” is a fundamental paradox. EVA is a much better investment tool even though it lacks the operational capacity of residual economic profit. are subject to debate. the firm will suffer the consequences of a lower economic profit. a series of improvements. with this definition of “residual economic profit”. the definition of growth factors. Notwithstanding those discrepancies. and subject to debate. which if violated. A firm is in business to generate income. it is hard to imagine an improvement in this version of economic profit and not observe parallel improvements in other methods like EVA and the capital dynamic. that a small decline in residual economic profit may even lead to an increase when calculated as EVA.366 will use this opportunity to load up on debt and move away from the optimal capital structure: a large acquisition that pays off rapidly will move the company back to the target. The parameters are so strict. it is predicated on generating income without regards to growth. The stringent parameters strictly constrain corporate performance to a set of actions. usually with a higher stock price in tow. will lead to stock price appreciation. but if too much is retained and reinvested. will diminish economic profit. For that reason. it treats growth as a cost and not a source of future income. a scenario that will diminish equity and later allow it to respond to improved earnings. Less wealthy firms will undergo a downturn that expands liabilities but contracts assets. For that reason it is more realistic for sensitivity analysis than optimization. Thus. Besides the inadequacies of using a dividend growth model for determining the cost of equity. MODEL BACKGROUND . .G). new shares. These shifts will cause concurrent changes in residual economic profit and allow us to observe potential year over year improvement. we obtain the amount of retained earnings. [[(number of shares outstanding) x (Dividend per share)] / Net income ] from the number. Any change in dividends causes a shift in retained earnings as does changes in net income. dividends and net income. We further decompose the growth rate. The expected rate. It is possible to use the more exacting calculations used in EVA models. the retention ratio is determined by subtracting the calculation. i. “1”. and G is the growth rate. “K” becomes: K = (D1 / P) + G. operating income. Mathematically. If we multiply this number by net income.367 Except for allowing changes in the number of shares and amount of dividends. the standard Gordon model is the basis for the “opportunity cost”.. interest deductions etc. The standard methodology of economic profit applies: we merely multiply our derived expected rate of return by stockholder’s equity. We derive an “expected rate of return” by shifting the variables in the equation. . where P is the current price of the stock. P = D1 / (K . the residual economic profit ascends or descends.[(opportunity cost) (stockholders’ equity)] The concept of the model is to make changes in equity by changing existing shares. “G” into a product of return on equity and retention. Net Income . The dividend per share is the current dividend and not D1. D1 is the next expected dividend. As ROE and retention change. but the concept is best exemplified in the investor-friendly format of: the capital dynamic. The idea is to improve this figure year over year and determine the reasons it might not.e. and subtract this product from net income. which represents the next expected dividend. B31 COLUMN E E1 (ENTER) E2 Equals E1 . Total Shares Total Current Dividends Retained Earnings Payout Ratio Retention Ratio ROE Book Value / Sh.368 MODEL SET UP Table 14-1 ROW 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 COLUMN D D1 D2 D3 COLUMN A INPUTS Net Income Current Dividend Old Shares Outstanding New Shares Outstanding Old Book Value / Sh. DERIVED Old (Last) Equity New Issued Equity Book Val.B24 E3 Equals E2 / E1 Equals B5 * B7 Equals B6 * B10 Equals B5 + B6 Equals B4 * B16 Equals B3 . Growth Rate New Issue Price / Sh. Market Price/Sh.B17 Equals B17 / B3 Equals B18 / B3 Equals B3 / B24 Equals B24 / B16 Equals B9 * B4 Equals B14 + B15 + B18 COLUMN B (ENTER) (ENTER) (ENTER) (ENTER) (ENTER) (ENTER) (ENTER) (ENTER) . Expected Dividend Total Stockholders' Equity ECONOMIC PROFIT Growth Expected Dividend Yield Expected Rate of Return Total Cost of Equity Residual Economic Profit LABEL Capital Long-term Debt LTD/CAPITAL Equals B20 * B21 Equals B23 / B8 Equals B28 + B29 Equals B24 * B30 Equals B3 . 1 .62.44. in the next section. last periods shares = 1377. THE CASE OF: CAN YOU TOP THIS? . Simply divide last year’s stockholders’ equity by the number of shares outstanding during that period.25. this period’s new retained earnings = 13179.1377.13179. Example: Last period’s equity = 52731. Difference in the number of shares = (1646. or it can be calculated as a trend. Number of years between interim = 4. The calculation is as follows: (1. The factor can be an estimate for the next year.8 = $38.8) = 268. Last period’s shares = 1377. Inverse of this number = 1 / 4 or 0.74.25 = 1.376/share These three calculations stand as inputs into the next section.1811. Take this year’s ending equity and subtract both retained earnings for this year and the total value of last year’s equity.8. The proper figure to input is 1.44 / 0. three separate calculations are made outside of the model for entry into the input section: • 1) Determine the growth rate of dividends. This is a factor set up as (1 + Decimal percentage). this period’s shares = 1646. Calculation: 52731 / 1377. This period’s equity = 82646.1811 2) Input a book value per share for last year’s (period’s) equity. For example. which is the difference between this years total shares and last years total shares.74) • 0.38 Divide by the change in shares: 16735. Example: Last period’s equity = 52731.369 MODEL ADAPTATIONS This model can be used as a “stand alone” for sensitivity analysis or adapted for comparisons between years.38 / 268.27 / sh. • 3) Input an issue price for new shares if any. Divide that figure by the number of new shares outstanding.3 Calculation: 82646 -52731 . Current dividend = 1. When adapted. we calculate a geometric mean over five years: Dividend five years ago = 0.3 = $62.1.8.62 =16735. At this juncture.44 . The decision was sound. earnings had grown at about a 40 % pace while dividends had grown at about 20 %. The disparity caused a build up of retained earnings so that it was next to impossible to improve on the economic profit figure for 2005. With record profits and no end in sight. they had hit a bit of a wall: in the past four years.36 1. the decision could have been to buy up a lot of stock or give a special dividend. No amount of income could have beaten the discrepancy unless retained earnings were severely restricted. On the other hand. ConocoPhillips was a favored company in a most favored industry. and this risk was not insurmountable. COMPARING SPREADSHEETS The following figures were applicable to ConocoPhillips residual economic profit in 2005: Table 14-2 2005 Net Income Stockholders' Equity Total Dividends Market Price (End of year) 2006 Expected Dividend QUANTITY 13529 52731 1639 116. however. Rather than attempting to beat the phenomenal 2005 figure. it was time to move away from its optimal capital structure and form a new optimum. ConocoPhillips used this “wall’ as an opportunity to expand. they decided to raise even more equity and buy Burlington Resources. If more income was not the solution to improving economic profit. ConocoPhillips decided to split the stock.370 ConocoPhillips was part of the run away oil industry in the new millennium. but ConocoPhillips opted to move away from their optimal capital structure and raise billions in debt. the stock price had topped $110 per share and was still growing. By 2006. In fact they also issued new shares of stock which they automatically began buying back in 2007. By 2005. In that case.371 Table 14-3 2005 Retention Ratio ROE Expected Dividend Yield Expected Rate of Return Residual Economic Profit QUANTITY (13529 .2379 13529 -[ (0. we can change the three main parameters. the company raised a substantial amount of additional capital and only raised long-term debt to capital by a small amount. Spreadsheet # 2 shows the solution to ConocoPhillips Phillips problem. (shares are being taken off the market) and we had Goal Seek change cell # B6. newly issued shares.8788) + 0. economic profit would have risen substantially.0124 = 0.757(Million).141 .525.2566)(0.1639) / 13529 = 0.97 per share for a total difference of 4896. Alternative decisions such as raising net income or buying back shares would not have worked as exemplified in Spreadsheet # 3.82 in residual economic profit is not nearly as large as the 2005 figure. We then set cell # 33 to 984. We merely go to the tools section in Excel and click on Goal Seek. dividends. We would expect that we would obtain a negative number in that cell that would diminish stockholders’ . Spreadsheet # 1 shows the input for ConocoPhillips in 2006 without any changes.2379)(52731) =984. The 416. That decision would have taken considerable cash that was tied up in the purchase of Burlington Resources. Raising a lot of capital without great risk will bode well for the future In this spreadsheet.8789 13529 / 52731 = 0. However.29 can be inputted as a comparison figure.2370.36 = 0.384 or 2.29 by changing one of the three cells that have the key variables.29 Thus. we raised the issue price to the market price of approximately $71. if ConocoPhillips had paid a special dividend of $2. the current dividend.2566 1. we changed the number in cell B4. According to the calculation.95. the approximate residual economic profit of 984.0124 (0.44 / 116. net income or the amount of shares. The “bar” may be set so high. Nevertheless. (Back to Table of Contents) . Since we obtained a nonsensical number. we do not recommend that a company adhere to the residual economic profit theory unless it is in the firm’s best interests and coalesces with set goals. that it may be next to impossible for a firm to beat a prior figure.372 equity. the decision was not valid. and may solve problems once it is developed into a cohesive strategy. ConocoPhillips displayed a masterful end around by changing their capitalization entirely. economic profit theory exhibits an array of options. the same impossible result would have been obtained. To conclude this section. Had we tried to change net income. 18 416.62 0.023638 0.3 38.384 13179.376 52728.188158 50.159476 0.8 268.95 1.27 71.41 16735.218405073 15550 1.4972 0.152436 0.8225 .1811 62.847564 0.44 1377.20564 1.183114 15133.5 0.48 1646.1 2370.373 SPREADSHEET # 1 A 1) INPUTS 2) 3)Net Income 4)Current Dividend 5)Old Shares 6)New Shares 7)Old Book Value / sh 8)Price/sh 9)Growth Rate 10)New Issue Price /sh 11) 12)DERIVED 13) 14)Old (Last) Equity 15)Newly Issued Equity 16)Total Shares 17)Total Current Dividends 18)Retained Earnings 19)Payout Ratio 20)Retention Ratio 21)ROE 22)Book Value / sh 23)Expected Dividend 24)Total Stockholders' Equity 25) 26)RESIDUAL ECONOMIC PROFIT 27) 28)Growth 29)Expected Dividend Yield 30)Expected Rate of Return 31)Total Cost of Equity 32) 33)Residual Economic Profit B D Capital Long-term debt LTD /CAP E 105737 23093.700784 82643. 242292233 0.194089 48.86 0.1 4896.314864 0.67125 3.048826 0.132978 0.48 1646.41 16735.3 38.685136 0.29 .374 SPREADSHEET # 2 INPUTS Net Income Current Dividend Old Shares New Shares Old Book Value / sh Price/sh Growth Rate New Issue Price /sh DERIVED Old (Last) Equity Newly Issued Equity Total Shares Total Current Dividends Retained Earnings Payout Ratio Retention Ratio ROE Book Value / sh Expected Dividend Total Stockholders' Equity RESIDUAL ECONOMIC PROFIT Growth Expected Dividend Yield Expected Rate of Return Total Cost of Equity Residual Economic Profit 52728.974388 1377.1811 62.71 984.25382 0.141 10653.376 Capital Long-term debt LTD /CAP 105737 25619.181804 14565.75 15550 2.51305 80117.8 268.27 71.95 1. 41 -180031 -1124.041575607 -0.375 SPREADSHEET # 3 INPUTS Net Income Current Dividend Old Shares New Shares Old Book Value / sh Price/sh Growth Rate New Issue Price /sh DERIVED Old (Last) Equity Newly Issued Equity Total Shares Total Current Dividends Retained Earnings Payout Ratio Retention Ratio ROE Book Value / sh Expected Dividend Total Stockholders' Equity RESIDUAL ECONOMIC PROFIT Growth Expected Dividend Yield Expected Rate of Return Total Cost of Equity Residual Economic Profit 52728.44 1377.95161 1.0799 2.16 38.29 .36 -1619.95 1.08 -0.8 -2502.15589 0.700784 -110133 15550 1.13226 14565.1811 71.023638 -0.14119 97.27 71.104121 -0.71 984.08 17169.10412 1.95 Capital Long-term debt LTD /CAP 105737 215870. we can use many concrete measurements to confirm our decision. we are pricing the risk of one firm’s equity and then comparing it to another. Of course. but once it reaches the upper limits of its range for a considerable amount of time. That difference in boundaries is analogous to how the cost of equity works in corporations. . using return as a criterion. rain can be expected. Thus. If that concept seems abstract. The capital structure analyst is less concerned with static values than with measurements of dynamic change. the analyst must be able to tolerate imprecision. THE TOLERATION OF IMPRECISION Since the cost of equity changes from day to day. but the crux of analysis revolves around determining the change in a risk based opportunity cost . The premium is placed on determining a cost that will reflect the reaction of the firm to the earnings of other companies with similar risk. it might be hard to conceive of a theoretical number that can shift from day to day and then use it as a decision-making tool. he or she will most probably not determine an exact figure for the cost of equity.the cost of equity. but “deadly” accurate when it comes to determining whether the figure is growing or declining. the limits for Seattle will be much different from those in Phoenix. For example.376 SECTION III: REAL WORLD CASES 15 ANALYTICAL TOOLS: PRACTICAL APPLICATION Capital structure is dependent on the cost of equity. For those who are educated in the sciences. one firm does very well within a range of eleven to thirteen percent. However. then compare it to a humidity index in weather forecasting. the index ranges between two numbers depending on location. but for another company such numbers will spell disaster. an accurate measurement of 10 % with a misjudgment of a 1 % decrease. it may be better to be imprecise with a concrete value. In essence. might be less helpful than an imprecise guesstimate of 13 % but with a very precise growth estimate of 2 %. the stock may be over priced or the company may be out performing the market in terms of earnings. can lead to a correct investment decision . . For capital budgeting.if the forces that move a firm toward the optimal capital structure are strong enough.or in other words.377 For this reason. For example. we mentioned that the difference in the various methods of determining that cost is the result of the market and the company not being in equilibrium . which would benefit from a precise measurement of the cost of equity. In the chapter on the cost of equity. the inverse of a P/E ratio. using a dividend discount model should not change a decision even if it determines a cost well above or below the market determination. For a non-growth stock. The analyst can detect movement toward the optimal capital structure as long as the methodology of determining the cost of equity reflects an accurate gauge of changes in the cost of capital. savvy analysts can exploit this difference by observing that the “expected” return. Since the same forces affect each method similarly. if a company has a return on equity (ROE) that is much greater than the market for a number of years. even using the simplistic proxy. most experts suggest using a consensus method that will balance an earnings-derived cost (dividend discount model) with a market derived one (CAPM). We illustrate this concept by using radically different methodologies to arrive at the same decision. Their conclusion would be to buy the stock because it is out performing companies of similar risk . In fact. and its cost of equity is determined through a dividend discount model. earnings per share / price per share. determined from the discount model. Interest rate changes and the proportion of debt to equity will be the prime factors that change each respective cost. period to period changes in EVA are more valuable than the absolute size of EVA.its risk/return profile is much better than comparable firms. is greater than the ‘required” return as determined through the CAPM. the figure will be significantly higher than one determined through the capital asset pricing model (CAPM). This is technically improper. • 2. in order to give it a “Federal Reserve bias”. we build a conservative approach: • 1. • 3. The proper methodology to determine the CAPM involves obtaining the five-year averages for both the market and the risk-free rates. However. especially shifts in Federal Reserve policy. Thus. . However. Just as we use five years of data to determine the CAPM. researchers have determined that investors normally need a five percentage point gain of equities over the risk-free rate in order to invest. we may sometimes use the current risk-free rate in our calculations. risk is to be assessed over the entire period of the regression. even if the market was surging.378 ERRING ON THE SIDE OF CONSERVATISM When using different methods to determine the cost of equity. five percent should be the minimum risk premium. we use a moving average of five years of return on equity and retention ratio data. we need to determine a growth rate before determining the cost of equity. Such an average will smooth the data and prevent a particularly inordinate year from determining the cost of equity. As long as the period to period change is similar. • 4. The worst case scenario is using a low cost of equity for one company and comparing it to another using a higher estimate. We can use the highest determination of the cost of equity over lower cost calculations. What if that market rate has been abnormally low for a number of years? That would give a bias to the downside. although larger numbers can be used as appropriate and will realistically escalate this cost. In the CAPM. the ten year treasury rate is no where nearly as volatile as the market rate. we can compare any absolute figures for EVA with the highest cost possible. In a short-term dividend discount model. To combat such a skew. the risk premium is determined by the risk-free rate (ten year bond) subtracted from the market rate. and using a current rate allows us to spot potential trends. first figure) . but then we have to view any estimation of next year’s dividend with suspicion. Experts advise using the average rate during the period of regression. THE CAPM: The preferred method is the capital asset pricing model because it brings so much market information to the table. it will be automatic and effortless and one need not be a trained statistician.1. It may be conceptually incorrect to use this model with all types of dividend paying stocks.Risk-free rate)] Once the student/investor goes through the procedure a few times.gov. THE GORDON MODEL: While the Gordon model was developed specifically for use with constant dividend stocks. [(last figure . its ease of implementation makes it ideal for the investor. The regression will produce a Y intercept and a beta. • a) We estimate next year’s dividend from a five-year growth trend: (Last Dividend / Dividend Five Periods ago) ^0.1)]will yield a rate. • c) We do a regression between a firm’s five-year monthly stock price and the S & P (percentage increases.federalreserve. by the number of years in the regression (five) to obtain an average.379 BRIEF METHOLOLOGIES FOR DETERMINING THE COST OF EQUITY 1. • a) We determine the average risk-free rate by downloading relevant ten year treasury data from a web site like www.25 . • b) We determine the market rate for the period by downloading monthly data for the S & P over the last five years. not absolute figures). • d) We assemble our cost of equity: Risk-free rate (step a) + [(beta from step c) (Market rate from step b . This equation will yield a percentage value for . to assume constant growth in the short-term will not produce a large disparity and may even be actualized. and each of these will be examined for changes because they are separate types of risk indicators 2. We divide this overall percentage change. it mirrors many of the interactions found in the relationships between CAPM variables. We then take the five year average to determine the growth rate. but a periodic difference will reflect some basic change in the cost of capital. THE E/P EARNINGS / PRICE METHOD: In a pinch. a firm must go to the credit and equity markets to raise adequate capital. earnings/price as a substitute measurement. the same forces work on each method: if retained earnings are insufficient. Using this ratio usually underestimates the cost of equity and thus overestimates EVA. becomes implicit in the behavior of beta in the CAPM. [1(dividends paid/net income)]. The 0.25 exponent is the inverse (1/4) of the number of periods between years (4). However. this figure will be well below or above the figure obtained through the CAPM because the two will not be in equilibrium. we determine another yearly growth rate and drop the first data point to make it a five year moving average. but should not expect premium results. • b) We determine the average midrange price for the stock over the appropriate period by adding the high and the low and dividing by two. This product is our growth rate for each year. Normally. The growth rate that is explicit in the Gordon model.380 yearly growth. For the next data. data is not always available and the Gordon model is a good substitute. • c) We take a five year history of the return on equity and the retention ratio. For example. • d) We assemble our cost of equity: (Next Dividend Estimate / Current Midrange Price) + Growth rate. Since it is so . We add this method because it rationalizes the use of P/E analysis and it works well if a company is not a growth stock and exhibits strong movement toward an optimal capital structure. 3. but E/P rises forty percent.at least until the market factored in the increase. and multiply the two for each year. the analyst can resort to using the P/E inverse. The growth rate is a five year moving average. we can interpret the cost of capital to be well behind the acceleration in earnings and the firm would be a good investment . Although the CAPM is the preferred method. if earnings rise twenty percent. the indicator becomes superfluous and investor emotion starts controlling the market. growth stocks in particular seem immune to such analysis because capital flows into them so erratically. P/Es can be distorted and become meaningless.381 quickly calculated. we can clear up some of the eccentric behavior that P/Es have exhibited over the years: Raising debt lowers a P/E because it raises the cost of equity (E/P). and so the investor must weigh the acceleration factors of the two just as with the cost of equity and net income. EVA is a mirror measurement of a proper “Peg” application when one considers that it is optimized while net income is rising and the cost of equity (read E/P) is declining. Individual P/Es are often compared to the market to check if a firm is over sold or over bought. a rising P/E will indicate a potential investment if earnings are growing even faster. the P/E is a proxy for the growth rate that is needed to maintain price. While many analysts still use the P/E approach. it is not proper to consider either E/P or P/E as a static measurement. Higher inflation and interest rates will lead to a lower P/E and a higher cost of equity . measuring the performance of the inversion should not seem unusual and may clear up some of the idiosyncrasies observed over the years. and allowing them to fall precipitously when investors move on. the “Peg” ratio of P/E in the numerator to actual earnings growth in the denominator would be a “buy” indicator if it was well under one and a “sell” if it was over one. However. they may predict company behavior. but as soon as that boundary is violated. a blind faith in P/E analysis can create havoc because in some markets. Again. Within a specific range. it can be used to identify favorable investment situations that warrant further examination. Unfortunately. Putting P/E in the domain of the cost of equity. In many markets. Often. For example. pumping up prices when a sector is favored. A greater variability of earnings is the product of more financial leverage and indicated by a low P/E or a higher cost of equity.E/P. In fact. but it should never be used in isolation. only when one examines the dynamic relationship between earnings and P/E will such an indicator be useful. This discrepancy can best be observed by . In 1969. A more thorough check may reveal pricing problems. not the static version) in the difference between ROE and the cost of equity percentage. but it is applicable to non-dividend paying firms as well.cost of equity %). for not only does it create more uniformity. Thus.58 and so the Gordon model looked something like this: (0. At the time. they were the premier growth stock. Just as a firm’s IRR establishes the “hurdle” rate for capital budgeting.08 / 58 = 3.53(growth) = 54 %. The ratio Net Income / Stockholders’ Equity is the maximum rate at which EVA will be zero. Without knowing the context of the price and earnings behavior. a darling of the 1960s and now a struggling company. The danger is even more emphatic to those who “value” invest in low P/Es. Industries with characteristically higher differences will not be favored over those with smaller differences unless their dynamics are better. EPS or asset value. The E/P rule of thumb would calculate the same cost as : $2. In this example. The expected dividend was 0.58 / $58. There is.382 comparing the Gordon model to E/P for a growth stock. when growth is above fifteen percent it is almost de rigor to use the CAPM. (ROE .whether it is sales. Any cost of equity percentage below this rate will produce a positive EVA.00) + . one concrete figure that can be used as a benchmark when evaluating the cost of equity. Consequently. over at least two periods.08 per share and the current price was $58. growing at a rate of over fifty percent. THE HURDLE RATE The danger of static investing is to blindly infuse capital when a firm is at a specific. investors may invest in a stock simply because it has the lowest P/E in the industry. we use Xerox. As the reader will notice the judgment to use the CAPM cost of equity as much as possible is derived from a basic need for uniformity. movement toward a more optimal capital structure is measured by the change (again.00. however. earnings were $2. the rate of change in the difference . too much debt or even a corrupt but thoroughly entrenched management. measured level .59 %. the return on equity (ROE) establishes a hurdle rate for the cost of equity. Thus the premium is for analysts who can foresee changes in earnings. While most analysis concentrates on earnings. The final equation is: Net Income . stockholders’ equity. the . it too has some predictability since it lags behind earnings and follows both interest rate behavior and the proportion of debt to equity. a firm who usually has a seven point difference will be highly rewarded if it shifts to an eleven point difference. And . However. However.including preferred stock and short-term debt. On the other hand. In fact. we do not figure the weighted average cost of capital (WACC) nor do we figure what is termed “NOPAT” or net operating profit after taxes. equity changes are implemented more gradually because company officials exercise at least some control . THE EVA / CAPITAL DYNAMIC We refer to our investor’s version of economic profit as the capital dynamic but we may use it interchangeably with EVA because it yields the same result. it will not reward a firm that meets the criteria.[(Percentage Cost of Equity)(Stockholders’ Equity)]. if Wall Street expects a company to have ten percentage points difference. Just as with EPS. THE WEIGHTED AVERAGE COST OF CAPITAL (WACC) It is extremely difficult for the average investor to determine an average weighted cost of capital because a precise figure requires the enumeration of each percentage of capital at each specific rate . Capital flows into unexpected growth. although it may be essential to have these numbers on hand for comparative purposes. We merely subtract the product of the percentage cost of equity and stockholders’ equity from net income. We then compare this derived figure with the next period and determine the percentage increase or deficiency. or the cost of equity. In our approach. and operating efficiencies. correctly forecasting a stock issue can be just as profitable.383 between ROE and the cost of equity must be increasing.gleaning information from company officials about taxes.over both new issues and the rate of earnings retention. sales and sector volatility make earnings so difficult to forecast that analysts often resort to “guidance” .although the cost of equity can be volatile. plans for long-term debt. 384 professional analyst should ordinarily have knowledge of this breakdown from a 10K or prospectus and can easily determine any changes on a spreadsheet. Dividing interest expense into debt is not an accurate substitute. A brief example: Table 15-1 TYPE OF CAPITAL SHORT-TERM DEBT LONG-TERM DEBT PREFERRED STOCK COMMON STOCK TOTAL AMOUNT 5 35 10 50 100 PERCENTAGE 5 35 10 50 100 . We know from previous chapters that the WACC is the summation of the products of the various costs of capital components and each percentage of the component type in the capital structure. The analyst is cautioned to use the current interest rates that are paid on long-term and short-term debt respectively.figured by dividing the preferred dividend by the issuing price -that is net of any flotation cost. Preferred stock is a perpetuity . The equation is Cost of Preferred = Preferred Dividend / Net Issuing Price • The cost of both short term debt and long term debt is (interest rate) (1-T) with T representing the effective tax rate. For our purposes: • • Retained earnings and common stock are priced at the derived CAPM “required” rate. 2 %.95 0.09 (1-. Even with the most precise calculation.4) = $8. upward .082 or 8.054)] + [(. interpretation of its movement is ambiguous.054 0.11 0.1)(.06 0. However.1066 Summing the products of the relative percentages and the costs: [(.35)(. One uses the WACC warily.5)(. In a market with rising interest rates (a normal yield curve).036 0.0425 0.00 $100.09 0. but like the WACC. depending on the correct proportion of debt to equity in coordination with interest rate changes.08)] + [(.08 0.0425)] = COST 0.4) = .385 Table 15-2 MARKET COMPONENT PREFERRED DIVIDEND PREFERRED ISSUING PRICE MARKET RATE BETA 10 YEAR TREASURY (RISK-FREE RATE) ONE YEAR BANK NOTE (SHORTTERM DEBT) 15 YEAR AA RATED BONDS (LTD) EFFECTIVE TAX RATE RATE $8.06 (1-. any upward movement does not certainly mean that a company is moving away from that structure.00 .95)(.036)] + [(.00 / $100.11-. a movement downward of the WACC coupled with an upward movement in net income will most likely indicate a move toward a more optimal capital structure and a higher stock price.00 0.0425 + [(. Minimization of the cost of capital can occur both downward and upward.1066)] = .4 Table 15-3 TYPE OF CAPITAL SHORT-TERM DEBT LONG-TERM DEBT PREFERRED STOCK COMMON STOCK FORMULA .05)(. One reason that we concentrate more on the cost of equity is that it gives us more focused information. If the Gordon Model dictates a twelve percent expected return versus an eleven percent CAPM-derived “required” return. Since these respective costs of equity will be similar only during brief periods of equilibrium. the savvy investor rushes to sell the stock because the market will certainly reprice its risk. at the top of the market. if the dividend discount model is well above the required return in the market. . COMPARING RISK: JUSTIFICATION FOR TWO COSTS OF EQUITY Rather than breeding confusion. Likewise. The CAPM offers a real-time indicator that is far more comprehensive than most technical analysis because it brings in exogenous variables like interest rate and market risk. the usual discrepancy in values between the Gordon Model (or any dividend discount model) and the CAPM offers an opportunity. For example. but the rate of change in earnings should be even greater. there will be strong pressure to unite the two costs. and the company follows up with just two percent growth (ROE times retention rate). but if the correlation in the original regression is large (market index and stock price). stocks immediately pick up value and certain specialists will profit from this move. This is not to say that a stock cannot languish despite a large disparity. there is far greater profit in capital structure foresight: the ability to coordinate several variables that indicate capital flows and the gauging of the risk that we encounter. Moreover. and is not just dependent on price and volume. Such a lag creates a “look ahead bias”. the disparity reflects the internal dynamics of a company either out performing or under performing the collective market. CHANGES IN THE CAPM When the Federal Reserve cuts interest rates. the cost of equity will have risen significantly. there may be upward pressure to buy the stock such that both market and dividend discount model are in equilibrium. the CAPM gives us freedom from the evaluation of financial statements which are not published until two months after the fact. consider a company who is close to equilibrium. However.386 movement cannot be interpreted as essentially negative because it must be examined in the context of earnings acceleration. If an alpha is quite large and growing. observing changes allows the investor to gauge non-systematic risk. To elaborate on “alpha”: although alpha is not part of the CAPM. If we were to average them over the period of the regression. but may also indicate the direction it takes. A change in beta.sometimes before accountants even finish the tally if rumor mills are strong. if the market is exceptionally high during six months of one year. Any radical change in market. a company is not only less dependent on the market. • Changes to watch include: 1. interest rate or beta will create more variation in the cost of equity. while conventionally improper. • As a comparison figure. the student/investor is encouraged to do another regression with daily data over a one year period and to use current risk-free and market rates.387 any earnings information we glean from a statement is not up to date enough to act upon. more weight would be given to periods of sustained volatility. its stock may improve even when the market is down. Every quarter the analyst can update the regression with three new monthly points. 5. This update allows the investor to examine changes in alpha. 2.and . The outcome will be a much more volatile figure that is mostly determined by conditions in the current stock market. However. the CAPM would be skewed in that direction. While this method is insufficient to price risk. 3. Such favorable status is usually temporary but may . The market will factor in earnings information almost instantly . 4. dropping the three oldest ones to create a moving average. it will allow the individual to observe pressure on the cost of equity. A change in R squared and (1. making the market index and risk-free rates yearly averages rather than five-year averages. will give the cost of equity a “current bias”. beta and the market. The following adaptations may work for the analyst: • The proper methodology is to use sixty data points from monthly index data and then regress a firm’s stock price change against the market’s change over a five-year span.R squared). A change in the market. A change in the interest rate. For example. A change in alpha. the greater is the economic profit given the internal dynamics of capital structure. the stock may be unstable. he or she will find that many appreciating stocks will have a similar comparative capital dynamic. which is a difference operation. one could use the quotient form of the capital dynamic. and so each market seems to put a premium on reaching a specific number . “the comparative capital dynamic”. into net income and derive a ratio. if (1-R squared) is growing and alpha is not. the generic. upward pressure on the stock may be followed by subsequent downward pressure. that is the product of stockholders’ equity and the percentage cost. but is unsure about the significance of actual size.388 happen in scenarios where the firm receives special treatment (tariffs). reacting violently in a calm market or not at all during periods of market appreciation. On the other hand. they may not perfectly inform us if a . publicly available algorithms can suit our need for estimation. the firm finds it difficult to follow such a stellar performance.as it rewards the company with a higher stock price.5. we divide the total cost of equity. The investor has determined that each is about to increase their respective EVAs. attempting to debunk their efficacy. The higher the ratio. if the student/investor observes this number over time. to a much larger number like 3. In fact. In the comparative capital dynamic.5. THE COMPARATIVE CAPITAL DYNAMIC The student/investor may face a dilemma between investing in two good prospects in separate industries. When one company goes from a smaller number like 1. which this author terms. or has special pricing power (gold mining when financial confidence is shaken). Remember that the capital dynamic is an investor friendly version of EVA. In this case. one a small manufacturer and the other a large multinational. THE MARGINAL BENEFITS EQUATION Many financial institutions will guard their probability models like state secrets which have led some academics to examine them in “deconstructionist mode”. Continuing with the theme of adaptation. The ratio Net Income / Total Cost of Equity can be used as a comparative indicator. With an algorithm that matches the specific probability of default for the company and industry. Of particular significance is the behavior of the default probability. and the default probability declines. with a decline in the long-term debt to capital ratio . we seek to increase the figure year over year. The left side of the equation is merely the outstanding long-term debt multiplied by the average effective tax rate for the industry. If more debt leads to a rapid increase in cash-flow. we use a default probability model as the prime variable in the determination of the cost of bankruptcy. we only need to examine the behavior of the components and to observe that the function is increasing. In a rising market. if debt increases. This author analyzes performance based on the fiscal year of the company simply because management strategy is exhibited in the behavior of default probability components over that period. this scenario is ideal because it represents both asset growth and earnings growth. • 2. When the tax benefits of debt exceed this cost. market value is created and the stock appreciates.more return in the domain of less risk.389 firm is about to become insolvent. However. and the stock will appreciate. it must be accompanied by a large debt issue and only a small stock increase because there will be countervailing forces in the next year to balance the equation. the default probability decreases. but our methodology is far different from the decision to grant credit. Both stockholders’ equity and long-term debt to capital should be components of that algorithm. for the evaluation of movement toward the optimal. and the stock will under perform. In essence. The key to the relationship is the interaction between default probability and debt. (Long-term debt) x (Effective Tax Rate) . if a debt issue languishes and requires more investment with little return. Like the EVA/Capital Dynamic. In most cases. it is because earnings pressure is positive and the company has added equity to buffer the debt to equity ratio. optimization of debt to equity occurs when the function is maximized and the first derivative is equal to zero. If it increases at all. To create a working marginal benefits equation: • 1. the probability of default will rise. However. Ohlson and Zmijewski have all published usable algorithms. a generic algorithm can display the increase in risk from year to year. We use a probability default model to determine a “ball park” figure for default.7182818. Table 15-4 COMPONENT (ZMIJEWSKI 'S) Intercept Total Liabilities / Total Assets Current Assets / Current Liabilities Net Income / Total Assets VALUE COEFFICIENT -9. the investor must calculate the “tangible book value per share”. [1 .06 Logit probabilities are expressed in logarithms.384 0. we will show the student/investor how to transform the parameters into a logit expression and then a probability. but this author prefers Zmijewski’s because it is economic and captures default risk in a few expressions. we also must . we subtract intangible assets and unamortized debt from total assets.. The full function is: (Total assets . In order to decipher how much of market value is intrinsic value and how much is made up of assets that are unclaimed by creditors.Intangible assets .390 • 3. We divide this figure by the total number of shares outstanding and determine the “tangible book value per share”.) by an exponent. Since we are not doing credit analysis. To determine the right side of the equation.069 -1. Merton.(Tangible book value per share / Market value per share)] x [(Number of shares outstanding) x (Market value per share)] • 5.. We determine the difference between current market value and tangible value to determine the amount of potential loss in case of bankruptcy. When we substitute Euler’s number. In the following methodology. We can algebraically eliminate the logarithm and replace it with the EXP function which is a command to multiply Euler’s number (2. Altman. Shumway. • 4.479 6.Unamortized debt) / Number of shares outstanding. The mechanics of this operation are as follows: Ln [ P1 / (1-P)] = Xiß P1 = 1 / [1 + EXP (-Xi ß)] The intercept is also negative because it is inferred that it has a coefficient of “1”.[(Probability of Default) x (1(Tangible book value per share / Market value per share))x (Number of shares outstanding x Market value per share)] If the stock price falls (perhaps as a reaction to the economy) one can read a “false increase” into this equation so it is important to understand the interaction between the component parts. The basic premise is that the value of a levered company is greater than an unlevered one because interest is tax deductible. and using the algorithm to observe changes in risk relative to debt and stock price makes it a valuable tool. if current assets / current liabilities is “2”. LEVERAGE STATE ANALYSIS . The final expression is the difference between left and right sides which we expect to see increase. This same function can be used to find a target proportion of debt to equity as long as the proper default algorithm is used.391 make all the components in Zmijewski’s algorithm negative.479. “throwing the baby out with the bath water” gets us nowhere. Thus. In many financial textbooks.(Long-term debt x Effective tax rate) . the author will refer to this situation as “marginal tax benefits equal marginal bankruptcy costs”. it should be quite obvious to the mathematician that the prediction of bankruptcy is not an exact science!. If a firm is still adding debt even as it increases earnings and equity. the function maximizes when the first derivative equals zero. With so much of the expression dependent on the relatively large intercept of -9. the probability of default will decrease which will put upward pressure on the stock. • 6. it becomes -2. We determine the cost of bankruptcy by multiplying the default probability by the potential loss or: (Probability of Default) x [1-(Tangible book value per share / Market value per share)] x (Number of shares outstanding x Market value per share) • 7. However. Therefore. however. compensate shareholders with net income that is generated from lower interest rates and increased demand for its products. sector and transition. for example. Companies that produce the most output during a recovery. the amount of capital. when their respective sectors are favored. Minimizing the cost of capital. but does so with a minimum of risk that will sustain the acceleration of earnings for as long as possible.392 Without question. is extremely important to firms who do not fund with debt. As the economy improves. Operating leverage. the same firm may want a lower operating leverage during a sector downturn. The premium is to search for a state that will not only generate more earnings. Equity is built by either attracting investors to the stock through higher earnings. will guarantee higher profits because demand is both stable and high during that period. a premium is created for equity funding. the cost of capital can actually rise while a firm is minimizing it. While other analytical tools are primarily concurrent indicators of stock price. or retaining those same earnings and decreasing the proportion of debt to equity. the anticipation of a profitable leverage state will be the most formidable weapon in your arsenal. and how they interact with each other in the business cycle can be perplexing. This volatile dichotomy between the cost of capital. Their greater financial leverage allows them to finance with less equity. When the economic outlook declines. the Federal Reserve lowers rates. Between these two extremes are intermediate positions that are based on phase. the Federal Reserve raises interest rates making long-term debt more expensive. often requires a different leverage state for each phase of the business cycle. The logic chain behind leverage states is basic. a leverage state that transitions from requiring more capital to paying off investors with higher profits is predictive. However. A higher relative operating leverage for these firms. rewarding the wealthiest companies who can afford the most debt with a lower cost of capital. The two most basic indicators are: . and the share price escalates. like the financial leverage ratio. the economic context of a change in the ratio is paramount. Thus. if more debt is incurred. this ratio begins to shrink .even if by a minuscule amount. They are: • 3.I. The full function is: EBIT / EBIT . If earnings are high. It is significant because it affects financial risk. For example. % ∆ EBIT / % ∆ Sales . Nevertheless. This ratio is formed by subtracting interest expense from earnings before interest and taxes (EBIT) and dividing it back into EBIT. the change in ratios is .While this ratio is often used interchangeably with operating leverage. a rising operating momentum will buffer that same expense. it is only the same function when a firm is in equilibrium . The long-term debt to capital ratio. an increase in this ratio will actually entail lower capital costs with a typical earnings increase in subsequent years. it offers the investor a measurable function that indicates a short-term trend in earnings. they will be retained once dividends are paid and the LTD/CAP ratio will decrease. This ratio will measure the proportion of debt to equity in a meaningful way.its ability to pay off interest expenses. The financial leverage ratio.a rare occurrence considering modern corporate volatility. a simultaneous shift upward in both ratios (more interest and more debt) along with a favorable earnings forecast will be a precursor for a transition to proportionally less debt and more earnings. A shift to a smaller ratio often signals a movement toward an optimal capital structure with less inherent risk. Analysts might recognize this expression as the inverse of the Du Pont equation simile EBT /EBIT which is a component part of the return on equity (ROE). • 2. Two secondary indicators also exist. when a company takes on greater debt. Although the absolute value in total leverage may be similar for each situation. Likewise. The probability of default decreases as more operating earnings cover interest expense. Operating Momentum.393 • 1. If interest rates are low enough. The important point to realize is that as earnings increase. an operating momentum that suddenly decreases will increase the company’s default risk . but there is no formula. If the company has taken on leverage and analysts predict comparatively higher earnings. an investor can observe the current most profitable sector in the economy. more vendor contracts and more short-term credit signals a potential increase in business. It is true that more current liabilities will decrease working capital and lead to a higher probability of default.394 indicative of operating margins. this is a situation in which the investor can depend on analyst’s forecasts. Firms who use debt strategically want to incur debt from a position of strength. That is not an invitation to invest as soon as the analyst observes a lower LTD/CAP for one quarter. they become “accounts receivable”. and current assets again balance current liabilities. The asset to capital ratio. it will be too late for the shareholder to invest because earnings would have been increased and share price would have moved concurrently. Short-term debt is a legitimate source of financing that may lower the cost of capital. While much is made of the asset to equity ratio. and the investor can look forward to a higher share price. Secondly. when this occurs. this ratio may be indicative of a “pick-up” in business. No firm wants to incur debt from a position of decreasing margins. but it is equally valid that more labor. • 4. if credit standards are lower in the overall economy. the payoff is expected to be rapid. Short-term debt is also a precursor to more sales. What was the leverage state before it became profitable? What is its leverage state now? . Since the investor cannot depend on a leverage state to provide momentum (investing in a state where earnings are high) he or she must anticipate the transition from a “debt” state to a more profitable state. In fact. When such sales are actuated. Gauging risk by observing the behavior of the financial leverage ratio and long-term debt to capital ratio over a number of quarters is standard. However. the company may actually be moving toward a higher cost of capital in that scenario. the proportion of long-term debt to capital is more trend worthy because firms must add or subtract capital in large increments to be cost effective. However. 395 Not all companies in a sector will fall into the same pattern, but if the student/investor is tenacious enough, he or she will derive a solid indication of how these leverage state components should be coordinated. The rationale for a coherent strategy is that each economy prices risk and return in a specific way; the proper combination of leverage factors will lead to the highest point on the “efficient frontier” - the highest return per unit of risk. In essence, there is one state that will encourage the minimization of the cost of capital per unit of income above all others. There will also be a state that actuates the movement toward an optimal capital structure. The investor’s imperative is to be in the game early enough to capture the acceleration of earnings. Lastly, observing “inside” activity is a must. When a leverage state is solid, company executives will begin to accumulate shares. Although massive selling activity is often for tax purposes and is not a “sell” indicator, the purchase of more shares is a “buy” indicator - if the analyst examines the leverage state first. These investments are far more lucrative at the beginning and mid stages of a recovery/expansion than in the later stages, at the top of the market. The investor should avoid buying shares if earning acceleration has already occurred, but might consider investing when the company has both insider and debt activity occurring simultaneously. Investing after earnings have already been actuated can occur when executives make “good faith” investments, but unbridled optimism is not the purview of an objective analyst. THE LOOK AHEAD BIAS The difference between investing in leverage states and other fundamentals like earnings or sales is that the leverage state lacks what is termed, a ‘look ahead bias”. Frequently, investors make decisions after receiving reports on earnings or another fundamental and fail to realize that the firm is in a constant state of flux; the report refers to information that was received long ago and may no longer be actionable. Moreover, any screen or strategy that is constructed on that basis may be obsolete because the market will no longer react to such information in the same way. Fundamentals are concurrent 396 indicators; they perform concurrently with the stock price. On the other hand, leverage states are predictive indicators because they are developed well before a price is expected to rise. In fact, there is a ten month “window of opportunity” in each fiscal year because a 10 K will come out about two months after the year is over. So called ‘smart money” investors may be adding shares at this time, but the investing public will be unaware until earnings begin to escalate. The predictive nature of leverage states versus concurrent indicators like EVA is that a leverage state will foreshadow the direction of the cost of equity in relation to earnings. While it is difficult to predict precisely when profits will rise or even if a sector will be strong at all during a business cycle, we know that beta will react to a change in the proportion of debt to equity. We also know when the Federal Reserve is lowering rates or raising them and that they move in a discernible trend. When the cost of equity becomes low enough compared to the change in earnings, the equity market begins to rise. Therefore, a leverage state is simply an extension of the same capital structure theme of earnings changes in relation to the cost of equity; that state which encourages earnings to rise fastest in terns of the cost of equity, will be the state that offers the most return for the least amount of risk. At the same time, it will be the state that moves the firm toward an optimal capital structure which maximizes the price of the stock. MICRO ANALYSIS: QUARTERLY OBSERVATION Few techniques are fraught with more risk than quarterly extrapolation of performance. Although we can detect changes in earnings and sales from period to period, short-term predictions are very susceptible to random volatility, simply because factors outside of the model may have as much affect on the stock price as the internal dynamics of the company itself. Consider the 2007-2008 credit crunch; firms that had excellent earnings and credit were damaged by speculation in the housing market. A decrease in the cost of equity over two quarters may have been a comparative advantage, but it could not prevent a stock from falling. Moreover, there is no fundamental that is so stable that it offers a reliable prediction variable. Even changes in the financial leverage ratio, which 397 can be minute, must be accumulated over a series of quarters to be confirming. Thus, trying to be “in the game” early, by anticipating a favorable leverage state, can be costly. For example, consider a company who goes through three quarters with lower LTD/CAP and lower financial leverage. The investor makes a move and accumulates stock, only to find out that in the forth quarter the company takes on an unsound acquisition and pays for it with an equity issue. The hallmark of quarterly analysis is to anticipate a large jump in EVA before it occurs. Professional analysts constantly forecast earnings and so a near-term outlook is almost always available. However, predicting changes in the cost of capital and stockholders equity may be vexing. While management has incremental control over equity and desires smooth changes, there is no guarantee that the growth rate will not suddenly become eccentric. Financial professionals know that the stock price is a multiple of book value and may want it to conform to industry averages, or they may want to exercise options in anticipation of further growth. On the other hand, the cost of equity will mirror the change in interest rates and stock prices, as well as the proportion of debt to equity within the firm. However, it lags the performance of earnings and is difficult to measure on a short-term basis. The cost is not always reflective of immediate equity risk because it is calculated over a sixty month period, but observation of the risk premium, the difference between the market index and the ten year bond, can indicate the direction of change. Again, the premium is placed on earnings accelerating faster than the total cost of equity, and the next sections will focus on the attempt to predict such a phenomenon. NAIVE EXTRAPOLATION In naive extrapolation, we compare the total cost of equity five quarters ago to what it is now, and derive a growth rate. We then compare analyst’s estimates of earnings as a growth rate to the growth rate in total cost of equity. We place the change in earnings in the numerator and the change in the total cost of equity in the denominator and create a percentage change version of the comparative dynamic. If it is over “1”, we anticipate the 398 EVA/capital dynamic to rise and the stock price to do likewise. In fact, we can even make concrete predictions of EVA if we multiple the growth rates by last period’s base of net income and total cost of equity. However such predictions do not anticipate transitional changes in equity and earnings that may be totally skewed. Analysts can often change forecasts once “guidance” is received from the company , but the average investor does not have such recourse. A naive extrapolation is very much like proclaiming that the weather tomorrow will be just like it is today. Four out of five times, that prediction will be correct but twenty percent of the time it will fail because it does not anticipate the transitions from sunny to stormy or vice versa. We apply the exponential growth rate technique (AKA geometric mean) over a span of five quarters although four periods is acceptable if data is unavailable; in this case, we are looking for a confirmation, not a decision indicator. Growth rate extrapolation is accurate for a stable dividend, but is not appropriate for volatile measurements unless it is used as a comparison to a different but related ratio. For example, extrapolating a ten percent growth rate for sales and assuming that sales are expected to rise by ten percent is a misuse and will be inaccurate. However, if we extrapolate the ten percent sales rate, and then extrapolate a thirteen percent increase for variable costs, the relation between the two is significant. To use the growth rate technique, we need both the percentage cost of equity from five quarters ago and the balance sheet item, stockholders’ equity. We also need the current data for those items. We then multiply each percentage cost of equity by each entry for stockholders’ equity. We then make a ratio between the latest data in the numerator and the historical data in the denominator, and multiply by an exponent that is the inverse of the number of periods between the data entries. In the case of five periods, the number between periods is four and the exponent is “1/4” or 0.25. If the period length were twelve years, the number between periods would be eleven and the exponent would be “1 / 11” or 0.0909. To illustrate the technique, examine the following data: 399 Table 15-5 VARIABLE % COST OF EQUITY STOCKHOLDERS' EQUITY CURRENT 10.6 % 100 5 QUARTERS AGO 9.4 % 84 We determine the respective products and use the results as the components of the ratio. (.106)(100) = 10.6 and (.094)(84) = 7.896. (10.6 / 7.896) 0.25 = 1.0764. As this is the growth factor, we subtract “1” to obtain a decimal percentage = 7.64 % per quarter. When we examine analysts’ earnings estimates, they will usually be an EPS figure for a year, and so we need to determine that growth rate and divide by four to put it on a quarterly basis. For example, last years EPS was $1.00 per share and the forecast is for $1.19. (1.19 / 1.00) equals a nineteen percent year to year gain. When we divide by four, we obtain a 4.75 percent quarterly gain. The final ratio is then 4.75 % / 7.64 %. The conclusion is that the total cost of equity is growing faster than earnings and so this would confirm a negative evaluation on the stock. However, if other indicators point to a banner year, the growth rate study should be disregarded. The premium is on judgment . The percentage cost of equity naturally rises until the next downturn, but the company can lower its beta in the interim and buffer some of that increase. If the firm is now in the process of buying back stock, the prospects would be better than our growth study concluded. EARNINGS PRESSURE The art and science of forecasting earnings is difficult. Perhaps the best methodology begins with a demand forecast for the industry that considers the direction of the economy. Next the market share of the individual firm must be examined. Finally, the internal dynamics of the firm itself are considered. The complexity of this feat creates variability in forecasts because it is essentially a mathematical version of Murphy ’s Law: more variables in the model can lead to more inaccuracy especially if those variables are 400 independent of each other. However, analysts do form consensus opinions and it is much better to use their collective judgment than depend on naive extrapolation. If earnings are a “hit or miss” proposition, any attempt to achieve precision without a thoroughly vetted model will surely be a “miss”. When we use naive extrapolation of earnings, it becomes a benchmark; it is simply an average growth rate over a set period. We can add this extrapolation to our own fundamentally derived estimate, and then compare both of these estimates to analysts’ expectations. Again, we always use our estimates as a confirming indicator. Any forward looking inputs for earnings should be from the analysts and not from extrapolation The first procedure is to construct an estimate from fundamentals. This will yield a percentage increase that would occur if the relationship between the inputs remained unchanged. One adaptation - we use the book values of debt and equity rather than the market value as an input. We are looking for a “ball park” bench mark and not a forecast. To construct this estimate, we need only to plug a list of fundamentals into a straight expression: Table 15-6 VARIABLE A) PAYOUT RATIO B) RETURN ON ASSETS (ROA) FORMULA Dividends / Net Income Net Income +( (Interest expense)(1-tax rate)) / Total Assets C) TOTAL DEBT / STOCKHOLDERS' Book value of debt / Book value of equity EQUITY D) INTEREST RATE ON TOTAL Interest Expense / Interest Bearing Debt DEBT E) TAX RATE Decimal effective tax rate In the expression, it is much easier to work from right to left: (1 - A) x (B + (C x (B - (D x (1-E))))) 401 When we parse this function, the factors that improve an earnings outlook are quite obvious: less interest and more debt, but a greater return on assets as well. The nature of the function is to create tension between its components, because elements with negative correlation are added to or multiplied against each other. Like the cost of equity as determined by the Gordon model, much of the growth comes from improving the retention ratio - (1-A). Once we have a figure from the fundamentals, we can do a five-year exponential growth rate on earnings and determine how earnings would grow if they followed a trend line. Again, this is the same function that we used in quarterly extrapolation, except that we use years instead of quarters as a period. We now have two of our own “earnings estimates” to compare with professional analysts’ estimates. On Wall Street, when a company performs above analysts’ expectations, they are rewarded with an increased share price. The objective in this exercise is to compare analysts’ forecasts with our mathematical determinations. When analysts’ estimates are above both our figures, we interpret that difference as a potential buy indicator because there may be upwards pressure on earnings. When analysts’ estimates are below both our figures, it is a potential sell indicator, because there may be downward pressure to under perform earnings trends. If analysts’ estimates are in the middle, we interpret the situation as neutral and glean more information. Our estimates are derived from trends and fundamentals and lack the extensive information analysts’ are privy to. However, our calculations also stand as a benchmark because they are objective and mechanical, and form the foundation for comparison. (Back to Table of Contents) 402 APPENDIX: DIVIDEND DISCOUNT MODELS Since many analysts “go by the book” and use the various dividend discount models to determine a stock’s “fair value”, we include an explanation of a basic model. However, the models require extrapolation on growth rates, sometimes years into the future. The reader is referred to Burton Malkiel’s A Random Walk Down Wall Street to observe some of the oddities of this process. In fact, most investment banks still use some form of discount model to evaluate stocks because the models do indeed have legitimacy when the market is fairly valued (about the middle of a business cycle). The theoretical background of the models is sound. A stock is valued at the present value of dividends that are anticipated to be paid in the future; that value is based on the growth rate of the company and the cost to borrow money - the weighted average cost of capital (WACC). The “chink” in the armor comes from anticipating growth rates and the cost of capital which are so volatile that they often defy prediction. Moreover, the economy may be in a phase when stocks are over or under valued, making empirical application less than viable. The reader is referred to a fundamental finance text to understand the concept of present value: the basic concept revolves around the knowledge that a dollar received in the present is worth more than a dollar received in the future, because the present dollar can earn interest. Any amount in the future is discounted by a factor of (1 / (1 + borrowing rate)^number of periods in the future). In this case, the “borrowing cost” is equated with the cost of equity. Thus, our growing dividends are discounted by one plus the borrowing rate multiplied to the power of the number of future periods. We sum all of the present values over the period of one growth rate and then add this figure to the present value derived from the next growth rate. PRICE = [ Σ D(1+G1) T-1 T N-1 N / (1+K) ] + [ (D(1+G1) / (1+K) ) x ((1+G2) / (K-G2))] The following table will itemize the variables. Of particular interest may be the function of “T”. T is a command to do the calculation in numerical sequence until the number specified by “N” is reached. For example, if N=3, then the respective denominators in that 403 calculation will be in sequence: (1+K) , (1+K) , (1+K) . The reader should also notice that (N – 1) represents the years of growth and should not confuse that figure with “N”, which is one period after the number of years of growth is over or (Years of Growth + 1). Table 15-7 SYMBOL Σ D K N-1 T EXPLANATION A command to sum the sequence of calculations The present dividend The cost of equity The number of years that growth will occur at a specific rate (G1) Starting at T=1, a command to do the calculation in integer sequence until "N" is reached The growth rate that occurs during "N 1" years Number of Years of Growth + “1” The growth rate that occurs after "N 1" years 1 2 3 G1 N G2 Dividend discount models can be quite involved and complex, enumerating three or four different growth rates at a time. Approach these with caution. (Back to Table of Contents) 404 16 KIMBERLY CLARK - “TOO MUCH OF A GOOD THING” Economic Profit and Marginal Benefits Analysis The recession of 91-92 was followed by one of the longest bull markets in history. The speculative excesses that helped fuel the economy encouraged the issue of massive amounts of equity. Although profits were rising, optimism was rising faster, and the large issues did nothing to dilute an already overheated market. Once the bottom fell out in 2001, shareholders (especially those who had invested heavily in NASDAQ stocks) were left holding the bag. Tech stocks that had been trading for over one hundred dollars a share could often be bought for less than ten dollars. In fact, as of this writing (2008), NASDAQ trades at about one-half its value in 1999. UNDERPINNING 1: POSITION IN THE BUSINESS CYCLE To those who saw this debacle coming, good investing sense led them into defensive sector stocks, mostly consumer staples, healthcare and household products, whose demand would outpace a lagging economy. Kimberly-Clark (KMB) was a large, well-known paper products company with more than thirteen billion dollars in revenue. Its beta of 0.44 made it a perfect candidate for portfolio rebalancing in place of higher beta tech and telecom equipment stocks. As a low beta household products company, it was positioned well in the business cycle - which is the first of four basic underpinnings that determines our investment analysis. Its profits were stable enough to take advantage of financial leverage, but its equity risk was low enough so that debt was not excessive; Kimberly-Clark could take on more leverage, especially in a low interest rate environment. The firm’s position gave it a favorable combination of cost and demand factors that would yield a strategic advantage; position in the business cycle is so important because of its integrative effect on all other factors. When a firm has the right combination of 405 operating and financial leverage the interface between sales and capital is strengthened. Hence, in a downturn, those firms with small but steady growth are less risky. UNDERPINNINGS 2,3, AND 4: OPPORTUNITIES FOR ANALYSIS Most firms are integrated with the greater economy through operating and financial risk. During each phase of the cycle, some firm will possess a combination of extraordinary factors that helps it dominate others, and when the economy changes, those factors recede. However, outside of sector rotation, there is little the analyst can do to predict a firm’s reaction to a changing market Without knowledge of the forces that move an individual company, our decisions will be limited to price and volume movements. Although macroeconomic conditions have the greatest impact on corporate activities, we need at least three other tenets to govern our analysis - a methodology that corroborates a firm’s possession of “favorable factors” if you will. The other three underpinnings are: shifts to an optimal capital structure; potential increases in comparative economic profit; and increases in sales and earnings potential. Researching corporate history for averages in capital proportions can be painstaking. There is no mandate that a company will perform at that average or that the average is even relevant given the contingencies in the current economic cycle; interest rate changes can shift the optimal proportion. However, certain patterns of leverage changes are correlated with performance, and any change in the probability of default is usually accompanied by a consequent change in stock price. Since default is a function of probability, it is quite difficult to configure the proper level of debt to equity deterministically; no two default functions are alike. For that reason, we rely on several measurements, realizing that the emphasis should be on detecting movement toward an optimal proportion rather than on calculating a decisive ratio. One of the corroborating techniques is to analyze economic profit for the potential to increase - which may not be so “esoteric” as it seems. When we invest on the basis of 406 earnings, we base our judgment on one variable - income - which is the projected outcome of coordinating costs, sales and type of industry. With economic profit analysis, we have an interaction of at least three variables which also represents the collective analysis of numerous components - beta, retained earnings, new issues and taxes. Our judgmental risk becomes diversified. We can be wrong about net income, for example, but precisely accurate about a rise in the cost of equity and still maintain some predictive capability. Statistically, we know that a large move away from the mean, will revert in the opposite direction because each industry sets a pattern for dividends, retention and the amount of equity a firm can safely issue. Balance sheets need to “balance”, and we use that conformance to examine changes in economic profit that show shifts in a firm’s level of risk. Because of their inherent volatility, changes in sales and earnings are the most risky predictions that analysts will make ; most professional analysts will receive “guidance” from both their industry and target corporation. In capital structure analysis, we try not to make earnings forecasts ourselves, but will rely on consensus opinion to provide inputs for sensitivity analysis in economic profit models. We are much more concerned with the changing pattern of operating risk that occurs when sales and earnings change. And we attempt to integrate sales and earnings changes into marginal benefits analysis by treating them as risk factors. In any default model, earnings becomes the linchpin in determining how much debt is permissible; we recognize that maximizing earnings may increase risk to intolerable levels, creating a rebound effect that undermines growth in future periods. Thus, we attempt to observe sustainable levels, amounts that can optimize capital structure and yet create shareholder wealth. THE LEVERAGE STATE The student/investor may estimate Kimberly-Clark’s optimal target structure from the chapter on marginal benefits modeling. From that model, we “guesstimate” KimberlyClark’s target to be approximately 26 % long-term debt to capital, but recognize that the 2 percent for operating income. those firms who can most afford new debt will have a big advantage. the cost of capital will be less expensive. However. we must make an analysis of both EVA and marginal benefits. Thus. . Since this investment is a non-speculative portfolio rebalancing during a downturn. especially in the realm of interest rates and capital. whose profits have been exceptional and who hovers near an optimal capital proportion. the parameters fit our needs perfectly. the importance of objective analysis cannot be underestimated. we attempt to establish that Kimberly-Clark’s low risk stems from its proximity to the optimal capital structure and that only a high. however. This ratio of 5. 1. we might look for a firm who is about to enter a proportionate equity building cycle.407 constraints are artificial. many firms were coming off of long stretches where funding was done primarily through retained earnings and equity issues that flooded the market with new stock. We also establish the five year geometric average growth which is 5. Nevertheless.risk move away from that target and then back again.2 / 6. In a normal market. Operating Risk To establish a measure of operating risk. we use the firm’s operating momentum: (% ∆ Operating Income / % ∆ Sales) which gives an approximation of operating leverage. a company that is going to lower its debt ratio while profits are increasing. and later. would greatly appreciate the stock.33 equals 0. Kimberly-Clark. would be a perfect choice for an investment.33 percent for sales. we need a breakdown of their current leverage state .82 which is comparatively low. If the Federal Reserve begins to lower rates. We then compare this figure to the most recent momentum figures for 1999 and 2000. In 2000. and 6. 085 -0. would be justified by the lower taxes paid on income and the potentially lower interest rate that would decrease capital costs.82 Kimberly-Clark’s operating momentum has been buttressed by a high operating income which is in the process of reverting to its mean. The expectation of an economic downturn would further dim prospects for a higher income.77 = 3. Table 16-2 2.5 = 1.7 2000 3203 13982 13. it still reflects a smaller probability of default when all other factors are held constant. The tax effects of taking on debt.34 % 5.408 Table 16-1 Kimberly-Clark Year 1998 Operating Income 2320 Sales 12298 Operating momentum Change 1999 / 2000 Geometric Average 1999 2815 13007 21.33 = 0. it is always a “crapshoot” anytime debt is raised proportionately. income must be increased enough to lower the probability of default or the firm can meander .34 / 5.84 -50.with more debt making up the shortfall. Despite the high- . However.087 2000 222 2844 1. Financial Leverage Ratio KMB Financial Leverage YEAR Interest Expense EBIT Financial Leverage Ratio Change 1999 / 2000 1999 213 2654 1. and so we would expect a lower operating momentum and a smaller increase in EBIT.18 % Although the change in the financial leverage ratio was slight.78 / 7.2 / 6. more analysis is required because there is equivocation in both the movement of the cost of capital and in earnings. Proportion of Debt to Equity Table 16-3 KMB Debt / Equity YEAR 1999 2000 1999 / 2000 Change % LTD / CAP 27. For example.9 A nine year average showed that Kimberly-Clark’s average long-term debt to capital was 23.7 7767.15 % Long-term Debt 1926. financial leverage ratios may decrease at the same time that debt becomes a higher proportion of capital. however.75 .75 % appears to be in the “right” direction. the higher EBIT allowed the firm to both increase its total amount of long-term debt (and tax advantages) and reduce its financial leverage ratio simultaneously. interest expense may be kept artificially low until a “day of reckoning”. the movement to 25. and the analyst would properly depend more on economic profit analysis to make a determination. It is.409 interest rate environment of 2000. In that case.3 Capital 7019. 3. When the two move in opposite directions. necessary to observe the financial leverage ratio in the context of proportional changes in long-term debt.6. Kimberly-Clark moved to reduce this ratio.35 %.1 5767. when a firm takes on zero coupon bonds.6 2000.6 Stockholders' Equity 5093. The approximately 26 % optimum that we determined from marginal benefits analysis is a “ball park” figure based on averages from the past five years. such anomalies represent a shift in balance and are difficult to interpret. but the margin of tolerance is so small that we need to depend on other measurements to confirm .44 25. the leverage state does little to reflect the inherent risk. Thus. but must be accompanied by a low cost of equity to be effective. firms who would otherwise raise long-term debt delay purchases and new projects until trends begin to unfold. there was simply no need to use this ratio as an adjunct. . Raising short-term debt can be a temporary substitute for long-term funding in periods of uncertainty. • d) It may signal greater vendor activity and potential sales increases. 4. taking on less debt when interest rates are high encourages optimality because more income is freed for shareholders in the form of dividends or stock buy backs. Kimberly-Clark raised its assets by 12. From a business cycle perspective. With such proximity to its optimal capital target. the increase in assets to capital was minimal and probably was not great enough to cover any capital shortfall. The higher amount of retained earnings becomes the source for funding. it can only be viewed as a “plus” and not the insolvency measurement that might occur in leaner times. Nevertheless.98 percent and capital by 10. any analysis would be incomplete. Without knowledge of EVA and the cost of equity. Since Kimberly-Clark raised this ratio from a position of strength (close to an optimal capital structure). The Assets / Capital Ratio The importance of this indicator cannot be dismissed.66 percent. we can begin to classify the investment itself very low risk with a high probability of exceeding the return on a ten year treasury note. The asset to capital ratio helps in at least four other ways: • • • a) It signals a potentially greater return on capital b) It may signal an increase in free-cash flow c) It may help limit the amount of external financing by covering “shortfalls’ in the capital budget.410 it. and less is paid out in more expensive interest. 23 % In the last section.23 percent. we described capital as a 10. a firm needs the correct amount of equity in the domain of two other variables . Already. not only because the market is “overheated”.9 Assets / Capital 1. we can see that potential for an EVA increase is greatly diminished. Observe the following comparison for percentage increases to better gauge the situation: Table 16-5 Percentage Changes YEAR 1999 2000 Net Income 1668 1801 Percentage Change 7. THE EXTREME CONSENSUS METHOD . even a near risk-less stock will undergo a percentage cost of equity increase Thus.the cost of equity. the cost of equity will always be raised toward the end of a business cycle.1 5767.3 Percentage Change 13. However.66 percent change. but because the Federal Reserve raises rates to combat inflation.7 7767. we have a set-up for the “perfect storm”: the company does everything right. and net income. the requirements to improve EVA are stringent.411 Table 16-4 Assets / Capital YEAR 1999 2000 Assets 12816 14480 Capital 7019. because we would most likely need a decrease in the cost of equity to accomplish it.83 1. but the situation is untenable because retained earnings have built up to a high level right at a point where they are most expensive (comparatively). and now we can contrast it with the change in equity which was 13.97 % Stockholders' Equity 5093.86 CHANGES IN ECONOMIC PROFIT Despite a move to lower risk with greater equity. or growth. “P / E”: it is volatile and difficult to interpret and yet depends on the market. and more importantly. 1. price makes a small difference in the cost of equity. The chain of logic for this combination is as follows: both the “rule of thumb”. we utilize what this author terms “the extreme consensus method”. Gordon Model Determinations To make determinations with the Gordon model. the cost of equity for the Gordon model would reduce to (D1 / P) + G. An average price to compare with that dividend for the years. In reality. the two methods form a unique consensus: both market driven and dependent on internal dynamics. E / P is quite like its inverse. net income and stockholders’ equity. If we average the two percentages. while in the E /P method it is a determining factor. the two methods produce extreme comparative costs which are hypothetically linked by the difference between market and book values. but originally formed from the same function (with assumptions). a gauge of risk. When used alone. in the Gordon model. If we make the assumptions that “G”. Prospective dividend growth for 2000 (already known) and 2001. E / P method and the Gordon model can be derived from the same equation of P = D1 / (K – G). then ROE would reduce to E / P when both the book price and market price are equal. . 2. 1999 and 2000. and we equate the cost of equity with ROE. we can get a fairly comprehensive idea of what the cost of equity actually is. we need three measurements: 1. is the product of return on equity (ROE) and retention. On the other hand. The relationship between the component parts of each method is stable and will display changes in equity risk when the measure ascends. However. 3. The Gordon model is more fundamentally driven because of its dependence on the book values. Together. An average of growth over the past five years to be determined by the product of ROE and retention.412 To obtain a cost of equity figure. 2174 0.001 0.1806 Next.0799 0.205 0.413 Table 16-6 1999 Growth Year 1995 1996 1997 1998 1999 Average Retention Ratio 1 0.345 0.132 0.244 0.132 0.2257 0.345 0.249 0. Few people buy low and sell high and investors tend to buy spasmodically. However.001 0.371 0.54 0. increasing volume as a stock climbs higher.68 ROE 0. The average price in terms of range may be better than the mean as an indicator of the potential distribution of a stock because volatility is expressed throughout the range.244 0.371 Growth 0. A strict fundamentalist would certainly choose the latter method.67 0.39 0.2174 0. and it may be more proper to match the end of period price with the end of period earnings figure.248 0. while an analyst who was fixated on the process of change would pick the average range method. earnings are accumulated throughout a period.39 0.205 0.63 0.0799 0.322 Growth 0. an average range price is determined for the purpose of deriving an E / P figure and for input into the expected dividend yield. . “D1 / P”.63 0.1349 Table 16-7 2000 Growth Year 1996 1997 1998 1999 2000 Average Retention Ratio 0.67 ROE 0.54 0. 12. We now have two next expected dividends: $1.08 which is the next expected dividend for 1999. and $1. 1.185 57. Growth extrapolation tends to work well with dividends because firms pride themselves on steadiness. To obtain an expected dividend for 2001 that will be applied to the year 2000 (the next expected dividend).92) 1/4 = 1. The D1 / P figure for 1999 is thus.0408) = $1. paid in 2000.08 / 0.08. it is 1.0408 which becomes the growth rate multiplier.56 73.12 / 57.the actual dividend of $1. The expected dividend for 2001 is $1.92. they are:$1.25 AVERAGE 57.08.185 = 1. 0.89 % and for 2000.94 %. In respective order. we need to extrapolate a growth rate and multiply it by the actual current dividend.12 which is the next expected dividend for 2000.08 / 57.625 Since our hypothetical analysis begins in 2000.414 Table 16-8 Average Range Price YEAR 1999 2000 LOW 44.81 42 HIGH 69. These figures are then added to the respective ROE figures to complete the Gordon model: .625 = 1. We then make a ratio out of the dividends and use the inverse of four (1 / 4) as an exponent: (1. We need three pieces of data to obtain this rate: the current dividend. and 4. we have already obtained a prospective dividend figure for 1999 . the dividend five years ago and the number of years between five years.08 (1. “Rule of Thumb”.34 Mid-range Price 57.4 % 3. and the internal dynamics of the firm simultaneously.1359 0.34/57. . However.09 3.0189 + 0. with non-growth stocks that are relatively stable (most DOW components) the student/investor should observe that extreme consensus is a viable option. Earnings to Price or E / P Determinations Once the Gordon model calculations are completed. the CAPM would be used to derive a cost of equity.96 % 2. the rest of the method is remedial.0194 Growth (ROE x Ret. We merely match earnings per share for the year with the mid-range price.0189 0.625 E/P 3. The combination of the market derived E / P and the book value driven Gordon model can create a working cost of equity because they are functionally related. Table 16-10 E / P Analysis Year 1999 2000 EPS 3.1806 EQUATION 0.415 Table 16-9 Gordon Model Year 1999 2000 Next Yield (D1 / P) 0.625 = 5. The objective methodology is that which best captures both the changes in the market.1806 Expected Rate 15.09/57.0194 + 0.8 % In a normal analysis.185 = 5.1359 0.185 57.49 % 19.) 0. ECONOMIC PROFIT At this juncture.77 1058.1996)/2 12.8 % . Table 16-12 Economic Profit Year 1999 2000 Change 1999/2000 KimberlyClark Net Income 1668 1801 Cost of Equity 0. At these times. greater volatility.054 0. and an over-heated market pushed the cost of equity up in 2000.17 .1045 0.058+0. Kimberly-Clark’s low beta.1 5767.88 % Higher interest rates.1549 0.058 Gordon Model 0.416 3.1996 EQUATION Cost of Equity (0.45 % (0.054+0. However. as we shall observe in the next section. we merely plug in the numbers: Net Income .1549)/2 10.1288 Stockholders' Equity 5093.3 Capital Dynamic 1135. it may be better to move “sideways” than either up or down.[(%Cost of Equity)(Stockholders’ Equity)]. Create a Cost of Equity We now average the two figures for each year and compare: Table 16-11 Cost of Equity Year 1999 2000 E/P 0. the calculation of the capital dynamic is remedial. even the best companies may have limited options when constrained by the sequential economic decline of companies around them. low probability of default and adherence to a target capital structure were a saving grace. Comparatively.6. a 78 million dollar increase in net income would have provided the impetus for a greater increase in economic profit. OLD EVA = New Net Income [(New % Cost of Equity)(X)]. Kimberly-Clark would have increased economic profit with a maximum equity increase of only 71.82 . there was no way that the firm should ration capital to make a “paper profit”. and would have impeded financial flexibility. Thus. If the additional 748. We solve for “X. Declaring a special dividend would have diminished retention.72 feasible? Absolutely not.77 = X .59. Given the higher cost of retained earnings and the possibility of an economic downturn looming on the horizon.417 In the chapter on economic profit optimization. From the net income side. Then “X” = 5164. Hindsight can give us an idea of what the maximum equity could have been given the same increase in net income.3)].72 versus the 674.1288)(X)]. but the best defense was the route that Kimberly-Clark actually took: they began a series of share buybacks that not only diminished the book value of equity. it is probable that Kimberly-Clark may have raised much more capital than required. but took shares off the market as well TOO MUCH OF A GOOD THING .77 = 1801 .72). X = 1878. we stipulated that earnings are the liberating force for equity.1288)(5767.3 .” 1135. (5767.5093. Kimberly-Clark must raise a minimum amount of new capital.2 that actually occurred: (5164.1 = 674.2 that was raised is going toward projects with a positive net present value. raising the payout ratio would have committed the company to future dividend payments that may have been unwieldy.[(0. A superficial examination indicates that nearly all of its expanded sales could have been funded with less retained earnings.1 = 71.82 Thus. To fund existing sales.[(0. However.5093.2) Was the 71. we can determine the shortfall by using the same sensitivity analysis and making net income the “X” variable: 1135. type of industry. and honoring contractual demands can all stymie the best intentions of management. buying up shares and increasing the demand for equity. the firm had such a solid foundation.418 And herein lies the problem of boom and bust cycles. Since Kimberly-Clark was growing earnings at a 7. It also reflected the need to compensate shareholders with some compendium of benefits rather than retain earnings at a high price. While most models cover major variables like assets and income. market) are so uncontrollable that any improvement in the measurement becomes the goal. MARGINAL BENEFITS ANALYSIS No default model is perfect. While it did little to indicate movement toward an optimal capital structure (the firm was almost there) it reflected the high price of equity and the interface with the greater market. Thus.97 percent pace and not the 12. rather than some specific amount. avoiding some of the speculative excesses of its peers. each is developed in a different period which determines the overall effect of each variable. While some companies do commit financial management to risk-less. the market did not reward the firm with a stock price increase. Kimberly-Clark was “victimized” by the economic reality of an over-heated market. The cost of equity skyrocketed because investors were attracted to higher profits. Companies can and do get “painted” into corners. Historic allocations of capital. stable. The “internal” financing that KimberlyClark was implementing through retaining earnings was subject to the same market turmoil as other sources of funding.and represented a low risk component in any portfolio. timeliness in the business cycle. that its stock did not decrease either . Kimberly-Clark managed the situation as it arose. With record sales and profits.64 percent that was needed. there was little that Kimberly-Clark could do in 2000 to increase it. The return on equity was over 37 percent in 1999 versus the still very high 33 percent in 2000. However. a model developed in the 1970’s might emphasize asset values and . The capital dynamic proved to be a valid measurement of corporate risk. it was only as cost effective as comparisons would allow. many factors (interest rates. increases in EVA. Since EVA was so high in 1999. and Kimberly-Clark fit that bill BASIC METHODOLOGY For detailed information. see the chapter on marginal benefits analysis. at the juncture of 2000 . economic profit. The slight changes in marginal benefits. while a model created in 2008 would emphasize sub prime exposure. This is the point where the first derivative of the function is equal to zero and is the prime determinant of an optimal capital structure. 2) corroborating methods and analysis that confirm a finding and 3) the ability to observe change in the measurement. and the market price of the stock confirm the correlation value of the measurements. we can gauge year to year . the user must have: 1) some tolerance for imprecision as the derived figure is an estimate and not a decision tool. In effect. As in the economic profit analysis. the soundest judgment was to seek a stock that would minimize risk.tentative but stable. However. the years 1999 and 2000 were compared by entering the necessary fundamentals into a function. We measured the cost of bankruptcy and the inherent tax advantages in the decision to use debt by forming a marginal benefits equation.2001. We then observed any improvement when the subsequent period measurement was larger than the initial. As with the cost of equity. The student/investor sets up a function that equates the product of long-term debt and the tax rate. Altman. They are accurate enough to detect large moves in the probability of default which is their primary purpose. the difference between the measurements was slight enough to be inconclusive but matched the performance of the company . The function maximizes when the incremental change on both sides is equal to zero. The better generic models (Ohlson. Merton KMV) will agree within a few percentage points as to the probability of default but will not evince the same amount of accuracy throughout the entire range. The state of the company seemed to be in a “holding pattern” that would neither confirm nor deny it as an investment vehicle. Shumway.419 inflation. with the product of a projected amount of loss and the probability of default. However. 1 580. Since the word “loss” is subjective. requiring multiplication of the tax rate and the amount of long-term debt. we can also test the marginal benefits of interest in the same function and use the result as the numerator in an immediate benefits ratio. [(Interest Expense)(Tax Rate)] . the tax advantages of interest expense will be less than zero which will signify that the company should not have debt. TAX BENEFITS FOR KIMBERLY-CLARK The tax benefit calculations for Kimberly-Clark are the most remedial.29 Interest Expense 213 222 Long-term debt 1927 2000. In essence. we can look for improvement. there is room for interpretation.420 improvement simply by observing whether next period’s marginal benefits are larger than the initial period’s.38 578. an estimated optimum occurs when the ratio is maximized. we have two functions to look at: 1. but each default is a legal construct with unique requirements. Moreover. Table 16-13 Tax Benefits Year 1999 2000 Tax Rate 0.[(Probability of Default %)(Amount of Loss)] = X2 We look for improvements in X1.9 64. The calculation is replicated with interest expense. In certain cases.[(Probability of Default %)(Amount of Loss)] = X1 2. [(Long-term debt)(Tax Rate)] . we look for an improvement in X2 / X1.3 0. When we divide this number by the original marginal benefits figure. but more significantly. Our .6 (Tax)(Interest) (Tax)(Debt) 63.174 AMOUNT OF LOSS FOR KIMBERLY-CLARK Creating a generic amount of loss for input into a bankruptcy cost is a purely experimental endeavor. and then dividing by the number of shares outstanding.443 THE PROBABILITY OF DEFAULT Most default probabilities are very good indicators of risk. We algebraically eliminate the logarithm and solve for the probability. In effect. Tangible book value is constructed by subtracting all intangible assets. we need an algorithm which will curve upwards.421 construction of a “loss” is based on the loss to shareholders above a tangible asset value which is assumed to be the property of creditors.04 Market Price/sh. it is an average between the low stock price and high stock price for the year. Ln [P1 / (1-P1)] = X1B where .185 57. it works well.8 539. 7. and yet optimize in the domain of earnings. displaying increasing risk at an increasing rate. The Zmijewski model is simple to apply. goodwill and unamortized debt from assets.09 27276. Thus. but few will be both accurate and flexible enough for capital structure modeling. The market value per share is determined by the same method as in previous sections. few possess the inherent flexibility of this model. The construction of this function is (1-(Tangible Book Value / Market Value)) (Number of Shares Outstanding x Market Price per share).625 Number of Shares 539.22 Amount of Loss 27025.12 7. The basic logic behind the function is that the product of parameters and fundamental ratios forms the logarithm of a probability of default. In an experimental mode. and although there are more accurate default algorithms. The unique requirements of capital structure are variables that emphasize the consequences of increasing debt by allowing solution for long-term debt or equity. Table 16-14 Amount of Loss Year 1999 2000 Tangible Book/sh. 57. 422 P1 is the probability of default. X1 are the fundamental ratios.12438 Default Probability 0.479 Table 16-16 KimberlyClark Year 1999 2000 TL / TA 0.00129 = 0.069 -1.8286 NI / TA 0. and B are the coefficients of the algorithm. where we give negative values to the fundamental ratios. we turn the equation around and input P1 = 1 / 1 + EXP [-XB].129 % THE COST OF BANKRUPTCY To calculate our experimental cost of bankruptcy. To obtain a probability.384 0.13015 0. .45405 CA / CL 0. The following table contains a definition of the fundamental ratios and the coefficients of the algorithm.126 % 0. we multiply the amount of loss by the probability of default. Table 16-15 Zmijewski Default NAME TL / TA CA / CL NI / TA Intercept FUNCTION Total Liabilities / Total Assets Current Assets / Current Liabilities Net Income / Total Assets NONE COEFFICIENT 6.00126 = 0.92616 0. The result would be a figure that the shareholders would lose in the event of liquidation of assets.06 -9.45045 0. 187 Marginal Benefits 544.09 27276. each bankruptcy cost is subtracted from each tax advantage and the periodic figures are compared. we can also observe whether the immediate tax effects of interest are growing in comparison to the overall tax advantages of debt.052 35.848 29.187 MARGINAL BENEFITS The primary objective in comparative analysis is to test whether the tax advantages have grown relative to bankruptcy costs.048 544. there is little change in default.1 580.443 Cost of Bankruptcy 34.05486 0.where this ratio is maximized.9 and 64. but the Zmijewski fundamentals optimize in that domain .05357 The interest benefits were derived by subtracting bankruptcy costs from the tax advantages of interest (63. Table 16-18 KimberlyClark Year 1999 2000 Tax Benefits of Debt 578.174 Bankruptcy Costs 34. while a ratio was formed by dividing this figure into the marginal benefits of debt. Some algorithms will not permit this calculation.38 for 1999 and 2000 respectively). As would typify a company that is hovering around an optimal target. All indications from both economic .00129 Amount of Loss 27025.193 Interest / Debt Benefits 0.423 Table 16-17 Kimberly-Clark Year 1999 2000 Probability of Default 0.052 35. tax benefits or the amount of loss.00126 0. With the Zmijewski algorithm.987 Interest Benefits 29. 71(X1) + 0. ALTMAN’S Z SCORE: BOOK VALUE VERSION The standard Z score is set for market values and is widely available. it is unsurpassed. but as a simple measure of changing risk. Its linearity makes it untenable for use in capital structure models.10 (X3) + 0. The basic function is: 0. However. staying at the target would provide minimal risk with the possibility of special dividends and buybacks as well as the natural appreciation of the regular dividend.847 (X2) + 3. perhaps no default measurement has stood up better than Altman’s Z score.998 (X5).if indeed Kimberly-Clark wanted to appreciate its stock. CONFIRMATION As corroboration of our findings.420 (X4) + 0.424 profit and marginal benefits analysis would point to the necessity of moving away from an optimal target to seek more risk . We merely plug in fundamental ratios and then add up the score. and so the book value version is more conducive to observing the effect of capital structure variables. market values get inflated. Such strategic decisions cannot be taken lightly because future prospects need to be weighed against the potential temporary diminishment of shareholder value. A “Z” Score is the summed products of ratios and coefficients where a larger score is better because it indicates a lower probability of default. The following table describes the component ratios: . However. Any increase in Altman’s Z Score can be matched with corresponding default probabilities from other methods to see if they confirm one another. In the year of a debt issue. a large amount of assets will remain under performing Without a new infusion of debt. the firm must begin paying back and lowering the proportion of debt to equity.or begin paying off rapidly and increase its stock price. Although a firm can issue more debt in subsequent periods. there will be no tax benefits to counteract the change in default probability and the result will be a stock that falters. but greater risk occurs in the next period. Thus. At this point. . the Z score is both mathematically sound and integrated with capital structure theory.425 Table 16-19 ALTMAN'S Z SCORE COMPONENTS (BOOK VALUE VERSION) X1 = Working Capital / Assets***Working capital is current assets minus current liabilities.if the wrong projects have been implemented . the risk / return ratio for the investor rises because the firm can either wallow in low profitability . X2 = Retained Earnings / Assets X3 = EBIT / Assets X4 = Book Value of Equity / Book Value of Liabilities X5 = Sales / Assets Even if each element in the numerator stays strong and stable. if earnings do not improve enough to increase their respective numerators. increasing debt will increase each denominator leading to a lower Z score. tax benefits will balance the increase in default probability. 3 6574.6 13007 12816 Table 16-21 2000 FUNDAMETALS Working Capital Retained Earnings EBIT Book Value of Equity Book Value of Liabilities Sales Assets AMOUNT -784 7982 2844 5767.426 KIMBERLY-CLARK’S Z SCORE Table 16-20 1999 FUNDAMENTALS Working Capital Retained earnings EBIT Book Value of Equity Book Value of Liabilities Sales Assets AMOUNT -284 6764.6 13982 14480 .1 5772.6 2654 5093. The student/investor will notice that Altman’s figure agrees with Zmijewski’s.877 0.at the beginning of a downturn.364 For the year 1999. they need to make a move away from the optimal target and then back again. The firm is correct to play it conservatively which makes it an attractive addition to any portfolio.4567 can be compared with 2000’s number.427 Table 16-22 KimberlyClark Z Score YEAR (X1) (X2) (X3) (X4) (X5) Z Score 1999 -0. However. the number 2.528 0. indicating a higher probability of default. INVESTMENT CONCLUSION Kimberly-Clarks financial acumen is admirable.022 0. consider the timing of such a move .364.196 0.015 2. If it seems unusual not to take the risk needed to appreciate the stock. and few companies would accomplish that better.4567 2000 -0.882 1. The “plan of the day” in early 2001 was to diminish risk.054 0. evidencing slightly more risk. which declined to 2.996 2. They know how to gauge risk.551 0.207 0. in order for the stock to appreciate. (Back to Table of Contents) . he or she will have more options and greater success. the firm must overhaul its capital structure through acquisitions. it is purely a mathematical exercise devoid of numerous outliers such as sector rotation. However. When we compare that risk to other companies.428 APPENDIX: EXTRAPOLATED RISK . and stockholders’ equity.When “Normal” is too risky Investing on the basis of forecasts is tenuous. Businesses that adopt a contingency plan are flexible enough to “shift gears” when conditions change and therein lies the secret of corporate longevity. We make a ratio out of the last year in the trend (current) and divide it by the first year (five years ago). again.” Thus. “This is the risk to economic profit if things keep going as they are. Since the cost of equity is less affected by company trends. additional equity. some firms create an environment where economic profit cannot grow. we drop that figure and use a more typical year. the interaction between retained earnings. and capital formation is quite strong. the correlation between capital. We then use the . The construction of the geometric mean is simple. if the investor can adapt a contingency plan to each forecast. We look at each element over a five year interim. However. the economy. and it is in these cases that extrapolated risk is valuable. stock buy backs or even divestiture Just as we extrapolated next year’s dividend in the Gordon model. we are not forecasting per se. we are merely stating. If we spot a bad earnings year as either the first or last year in the period. capital. At this point. or the cost of capital. too much equity issued at too high a price will show up as a trend that can cripple the capital dynamic. we leave it as is. the emphasis is not on precision but to produce a growth figure that encompasses the trend in that element. and recognize its inherent fragility. However. contingency planning will help diversify a firm’s actions and lowers its overall risk When we use the geometric mean to extrapolate the three components of the capital dynamic. we can find a trend line for net income. We are following the trend line formed over a number of periods and determining the risk to economic profit. net income and equity is very high. The risk measurement is 1916.429 inverse of the number between periods as an exponent.6 / 6222)^0.1111. leaving the cost of equity as is.99 = 1125. Table 16-23 KimberlyClark Year 1996 2000 Net Income 1404 1801 Stockholders' Equity 4483 5767 Capital 6222 7767.1288)(5767)(1. This figure is a growth factor.25 GROWTH FACTOR 1.0642) . The student/ investor needs a calculator to use the fractional exponent. we divide the extrapolated cost of equity into extrapolated net income. In the case of a five year period.057 The extrapolation of the capital dynamic: 1801(1.25.62 . If the period were ten years. we plug in a company’s figures into the geometric mean function and then do a capital dynamic calculation on the extrapolated figures.[(0.0649)] = 1916. The growth function is: (Current Number / Number X periods ago) 1 / (X-1) .99 = 2. At this juncture.423 .6 Cost of Equity N/A 0.1288 Table 16-24 FUNDAMENTAL Net Income Stockholders' Equity Capital FUNCTION (1801 / 1404)^0.790.25 (7767. the number between would be nine with an inverse of 1 / 9 or 0. If we subtract “1” we obtain an average yearly percentage growth. To obtain a risk calculation for comparison.0649 1.0642 1. the number between periods is four and the inverse would be 1 / 4 or 0. this figure is the comparative capital dynamic for that company.63.62 / 790.25 (5767 / 4483)^0. And that is the crux of extrapolation . the economic profit may rise. and an extrapolated figure provides information on what a “normal” increase would be.to observe what is normal.Equity = Long-term debt = 2069. most firms desire a stable equity because the market price of the stock will be some multiple of the book value and stability helps maintain the difference. However. (Back to Table of Contents) . low beta firms who have steady incomes (like Kimberly-Clark).0649) = 6141. If capital is growing faster than equity. The technique is more applicable to low risk. with a risk factor of 1801 / 742. We multiply the growth factors by the fundamentals for capital and equity and then subtract to obtain an extrapolated figure for debt.424. A second implication of the analysis is the rise in long-term debt.28 Extrapolated Capital = 7767.21. a virtual tie that is described by the rate of growth of net income and equity.057) = 8210. Table 16-25 EXTRAPOLATED LONG-TERM DEBT Extrapolated Equity = 5767 (1. debt is not an absolute measure of risk. the firm is taking on debt.430 The current capital dynamic is 1801 .[(0. and decide whether “normal” may be too risky.07 Without reference to many other figures. but we can at least observe what a “normal” increase would be.1288)(5767)] = 1058. If the cost of equity goes down in 2001.6 (1.35 Capital .79 = 2. By observing capital structure decisions from trough to peak. performance is captured in both financial statements and stock charts. However.431 17 FULL STEAM AHEAD: AN ANALYSIS OF CONOCOPHILLIPS. Most cycles are distinguished by the interspersing of several sectors. the measurements are forthright. That ConocoPhillips would establish itself as a premier player in an industry that would dominate an entire business cycle was simply unforeseeable.2006 When a firm has earnings that are accelerating much faster than the market rate. While the price fluctuations of any natural resource can cripple commodity related businesses. However. industry players want to “lock in” a lower price and may profit from these decisions. The risks of growth after Phillips Petroleum merged with Conoco in 2002 were apparent.and ultimately. Inevitably. THE CONTEXT Rarely do we have a chance to associate a specific company with the performance of an entire business cycle. As oil prices rise. Not only do shortages have far reaching macroeconomic consequences. and the next sector to have earnings acceleration will . volatility in the oil market causes comparatively more risk. hedging oil related commodity prices is an important part of the business. less profit. signs of incipient and sustainable growth are more elusive. but an oil glut creates less competition within the industry itself . Any threat to supply. nothing prepared investors for the steep rise in oil prices that were a combination of both political risk and global demand after the war started. the student/investor can correlate individual corporate behavior with the forces that work on the greater economy. Consequently. will propel futures prices upward. each of which dominates one or two phases. or the war in Iraq. 2002 . Investment success was not guaranteed in an economic environment plagued by both recession and the preparation for war in the Middle East. whether it is it from OPEC. the fall from grace occurs when capital costs begin to eclipse earnings. Several risk related “hurdles” were deftly handled by ConocoPhillips: • 1. public outcry about “Big Oil” became large enough to warrant post-Katrina hearings in Congress. the former assistant Secretary of State sat on the Board of Directors at ConocoPhillips. To say that the oil companies had “friends” in Washington may be an unfair criticism or an understatement depending on to whom one speaks. For ConocoPhillips. ConocoPhillips was formed in the aftermath of the 911 attacks and increased tensions in the Middle East. leaving a recession trailing behind.the worst case scenario. the stock market “crashed” in 2001 and 2002. for example if oil were to be shut off . Oil companies manage risk as well as the best financial institutions simply because the nature of the business is so speculative. For lack of a better word. Phillips Petroleum did not have excessive debt on its balance sheet. they needed several plans to negotiate through the risk of Middle East turmoil: a plan. Exploring for oil is perhaps the archetypal risky business and calls to mind such images as Texas wildcatters and “instant” wealth. India and China were to become major economic players on the world stage and emerging markets began to grow at a furious pace. Secondly. Equity based companies were especially hard hit. • 4.432 be the “rising star”. In fact. Oil became a political issue. nor had they issued a lot of equity as some companies did in the late 90s. However. the meteoric rise in earnings was punctuated by five years of solid risk management that exploited a growing economy. In fact. Any . • 3. • 2. The demand for oil was no longer an isolated domestic concern. In the case of the merger between Phillips Petroleum and Conoco in 2002. Oil prices began to rise well beyond the point warranted by the risk in the Middle East. the risks were far more complicated than mere speculation: they had to foresee the economic recovery that was foreshadowed by a surging stock market in March of 2003. On the other hand. The Federal Reserve began to lower the federal funds rate to levels not seen in fifty years. the Federal Reserve lowered interest rates. When . In order to pay for the war in Iraq. Workers did not cause higher inflationary expectations and employers did not “bid up” the price of wages.433 extremes in price would entail more risk because changes in demand would have an exaggerated effect on variable costs on the upside or alternatively. A chain reaction occurred that became “the credit crunch” of 2007-2008 with a combination of high rates of foreclosures. The Federal Reserve lowered interest rates to historic lows and then kept them there. The government went on a spending spree. but information service jobs moved out of the country. This was a real estate speculator’s dream. In the mean time. a computer glitch can be fixed from India as well as from San Francisco. this recovery was more unusual than others in several respects: • 1. low beta stocks began to recover followed by riskier consumer discretionary stocks. Not only were illegal immigrants competing for low wage jobs. The crux of this anomaly lies in the alternatives proffered by a global economy. When housing prices soared. the dollar sank to comparative lows. • 3. and the firms that could most afford debt took advantage of lower capital costs. In the “normal” sequence. on unused capacity if the recovery were short-lived. and there was little wage/price pressure at the anticipatory level. the bubble began to burst because corporate earnings could not provide the equilibrium that would match higher home prices with worker income. manufacturing jobs moved to China where inexpensive labor could make a widget. Employers were slower to hire. The economic recovery that began in 2003 was classic in many ways. Anybody with a mortgage was refinancing at a lower rate. • 2. However. Interest sensitive stocks took off with housing leading the way. and housing derivatives (collateralized debt obligations or CDOs) that were worthless. Moreover. and then send it back to the United States. defaults on loans. the United States government floated large bond issues to China and ran up huge deficits simultaneously. like Michigan. Additionally. remained depressed throughout the “recovery”. Some areas. the U. began to fall by the wayside. not on the qualitative basis of better products and services. but it was the foresight to manage risk that allowed them to continue to profit. ConocoPhillips may have been in the “right place at the right time”. earnings were inflated above their true productive value. Income Sales Assets 2001 4937 25030 35217 2002 4953 57201 76836 2003 12638 105097 82455 2004 18713 136916 92861 2005 28297 183364 106999 2006 36704 183650 164781 . the case can be made: Table 17-1 YEAR OP. If the student/investor will observe just three fundamentals from this period (2001-2006). On the other hand. The stage was set for a recovery and expansion that were highly skewed. export market picked up.while others. and American goods were a relative bargain. areas of heavy construction like Las Vegas profited enormously from the housing boom. The greater risk in the oil industry was thwarted by counterbalancing it with less risk as a merged company.434 multinational “profits” were translated from an ascending currency to one that was depreciating. THE DECISION The decision to merge Phillips Petroleum with Conoco was a masterpiece of both timing and operating synergy. the economy favored two sectors throughout the business cycle . their collective response to a changing economy was integrated into specific capital structure decisions that positioned them so well. In fact. All sectors in the economy did not participate nor did the entire working population. like newspaper publishing.S.oil and housing .. but because the dollar was at an all time low. While such cyclical dominance does not bode well for the respective futures of these industries in the next cycle. On the other hand. In effect.714 in 2005.7107 in 2001. By the fourth quarter of 2002. was just 0. This increased productivity created a more even cash-flow. One benchmark of any investment is the time it takes to return it. allowing greater flexibility in funding but without changing the leverage characteristics toward greater use of debt. When any company purchases a large amount of assets. That scenario alone could have devastated shareholders because the remainder of the economy was beginning to “heat up”. ConocoPhillips “recovery” almost perfectly coincided with that of the greater economy. The slightly increased leverage from the combination of balance sheets reduced the potential for share issues as well as increasing tax benefits. Asset turnover (sales/assets). What happened? The character of the merged company was radically different than the old Philips Petroleum. notice that Phillips’ operating margin in 2001 was 19. there is usually a lag between the integration of those assets and profitability. while it was 1. Besides the tremendous surge in growth after merging with Conoco. the ConocoPhillips merger increased return. In the case of the ConocoPhillips merger.72 % (Op income/ Sales). the company would have almost “missed the boat”. any rising default probability quickly began to dissipate by the fourth quarter of 2002 as sales and profits picked up. It is at this stage. From a technical standpoint. operating income was surging. however. that stock prices decline or become stagnate because capital is tied up in a project with little current yield. Had the lag between asset purchase and profitability been a year or two. In fact. PRICE PERFORMANCE . failing to participate successfully in the first phase of the expansion cycle. in 2005. With a small sacrifice in marginal profit. it was only 15.435 The 2001 figures are for Phillips Petroleum alone.43 %. the returns were almost immediate. almost two and one half times as large. but reduced risk which is a rare and formidable combination. setting the newly formed company up for more profitability to follow. the merger decreased the capital intensity ratio (Assets/Sales) requiring less fixed assets per dollar of income generated. at the peak of growth. 99 116. but in this case.83 116. The task for the investor is to forge the transition between “debt state” and “profit or equity state” by recognizing the shift in capital structure. . debt was never prohibitively expensive.12 49.57 86.73 65.39 65. equity states were generated by more retained earnings and an attempt to maximize the efficiencies of the merger.61 END 48.51 32.33 65. the traders who speculated on the basis of momentum won the “jackpot”: the naive investor who keeps “playing the same number” was rewarded by an oil market that soared.71 ANTICIPATING PERFORMANCE: LEVERAGE STATES ConocoPhillips displays the classic progression from a debt oriented state in 2002 to a sequence of profit generated equity states in 2003 to 2005 back to a debt oriented state in 2006. assumptions about stock prices can blind us to the reality of defied probabilities.57 90.61 38.95 LOW 44. However.11 60.436 In the majority of business cycles a higher stock price reflects sector dominance which may span an interval from six months to two years.34 19.03 HIGH 63.36 73. We assume that a firm like ConocoPhillips can move toward an optimal capital structure for only a short time.31 64. While interest rates were increasing. Table 17-2 YEAR 2002 2003 2004 2005 (SPLIT) 2006 START 60.48 84.07 % CHANGE -19.66 45. which effectively reduced risk.36 71.92 32.11 56.78 84. Additionally. when acquisitions were made.6853 2006 1. This “miracle” of market timing produced a net gain in beta at the only time in the cycle when such a gain would be advantageous .5973 1.853 1.00519 1. first with Conoco in 2002. LTD/CAP % Op Mom. That deal required more equity than debt. were characterized by steady operating momentum within a narrow range.5345 2002 1. Assets/Cap 2001 1. but reduce the other risks associated with financial leverage. the stable relationship between sales and operating profit was upset because these new assets must be integrated into the existing structure.129 39.84 190.94 1.5864 2003 1. but it was soon buffered by profit generated equity so that long-term debt to capital decreased.0734 37. The merger was consummated with more debt. layoffs etc. operating momentum was raised during the initial year of the expansion (2003) even as financial leverage was lowered.0305 21.0766 32. In essence. and later with Burlington Resources in 2006. To ConocoPhillips’ credit. and will not “ratchet up” the probability of default. the “debt” years were characterized by increasing operating momentum which signifies that additional interest expense can be paid.437 Table 17-3 YEAR Financial Lev. This decrease allowed greater flexibility in the use of debt when ConocoPhillips needed to make an acquisition as they did in 2006 with Burlington Resources. However.52 -0.5504 The most profitable years (2003-2005).59 1.6261 2004 1.22 1.6265 2005 1.0179 16. although ConocoPhillips more than doubled existing long-term debt to .48 1.17 1.05 0.51 1.03 25. ConocoPhillips was able to raise beta during a period of market expansion. requiring restructuring charges.when the entire market is rising and sector performance is not yet critical Also noteworthy is how the merger left Phillips Petroleum’s’ debt structure relatively untouched. However. the following quarterly data indicates that speculation on ConocoPhillips’ leverage state would prove to be the correct and obvious choice. With that caveat. While leverage state investing offers better choices.: . The consequent LTD/CAP ratio went up to only 21. QUARTERLY LEVERAGE RESULTS The rational connection between speculative gambling and probabilistic investing is greater today than at any time in financial history. it is a crap shoot to assume that it will behave similarly in weeks four and five. It does the student a disservice to deny that speculation is sometimes lucrative and that “high rollers” exist on Wall Street who make it a professional career. as it was in 2006. The diminished ratio would have been more of a “judgmental negative” had the company or sector been suffering and the stock languishing. ConocoPhillips no longer needed to hedge the risk in long-term debt as they did in 2002. when it is decreased. only to make a bad debt laden acquisition in the third quarter.438 23091. Not all companies work to optimize their capital structures. However. the mathematics of short-term decision making favors the variation caused by uncontrollable random factors. In cases where both debt and operating momentum are increased. profits had been surging and the long-term debt to capital ratio was relatively low. anticipating a state (and the transition between debt and profit) is fraught with the same pitfalls as any short-term investment: the company will have higher profits for two quarters. Options trading.84 % from a diminutive 16. greater assets to capital acts as an additional plus for the investor rather than a negative indicator.94 %. Such increases will escalate the return on capital (ROC) and can offer a less risky source of financing. currency swaps and exotic derivatives have gained legitimacy because their proper use offers risk reduction and protection that an investor might not have otherwise. If one invests in a stock that is trending for three weeks. 32 4. the investor would be left “holding the bag’ with less long-term debt but a greater financial leverage ratio because the ability to pay interest expense would be undermined.1186 39.5864 1st QTR 2003 3651 27077 1.7308 Any fundamentals that mimic this type of relationship are not to be interpreted as worthy of speculation.31 N/A 1. we can also interpret the quarterly regression in 2003.145 37.55 1. with three data points dropped and replaced as a new quarter is added. it would be improbable that ConocoPhillips would begin incurring greater long-term debt to capital after reducing it in quarters one and two. INTERPRETING REGRESSION Without a full establishment of the securities market line (CAPM). These regressions are the precursors to the determination of a cost of equity. The firm is reducing its probability of default as well as lowering the financial leverage component of beta.439 Table 17-4 QUARTER Operating Income Sales Financial Leverage LTD/CAP % Operating Mom.0715 31.6196 4th QTR 2002 2065 20688 1. After taking on debt in 2002 to consummate the merger. Assets / Capital 3rd QTR 2002 1058 15678 1.057 2.48 1. However. we observe earnings pressure coincident with a decline in longterm debt to capital and the financial leverage ratio over more than two quarters. They are simply five year regressions between the percentage change in the stock and the market percentage. if profits began to falter in the third and fourth quarters of 2003. Such an assumption is problematic. We are searching for changes in R squared or .742 2nd QTR 2003 2756 25595 1. Such a transition state of “half and half” is much more difficult to read and is one reason why speculation should be avoided.0607 32.757 2.978 1. That anomaly was the outgrowth of price and demand effects that were in the oil companies’ favor at the time. will have a higher alpha. beta.288006 0.6477 (X) CORRELATION 0.995 % 4.5256 + 0. but depended on their position in the oil industry “pecking order”. and the correlation coefficient simultaneously.440 alpha that would indicate less or more dependence on the market.9668 + 0. Any industry that produces an especially valuable commodity that is in constant demand. and changes in beta which would be attributable to greater leverage.4405 + 0.2 % 5.3057 + 0. A “cartel-like” aura formed around oil which seemed to immunize it from cyclical downturns: The stocks would rise even if the market did not. while the market was increasing. while the other number is alpha. However. although public fears of collusion were forever imminent. In effect.408869 ACTUAL MKT % -9. The correlation is “R” and not R squared.was increased. ConocoPhillips increased alpha.237738 0. .3408 (X) Y=0.11 % 9. but the actual market changes for that period are posted.35043 0. As this initial recovery “surge” in the market began to sputter. ConocoPhillips began reducing risk to the point where profits were no longer systemic. non-systematic risk .03 % 7. Table 17-5 PERIOD 2002 1st QTR 2003 2nd QTR 2003 3rd QTR 2003 4th QTR 2003 LINE Y=0. and commands a high price.the correlation coefficient.5226 (X) Y=0.3331 + 0. the increase in beta is derived from increasing operating income relative to the market. Since the firm was reducing its debt to equity ratio at this time. The higher alpha indicates that the company is going to produce a return even if the market returns nothing and is a function of the integration of company with industry.3554 (X) Y=0. ConocoPhillips would continue to reduce financial leverage as well as its dependence on the market . The regression line itself uses five years of market data to establish a beta. “R”. the number that precedes “X” is beta.246158 0.as measured by alpha.3814 (X) Y=0.65 % In this display. and three out of four stocks will go up during a bull market. Table 17-6 YEAR 2002 2003 2004 2005 2006 BETA 0.331477 0.0725 0.505491 0.05 0.05 0.647655 0. the interaction between earnings. In this period.05 0.05 RISK-FREE % 0. and so that figure was established as a floor. The investor wants a comparative risk figure. Below are the tallied costs of equity for the years 2002 .048 Cost of Equity 0. • THE CAPM.0401 0. starting in 1998 for the 2002 figure and comprising 2002 to 2006 for the latter figure.0461 0.0427 0.0839 0. A monthly five year regression was performed.0682 0.the CAPM.05 0. for investment purposes.0429 0.2006. The average riskfree rates are tabulated as well.441 ESTABLISHING A COMPARATIVE COST OF EQUITY Rather than depending on one method of determining the cost of equity.823431 0. the market risk premium never exceeded five percent.0652 0. idiosyncrasies creep in that are identified with market skew or the specific situation of the company. However. we believe that the CAPM offers the most equitable method for most firms because of the tendency to depend on the market: Nine out of ten stocks will go down during a bear market. We show that a surging company like ConocoPhillips is resilient enough to harbor radically different costs of equity and still establish an increase in EVA. an opportunity cost that is best derived from a sample pool that comprises all stocks .690769 RISK PREMIUM 0. price and market can best be shown with three.08254 . When using the Gordon model or the E / P method. but is not a reliable measure of comparative risk with the market.58 75.442 • THE GORDON MODEL: The Gordon model seems to price the cost of equity too high. For that reason alone.617 0 0 1999 609 4599 13. but establishes a measure of performance above the market if the firm is doing comparatively well. The student/investor should be aware that the product of these two figures is added to the “expected” dividend yield .66 87.19 % 12.66 76. under performance will decrease it.55 2006 15550 82646 18.41 24. Table 17-7 YEAR Net Income Equity ROE Retention Ratio Growth 1997 959 4814 19.89 22.74 8.15 2005 13529 52731 25.82 85.51 5.the dividend expected in the next fiscal period. but unlike the CAPM. more earnings will increase the cost of equity. it can be used as a comparative figure with the CAPM. The reason that the measurement is placed well above that derived by the CAPM is that it establishes an internal benchmark for the cost of equity dependence on both the return on equity and the retention ratio for earnings -which may not reflect market averages.47 2004 8129 42723 19.56 81. The following tables give ROEs and retention ratios for the appropriate years.364 % 0% 0% 2003 4735 34366 13.62 10.36 16.88 2001 1661 14340 11.06 .76 2000 1862 6093 30.77 Table 17-8 YEAR Net Income Equity ROE Retention Ratio Growth 2002 698 29517 2. Like the CAPM.92 % 68.that is .03 84.24 43.84 16.59 % 1998 237 4219 5. The uncertainty of this measurement is derived from the volatility of the market which .39 % 2003 11.89 $1.443 Table 17-9 YEAR Average Price Next Dividend 2002 $54.965* $1. and the relationship between earnings and price carries a lot of information about the cost of equity.59 2005 12.18 2005 112.64 • EARNINGS / PRICE OR “E / P”: In any irregular market where stocks are either under valued or over valued (just about all markets).55 1. it is the capital structure equivalent of a P/E ratio. they are willing to bid up the price of the stock and the firm can easily attract equity.70* *Price: Indicates Split *Next Dividend .44 2006 $64.Indicates Author’s Projection We then assemble the growth information in a five year moving average and add each figure to the modified dividend yield for each year.85 2006 15.44 $0.20 $0. Table 17-10 YEAR Gordon Model 2002 9. However. Equity issues become profitable because the greatest amount of capital is raised for the least amount of shares issued.90 2004 $77. When investors value one stock’s earnings above another.6 2004 13.82 2003 $55. the E / P “rule of thumb” fails to work. 66 64.0652 -4.64 % 2005 0.02364 0. They are very similar measurements.55 112.2 1.8 77.72 54. reach a peak. and using this model with certain growth stocks or in times of excessive speculation.49 DIFFERENCING ROE AND THE COST OF EQUITY Table 17-12 YEAR ROE Cost of Equity ROE .89 7. That sub-cycle is analogous to how a firm attracts more capital with better earnings.Cost of Equity 2002 0.55 7. the relationship is not mathematically precise.57 % The performance of this measurement mirrors stock performance as much as the percentage gain in EVA.45 2006 9. Although the price changes on a daily basis.1882 0.1366 0.0825 10.45 55. a decrease in this indicator (or EVA) does not necessarily signify a decreasing stock price.1903 0.0725 6. although the correlation is . the risk of equity is more stable. will badly skew any comparison to companies of similar risk.32 % 2003 3.0839 10. The basic logic behind this correlation is that both price and earnings rise in tandem throughout a typical business cycle.22 2004 5.45 2005 9. and then fall when the risk of lower earnings eclipses price.0682 18.444 makes the pricing of equity unreliable as an opportunity cost.84 % 2006 0.44 6. Nevertheless.905 8.16 % 2003 0. in a surging company. Table 17-11 YEAR EPS PRICE E/P 2002 0.41 % 2004 0.2566 0. even this unreliable measurement will create a ball park figure for percentage increases in EVA because it tends to rise and fall similarly to the CAPM and the Gordon model. Unfortunately. and dividend discount models like the Gordon model will express an “expected return” that is more dependent on the internal dynamics of the company.35 2004 13.87 % 2003 11. Notice though.39 5. we recognize that the CAPM expresses a market comparative “required return”.03 2006 15.59 8. Since it is a concurrent indicator of stock performance.52 % 2. this indicator is similar to EVA and the ROE differencing that we displayed previously.254 7. the investor does not know if the figure has peaked and if the . the Gordon model) we can use the difference to compare the two risks .25 4.82 6. Table 17-13 YEAR GORDON CAPM DIFFERENCE 2002 9. while the CAPM is more dependent on forces in the market. similar to EVA. the Gordon model calculation will often be below the CAPM derived cost of equity. If the market is satisfied that even a decrease (as in 2006) will outperform the market. CONTRASTING THE REQUIRED RETURN WITH THE EXPECTED RETURN To take advantage of the difference between the CAPM and the Gordon model.445 strong. if we subtract the CAPM cost of equity from the dividend discount model (in this case. there is often upward pressure on the stock.64 8. we are measuring the acceleration of earnings in comparison to the cost of equity.39 % 6. that the figure in 2006 was above that in 2004 and 2005. This lack of equilibrium can be an investor’s best friend.6 7.386 While the difference does not represent a perfect rendition of pressure on the stock.85 6. When internal dynamics out pace the market. When a stock under performs the market. equity capital may still flow into the company providing that the near-term outlook is optimistic enough.2 2005 12. The growth factor in the Gordon model is derived from earnings and retention.market and internal. Effectively. A positive increase in EVA does not always translate to movement toward an optimal structure. At that time. but the size and stability of EVA can be a force that guarantees a growing dividend. both E/P and the Gordon model are valuable comparative tools to quickly locate a prospective investment. But . While bankruptcy costs are almost always minimized when the cost of capital is minimized. there may be times when the capital structure cannot conform to the market. EVA/capital dynamic would penalize the company for not taking on the debt. Moreover. if interest rates are especially low. Ex-post evaluation of several years of EVA increases is a proven indicator of viability. because it is dependent on the cost of capital rather than the cost of bankruptcy. the CAPM should be chosen to calculate the cost of equity. In this regard. Thus EVA would decline even as the firm attempts to minimize the cost of bankruptcy. . In this case. rather than an academic illustration. a full regression is done and a precise assessment is made. For example. we are more concerned with the movement of EVA that its absolute size. the three methods of calculating the cost of equity reveal the same earnings/risk pressure. THE EVA / CAPITAL DYNAMIC Institutional investors such as pension funds must diversify risk and search for companies who have stable cash-flows. Not only do increases correlate with positive changes in stock price. the investor needs to follow the market imperative and recognize the signs of a market “top” (see the chapter on the business cycle) as well as examining leverage positions. For a surging company like ConocoPhillips. the probability of default would be too high and would eclipse the need for more tax benefits.446 stock will subsequently decline. In essence. a company may need to shed some debt instead of taking on new debt at a lower rate. in capital structure.a judicious use of E /P in firms with a P/E right around the market average can identify companies that might warrant further examination. because the tax advantage would outweigh the decline in net income. However. if the required decision is investment grade. if stockholders’ equity balloons to new heights.0652 -1227 2003 0.0682 9933 2006 0. This high benchmark will place the calculation very close to ROE and produce a small EVA.0725 2243 2004 0. the standard EVA calculation will reveal that the company must retain the correct amount of earnings and pay out a large enough dividend. this method will exaggerate the increase. periodic comparisons of this method indicate the competency of financial management because there is a low threshold for error. Similarly. if dividends grow at an exorbitant pace in comparison to price and earnings. However. The reader will remember that the calculation is: Net income -[(% cost of equity)(stockholders’ equity)] Table 17-14 YEAR 2002 Net Income 698 Stockholders' 29517 Equity 2003 4735 34366 2004 8129 42723 2005 13529 52731 2006 15550 82646 Table 17-15 1.0839 4545 2005 0.0825 8732 .447 In the case of the Gordon model. CAPM YEAR CAPM EVA 2002 0. the deficiency will be revealed in a smaller EVA calculation. For example. The following tables itemize net income and equity for ConocoPhillips and then apply the three different methods to determine the EVA/capital dynamic. in comparison to net income and stockholders’ equity. they would have missed out on another thirty percent gain: the year 2005 turned out to be ConocoPhillips’ biggest yet with a stock split.1359 2323 2005 0. How would an .0939 -2074 2003 0. GORDON MODEL YEAR GORDON EVA 2002 0. Therefore. before totally dismissing the E/P indications of movement toward an optimal capital structure (an increase in EVA). by the end of 2004. and record sales and profits. he or she would have picked up on ConocoPhillips in early 2003. the company would have had two years of over thirty percent gains. NAIVE EXTRAPOLATION If the capital structuralist were astute. However.116 749 2004 0.448 Table 17-16 2. the probability of default was virtually unaffected. Had investors hesitated.1285 6753 2006 0. “smart money” investors would surely take profits. The marginal benefits function will later reveal that indeed the tax benefits of the acquisition of Burlington Resources far out weighed any increase in bankruptcy costs.the proportion of equity was building up through retained earnings. more in depth analysis needs to be done.0622 2597 2004 0. EARNINGS / PRICE (E/P) YEAR E/P EVA 2002 0.0745 4946 2005 0. The market soared and the firm was positioned to take advantage of the surge.0845 9073 2006 0.0132 308 2003 0.1564 2624 Table 17-17 3. At this point. The leverage position was solid in an expanding economy .0749 9360 The student/investor should notice how the CAPM and Gordon models both “punished” ConocoPhillips for issuing too much equity in 2006. Anytime that we compare quarterly performance one year apart. The Federal Reserve was still raising rates to hedge inflation. As in late 2002. the calculation is the same that we use for yearly periods: [(This Period’s X / 5 Period’s Ago X)^0. analysts would have perceived the inordinate demand for oil and extrapolated it into a usable forecast. the collective judgment of professional analysts offers some solace. the economy was heating up.25] . investors who watched ConocoPhillips would have been in for “the ride” in 2005. Therefore. While no indicator can guarantee that the firm will not begin accelerating equity through stock issues as soon as an investment is made. Although one cannot anticipate a positive corporate movement after years of stock gains with leverage theory alone. The increase in earnings was a 66. If we added that information to our own naive forecast. we are testing to observe whether or not earnings are growing faster than the total cost of equity. The naive extrapolation procedure requires us to find the geometric mean of growth for both earnings and the total cost of equity in the near term. the probability of an EVA increase is greater when earnings are leading equity. The war in the Middle East did not produce uncontrollable supply and demand issues .just a steady rise in oil prices. Thus. . the effect is the same as separating the fundamentals by five periods with four intervals between periods.449 investor play this dilemma? From a leverage perspective. Investors who judge primarily on the basis of probability (this author) would have taken profits. In this case.which could not go unnoticed even if the prediction were off by half. their respective economies were booming.42 % gain in 2005 . the probability was against another banner year. India and China were in the throes of competition for more oil. there was a confluence of risk factors in the firm’s favor.1 = Percentage Growth per Period. and so analysts’ consensus expectations become a benchmark for future EVA forecasts. Except for during an economic downturn. Cost of Equity: [(3584/2492)^0. but stockholders’ equity is a cumulative total. The major advantage of naive extrapolation is that we can gauge the relative size of component changes and then compare them to analyst’s earnings forecasts.52 The reader will note that net income accumulates throughout the year to determine an annual figure. naive extrapolation mirrors the size of the increase in EVA and assumes that next year’s EVA will be like this year’s. For example. we use the cumulative yearly totals and treat them as a growth progression made over four periods.51 % 3.51 =1.0725 2492 4th QUARTER 2004 8129 42723 0. financial management does not want equity to radically shift and potentially dilute the price per . if the fourth quarter in 2003 had a restructuring charge. While earnings can be volatile. Rather than resort to using the third quarter figure. simply using the yearly figures for 2003 and 2004 will smooth out any errant noise. Unfortunately. earnings might be negative for that quarter only. Comparative Ratio: 14.1 =14. we never assume that the cost of equity will decelerate. Since a quarterly net income figure might not be part of an active trend.1 = 9.47 / 9. Net Income: [(8129 / 4735)^0.0839 3584 The three calculations are as follows: 1.25] .47 % 2.450 Table 17-18 QUARTER NET INCOME (Cumulative) EQUITY (Cumulative) CAPM % TOTAL COST OF EQUITY 4th QUARTER 2003 4735 34366 0. the cost of equity is more stable in the near term.25] . ConocoPhillips issued the precise amount of debt to forestall an increase in the probability of default and increase tax benefits at the same time. periodic changes are made when it is cost effective to issue new stock. However. ConocoPhillips issued too much equity and had the capacity to increase debt. but remains a concurrent indicator. EVA always optimizes on the basis of the lowest cost of capital. more debt would have would have undermined the increases in assets and net income. from the perspective of marginal benefits. THE MARGINAL BENEFITS FUNCTION The market can temporarily misprice risk causing a divergence between the cost of capital and the probability of default. and ConocoPhillips effectively optimized long-term debt in the domains of tax . The need for balance distinguishes the function. increasing both long-term debt and equity. the marginal benefits function anticipates new dynamics because it maximizes when the change on each side of it is equal. However. In fact. ConocoPhillips bought Burlington Resources.451 market share. and does not portend a change in the income/capital relationship. The institutional imperative would have been to issue more debt. it should be avoided. for example. In 2006. ConocoPhillips issued stock as a method for purchasing Burlington Resources. From an EVA standpoint. If long-term debt declines. that figure declined. On the other hand. Over the long term. In essence. However. such disparities will work themselves out. more debt will incur a higher cost and increase bankruptcy costs simultaneously. Although the potential to increase the EVA/capital dynamic was apparent. This analytical dichotomy between EVA increases and the marginal benefits function was illustrated in fiscal year 2006. and those changes are unpredictable. we would look for a reduction in either stock price or default probability on the other side. only those who spoke to company officials would have predicted the large equity issue. if lowering the cost of capital increases the cost of bankruptcy. With ConocoPhillips low longterm debt to capital ratio of approximately twenty-one percent. a rising marginal benefits function in the domain of a declining EVA may have upside potential. 479 6. for help with the mechanics.421) = 4529 2006: (23091) x (0.421 53093 16439 2006 0. To reiterate the equation: Marginal Benefits = (Tax Benefits) – [(Default Probability) X (Amount of Loss)].451) = 10414 .0944 129.06 Table 17-20 YEAR Total Liabilities / Total Assets Current Assets / Current Liab. Table 17-19 2005 . ESTABLISH THE TAX BENEFITS 2005: (10758) x (0.4962 0. we will go through the step by step calculations for 2005 and 2006. respectively.9182 0.452 benefits and default probability.2006 DATA ZMIJEWSKI FUNCTIONS Intercept Total Liabilities / Total Assets Current Assets / Current Liabilities Net Income / Total Assets ZMIJEWSKI PARAMETERS -9.9484 0.1 (without split) 1706 23091 0. Net Income / Total Assets Midrange Stock Price Number of Shares (millions) Long-term Debt Tax Rate (decimal) Unamortized Debt Intangible Assets 2005 0.1264 98.215 1455 10758 0.4912 0.069 -1.384 0. The reader is referred to the chapters on Analytical Tools and Capital Structure.451 80940 32439 1. To exhibit these concepts. cost overruns coupled with slow integration will require .060)))] = 1 / [1 + (EXP 6.16 By maintaining adequate earnings and issuing the right amount of debt.91 2006: 10414 .3818874)] = 1 / 592.4912)x(6.169 % 2006: 1 / [1 + (EXP (9.060)))] = 1 / [1 + (EXP 6.16439 .((0.13 / 129.(25.215) = 142. Such a task is difficult to accomplish so soon after a major acquisition because it entails the rapid assimilation of the other firm’s assets while increasing profitability. If less debt is incurred in the next year.6 2006: Potential Loss [1 .384)) .001689069)) = 4350. ESTABLISH THE PROBABILITY OF DEFAULT 2005: 1 / [1 + (EXP (9.00169885 = 0.453 2. the premium will be on increasing stock price while decreasing the probability of default.((105437)x(0. ESTABLISH THE TANGIBLE BOOK VALUE PER SHARE 2005: (106999 .(0.32439 – 80940) / 1706 = 30.13 3.0944)x(1. Often.75 / 98.215)] x (142903) = 105437 2006 Market Value = (1706)x(129.069)) + (0.244.479) .3778)] = 1 / 588.9182)x(0.1) = 220.903 2005 Potential Loss: [1 .75 2006: (164781 .82 4.1264)x(1.6) = 168842.00169885)) = 10127.(30.169 % 5.042 = 0.4962)x(6.((168843)x(0.069)) + ((0. ConocoPhillips created a merger that did not increase the risk of default.9484)x(0. COMBINE THE EXPRESSION 2005: 4529 .1)] x (220244.384)) .6334 = 0.((0.479) -((0.001689069 = 0.53093) / 1455 = 25. ESTABLISH THE AMOUNT OF POTENTIAL LOSS 2005: Market Value = (1455 )x(98. THE COMPARATIVE CAPITAL DYNAMIC: GAUGING UPSIDE POTENTIAL The relationship between earnings and the cost of equity cannot improve indefinitely. To reconcile the risks of the merger would entail examining Burlington Resources past profits while itemizing fixed and variable costs and looking for synergy. The result? A higher proportion of long-term debt to capital. but that calculation cannot predict future profitability either.454 more capital and retained earnings will not be adequate to cover the deficit. In most cases. the only recourse for the analyst is to do a comparative history of the capital turnover ratio % ∆ SALES / % ∆ CAPITAL and ensure that the company is competitive. there is little ability to detect a downside shift in earnings and almost no ability to predict an equity issue. The actualization of the optimum remains a function of earnings accelerating faster than the total cost of equity. While we attempt to compensate for the information deficit with growth rate comparisons and leverage state analysis. We can conclude that none of the added asset value was reflected in EVA. An even more detailed examination would entail producing a time line for implementation. Throughout this text we have emphasized the point that stock price maximization occurs when the target capital structure is reached. a better method might be derived from using the comparative capital dynamic. Net Income / Total Cost of Equity is an adaptation of EVA that offers a chance to compare magnitude. When a firm takes on debt. The ratio of the absolutes. This is one more reason that debt laden leverage states tend to be sequential. ConocoPhillips’ improved potential was more apparent in the marginal benefits function. Just as effective.an informal type of “guidance”. or % ∆ Net Income / % ∆ Total Cost of Equity >1. is to weigh the estimates from several company sources . but prone to exaggeration. The reader should also note that the comparison between 2005 and 2006 entailed an adjustment for the 2005 stock split by putting the 2006 price on the same level. its comparative capital dynamic (CCD) tends to be smaller than when it . Although EVA increases are concurrent with stock price appreciation. Also significant would be the prevailing level of interest rates in the economy. each industry has a different standard of CCD based on the relationships between three specific types of returns: return on assets (ROA). In effect. In 2006. then the CCD will tend to be larger.9 32.34 2006 15550 6822 2.363 . Although the stock still appreciated.455 is growing through retained earnings. earnings must be growing to produce more retained earnings. both of which have very high CCDs. Later in the business cycle. more earnings will enlarge the CCD until it comes to a peak. If they are generally low. As the reader will observe from the data.51 % 2003 4735 2492 1. return on equity (ROE). the prime function of the CCD is to do intra industry comparisons. the firm acquired Burlington Resources with stock and debt that diminished the CCD.28 18.71 . much of the gain was from the particular circumstances in the oil market with rising prices and pent up global demand.61 2005 13529 3596 3. by default. where more financing is cost effective from an equity issue.76 39.27 32. and return on capital (ROC). the analyst must be careful about comparing different financial structures like insurance or investment banks.19. In the case of ConocoPhillips.92 2004 8129 3584 2. with a high turnover CCD such as a restaurant chain. Table 17-21 YEAR Net Income Total Cost of Equity CCD % Stock Price 2002 698 1925 0. two for one. the CCD hits a peak where earnings become more risky. and comparisons between intervals for a specific company. Nevertheless. While some inter sector comparisons can be made. the CCD peak was reached in 2005 when the stock split . if the analyst has an idea about the direction of earnings components with forecasts. We do not have to sell a stock if analysts’ forecasts are less than the time progression. The following equation is a “plug in the numbers” function that works easiest if done from right to left: (1-A) x (B + (C x (B .456 Since there are no industry standards. For example. another technique can produce a fundamentally derived forecast that compares favorably with naive extrapolation. if ConocoPhillips rose back to above a “4”. the investor might question whether the next year would be “bearish”. and . expense) x (1-tax rate)) / Assets Interest Expense / Interest Bearing Debt Effective Tax rate expressed as a decimal . Like P/Es. these figures rise and fall but have more stability and rationality behind them. Moreover. EARNINGS PRESSURE If we compare naively extrapolated earnings growth rates with analysts’ forecasts. the researcher is left to interpret what “too high” a number would be. but we need to recognize that changes occur after earnings acceleration peaks and we need to assess risks more closely. However. we obtain information that contrasts a time progression (naive extrapolation) with an integrated demand forecast (the analysts’). we always assess earnings pressure in the domain of analyst’s forecasts because consensus opinion is more informed than any time series can be.(D x (1-E))))) Table 17-22 FUNCTION A) Payout Ratio B) ROA C Interest Bearing Debt / Equity D) Interest rate on Total Debt E) Tax Rate EXPANATION Dividends Paid / Net Income Net Income +( (int. if accountants tell the analyst that the effective tax rate is estimated to be 39 % this year. When estimates undermine this growth rate.04 0. the equation can be put into a spreadsheet and a trailing twelve month figure can be produced for net income and the retention ratio so that each quarter the equation is updated with new information.457 The equation will produce a growth rate. If interest rates are going up by twenty-five basis points a quarter.451 . we assume downside risk and earnings pressure. If analysts’ expectations exceed this growth rate. we produce a concurrent growth rate and compare it with analysts’ estimates. The following was an assessment made at the end of 2006: Table 17-23 PRELIMINARY COMPONENT Net Income Interest Bearing Debt Stockholders' Equity Tax Rate Interest Expense YEAR 2006 15550 27134 82646 45.09799 0.451))/164781 27134 / 82646 1087 / 27134 0.1 % 1087 Table 17-24 FUNCTION A) Payout B) ROA C) Debt /Equity D) Interest Rate E) Tax Rate CALCULATION 2277 / 15550 15550+((1087)(1.3283 0. In essence. we look to it as a possible investment vehicle. which will be current when guidance about the direction of the components is given. that is one piece of information.1464 0. that is another piece of information. For example. However. as an investor.451 RESULT 0. 10495 or 10. Thus. any investor should do a thorough long term assessment because analysts’ estimates undercut our own growth figure. Earnings per share for ConocoPhillips were $9. (Back to Table of Contents) .495 %.458 Plugging these numbers into the equation yields a growth rate of 0.1 = -9.89 per share in 2007 for a decrease of (8.89 / 9.29 %.80) . Analysts’ estimates are $8. There is downward pressure on earnings.80 per share in 2006. 129 2003 12638 105097 82455 844 16340 34366 50706 4735 13. taxes.2006 Table 17-25 YEAR EBITDA SALES ASSETS INTEREST LTD EQUITY CAPITAL NET INCOME ROE % RETENTION% LTD / CAP FIN.62 32. The standard methodology is to use EBIT only.03 2005 28297 183364 106999 497 10758 52731 63489 13529 25.22 1.58 75.459 APPENDIX: SELECTED FINANCIAL DATA .0735 2002 4953 57201 76836 566 18917 29517 48434 698 2.0305 *The financial leverage ratio was determined from EBITDA (earnings before interest.66 76.84 25.52 % 1.05 1. depreciation and amortization).03 84.82 85.0766 2004 18713 136916 92861 546 14370 42723 57093 8129 19.17 1.66 87.36 21.89 16.364 0 39.0179 2006 36704 183650 164781 1087 23091 82646 105737 15550 18.94 1.84 1.CONCOCOPHILLIPS 2001 . LEV* 2001 4937 25030 35217 338 8610 14340 22950 1661 11.74 37. (Back to Table of Contents) . EPS and P/E. a company with more operating leverage but less financial leverage would be able to increase its capital dynamic by making large increases in net income. In another allusion to psychology. and we do so by making constant comparisons. Analogously. Early stages would favor wealthier companies who fund with debt and could take advantage of lower rates. The capital dynamic/EVA is by no means the perfect measurement either. indicating a “value buy” when the figure is comparatively low. “normal”. is often used to compare companies within industries. when we define one dollar as a “small” amount. our two dependable measurements. In theory. Since these companies would presumably be in different sectors. for better or worse has become the universal comparative measurement in corporate finance. which would have a comparatively small reaction to interest rate hikes. “reference dependence”: one dollar and ninety-nine cents is only meaningful. For example. Both of these measurements suffer from what behavioral finance calls. Thus. That relationship should increase the capital dynamic even though net income might be comparatively small.460 18 MICROSOFT VERSUS CONOCOPHILLIPS: COMPARING COMPANIES IN DIFFERENT INDUSTRIES Earnings per share. then a ninety-eight percent increase is considered “small” and sub-standard. a company with low operating leverage should have a steady enough income to afford more debt. and a ninety-nine percent increase as “large”. the phase of the business cycle would determine a firm’s respective cost of capital. it can compare two companies with different operating and financial leverages and put them on an even keel. give us little information to distinguish the quality of what is measured. the P/E ratio. . If ninety-nine percent increases pass for “normal”. we also need to define what is termed. and build up EPS through a more limited use of equity. Late stages in the cycle would favor the equity financing company. Its offshoot. dividends. On the other hand.producing a “virtual” symbolic variant of the “real”. One company spends billions of dollars combing the earth for a resource that literally fuels the world’s production lines . and even the holiday season. capital expenditures. generating income with equal parts of profit margin. Our contention in this chapter is that the adaptations of EVA offer more information than either earnings growth or P/E calculations. APPLES AND ORANGES: MICROSFT VERSUS CONOCOPHILLIPS Few companies are as different as Microsoft and ConocoPhillips. The following tables display the comparative fundamentals for the years 2005 and 2006: . The other spends billions to fuel those same lines intellectually . Both of these structures are representative of their respective industries. at different intervals. ConocoPhillips is comparatively well balanced. earnings . would dividing EVA by total assets be a better comparative figure than EPS? What would the results be if we put EVA on a per share basis? In fact. characterized by different amounts and volatility of cash-flow. there can be no universal.461 Several questions need to be addressed: accounting for the size of the company. Microsoft depends almost entirely on profit margin for its net income. Its shunning of debt financing has been legendary. Naturally.physically. and that they can be used to compare companies in entirely different industries. the two companies have entirely different capital structures in place. prices may track sales. asset turnover and equity multiplier. comparative indicator because the complexity of stock market price increases does not warrant it. 3596 = 9933 2006: 15550 .82 22.7345 = 23.73 2277 As detailed in the chapters on Capital Dynamics and Analytical Tools.462 Table 18-1 MICROSOFT YEAR Sales Operating Income Assets Net Income Stockholders' Equity CAPM % Cost of Equity Growth* Shares Outstanding Dividends Paid* 2005 39788 15416 70815 12254 48115 11.11 31. The growth rate of retention multiplied by ROE is adjusted for this payment.25 16.65 1639 2006 183650 36704 164781 15550 82646 8.7345 = 18.6818 = 8732 . Table 18-2 CONOCOPHILLIPS YEAR Sales Operating Income Assets Net Income Stockholders' Equity CAPM % Cost of Equity Growth Shares Outstanding Dividends Paid 2005 183364 28297 106999 13529 52731 6.0.07 10062 3345 *Microsoft paid a special dividend in 2005. ConocoPhillips 2005: 13529 .[(CAPM % Cost of Equity) x (Equity)] = EVA.55 1416.47 % x 0.06 1609.7 10710 36112 2006 44252 17375 69597 12599 40104 10. the EVA for both companies is determined as follows: Net Income .15 25.42 % x . 28 % Ultimately. but ConocoPhillips is by far the better company based on EVA per share.85 CONOCOPHILLIPS 9933/106999 = 9. On the other hand. Microsoft is the superior company based on value per dollar of assets. we can divide by both asset value and the number of shares outstanding. we face a similar dilemma that we had with earnings. The fundamental difference is derived from the quality of their respective assets and the method by which they are funded. Table 18-3 ASSETS 2005 2006 Table 18-4 PER SHARE 2005 2006 CONOCOPHILLIPS 9933/1416.6 = 7.463 Microsoft 2005: 12254 .73 % 8544/ 69597 = 12. is a twenty percent gain in a small EVA better than a ten percent gain in an EVA that is twice as large? To put the EVAs on a common basis.7 = 5.64 8544/10062 = 0.42 MICROSOFT 6889/10710 = 0.5365 = 6889 2006: 12599 .28 % 8732/ 164781 = 5. we are unsure of the magnitude. ConocoPhillips has many tangible assets because oil exploration produces a physical commodity requiring a heavy investment in machinery.01 8732/1609. Microsoft has comparatively little tangible asset value per share because it produces its income from intellectual property and places a premium on managerial and programming talent.30 % MICROSOFT 6889/70815 = 9.4055 = 8544 COMMON GROUND While we believe that the percentage gain in the capital dynamic / EVA will mirror a shift toward an optimal capital structure. The exaggerated growth cycle in technology during the . we need to ask ourselves the question. we are unsure how the market will value the measurements. For example. and more market dependent as well. The two companies face polarized risks as well. the “R squared” component of the regression. For ConocoPhillips. the capital intensity ratio is much greater for Microsoft. And . competition has produced enough substitutes (Linux. To illustrate this dichotomy. they buy back shares with retained earnings. . a market glutted with innovative software can reduce effective demand. Although the demand for a quality product remains strong. Alternatively. software production requires a level of expertise and technology that produces more fixed costs. For Microsoft. Additionally. requiring a higher percentage of fixed assets. In fact.although it is low and stable in comparison to most high tech firms. The correlation with the market. After the Conoco merger in 2002. ConocoPhillips has a balanced approach toward financing. but the local university probably has quite a few individuals who are capable of producing a commercial software product. Although oil production is associated with drilling and machinery. political turmoil and scarcity affect the potential supply of their product. keeping shares to a minimum and expanding with retained earnings and some long-term debt. This “democratization” of technology has afflicted both the personal computing and software industries with over production. Open Source Code) to undermine the pricing power of major software developers. the unique position of oil as a scarce and valued commodity has changed the dynamics of the cost of equity for most oil companies. the beta for its stock is affected almost entirely by operating risk. the investor would choose between the more risky growth of Microsoft and the steady dividend of Phillips Petroleum. the risks of legislation. Since Microsoft carries no financial leverage. At one time. consider the following: most individuals will never see an oil well in their home towns. the rise of oil companies have provided both rapid growth and dependable dividend income. Microsoft’s beta is about twice as large as ConocoPhillips’. while technology stocks have mirrored the peaks and troughs in the economy.464 1990s required Microsoft to issue a large amount of equity for which it has been penalized. 343046 BETA 0.373933 1. the “alpha” component has risen.465 has steadily declined. In effect.691 x Table 18-6 MICROSOFT YEAR 2005 2006 ALPHA 0.380048 REGRESSION Y=0. and will propel a high beta stock further upward: Table 18-5 CONOCOPHILLIPS YEAR ALPHA 2005 0. ConocoPhillips’ stock has taken on a “life of its own”. more debt raises risk.112178 R SQUARED 0.112 x While stock prices mirror EPS increases in the short-run.633924 -0.28+1. even as it becomes more dependent on the commodities market.374 x Y= -0. beta.343+0. while a higher beta is not synonymous with equity financing.1030512 0.32198 0.0766277 REGRESSION Y=0. while the non-systematic risk element.) how will an increase in EPS translate into an increase in market price? The greater amount of information used in EVA calculations eliminates some random . it is the size and stability of earnings that will help maintain those gains.28049 BETA 1. an examination of each firm’s regression lines for 2005 and 2006 reveals some of the truth behind their capital structures. It rises without being dependent on stock market volatility. economic outlook etc. The considerable adjustment for the amount per share and the number of shares outstanding makes a comparison between the two firms an exercise in probability.505 x Y=0.657499 2006 0. given a set of variables (market cap.657+0.505491 0. Moreover.690769 R SQUARED 0.634+1. Given the exigencies of their respective industries. but requires them to restrict the application of retained earnings to periods when the cost of equity is relatively low. their target structure contains zero debt. the need for diagnostic software that is contingent on the demand for a new operating system or chart making software that can tie into an Excel spreadsheet Thus. Since that constraint is tantamount to capital rationing. In this regard. Microsoft may face more volatility and a dilemma not encountered by companies with financial leverage: it can do all projects with a positive net .466 variation. without any financial leverage. each firm is well managed and made moves to optimize their capital structures. The market decides how it will value each security. the investor needs to look past the immediate demand for oil as well as the over production in software products. some arbitrary standard will be set: the market will decide that growth in some industries is more desirable and sustainable than in others based on patterns of demand and the cost of capital. Microsoft’s optimal structure will depend on the quality of its operating leverage and the management of its equity. the analyst needs to use the EVA/capital dynamic as a single arbiter. Microsoft can invest in both new software and the physical world it represents . for example. Consider. In effect.innovative applications multiply the demand for even more software. On the other hand. Moreover. Although the surge in ConocoPhillips’ stock mirrors the performance of Microsoft’s a decade ago. pricing in the risk from an array of economic factors. A change-over to alternative fuels would require extensive risk-taking and investment that would not immediately translate into profits. Microsoft is infinitely more flexible than ConocoPhillips. Like the auto-makers in Detroit. For myriad reasons. but the validity of prediction is still based on the dynamics of the market place. use of a comparative marginal benefits function is eliminated from the analysis because Microsoft takes no tax benefits. the future prospects for these firms may not be an extension of their present economic realities. During the recovery and expansion phases of a business cycle. oil companies are tied to the production process. it places a value on prospective growth. In the meantime.18 6889 2006 22.and face diminished returns. EVA is a concurrent indicator of stock price. In the first scenario. two unusual scenarios were encountered that created a delay. In either scenario.even during periods when the cost of equity is high .02 A direct comparison between both companies’ EVAs and stock prices displays a market reaction that is diametrically opposed to capital structure theory.98 8544 PERCENTAGE GAIN -4.9 percent decrease. a 12. a decline of 4.467 present value . and effectively “ration” capital. Microsoft’s EVA went up about twenty-four percent to 8544. Microsoft will move away from its optimal capital structure and EVA will be diminished. While ConocoPhillips’ EVA went down to 8732 from 9933.96 percent. their stock went up another nineteen percent.96 24.18 to $22. or it can restrict projects to periods when the cost of equity is low. speculation about the acquisition in an economic .71 -12. Normally.36 (split) 9933 2006 71. but their fiscal year stock price decreased from $24. increasing the size of the company by almost two thirds. but in these cases.95 8732 PERCENTAGE GAIN 18. A MARKET DISCONNECT AND EVENTUAL RECONCILIATION Table 18-7 CONOCOPHILLIPS YEAR 2005 PRICE EVA 116. ConocoPhillips bought Burlington Resources.09 Table 18-8 MICROSOFT YEAR PRICE EVA 2005 24.98. As evidenced in the previous chapter on ConocoPhillips. Nevertheless.1 from 1.40 a share. anytime growth rises at the same time that the CAPM percentage declines. there was simply no better place for investors to be than “Big Oil” mid-way through the decade.7 percent (modified for a special dividend). When net income rose only 2. upward pressure is placed on the stock.fiscal year 2006 closed with a slight decrease. Microsoft gave investors a special dividend of $3. Microsoft seemed like a great bargain. The 36 billion dollar pay-out compensated for a slightly reduced share price in 2005.8 percent in 2006. And yet . the stock began climbing in a delayed reaction that is a prime illustration of a stock that is out of equilibrium with the market. by 2007. The nineteen percent stock surge that occurred in spite of the decline in EVA was a triple-play combination of momentum. and low-risk leverage. The answer to this conundrum is that analysts who only followed earnings were unimpressed by the paltry increase in net income. the stock soared. In the second scenario. The combined force of a lower cost of equity percentage. the ”expected” rate of return from retention and ROE rose to 23.3. However. The movement to a more optimal capital structure was coupled with better earnings prospects with the release of the new “Vista” operating system. the reader will notice that Microsoft effectively reduced its beta to 1. ConocoPhillips was already buying back shares. too much equity at too high a price produced movement away from an optimal capital structure. coupled with the already reduced stockholders equity produced a higher EVA. However. Additionally. the large equity issue that helped pay for Burlington Resources diminished EVA.468 environment of rising oil prices would momentarily trump EVA.07 percent from a 2005 figure of 18. speculation. in 2006. and gained nearly fifty percent before the market peaked in 2007. When compared to the lower “required” rate of return derived from the CAPM. The comparison between the two firms illustrates three important principles: . investors were blinded to the fact that Microsoft was decreasing equity and buying back shares. From a capital dynamic perspective. the EVA/capital dynamic must be assigned a risk premium by the market. brokerages etc. we can examine different firms in different industries using the comparative capital dynamic (CCD). Like EPS. For example. the firm would . The nineteen percent price gain made in 2006 was not nearly as large as the thirty-nine percent gain in 2005. • 2. Unlike P/E which is vulnerable to speculative excess. implicit in the term “total” is the product of the CAPM percentage and stockholders’ equity. any gains must be compared to the industry outlook and the performance of a firm’s peers. In effect. Well managed. an EVA that rises on the basis of a lower cost of equity may be a prelude to a rise in the stock since prices more immediately respond to earnings. it is a ratio that measures relative size . the CCD is potentially more useful than P/E because it encompasses more information. if our current research shows only a handful of company have a CCD over 3. We merely divide net income by the total cost of equity (Net Income / Total Cost of Equity). “trusted” companies like ConocoPhillips can temporarily move away from their optimal target structures and make up the deficit with the” “promise” of prospective earnings. but still beat the market by a wide margin. however. Unlike EVA. Similar to the P/E. it compares a market benchmark to the fundamental earning power of the company. Although EVA is mostly an indicator that is concurrent with stock prices. rather than absolute size. then we would perceive that a firm’s risk increases at that point.469 • 1. • 3.) that are capitalized differently from manufacturing and service companies. By bringing in the concept of an “opportunity cost” in the form of the percentage cost of equity. As a risk based indicator. the CCD is a measure of magnitude. the CCD can be used to compare companies between industries and within sectors. the CCD is grounded in the mechanics of capital structure. banks.5. INDUSTRY COMPETITION: CHEVRON AND THE COMPARATIVE CAPITAL DYNAMIC (CCD) Excluding firms in the financial sector (insurance. are the dynamics of an increase. Another analogy to P/E mechanics. the CCD can be increased with a lower cost of equity. The following data for Chevron will help illustrate the many uses of the CCD. Just as lower earnings or a higher price increases P/E.53 11341 PERCENTAGE CHANGE 29. Table 18-9 CHEVRON YEAR Sales Operating Income Assets Net Income Stockholders' Equity CAPM % Growth % Shares Outstanding Dividends Paid 2005 198200 25679 125833 14099 62676 7. a higher net income.470 “retrench” before building a higher EVA again. or some combination of both.4 2197.31 2155. A higher EVA coupled with a lower CCD signifies a different relationship than if both increase together.6 .52 17.11 16.17 4456 Table 18-10 CHEVRON YEAR STOCK PRICE EVA 2005 56.77 9643 2006 73.41 18.81 3876 2006 210118 32427 132628 17138 68935 8. The “expected” rate of return. Net income would decrease. Each rose from around “2” to over “3” and then back again. they were also more market dependent with a higher R squared. the imperative was to be growing within the sector. The EVA delineated “risk premium”.17575 0. If the market is rapidly deteriorating. their stock rose only slightly above ConocoPhillips’. and the total cost of equity would . surpassed the “required” rate of return .564 x Y=1. the CCD seems to stay within particular boundaries for each market.in the “right” market.98+0. While Chevron pushed their EVA up approximately seventeen percent. The comparative capital dynamics were very similar for both oil companies. Like ConocoPhillips. paring down long-term debt to capital and expanding with retained earnings before it had a twenty-nine percent gain in the stock in 2006. by approximately nine percent.722 x Chevron is a well known.213414 REGRESSION Y=0. ConocoPhillips was perceived by investors as a company who was rapidly improving market share .980456 1. It followed the “sector imperative”. equally sized competitor to ConocoPhillips. Again. it expanded its EVA in 2006 to 11341. However both oil companies had a comparable beta which would be expected for firms with similar capital structures in the same industry.471 Table 18-11 CHEVRON YEAR 2005 2006 ALPHA 0. the natural correlation of stock price to fundamental. was not high.050042 BETA 0.in 2004. the CCD should decline as well:. derived from the CAPM. who had actually decreased EVA. Although Chevron had a higher “alpha” in their regression than did ConocoPhillips. which implied that the stock had upward pressure on it. Like P/E.05+0. the run up in stock price had been so rapid.563904 0.721824 R SQUARED 0. In contrast to ConocoPhillips. that it split its stock . Chevron increased EVA and stock price simultaneously. Notice how the market seemed to reward each company for reaching a “3” in CCD. In the following data. Table 18-12 CHEVRON YEAR 2005 2006 Net Income 14099 17138 Total CST.6 CCD 3. Eqty.02 CCD 2. they had their best year ever with a thirty-nine percent gain in price as well as a stock split. And when ConocoPhillips had a CCD of 3. .28 By maintaining a relatively high CCD of nearly three. Eqty 4456 5797 EVA 9643 11341 % Change EVA 17. but because the formerly high net income attracted more interest in equity issues.16 2. As previously mentioned. 3596 6818 EVA 9933 8732 % Change EVA -12. Microsoft encountered a delayed reaction when they surpassed “3”. Eqty.107 Table 18-14 CONOCO PHILLIPS 2005 2006 Net Income 13529 15550 Total CST.472 be overly high . not seeing a subsequent rise in the stock until the fiscal year was over and the final totals were tallied.not only because the CAPM percentage increased.28 3.76 2.956 Table 18-13 MICROSOFT YEAR Net Income 2005 2006 12254 12599 Total CST.09 CCD 3.76 in 2005. 5365 4055 EVA 6889 8544 % Change EVA 24. a side by side comparison is made with Microsoft included. In fact. maintaining the highest possible CCD with the lowest standard deviation. It is relative to both fundamentals and the market. recent retained earnings. for example. it would be helpful to calculate new retained earnings as a percentage of the total cost of capital. In this illustration. Thus. . The latter approach is more conservative and exacting. it we treat equity like the weighted average cost of capital (WACC). the effective result is a higher P/E and more risk. or we integrate preferred stock into our calculation by weighting it with its percentage and cost. Either we eliminate preferred stock altogether and use the aggregate figure for common equity. we assume that preferred stock is zero. An industry with a historically small ROE will have a much lower CCD than high-ROE industries like software. The emphasis would again be on the risk/return difference of mean-variance. PERCENTAGE OF NEW RETAINED EARNINGS If the comparative capital dynamic is declining and EVA is increasing. we will find old retained earnings. but it needs to be gauged by both its own performance and common standards for the industry. but they can potentially propel the firm into a leverage state in which risk is rapidly reduced by lowering beta. Any comparison between firms in an industry should reveal that the competitor with the highest CCD is optimizing capital structure better than others. the rate of growth of the cost of equity surpasses that of earnings. the CCD stakes out a position between P/E and ROE. old paid in capital. Not only will more retained earnings imply that net income has increased. That scenario is analogous to a firm that maintains its stock price despite lower earnings. When we decompose stockholders’ equity into its constituent parts. A restaurant chain. and emphasizes the weighted cost approach to the cost of equity. we can use the percent change in new retained earnings as an indicator of potential performance. recent stock issues and preferred stock.473 In the somewhat obscure territory occupied by financial indicators. and maximizing the price of the stock. will have a different optimal CCD than a steel company. The following example will emphasize the correlation with EVA. if new retained earnings were substituted for a stock issue. we saw that more retained earnings can detract from EVA when the cost of equity percentage is high (such as at the end of a business cycle). for example.55 % 3. Paid in capital (a stock issue) would not increase net income and would fully detract from EVA. there would be a net gain in EVA. In the chapter on Kimberly-Clark. because net income would increase.76 9933 2006 2277 15550 13273 82646 16. even in that extreme case.06 % 2. However.(Dividends paid / Net Income)) x (Net Income)) x (Percentage Cost of Equity) / Total Cost of Equity which will simplify into New Retained Earnings / Stockholder’ Equity. The rationale behind viewing new retained earnings in a positive light is that they are implicit in net income and will not decrease that side of the function. Table 18-15 CONOCOPHILLIPS YEAR Dividends Paid Net Income Retained Earnings Stockholders' Equity Retained Earnings / Equity CCD EVA 2005 1639 13529 11890 52731 22. Alternative sources like equity issues or debt would be even more costly .474 The calculation is ((1 .in either interest payments or flotation costs and dilution. the decrease in EVA can be considered the “least harmful possible” because capital is in effect being raised without any accounting costs. and then only a percentage of retained earnings would be subtracted.28 8732 (Back to Table of Contents) . Thus. As we saw in our observations of Fed-Ex and Staples. Our analysis was full of countervailing balances. the cost and the type of funding. Rarely should risk analysis be pursued in isolation.Fed-Ex had pricing power that Staples lacked. the demand variable. it has both risk and return components that sometimes offset each other. but also on an important qualitative factor . but it also affects capital structure by determining the source. The validity of the argument was based on the size of the return.although it entailed more risk. that ratings agencies like Standard and Poor’s and Moody’s use it as a prime component to determine credit ratings: (NEXT PAGE TABLE) . operating income is so crucial to capital budgeting.475 SECTION IV: CORRELATION AND PROBABILITY STUDIES 19 OPERATING INCOME CORRELATION STUDIES Operating income is the linchpin of capital structure. and we came to the conclusion that the income stream from Fed-Ex was more desirable than that from Staples . In fact. Operating income interfaces with sales. 1 10.75 17. equity becomes proportionally less expensive than debt. Note that six of these top nine key ratios contain some adaptation of operating income as a main component. they would need to cover interest by an average of 16. Originally culled together as one investor’s attempt to capitalize on a bull market. However. 2: As operating income .2 22.9 0.5 11.3 Therefore.66 times in the preceding three years.59 1.33 2. NAME: Comparison Effects of Changes in Operating Income and Long-term Debt on Next Year’s Mid-range Stock Prices PREMISES: 1: In an economic environment of both rising markets and rising interest rates.1 56.476 Table 19-1 RATIO (Percentages where Appropriate) Pretax Interest Coverage Pretax Fixed Charge Coverage Funds Flow Interest Coverage Funds from Operations/Total Debt Free Operating Cash Flow/Total Debt Pretax Return on Total Capital Operating Income to Sales long-term Debt/Capital Total Debt /Capital AAA 16.2 BB 1. if Company X wants to float a bond issue and receive an “investment quality” rating of AAA in the given year.7 23.7 49.66 6.2 24. The following study is a pragmatic exploration of the relationship between operating income increases and long-term debt. At the end of the chapter some key statistical concepts will be discussed to both criticize this study and propose alternative hypotheses.6 51. what it lacks in scientific legitimacy. it makes up in timely effectiveness.98 134.9 9. the student should be aware of methodologies by which the study could have been improved.39 22. 1999). but any duplicate industries were deleted. hereunto referred to as . 3: A combination of higher operating income and less use of debt encourages a firm to maximize its stock price. the measured operating and financial risks would skew any meaningful result. In a similar manner. DATA POINTS AND STRUCTURE Twenty four companies were picked from the S & P 500 at random. because their respective capital structures are unique and not representative. if operating income did not change at all. no macro-variables such as interest rate changes or risk premium spreads were included Unique to the study was the measurement of changes in stock price. i. it measures volatility as much as it measures performance.e. creating one hundred and eighty-two separate data points. Mid-range price is merely the average between the high and low price for that year without regard for timing. Companies that had no change in the long-term debt to capital ratio or had no longterm debt at all were counted as data points that decreased long-term debt to capital. Both financial institutions and utilities were left out of the study. Seven to eight years of fundamental increases were tabulated for each company.. lowering the overall price effects in their respective categories. While the study was conducted during a period of relative economic prosperity (1990 . The subsequent year’s mid-range price was compared to several fundamental variables that had been segmented according to whether year to year operating income was increasing or decreasing. Consequently. but also captures any inherent trend when used collectively. companies can use less debt and more retained earnings to lower the cost of capital.477 rises. based on increases and decreases in both operating income and long-term debt to capital. CATEGORIES Data points were separated into six distinct populations. it was counted as a decrease. it was later determined that they were actually antagonistic to the researcher’s premise. Although these “all-equity” data points might skew results. the correlation value is scaled from -1 to 1 with greater association represented at the extremes. that is . The populations were identified as: 1) Operating income increases 2) Operating income increases. However. the probability associated with each correlation must be read as “proof of the null hypothesis” that no association exists. and standard deviations of each fundamental were calculated to provide a scope of the skew of the distribution. All values with a ninety-five percent significance are reported. Data is simply ranked from lowest to highest and a standard linear regression is performed on the ranks. but the reader must recognize that the study is invalidated by small sample size and is meant to identify trends. LTD/CAP decreases 4) Operating income decreases 5) Operating income decreases. not absolutes. .478 “LTD/CAP”. this ordinal method will produce associations that are curvilinear and reveal relationships that might not be apparent in normal least squares-type regressions. medians rather than means were used to identify averages within the population. LTD/CAP increases. the more valid the association. Analogous to linear regression. The Spearman rank correlations are associations made by regressing the ranked order of the data against each other. Each population was analyzed in terms of descriptive statistics to enable comparisons between population groups and Spearman rank correlations that extracted relationships within each group.the lower the probability. LTD/CAP decreases 6) Operating income decreases. Since the sample size was limited. LTD/CAP increases 3) Operating income increases. profit margin. the financial leverage ratio (EBIT/EBIT. sales. (EDS) Ceridian Corp. (NUE) Cendant Corp. Inc. and was regressed against the mid-range stock price change of the next year to observe any discernible effects. (BCE) McKesson HBOC (MBK) Halliburton (HAL) Home Depot (HD) Black and Decker (BDK) Hilton Hotels (HLT) General Electric (GE) Charles Schwab (SCH) FUNDAMENTAL VARIABLES Variables derived from financial statements include five variables that encompass Du Pont .type analysis that would give some indication of effects of the return on equity or ROE. These are yearly increases/decreases in operating income. capital (the sum of equity and longterm debt). long-term debt to capital (LTD/CAP). All yearly changes are reported on a percentage basis. operating momentum. (DBD) Emerson Electric EMR) H & R Block (HRB) Pharmacia Corp (PHA) Electronic Data Sys.05 (95 % confidence) was encountered. and asset turnover. Seven variables that affect the cost of capital are also included. Each variable was regressed against the change in next year’s mid-range stock price and reported if a probability under . (PD) Advanced Micro Dev. (AMD) CVS Corp. Inc. capital expenditures. Each variable is an increase or a decrease. (CEN) Tenet Healthcare (THC) McGraw-Hill (MHP) Phelps Dodge Corp. Descriptive statistics are included for the Du Pont-type ROE variables as well as for the cost of capital variables with the exception of the interest rate on total debt which is deemed beyond a firm’s direct control. total debt. long-term debt (LTD). EPS. the capital multiplier (assets/capital). (CD) BCE. If a percentage change was judged to be overly skewed .Interest) and the interest rate on total debt (interest expense/total debt). (CVS) Nucor Corp.479 COMPANIES IN THE SAMPLE Table 19-2 Enesco Group (ENC) Clayton Homes (CMH) Compaq Computer (CPQ) Diebold. These are: yearly changes in assets. and the number of shares outstanding. Other variables included in the study are yearly changes in: cash. The term “leverage” in “total leverage” is derived from multiplying the financial leverage ratio by operating momentum and not indicative of “true” operating leverage. ALL VARIABLES Table 19-3 Next Mid-range Stock Price Operating Income Sales Outstanding Shares Interest Expense Assets Capital EPS Interest Rate ( onTotal debt) Total Debt Asset Turnover Profit Margin Capital Expenditures Cash LTD/CAP Financial Leverage Ratio Capital Multiplier (Assets/Capital) Total Cost (Sales-Op Inc.000 percent decrease). . absolute measurements that indicate a scalar number rather than a change. which is not measured. These were included to observe whether the size of these ratios affected stock price.480 (such as a change involving a 10. it was eliminated at the discretion of the researcher.) Operating Momentum Long-term Debt Long-term Debt/Total Debt Total Leverage Total Leverage -absolute Operating Momentum Abs The last ratios are concrete. INCREASING LTD/CAP Table 19-5 Variable Sales Operating Income Profit Margin Capital Multiplier Asset Turnover Mid-range Price Assets Capital LTD/CAP LTD Total Debt Financial Leverage Ratio Median 14.6277 14.09 -0.59 51.58 740.0268 1.481 STATISTICAL RESULTS (Percentage Increases) INCREASING OPERATING INCOME Table 19-4 Variable Sales Operating Income Profit Margin Capital Multiplier Asset Turnover Mid-range Price Assets Capital LTD/CAP LTD Total Debt Financial Leverage Ratio Median 15.1613 Standard Deviation 25.41 110.0359 16.67 -0.83 -2.11 30.56 16.05 20.38 215.28 75.7 24.06 39.6 254.15 3.25 13.98 20.64 -7.5 0.48 29.8413 Standard Deviation 25.34 INCREASING OPERATING INCOME.66 46.033 -0.11 8.39 17.85 70.84 67.05 2268.7 16.83 -1.22 20.2 69.78 4034.9 47.73 -0.58 21.56 40.57 8.09 .39 17.26 1296.6 7.32 22. 82 17.1 36.8 173.29 30.16 2.43 1.57 15.17 -4.51 20.8 29.6 -22.96 0 -1.83 21.87 13. DECREASING LTD/CAP Table 19-6 Variable Sales Operating Income Profit Margin Capital Multiplier Asset Turnover Mid-range Price Assets Capital LTD/CAP LTD Total Debt Financial Leverage Ratio Median 16.72 DECREASING OPERATING INCOME Table 19-7 Variable Sales Operating Income Profit Margin Capital Multiplier Asset Turnover Mid-range Price Assets Capital LTD/CAP LTD Total Debt Financial Leverage Ratio Median -2.38 Standard Deviation 25.46 1.0992 Standard Deviation 17.58 -11.34 47.71 0.97 25.28 13.52 8.25 32.14 16.06 -17.93 267.92 42.63 12.49 4.66 34.44 32.482 INCREASING OPERATING INCOME.44 -7.5 15.1 22.21 20.62 24.46 26.8 .45 -1.5 -1.47 1.68 5.34 2.99 66.99 73. 37 1. DECREASING LTD/CAP Table 19-9 Variable Sales Operating Income Profit Margin Capital Multiplier Asset Turnover Mid-range Price Assets Capital LTD/CAP LTD Total Debt Financial Leverage Ratio Median 0.3 DECREASING OPERATING INCOME.12 11.43 32.27 19.35 55.85 35.51 17.31 2.76 25.75 Standard Deviation 19.INCREASING LTD/CAP Table 19-8 Variable Sales Operating Income Profit Margin Capital Multiplier Asset Turnover Mid-range Price Assets Capital LTD/CAP LTD Total Debt Financial Leverage Ratio Median -5.17 29.02 -7.53 14.87 113.097 -8.18 5.52 -2.09 6.483 DECREASING OPERATING INCOME.41 16.53 239.55 -8.78 -1.23 -14.82 3.53 26.82 30.61 0 Standard Deviation 16.68 80.32 22.78 27.12 47.5 7.37 -1.07 1.48 24.8 34.38 16.69 131.35 2.24 14.78 1.67 -14.0192 -11 -29.32 16.05 . 006968 0.) Correlation -0.238255 -0.351063 0.012722 0.049177 0.484 SPEARMAN RANK CORRELATIONS: NEXT YEAR’S MID-RANGE PRICE INCREASING OPERATING INCOME (N=130) Table 19-10 Variable Assets LTD/CAP Financial Leverage Ratio LTD/Total Debt Correlation 0.021278 DECREASING OPERATING INCOME (N=52) Table 19-12 Variable Interest Rate on Total Debt Operating Momentum (Abs.242514 Probability 0.010717 0.013494 0.006339 0.172891 -0.033053 INCREASING OPERATING INCOME.) Total Leverage (Abs. DECREASING LTD/CAP (N=90) Table 19-11 Variable Assets Capital Financial Leverage Ratio Correlation 0.014102 .187093 Probability 0.338513 Probability 0.340562 0.015833 0.261687 -0.211257 -0.282562 0. INCREASING LTD/CAP (N=40) NO SIGNIFICANT CORRELATIONS INCREASING OPERATING INCOME. or does it stem from the size of the capital outlay precluding the use of equity? At times.485 DECREASING OPERATING INCOME. INCREASING LTD/CAP (N=21) Table 19-13 Variable Interest Rate on Total debt Operating Momentum (Abs) Total Leverage (Abs) Operating Momentum % Total Leverage % Correlation -0. slow.002895 0.57013 0. it is important to look for undocumented variables (like rising interest rates) that occur outside the domain of the study. If such integration proves risky.009083 0.616883 0. and ultimately unprofitable. Since correlation shows association and not causation. and retained earnings would not usually be large enough to cover the deficit.006966 0. the type of funding can not be held liable. many debt issues go to fund large acquisitions that must be integrated into the corporate structure in subsequent years. Debt issues are not destructive in and of themselves unless they are being substituted for less costly sources of capital. is it the fault of long-term debt.554545 0. Probability and economics make strange bedfellows if only because of the existence of so many other variables that can refute a hypothesis. If this situation occurs regularly.631169 Probability 0. For example. .484416 0. It is not the intention of the researcher to “implicate” long-term debt as the main culprit in profitability loss.002153 DECREASING OPERATING INCOME. a large issue of stock can do damage to share price merely on the basis of supply and demand. DECREASING LTD/CAP (N=31) NO SIGNIFICANT CORRELATIONS INTERPRETATION There are several caveats to observe in this study. foresight was lacking on the part of management.026056 0. Additionally. In the highly inflated market of the 1990s. the Du Pont-type fundamentals that comprise ROE are much higher for that category. long-term debt will be more expensive than short-term debt because investors need to be compensated for the inflation risk of a longer maturity. which is an acceleration of earnings after an initial increase. While we are unable to draw firm conclusions from the data. Thus. companies would have desired to exercise stock options and issue new equity because they could have done so without paying the penalty for diluting the shares. Paying lower interest expense automatically raises EPS. • More operating income and less interest expense can lead a company to do more equity financing through retained earnings. The increasing operating income. • Greater operating income has the potential to be distributed as a higher dividend. all other variables being equal • Raising the proportion of long-term debt to equity will increase the cost of that same equity • In a normal market. our premises are largely unproved and our results are inconclusive. our statistics can be misleading because they occurred during a period of rising interest rates and relative prosperity which would have created a propensity for equitytype financing. compensating shareholders and raising the stock price. many such issues rose “automatically”. although the study raised many questions and pointed us in the “right direction”. The data itself displays a classic argument for minimizing the cost of capital. we can use several known propositions to support it: • • Long-term debt obligates a company to paying interest expense. • Operating income increases can lead to “momentum”. decreasing LTD/CAP category in particular exhibited a midrange stock price difference of as much as eighteen percent. .486 Moreover. The researcher Joseph Murphy. both decreased long-term debt by about one percent. one will find a large difference in the return on equity factors like sales and profit margin as well as substantially different capital outlays.64 % vs. but may rise proportionally because there are less retained earnings to buffer it in the capital structure. this phenomenon is especially apparent during rising markets. decreasing LTD/CAP data. Further substantiated by both Robert Levy and Robert Hagin. Besides rejecting the null hypothesis for the increasing operating income. who studied three hundred and forty-four companies in twelve different industries found no discernible correlation between year to year earnings data. and capital. One finds .96 %. earnings. 2. It merely encompasses a six month increase in earnings once earnings accelerates faster than the market in the previous six months . Any decrease in earnings is derived from an inefficient use of capital. we find that companies who increased both operating income and LTD/CAP had adequate sales.and it is essentially probabilistic. On the other hand. and actualized by a combination of less sales and higher production costs. not deterministic In this study. Manown Kisor and Van Messner found a pattern that they termed “increased momentum in relative earnings” which is composed of an accelerated rate of change of earnings in comparison to the performance of the entire market. 13. Observing the long-term debt increase in both data sets reveals that there is little significant difference. the last proposition is questionable. but performed less well than companies who increased operating income and decreased LTD/CAP. it will not propel a distressed company into “the black”. Evidently. It seems that while lowering capital costs will further enhance an already profitable company. If one compares the data sets of pure operating income increases to pure operating income decreases.487 Stated as fact. an autocorrelation of one year’s earnings regressed against another year’s proved no correlation between the two sets of data. With further data separation. a few other comparisons were potentially significant. more long-term debt does not drive the decrease in operating income or the consequent ROE factors. Since this study was limited to numerical differences and not qualitative ones. Since S & P 500 companies need substantial capital to grow. Therefore.such as only one segment of one industry-to be valid. Such large purchases are normally made with combinations of equity exchanges and bond issues. A large amount of cash on the balance sheet is far different from the integration of a competitor into one’s asset structure. only to be removed in the next year because the capital turnover was not rapid enough. Unfortunately. retained earnings and cash prove an inadequate source of funds when purchases are literally in the billions of dollars. financial data is heavily skewed as evidenced by the large standard deviations in the descriptive statistics section. it lacks statistical validity because it fails to prove or disprove a premise. the quality of assets must be examined. If more long-term debt does not “drain” a company of profit. In fact. How do we know whether differences were obtained on the basis of size or capitalization? Researchers compensate for this bias by setting up a “control” group and comparing the populations . METHODOLOGICAL CRITICISM Although a study of this magnitude can pose valuable questions and give researchers leads that may yield some answers. While separating the population into unique categories helps to accentuate the differences. Such distributions require a much larger sample size and a much narrower focus .and yet that too may lead to difficulties . and yet the ROE factors for the increased LTD/CAP set significantly under perform. along with cash. then some other factor must be working to limit performance. The data does not reveal the peculiarities of acquisitions. it is typical that speculative cash goes into an acquiring firm’s stock if the purchase is considered a “smart” move. The “synergistic” effects of an acquisition are not automatic. Acquisitions of target companies need time to be integrated into the acquiring firm’s asset structure which puts a strain on existing resources and may lead to some under performance.488 similar increases in assets. the sample comprised only large well-known companies leaving out small caps and private businesses. we can only speculate on the nature of the reason behind them. Since most economic hypotheses must adapt to a changing market. if one sells homes at the beginning of an expansion. For example. Companies with steady incomes. To make sure the measurements are not biased. the period in which the sample is located may bias the study. Since seven to eight years of data are extracted from each company. percentage changes are inferior to logarithmic changes because the base of the percentage will change and skew the results. far reaching. that sector may be profitable for several years in a row. it would be better to have as many companies as we have data points in the sample. The hypothesis was not wrong at the time of formation. Characteristics of measurement were also problematic. Lastly. While statistically valid studies sometimes have too narrow a focus to be of interest to the layperson. which is the same measurement as an EPS that goes from four dollars to six dollars. For example. our hypothesis about debt and tax deductions would be obsolete if the nation went to a national sales tax and away from an income tax. Analogously. biasing sales and income in an upwards direction. who can make the . Unlike “twin” studies in the biological sciences. the data may be serially correlated: A company may have a strategy over a period of time that skews the data in one direction. they often lose legitimacy when repeatedly tested against reality. Any recommendations should be derived from a few business cycles which may have a profound effect on the results .especially when they comprise income and debt. For example. both students and investors should be encouraged to experiment.489 because companies may prove to be dissimilar. In the pursuit of more information. Although it is not our objective to defend “bad statistics”. The Spearman rank correlations would have further emphasized this anomaly because they are ordinal based measures of magnitude with no reference to a stable base. a nickel change on an EPS of ten cents is a fifty percent improvement. ideas that “break the rules” may offer solutions when old methods become stagnant. but was replaced by a new development. few factors are controllable in the economic environment. but technically incorrect studies can propose hypotheses that may have later legitimacy. investors sell stock and buy bonds. or who lower their long-term debt to capital ratios. A rising market made equity funding more feasible and profitable because most companies’ stock was in high demand. Companies that proportioned their capital structures counter cyclically to the changes in interest rates profited more than those who did not. Since the interface between the market and the company is the interest rate. Many researchers have found that capital structure is firmly integrated with tax and interest rates. 3. The study was flawed and no ideas were legitimate. When rates rise. will be rewarded. scholarship that produces practical. Consequently. (Back to Table of Contents) . as they do in prosperous times to combat inflation. capital costs are determined by the effects of interest rate changes on the asset structure of the company. 4. When the risk premium is high. investors are more rewarded by equities. 2. while those with more operating risk must finance through retained earnings or equity issues. When the risk premium between debt and equity becomes very low. capital costs are often minimized because there is less of a risk to equity and the inherent cost of equity remains stable. Gradually. In essence. stock prices and profitability of the company. that entails observing all possibilities. some new social or economic construct would develop to replace it. Four tenable explanations for our data can be produced: 1. there is no real correlation between type of capital structure and stock price because the market determines which structure will be favored. useful ideas should be encouraged. When more debt is incurred at a lower rate. those companies who need less debt. Large debt issues will raise the cost of capital and be unprofitable in the near-term. would be penalized.490 best use of debt funding. Companies with low operating risk will be in a better position to take advantage of low rates and incur more debt. and in finance. The first and last hypotheses can be rejected on a historical basis. The case we make for leverage factors is that they are specifically actionable. not a forward looking one.491 20 CHANGES IN CAPITAL STRUCTURE AND THEIR EFFECT ON STOCK PRICES This chapter examines the same data that fueled the operating income study in the last chapter. Moreover. at these brief junctures.leverage factors are an inherent part of the economic system and defy change. and it is a lagging indicator. there are systemic factors that validate the possibility. and 2) whether financial leverage ratios have any predictive value considering that more risk may imply a higher cost of capital in the near future. return rises with risk. or financial leverage information to make extraordinary profits. the investor can use “stock screens”. one can argue that economic factors like inflation behave in a predictive. momentum. But . as it may during an interest rate hike. While it tends to confirm conventional wisdom on the subject. . and affect profits and assets. However. but if risk becomes too expensive. there may be brief periods where there are “arbitrage” opportunities when inefficiencies occur. As stated previously. Inflation may disappear for years only to come back rampantly. albeit unstable fashion. THE VALIDITY OF LEVERAGE FACTORS If indeed financial leverage ratios have no predictive value. it also opens up avenues of discussion in two other areas: 1) whether a large capital outlay in one year creates diminished capital funding in the next (capital rationing). Like earnings momentum. but the practical value is to observe the behavior of capital structure variables over an entire business cycle. the market changes and the factors that created such opportunities will no longer be operable. Thus. There is always some . It attempts to establish a hypothetical relationship between changes in stock prices and changes in capital. the statistical viability of such a study can be questioned because of small sample size and serial correlation. then return is diminished. it is because the market is supposed to factor in such information in an “efficient” manner. On the other hand. and low inflation all contributed to stock market increases. the end of the Cold War. One reason that the bull market of the 1990s was a good laboratory for studying capital structure was the stability of earnings. we can infer that earnings are rising faster than the cost of capital (specifically the cost of equity) when this situation occurs. but to get the reader to think in terms of possibilities . the cost of equity outpaces earnings and . but will need to move away from it once the prospect of higher prices diminishes. we have stated that the environment for such earnings increases is created by an optimal capital structure.492 combination of leverage factors that will yield the highest amount of return with the least amount of risk in any market. While there is always the tendency to interpret correlation as “cause and effect” instead of mere association. our objective in this chapter is not to prove a theory. CONNECTING THE DOTS: EARNINGS AND DIVIDEND GROWTH AND THE COST OF EQUITY Throughout this text. and difficult to develop into a coherent strategy.even if he or she so chooses. In fact a hypothetical comparative dynamic can be constructed for the entire economy with % ∆ GDP / % ∆ Risk Free Rate (10 Year Bond) In essence. we have emphasized our belief that stock appreciation is a function of earnings acceleration. a two term president. ways to refute this research. the gain from investment must outpace its cost. we may choose an asset class like gold or real estate to hedge against inflation. % ∆ Net Income / % ∆ Cost of Equity. Secondly. When interest rates rise and stocks ascend too rapidly in comparison to risk. but the sample population is made up of firms with primarily large capitalizations. the response to economic factors like inflation is reactive. The theoretical underpinnings for this scenario would be the change in each component part of the comparative dynamic. a peace dividend. When one firm’s earnings rises faster than other firms’ in the market. that firm’s stock will out perform those of similar risk.and . absolute size tends to buffer some volatility. Excess speculation in the latter third of the decade had a skewing effect on the results of this study. Thirdly. and the result is a diminished stock price. an effective use of dividend discount models should provide the same outcome as the CAPM. An alternative to this “top down” theory is to micro-manage capital structure through the dividend. albeit at a constant growth rate of dividends. the dynamics behind the model are informative. it is entirely possible to manage capital structure from the perspective of dividend growth and retention. but it will need to finance in the credit markets for lack of sufficient retained earnings. That tenet has made many corporations reluctant to cut dividends in the worst of times.493 the stock market withers. because of the negative information that such a move signals. In this case. because both methods concentrate on the relationship between earnings and the cost of equity. the dynamics are a microcosm of the “top down approach”. EARNINGS ACCELERATION . While we believe that a more accurate cost of equity is obtainable from the CAPM. This type of funding raises the cost of capital even as earnings are depleted. Those who espouse dividend theory believe that all stock valuation is inherent in dividend payments and that the proportion of equity in capital structure depends on the management of retained earnings. as measured by the cost of equity. the change in earnings is decreased while the change in capital costs is increased. because they directly relate the cost of equity to the price of the stock. The relationship between components of the capital asset pricing model (CAPM) is established by this balance between earnings and the cost of capital. but is subject to distortion in determining an adequate growth rate. The residual theory of dividends proclaims that unless projected earnings meet a specific level of return. Not only will such a company fail to compensate shareholders. Thus. The Gordon model offers an estimate of the cost of equity. they should be paid out to shareholders. without knowledge of beta. Nevertheless. In fact. the opportunity cost of retaining earnings is greater than paying them out. To obtain this “internal” cost of equity. but totally dependent on the severity of the decline. We obtain a “ball park” figure for the cost of equity. the internal return for the company is outpacing stocks of similar risk on the market. While the Federal Reserve desperately cuts rates. stocks can not recover until earnings again outpace the cost of equity in the opposite direction . the proposed dividend yield component. but when a downturn erupts. we algebraically turn the model around and see that Ks = ( Next Dividend / Stock Price ) + Growth. The same phenomenon occurs during an expansion: companies are not sure whether growth is stable enough to warrant a dividend increase and the proposed dividend yield shrinks to cause an upward lag behind earnings. While the Gordon model is only applicable to stocks whose dividends grow at a constant rate. ((ROE)(Retention ratio)). Ks is equal to the cost of equity and G is the growth rate for which we have formed the generic equivalent. we do not use it for valuation. Since we have made growth contingent upon earnings. Thus. both the amount retained and the return (ROE x Retention). earnings outpace the cost of equity. . or communicates expectations thereof.494 To reiterate the Gordon model: The current price of the stock = Next Dividend / (Ks – G). earnings deteriorate far more rapidly than the cost of equity. the cost of equity rises as earnings rise except that it lags behind the rate of earnings growth. At the same time. we can easily observe how the cost of equity has a direct. then we observe whether that stock is a good candidate for investment. In that case. This lag is a double edged sword: as earnings outpace the cost of equity. which is comparable to an IRR for that stock. (Next Dividend / Price) will grow because stock prices (the denominator) decrease.usually a six month process in recent decades. the growth component is shifting downward because retention and earnings are both lower. positive relationship with earnings. stocks rise. If it is higher than that obtained through the CAPM. The dynamics of this lag are implicit in the Gordon model. If a company delays in cutting its dividend. throughout the business cycle. In fact. or Assets/Capital should be correlated with stock prices is evident from the implications of these models. A higher percent of retained earnings will raise a firm’s equity position. Since inflation is a lagging indicator in itself. A paradoxical relationship between interest rates and stock prices occurs.but . It lets earnings growth outpace rate raising initiatives and the stock market rises. the Fed can not lower rates fast enough without causing dislocations in the capital markets. the amount of debt is implicit in the retention ratio. Analogously. stocks will be positively correlated with interest rates until rates are high enough to discourage investment. which regulates growth. and yet there are several equivocal variables.495 In the CAPM. Both inflation and rising interest rates can skew the results. In that case. in the CAPM. earnings will not put as much pressure on stock prices.they are reactionary and not forward looking simply because changing them entails so much risk. the two models create the foundation for correlation. the lag appears to be created by delays in cutting or raising interest rates. the connection between the two models lies in expectations: both the risk free rate and the proposed dividends of firms are functions of the collective expected economic outlook . at a market top. For example. but may undermine dividend growth if it is excessive. During a downturn. a higher beta is created from taking on more debt which will potentially raise the stock price in a rising market. and the Fed begins a campaign of rate cutting. stocks begin to be negatively correlated with interest rates. during the early part of an expansion when interest rates are low. STATISTICAL VALIDITY The proposition that changes in leverage factors such as the long-term debt to capital ratio (LTD/CAP). Together. Similarly. the Fed must be wary of stymieing business growth. the market factors in the real rate of inflation. In the Gordon model. At that point. more long-term debt may be correlated with higher stock prices. In essence. On the other hand. more debt might be correlated with higher prices because the market factors in the . the higher cost of debt may cause those variables to be negatively correlated. if inflation is high. but from an investor’s perspective. researchers in this subject must tread a fine line between the need for mathematical proof. the researcher needs a large. and earnings are rising. and the practical observation of “real” economic behavior. it may end up being valuable because it is viewed within a specific context. when the yield curve is ascending. In fact. any study that is so narrowly focused as to be statistically valid.except without further categorization. From a mathematical perspective. the results would be termed statistically inconclusive. But when it is large enough. inflation is in check. that firm will probably incur more long-term debt simply because it minimizes the cost of capital even as interest rates are climbing. there is no long term correlation. and then negatively correlated for one. unbiased sample. In order to achieve statistical validity. Most companies react to the same market conditions and act similarly when it comes to changing capital structure. changes in mid-range stock prices were regressed directly against changes in leverage factors on a percentage basis . In this case. the futility of a controlled experiment on leverage factors should be evident. Thus. may also be irrelevant. one has to gauge the value of that fundamental (depreciation) in terms of the effect of some outside factor . However.496 repayment of loans in depreciated dollars. This paradox has can blind researchers to the inherent value of some fundamentals. If. what is considered a “normal” business cycle creates an equilibrium in which risk is priced correctly. for example. Therefore. we need to view the leverage factors from the context of changes in the cost of capital which are specific for each company. Both the . for example. depreciation is highly positively correlated for two years. In capital structuralism. each variable was freely regressed against mid-range prices without reference to any sub category.a change in the tax law. A BRIEF STUDY Using the same data as in the operating income study of the previous chapter. factors outside of the experiment will change the relationship between variables. an economic environment will be conducive to forming trends among some leverage factors. If the cost of debt is ridiculously low for one company. Those variables that are significant at the 95 % level (a 95 % chance of association with the stock price) are displayed. In effect. Thus the study may evidence strategies taken by a firm over a specific period. The sample includes the same one hundred and eighty-two data points.497 change in the current mid-range price (the year of the leverage changes) and the change in the next mid-range price are described. This qualitative aspect has pragmatic value. . we can observe how capital structure changes affected stock prices over an economic cycle. The following table delineates the variables used in the study: Table 20-1 Next Mid-range Stock Price Operating Income Sales Outstanding Shares Interest Expense Assets Capital EPS Current Mid-range Stock Price Interest Rate (Total debt) Total Debt Asset Turnover Profit Margin Capital Expenditures Cash LTD/CAP Financial Leverage Ratio Capital Multiplier (Assets/Cap) Total Cost (Sales-Op Inc) Operating Momentum Long-term debt Long-term debt/Total Debt Total Leverage Total Leverage -absolute Operating Momentum Abs. Additionally. but invalidates some statistical conclusions. This divergence allowed the researcher to observe the immediate effects on prices and also the forward looking effects (if any). there are simply too many uncontrolled variables. some of the variables were regressed against each other if they verified some tenet of capital structure theory. any variable that has explanatory value in terms of capital structure may be exhibited. The same caveats also apply because several data points will represent one company (twenty-four companies in all) and there may be serial correlation in which one data point will influence a subsequent data point. but we cannot attribute cause and effect. representing the same number of fundamental changes as in the previous study. 05 or less. the less the two variables are associated. Each economic cycle has periods of complete incoherency when “normal” correlative values are suspended. This process confuses . Variables that increase the cost of capital will decrease the price of the stock. • 2) Secondly.498 THREE ASSUMPTIONS We interpret the results given three assumptions: • 1) We assume that there is no transitivity of correlation . unless proven otherwise. but alternatively. the 95 % probability level is stated as 0. Data is ranked from highest to lowest before linear regression is applied. that consistency does not always hold. In that case. we assume that the correlation is valid in year two. All regressions in the study use the Spearman rank correlation.one variable such as sales will mathematically determine another variable such as operating income. That assumption is only realistic if the two variables are deterministic functions of each other.if one variable is correlated to another variable. If assets are correlated to sales in year one.that is:. that is . The higher the probability. However. for example. we do not assume that the first variable is also correlated to the third. and that variable is correlated to a third variable. the variables will always be consistently correlated because total cost is subtracted from one to obtain the other. there may be connections that invalidate such financial truisms. Nevertheless. Stock prices rise when earnings rise. we assume that there is consistency in correlation. We use the study to defend these arguments. . in probabilistic associations. but the exhibited probabilities are the probabilities of no association. This method captures the associative value better than straight linear regression. Correlations are scaled from -1 to +1 just as in linear regression. • 3) We assume that conventional financial wisdom is valid. Thus. such as with stock prices and leverage. but does not capture the strength of the association as well. it is the preferred method because it tends to emphasize curvilinear relationships that might not otherwise be apparent. Moreover.499 students and investors who have been taught that “more” connotes a larger number. we are actually finding the probability that the” null set” (no association) is true. it is expected that no current period variable will be highly correlated with earnings in the next period. Therefore. where the flow of income is balanced with the change in equity. in the majority of observed cases. In EVA analysis. We often assume that the cost of equity will rise if a firm issues stock. The argument for an “efficient” market stems from futile attempts to predict market behavior. any magnitude of change would rest upon the amount of diminished risk. usually implying a movement toward a cost reducing capital structure coupled with an appreciable change in income. In this case. At that point. we should expect at least a . when net income is great enough to warrant large increases in retained earnings. income or capital will yield a large change in price unless it contributes to lowering the cost of capital first. Thus. no “exact” amount of sales. We also assume that the company will pay a penalty if it retains earnings instead of distributing them as dividends when the cost of equity is particularly high. A third expectation is a simple extrapolation of deductive logic which most financial professionals would recognize: prices are affected by earnings which contribute to retained earnings which would raise the proportion of equity to debt when earnings are high. the stock responds positively: net income is more than enough to absorb the increase in equity and the cost of equity is kept low because more retained earnings diminish beta. capital structuralism acknowledges the transient and unstable relationship between fundamentals and stock prices. there is little explicit awareness of the quality of equity changes. However. but that such a rise might not be malignant if the equity buildup was derived from retained earnings. the combination of fundamentals that achieves this optimum keeps changing. EXPECTATIONS It is expected that current year fundamentals will be more strongly correlated with stock price changes of the same period than with stock price changes of a subsequent period. When the cost of capital is minimized at varying levels. unless the debt begins to enhance returns.2 %. the tax benefits may remain the same but now the probability of default rises . The student/investor should be cognizant of the tax benefits of more long-term debt because the correlation between long-term debt and stock price turns from positive to negative. that we can separate variables into “capital cost” components and “profit “ components which are more correlated with the current price.76% probability of no association with the current midrange price. They each showed significant correlation when regressed against next year’s price.500 concurrently strong correlation between earnings and stock price and a negative correlation between long-term debt to capital and earnings.long-term debt to total debt. but little association when compared to current prices. this indicator exhibits the same behavior as many of the earnings associated fundamentals like sales. although both can rise at the same time under the proper conditions. an increase in LTD/CAP might imply that retained earnings have decreased. the financial leverage ratio and the interest rate on total debt . In fact the lag is so pronounced for any variable that directly affects the cost of capital.displayed the same “leading indicator” characteristic. three other cost of capital factors . assets and operating income. While long-term debt and capital in isolation both exhibit stronger current price correlations. In fact. once they are paired in a ratio. DATA (See Tables) INTERPRETATION AND RESULTS The behavior of the long-term debt to capital indicator (LTD/CAP) defied expectations. Once the debt is carried over to the subsequent year. when regressed against next year’s stock price. it displays the leading characteristic of the other capital cost components. While it had a 76. . However. The reader should notice that when total debt alone was regressed against mid-range prices. with the addition of long-term debt in the numerator. it had a probability of no association of 0. the correlation becomes stronger with next year’s mid-range price. it displayed a stronger correlation with the current mid-range price than the next. Alternatively. With a smaller risk premium (the distance between the risk free interest rate and equities). gains of one and a half times the average could be expected simply by choosing companies that were lowering the LTD / CAP ratio through more retained earnings. there is a combination of market factors (interest rate and equity market increases) that contributed to the lag.501 Table 20-2 COST OF CAPITAL COMPONENTS: MORE CORRELATED WITH NEXT YEAR’S PERFORMANCE Long-term debt / Total Debt Interest Rate on Total Debt Financial Leverage Ratio: Operating Income / Operating Income . in an already inflated stock market. such a market “inefficiency” was the mechanism that powered this research. the effect on LTD / CAP might have been less pronounced and long-term debt would have been less expensive. However. Moreover.Interest Long-term debt / Capital Table 20-3 PROFIT COMPONENTS (MAIN): MORE CORRELATED WITH CURRENT YEAR’S PERFORMANCE Operating Income Sales Assets Capital Profit Margin Long-term debt Number of Shares Outstanding From a capital structure perspective. we can interpret the lag as an increase in risk that is not factored into the stock until it fully affects the cost of capital. Again. since the data describes a “normal” market with rising interest rates (the 1990s). . we cannot pinpoint whether the proportion of long-term debt to capital raised the cost of capital. and to observe at least some stagnation in the stock. The reason? There is no concrete measurement that can anticipate a rapid turnover of capital and assets: the sooner a firm can turn the purchase of an asset into a profitable venture. there is some lag between purchase and the generation of income. In these cases. Only the financial logic inherent in capital structure. No retained earnings are available. In this interim. it is typical to see the proportion of long-term debt to capital rise for more than one year. the reader can observe another avenue of approach. For these reasons. the stock is not in demand and long-term debt to capital rises. the increase may be part of a pattern that lasts at least two years if not more. When we accept the premise that both earnings and EVA are highly correlated with stock price.502 increases in long-term debt to capital ratios could even indicate an expansive market if interest rates are low enough and earnings prospects are high. Since most large debt issues are for the strategic purchase of profit-making assets. Companies who are in a sector that has not been favored by the market may use long-term debt to fund fixed assets that maintain existing operations. even statistical inference among all the variables will not give a satisfactory explanation. HYPOTHETICAL CAUSATION While correlation is not to be confused with causation. along with observation of the investment process. the more investors will demand equity in that company. give out executive compensation (golden parachutes) and consolidate operations. but the company must have a source of funding or face the prospect of downsizing and selling assets. we need to examine the . the confluence of operations brings unanticipated problems and the need for another (albeit smaller) loan. At times. Sometimes. Again. the incurring of debt is not a strategic play. can yield an answer. Long-term debt becomes a tool to forestall insolvency until the company can increase income. If we coordinate the behavior of long-term debt to capital with our knowledge of EVA. firms will often take restructuring charges. According to our correlations. In the exhibit of earnings correlations. Such a rise in the cost of equity would be a function of a higher beta caused by more leverage .the increase in LTD/CAP. Consider shortterm debt as an illustration. In essence. the cost of capital does not always respond to increased risk.503 capital components of the EVA equation. By the process of elimination. For example. Short. if current LTD/CAP is not highly negatively correlated with next year’s earnings. and a year later. That leaves a rise in the cost of equity as the reason for a diminished EVA and stock price. There is no factor that some manager can “grab a hold of” and improve next year’s earnings. and the proportion of debt to equity is all encompassing. THE HAZARDS OF PLAYING DETECTIVE The connection between the cost of capital.term debt may expose a firm to the risk of default more easily than long-term debt (more frequent payments at varying rates). it is difficult to obtain mathematical proof of causation because the sample continually changes as economic conditions arise: in one market. The insistence on interpretation through capital structure changes can blind us to other explanations that may be just as legitimate if not more so. the same indicator will be negatively correlated to it. and yet it is less expensive. the student/investor should notice the strength of correlation for concurrent fundamentals like sales and assets. the correlation between capital costs and risk is not always positive. Unfortunately we may tend to “read” a thread into a scenario where none exists. In fact. the net income part of the EVA equation is unaffected. there is no variable that is strongly correlated with an increase in next year’s earnings. Such a market efficiency has created the “random walk” theory because few can successfully coordinate available information to make accurate predictions. the following table represents a straight . and then contrast these when they are regressed against next year’s earnings: fundamentals have no predictive value whatsoever. a stock will rise with more long-term debt to capital. stock prices. 2 % chance of being associated.0. and the reader can observe that no momentum was exhibited. Obviously.9816 .04087 R Squared = 0.008280 or 99. However.504 autoregression of one year’s earnings increases on the next.which the data substantiates: LTD/CAP has a negative correlation with earnings of -0. The proportion of equity naturally rises because more earnings attract demand for the stock as well.195179 and a probability of no association of only 0. Even in the depths of the depression. Much of the upward trend in stocks that occurred after lowering long-term debt to capital.04087 (X) The minuscule coefficient of determination reveals that there is no predictive linear connection between year to year earnings. no less a researcher than Alfred Cavello stated that if stocks went up for a year. and Robert Levy of American University paralleled this approach with stocks. Manown Kisor and Van Messner worked with earnings acceleration.9816 Coefficient = -0. While long-term debt to capital is heavily negatively correlated with current earnings.039136 with a 0. they were likely to follow with another gain. several researchers have found that if earnings are greater than the market over a six month period. more retained earnings are created that lowers the long-term debt to capital ratio. Table 20-4 AUTOCORRELATION WITH NEXT YEAR'S EPS PERCENTAGE Y Intercept = 22. In the late 1960s. they will tend to rise in the next six months as well. there will be some response from the investment that will propel earnings . more long-term debt to capital is not depressing next year’s earnings.625401 probability of no association.001671 Y = 22. it would appear that this leverage ratio is inversely correlated with earnings . When earnings are high. may be attributed to earnings momentum. Thus. However. although it may stop earnings momentum. Both found six month momentum trends following acceleration increases. the correlation of current LTD/CAP increases to next year’s earnings increases is almost insignificant at -0. any large acquisition or merger or even a change in fixed asset technology. we were not prepared to observe a decrease in capital for the subsequent year when interest rates were raised during the current year. sales/assets. All of these risks would be reflected by some stagnation if not depreciation of the firm’s stock price. Paradoxically. -which makes up the major factor on the right side of the equation. We know that on the left side of the equation. the sooner a firm can integrate assets into a profitable scenario. Perhaps the best explanation of the dynamics of proportionality is found in the marginal benefits equation. In the next year. If one examines a typical default probability algorithm like Altman’s. the student/investor will find that many components have “assets” in the denominator retained earnings/assets.505 slightly higher in the next year. current increases in capital have no significant . When these variables do not increase. operating income/assets. However. the probability of default rises. we need to increase the performance values that are contained in the numerator part of the default probability . as well as the current interest rate and next year’s capital.the sales. Although we assume from capital structure theory that more capital will increase risk and lead to higher rates. Thus. and the marginal benefits equation will decrease. etc. we also raise the asset portion of the default probability. not to mention the possibility that the acquisition may not match the firm’s operating characteristics. the more marginal benefits will rise. requires a “period of adjustment” before these assets can generate a profit. The intensity of that “response” seems to determine whether the stock will become attractive and the firm can begin raising equity. more debt will increase tax benefits. This negative correlation occurs in a context where current long-term debt increases are highly correlated with current increases in the interest rate on total debt. working capital -the basic return on assets. Regressions were performed between current capital and next year’s interest rate. As we raise debt. THE ARGUMENT FOR CAPITAL RATIONING By far the most curious relationship was found between capital and the interest rate on total debt. without an infusion of more debt. which is generally analogous to accepting projects with a positive net present value. we can hypothesize at least three reasons for this occurrence. By no means an exhaustive list. given a specific cost of capital and recognizing future cash flow in current dollars. any of these beliefs will undermine a firm’s value when carried to fruition because they will frustrate movement toward an optimal capital . a firm accepts all projects that have a positive net present value. In normal capital budgeting. it is trading the viability of future projects in order to meet some short term objective . • 1) Over the eight year study period. The shift in priorities is often termed agency factors or “agency friction”. 3) the fear of losing voting control by issuing more equity. especially when interest rates are anticipated to rise. • 3) Firms ration capital and cut back on projects when a large acquisition is made in order to concentrate on making it a profitable venture. because management acts as an “agent” for the shareholder. While financial professionals recognize capital rationing. Accordingly. If a firm does not fund all projects that would exceed its cost of capital. it does not maximize its value and partakes in what is termed “capital rationing”. In layperson’s terms. From economics.506 association with current year interest rates at all. In short. that proposition obligates a firm to take on all profitable projects. These factors may include the following: 1) the belief that raising more capital will immediately raise its inherent cost. the scenario is less familiar to the average investor. we know that a firm should produce to a point where marginal revenue is equal to marginal cost. anomalies occurred in the data that created a negative relationship between next year’s capital and the current interest rate on total debt. 2) the recognition of greater default risk once a large block of debt is incurred.which are often numerous when large mergers are implemented. 4) the belief that current projects are either very risky and / or understaffed and require more managerial attention. • 2) The greater risk that is implicit in higher interest rates encourages firms to fund projects in large blocks of debt. Net Income / Cost of Equity. more debt raises interest rates which raise risk. The quotient of economic profit is the comparative dynamic. and thus changes in capital become negatively correlated with current risk However. net . In the capital rationing scenario. The alternative is to ration capital and maximize profit given those constraints. This alternative concept (adequate funding for smaller projects) is a reliable assumption given the strong correlation between increases in current stock price and increases in current capital (a correlation of 0. While it is always a possibility. the cost of equity is the full cost of the CAPM percentage multiplied by stockholder’s equity. Without transitivity of correlation. the firm will have problems financing new projects. equity is not attractive unless the firm is generating sufficient operating income which is difficult to achieve so soon after an acquisition. In this ratio. these capital intensive projects must be followed by smaller. The rights of a bondholder are almost always prioritized over those of shareholders which sets up an environment conducive to capital rationing. the mid-range stock price might decline. Evidently. with increasing leverage. the hypothesis of capital rationing cannot be proved. Companies that incur long-term debt are essentially giving up financial control to creditors. If indeed new projects require less capital. while at the same time operating income (ROA) is temporarily depressed by the need to integrate new assets and pay additional interest expense.507 structure. the premise is based on knowledge of a firm's specific IRR and “project pipeline” to which few individuals have access. Under-funding prospective projects will lead to diminished earnings and the potential to misallocate the proportion of debt to equity. Since many long term projects are funded with long-term debt (cash flow matching). more manageable projects in subsequent years. we can also observe that like the current interest rate.418036 with a probability of no association of 0 %). current long-term debt to capital increases are negatively correlated with next years capital increases. Indentures often restrict the raising of new debt to a specific times interest earned ratio (TIE) and if retained earnings are non-existent. In industries that require both a high level of technology and a high level of debt. project deferment can lead to obsolescence because a firm can only compete with the newest innovations which tend to have a short “shelf life”. The lack of a significant correlation between any current factor and increases in next year’s earnings leads to the conclusion that raising the cost of capital is the prime reason for a lower stock price in the subsequent year. they need to decrease the cost of capital.508 income is slightly diminished while the CAPM percentage rises because of an increased beta. even though equity funding was not extensively used. or if the stock was especially attractive at a low price. The result will be a diminished EVA. this study would have needed to not only . such an occurrence would diminish beta. our conclusions are merely speculative. when we examine the next EVA. Without specific reference to a numeric cost of capital. and a declining comparative dynamic. but by the time they do. For example. This “hidden” cost of debt is seldom discussed when firms integrate their operating and financial risks. the projects they would have funded are no longer legitimate. The only factor that could possibly offset this decrease in EVA would be an immediate decrease in LTD/CAP that would be derived from retained earnings or equity issues. we find that: 1) net income is basically unaffected by last year’s capital allocation. 2) capital outlays may be depressed if long-term debt to capital was increased in the last period. and 3) the incurring of debt raises beta. net income may not increase enough to raise economic profit. the risk of creating unsold inventory became far too great. but would only be achievable if operating income were bolstered. the PC industry was changing so rapidly in the late 90s that even with a lower operating leverage. which increases the cost of equity. because operating leverage only measures the amount of fixed costs and not their quality. Since all profitable projects are not being funded. Such companies are hit with a “double whammy”. To obtain statistically viable data. By default. The firm fails to move toward an optimal capital structure simply because resources were channeled toward last period’s capital allocation and away from current profitable projects. but may not relate to 2015. and raising long-term debt may actually benefit the stock in subsequent years. the student/investor is encouraged to do research on which leverage pattern is working for the top sectors. taxing dividends at the same rate as capital gains makes those stocks more attractive. It was the workable pattern of the 1990s. the reader must be warned with the caveat that this study can not be overlaid on other periods and be expected to be relevant. Greater demand for equity in these companies may shift the cost of capital. but to determine the amount attributable to myriad factors . For example. Other shifts include changes in bankruptcy laws and tax legislation. .509 measure the different costs of capital for each period. Therefore.Cost of Equity).the Federal Reserve. The probability of default was different for the same amount of debt held in different years. Similarly. and to understand why it is profitable by observing the components of the capital dynamic: (Net Income . the internal dynamics of the company. Thus. Levels and changes of the risk premium. and other viable alternative investments may change the way a firm moves toward an optimal capital structure. technology shifts. the stock market. the shift to a lower level of interest rates in the new millennium also shifted the optimal capital structure. such as - .020864 0. capital multiplier. Additionally.041765 -0.767631 0.024704 0. Several items were not listed simply because they were not highly correlated .135234 0.g. All items with a probability of association of 95 % are listed.275874 0. that fundamental is listed as well.036176 0.031123 0.171175 -0.150517 0.000606 0.002579 0.17042 0.106845 0.418036 -0.230556 0.166469 0.001741 0. Profit Margin LTD/Total Debt LTD/CAP Financial Leverage Capital Multiplier Long-term debt CURRENT MID-RANGE Correlation Probability 0.222137 0.020146 0.190681 0.000164 0. if an item is significantly correlated in one year but not in the next or previous periods.251781 -0.153022 0.102052 0.038071 -0.002816 0.151118 Each item in the variable column represents a percent change in that fundamental.36263 0 0.039177 0.058892 0.414726 0.609878 0.004632 0.140295 -0.220211 -0.149717 -0. it is also listed.072362 0.007661 0. but if an item has comparable value e.068727 0.149316 0 0.510 SPEARMAN RANK CORRELATION WITH PERCENTAGE INCREASES IN CURRENT AND NEXT YEAR MID-RANGE STOCK PRICES Table 20-5 VARIABLE Most Significant Operating Income Number of Shares Assets Capital Interest Rate Sales Total Debt EPS Capital Expend.209008 0.022051 -0.331668 0.415331 0.009926 0 0 0.159849 0..57561 0.155414 0.443389 0.06787 0.172137 -0.044241 NEXT YEAR MID-RANGE Correlation Probability 0.042544 0.043669 0.197081 0. 253549 0. Correlation is not causation.045173 0.09137 0.067203 0.037876 0.611708 The effect of changes in the current year interest rate on changes in next year’s fundamentals: Table 20-7 NEXT YEAR Variable Operating Income Sales Capital Assets Total Debt Capital Expenditures Number of Shares CURRENT YEAR INTEREST RATE Correlation Probability -0.045664 0.401484 -0.082255 0.000044 0.304204 Out of several variables.573031 -0.17418 -0.167101 0. SPEARMAN RANK CORRELATION ON CHANGES IN THE INTEREST RATE ON TOTAL DEBT The effect of changes in current fundamentals on current interest rates: Table 20-6 CURRENT YEAR VARIABLE Variable LTD/CAP Capital (Equity + LTD) CURRENT YEAR INTEREST RATE Correlation Probability 0.297846 0. This point is .035859 -0. only changes in capital have a significant (95 %) chance of being associated with increases in the current interest rate.56886 -0.511 asset turnover and end of year balance sheet cash.10865 0. Their exclusion does not mean that they are any less important than the listed variables. but it appears logical that less capital would be demanded if the price of it has risen. just that they lacked correlation with stock price increases for the specific years with these twenty-four S & P 500 companies. The effect of changes in current fundamentals on changes in next year’s interest rate on total debt: Table 20-8 CURRENT YEAR VARIABLE Variable LTD/CAP Capital Financial Leverage Ratio Capital Multiplier NEXT YEAR'S INTEREST RATE Correlation Probability 0. but as soon as longterm debt is incurred. the incurring of long-term debt obligates the company to a series of interest payments that raises the risk associated with the cost of capital.165464 0.000278 0. Evidently.068533 0.392213 Notice that the relationship between current interest expense and income (financial leverage) is not significantly associated with next year’s interest rate.416949 0. but this hypothesis is statistically inconclusive: the effect may be due to Federal Reserve hikes and not the implicit increase in corporate risk. Moreover.290095 0.512 even more emphatic in light of the previous table’s indication of no significant current association between the two. there is an effect on next year’s interest. tend to force the interest rate upward in the next year. greater capital outlays in the current year.037738 0.065025 0. SPEARMAN RANK CORRELATION BETWEEN CHANGES IN TWO CURRENT FINANCIAL LEVERAGE MEASUREMENTS AND CHANGES IN NEXT YEAR’S FUNDAMENTALS . 019291 -0.135525 Only capital has a strong inter-period relationship with the change in the financial leverage ratio and it is negative.6636 -0.185999 0.119273 0.00365 0.191216 -0.29166 0.000065 The association between changes in the current financial leverage ratio and changes in next year’s fundamentals: Table 20-10 NEXT YEAR’S Variable Operating Income Sales Capital Assets Number of Shares Total Debt Capital Expenditures CURRENT FINANCIAL LEVERAGE RATIO Correlation Probability 0. The association between changes in the current long-term debt to capital ratio and next year’s ratio: .103109 0.489493 0.034869 0. It is not proved.963699 -0.513 The association between the change in the current financial leverage ratio and the change in the current LTD/CAP measurement: Table 20-9 CURRENT Variable Financial Leverage Ratio LONG-TERM DEBT TO CAPITAL Correlation Probability 0.197328 -0. We can only assume that firms are decreasing the cost of capital by minimizing next year’s capital allocation.05538 0.104523 0. 104414 0.224585 0. Notice that both assets and capital in the next period are significantly associated and that every correlation is negative.173104 0. Notice also.004555 -0.664871 While this author believes that incurring debt tends to encourage more subsequent debt.083311 -0.005142 -0.221585 0.TERM DEBT TO CAPITAL Correlation Probability -0.003137 0.029626 This table should convince the reader that growth in the subsequent year was stymied by an increase of long-term debt to capital in the previous year.034729 0.17629 0. that operating income is barely affected which would provide the . the stock price would naturally decline.026711 -0.191686 -0. The effect of changes in current long-term debt to capital on changes in next year’s fundamentals: Table 20-12 NEXT YEAR’S Variable Operating Income Sales Capital Assets Number of Shares Total Debt Capital Expenditures CURRENT LONG. but the amount of change is not related to previous changes.138209 0.968793 -0. There is little association between changes in the amount of debt in juxtaposed years. this study disputes that assumption.TERM DEBT TO CAPITAL Correlation Probability 0.514 Table 20-11 NEXT YEAR’S Variable LTD/CAP CURRENT LONG. although we do not test for the statistical viability of a debt increase by itself. Since Wall Street tends to prize growth above all else. Long-term debt to capital may still rise in the next year when debt is incurred in the current year. This disconnect can be observed in the triad relationship between long-term debt to capital. but the challenge for the investor is to determine the time and quantity of that rise. current long-term debt to capital changes have a strong negative association with current EPS changes. the value of the company immediately increases because of tax benefits. we assume that capital rationing is not being implemented. the exact recipe for an improvement in the marginal benefits equation and (most likely) in EVA. When a firm increases its long-term debt to capital. sales.515 foundation for no association with next year’s EPS. On the other hand. earnings and stock price. SPEARMAN RANK CORRELATION BETWEEN CHANGES IN EARNINGS PER SHARE AND CHANGES IN FUNDAMENTALS We display only the capital components in the correlation with current EPS changes because most fundamentals have a strongly positive correlation with current earnings and to exhibit them would be redundant. Current long-term debt increases have little association with current stock price. these rises would encompass a large gain in tax benefits as opposed to a smaller increase in default probability. It immediately appears that investors are speculating about new assets. We conclude that there must be some times when the current stock price rises strongly in the light of a long-term debt to capital increase. It is right to assume that assets. but current EPS increases are heavily associated with it. We cannot prove that assets and capital are intentionally kept low to improve the return on investment. capital and operating income will rise concurrently with EPS. Since large blocks of capital are more cost efficient than the same amount spread over several years. . but we know from our study of capital structure that marginal benefits are increasing. and yet this rise in stock price is disconnected from earnings. 195179 0.202482 0. The effect of changes in current fundamentals on changes in next year’s EPS: .004616 0. All fundamentals must be smoothly coordinated for such growth to occur. What strategic actions can we take this year that will lead to increased earnings growth in the next? As revealed in this study.810172 0.516 The effect of changes in current fundamentals on changes in current EPS: Table 20-13 CURRENT VARIABLE Variable LTD/CAP Capital Financial Leverage Ratio Capital Multiplier Interest Rate on Total Debt Current Mid-range Stock Price CURRENT YEAR’S EPS Correlation Probability -0.017927 0.0568 -0. but is in the domain of corporate marketing.00612 -0. there is certainly no “magic wand”.209008 0.351096 -0. and such research is not in the realm of investor finance.004632 The information that would be most helpful to corporate executives would be forward looking indicators of earnings growth.14146 0. In fact.069514 0. There can be no leading statistical measurement that will determine whether a company's sales will suddenly escalate. part of the problem with forecasting earnings and the need for company guidance is attributable to this lack of correlation as evidenced in the following table. While it may not be as simple as increasing or decreasing one or two variables.517 Table 20-14 CURRENT VARIABLE Variable LTD/CAP Capital Financial Leverage Ratio Capital Multiplier Correlation -0.091873 If there were some association between any current financial statement item and changes in next year’s EPS (outside of ordinary maintenance like capital expenditures).134557 NEXT YEAR'S EPS Probability 0. Firms must minimize capital costs to reach an optimal proportion of debt to equity.277712 0.039136 -0. However. it would have been exploited long ago.870044 0.086882 0. In this state.013119 -0. profits will be nearly maximized while the price of the stock will be maximized (Back to Table of Contents) .625401 0. this “analytical confusion” has set up the capital structuralist with the opportunity to coordinate several variables and attempt to distinguish the path of capital costs. it follows a chain of logic that most investors are capable of implementing. The market. Formerly. By late 2000. and investors anticipated big pay days just around the corner. threw the card counters out. Although many investors will treat Federal Reserve rate hikes like a dealer’s up card. By late 1999 many investors were putting money into Internet stocks which were mostly start-ups with little or no earnings. while those in the financial markets are not. increasing earnings that were still negative. “The New Economy” was touted by the financial press. even companies with good earnings began to see their stock prices plummet. Some of these stocks shot up to fifty or one hundred dollars a share. unlike the house. As soon as professional gamblers began applying Thorp’s principles. good black jack skills meant “hitting” or “standing” on the basis of the dealer’s up card. or more likely. a statistician by the name of Edward Thorp wrote a book that completely revolutionized the game of black jack. The probabilities in card games are always concrete. Relationships between variables like inflation. GDP.the yearly average gain approaches ten percent . a strategy that gave most casinos a three to four percent advantage. leading to an inherent unpredictability. the similarities with black jack are not extensive. An example of a variable that typifies this chaotic flux is the case of earnings in 1999 to 2001.but variable relationships and their consequent probabilities change dramatically. the casinos began losing a lot of money and either began using multiple decks. stable and calculable. foreign exchange. will give the investor a favorable upward bias over time . or at best. The technical analysis that allowed investors to anticipate and profit .518 21 PROBABILITY AND CAPITAL STRUCTURE In the mid 1960s. he or she would have a distinct advantage over the house. The book was called “Beat the Dealer” and touted the premise that if the player counted the amount of ten count cards. and the level of technology are constantly evolving. 519 off trends began to fail. The semi-strong form states that all publicly available information. most financial professionals adopt the semi strong form as the most rational tenet. In the weak form. What happened? The story of the tech bubble is historically complex. To explain this seeming injustice. and investors began cutting their losses. SCREENS . the hypothesis states that all past pricing data is fully reflected in a stock’s price and can not be used to predict future prices. While illegal “insider” trading is almost always profitable. Since insiders do make money off of illegal trading. trading on the basis of fundamentals gleaned from a prospectus is not. In essence. the “pie in the sky” Internet companies went the way of Florida land deals and Texas oil wells. from past pricing data to newly published sales figures. the “New Economy” was over because the market reverted to its mean. The bloated market could not withstand investor perception changes. The correlation between the last period’s earnings and the next stock price was no longer valid. financial academics proposed the efficient markets hypothesis which categorizes reaction to information in three basic “forms”: weak. but it is safe to say that the market began correcting itself before a recession was imminent. THE EFFICIENT MARKETS HYPOTHESIS Analysts have attempted to predict cyclical shifts for many years without much luck. Finally. semi -strong and strong. In fact. the “strong” form states unequivocally that no public or even private information is useful for creating a valid forecast. an entire body of literature exists that categorizes our ability to predict prices based on prior information. can not be used to predict future prices. projected future earnings can not be the foundation for an investment if they lack the production capabilities to back them up. and it is found that the market almost instantaneously prices securities before those with publicly disseminated information have a chance to react. Without an asset base that represented lower operating risk. higher prices and lower expectations for GDP. 520 Stock screens have an unreliable history for the very reasons discussed above: The market will change enough to make them obsolete after they appear to be working for a while. These interest rate-sensitive stocks will have high price earnings ratios when they should be bought. The full return of capital equation was composed of (EBIT/Sales) x (Sales/Assets) x (Net Income/EBT) x (EBT/EBIT) x (Assets/Capital). With double digit inflation and unemployment. When the market is trending. but investors move on to more profitable sectors as interest rates rise. ROE directly trades equity for debt while ROC maximization . We know from optimization models that ROE will be maximized at a capital structure of all debt. while ROC will maximize at a structure of all equity: given a constant operating income. but will stop their “magic” as soon as capital costs outpace the acceleration of earnings. “reading” the economy was difficult at best. because higher earnings did not translate into higher stock prices. low and volatile in the bear market. will typically do well right after a recession. screens like the PEG ratio can work very well. Three out of four stocks will rise in a bull market and nine out of ten will fall during a bear market. The illusion of beating the market stems from the upwards pressure of earnings that tends to create momentum when the economy is in high gear. We further decomposed Assets/Capital into (LTD/Capital) x (Assets/Short-term Debt) x (Short-term Debt/LTD). and low ratios when they should be sold. THE RETURN ON CAPITAL Our adaptation of the Du Pont equation yielded a return on capital. one had to gauge the industry to see if it were to become suddenly obsolete. The correlation between earnings and price is high and volatile in the bull market. or over sensitive to the rise in commodity prices that occurred simultaneously with an oil shortage. as an example. and strong and stable in the long run. Cyclical stocks. depressing the prices. rather than a return on equity. Another example of a failed earnings screen occurred during the 1970s. but then stagnate for the rest of the business cycle. which is the exact opposite of conventional wisdom! The reason for this anomaly is that earnings increase throughout the business cycle. they merely offer a more elaborate method of observing risk. debt may be relatively inexpensive and leverage ratios that entail increasing debt might flourish. They are not meant to forecast stock prices. In essence each combination forms a “leverage state” that has a characteristic profitability and foreshadows near-term capital changes. Companies do not want to issue stock each year because such actions dilute the price of outstanding shares and have prohibitively high floatation costs. at the beginning of an expansion.. we intuitively discover that the last three parts concern capital investment that becomes the foundation for increasing the first two parts. and viewable on an investor level. there is ultimately a tendency to increase the profit variables by a greater degree when the leverage ratios are decreased. leverage ratios that imply equity financing might be more profitable. When we look for changes in the leverage ratios (EBT/EBIT is merely the inverse of our financial leverage ratio.521 has no interest expense. Different combinations of changes in the ratios seem to have a favorable or unfavorable effect on risk/return factors and ultimately on the company’s stock. . there are many instances when both leverage ratios rise and profit ratios rise. Of course. EBIT/EBIT. lower interest rates. because the market reacts to the proportion of debt to equity based on the price of risk of each respective component. For example. On the other hand. after interest rates rise.e. Secondly. There is always a lag time between capital investment and profitability. They are both actionable on a corporate level. LEVERAGE STATES Consider the nature of capital investment.companies do want to make public issues when the price of the stock is high enough such that the issued number of shares is kept to a minimum. but identifiable patterns emerge in which associations tend to be negatively correlated. While the patterns seem like the old screens “dressed” in probabilistic garb. and intracompany fundamentals. In fact. If we examine the equation by its five parts. operating margin and asset turnover. Yet .Interest) we will find that they sometimes offset the “profit” ratios of operating margin and asset turnover. and the stock reflects this risk. and offer an interface between trends in the capital markets i. and even more debt issues. the momentum of the company’s respective sector will determine. Rarely will a pattern of alternating yearly debt increases with equity increases be profitable. and again the stock falls. it is indicative of an insolvent company and the stock falls. a higher market will mean both more stock issues and more retained earnings which will boost the proportion of equity to debt. Successful companies will try to minimize the years that leverage increases. and maximize the number of years an investment pays off. small. ROC factors can tell us if we need to increase debt or equity. because there is always some lag time. At the “omega” point. then the company is viewed as not taking on enough risk to be competitive. and changes in the ROC factors will be derived from the integration of the production cycle with the capital markets. On the other hand. the movement back towards the optimal target structure encompasses paying off debt with higher earnings and retaining those same earnings to improve the company’s equity position. When many increases in leverage occur over a number of years. it tends to do so for as long as its sector is favored by the market. In essence. the level (proportion) of debt in the capital structure will be derived from the steadiness of the production cycle and the type of assets it needs. a company will retrench.522 once a company either starts to pay off its debt or replaces it proportionately with equity. but they cannot set the optimal target proportion by themselves. but most movements will entail the acquisition of productive capacity purchased with a debt issue. risk reducing acquisitions. The key to stock appreciation is to move away from the optimal capital structure and follow it up by a number of small jumps back towards it. Usually. INDUSTRY AVERAGES: LEMMINGS VS LEADERS . In essence. a leverage state will react to the market in terms of price and risk. as well as the stock price. If too many equity increases occur. and momentum that will make changes in each capital component fall into a trend. In this case. and hopefully “recharge” the cycle with research and development. Companies can move away from the target with either debt or equity. how long high-demand for its products will last. costs.523 The type of industry responds to market demand by establishing average values for each ROC factor. becomes the financial leverage ratio. Assets/Capital. Net Income/EBT. We measure the change in long-term debt to capital. establishes a pattern of profit. including operating momentum. entire industries will be taking on debt or building equity. The cost of debt. Operating momentum and the financial leverage ratio establish an interface between production and the cost of debt. allowing analysts to make predictions. Lastly we measure the change in assets to capital. Although the tax retention ratio. many companies within a sector will share the same leverage state. the optimal target structure may change to conform to it. and each company must make a required amount of capital expenditures at the same time. we are actually measuring “a change of a change”. simply because their response to product demand is so similar. in the case of operating momentum. In fact. . LTD/CAP. Thus. Depreciation rates are similar. While no industry average is “set in stone”. Although it is possible to take risks when other firms are “resting on their laurels”. operating momentum. and the cost of capital. firms must manage their tax structures within the parameters set by the government. The amount of fixed assets and the inherent risks they entail. We change operating margin into a dynamic ratio. a company must examine its existing prerogatives – especially if it finds itself “out of sync” with the funding actions of the competition. depending on existing market factors. In fact if a major macro-factor like interest rates is anticipated to be at a certain level for a long time. EBIT/EBIT-Interest. EBT/EBIT. which is % ∆EBIT / % ∆Sales. and establishing definitive corporate behavior for each leverage state. are treated as an increase or a decrease. THE LEVERAGE STATE RATIOS We modify the ROC factors to make them more useable. and capital needs. is crucial to the marginal benefits equation. the market will determine which capital increases are most efficient. it is not included in the leverage state because it is the least subject to strategic change. When capital structure is involved. All of these ratios. ROC factors generally revert to their mean. Since the market changes. Lastly. Others beat the market by a substantial margin. As an example. Any alternative financing done with short-term debt actually becomes more expensive and raises the cost of capital. While sentiments were “bullish” during those years. investing on the basis of a leverage state without doing prior research. Moreover. each of these states will be valued somewhat differently in each business cycle. which rises with the proportion. and may reveal dynamics not encompassed by other measurements. . COMBINATIONS The student/analyst/investor is encouraged to search for meaningful combinations of changes in these ratios that precede future growth. some of these leverage states produced paltry gains if any.524 Long-term debt to capital (LTD/CAP) establishes not only the proportion of debt to equity. consider what might happen if one increases Assets/Capital when the yield curve becomes inverted and short-term borrowing rates are above long-term rates. such that the least amount of risk was combined with the greatest amount of return. but implies the cost of equity. an increase in this ratio can imply that the cost of capital is decreasing because of alternative short-term financing or on the other hand. will create an uncoordinated response to current conditions. some of these combinations occupied a high position on the “efficient frontier”. It therefore requires more extensive analysis to interpret correctly. the ratio Assets/Capital establishes the degree of current liabilities employed by the company. that debt is so pervasive that solvency is endangered. The market then ends up revaluing any leverage state based on this ratio. and evaluating the context of the market . This author’s combinations are only suggested because they worked over the entire decade of the 1990s. In essence. THE MECHANISM OF LEVERAGE STATES When we can grow a company by increasing both debt and equity and simultaneously diminish the probability of default. DECREASE + + - Assets/Capital became a legitimate categorical variable when firms decreased their proportion of long-term debt and financial leverage ratio because it indicated a substitute method of financing. INCREASE 4. assets to capital will act differently when the yield curve changes. INCREASE 3. . INCREASE 5. it successfully predicted yearly capital outlays when it was increased while operating momentum and the leverage ratios were decreasing.525 Table 21-1 LTD/CAP FINANCIAL LEVERAGE RATIO + + + + OPERATING MOMENTUM + + + + + ASSET/CAPITAL 1. However. a lower cost of equity. Additionally. extrapolation into other periods represents a specific danger. or both. Mandelker and Rhee decomposed beta on the basis of operating risk. without increasing the cost of bankruptcy. DECREASE 6. The foundations for this mechanism are the following: • CHANGES IN BETA: Both Hamada and the Miller/Modigliani team had established that beta changed because of increases in the proportion of debt to equity. DECREASE 7. Such a scenario is characterized by an increasing EVA. DECREASE 10. In fact. with the 1990s data. DECREASE 9. and is perpetuated by lower interest rates. INCREASE 2. we will maximize the tax benefits of debt. DECREASE 8. popularity quickly eliminates them. Market inefficiencies seldom occur for any length of time and when they do. • TRENDS AND THE OPTIMAL CAPITAL STRUCTURE: When a firm is close to its target capital structure. usually because one of the two components of operating momentum. EBIT or sales. it is most probable that others have discovered it . From an economic standpoint. the stock rises. these funds must be gauged by the comparative risk/return profile of the market.526 Thus financial risk and operating risk will change the risk/return characteristics of a stock. as even this “free” source of capital can be costly when the market is especially overheated. which is measured by changes in EVA. its profits are nearly maximized against the cost of capital. For example. will begin using it. Reversion of the leverage state components to their means is less. • CHANGES IN THE COST OF CAPITAL: Diminishing beta will decrease the cost of equity. • EARNINGS PRESSURE: Any time earnings accelerates faster than that of other firms in the market. he realized that default probabilities did not correspond to prices. and that it will begin to work less and less profitably. a strategy that brings a company near its target will be almost self perpetuating. is accelerating as well. because a firm needs to take less action to achieve an optimum. the generation of large profits will have diminished the need for more debt. if I trade online and find that a “double stochastic crossover” is especially profitable. When Michael Milkien studied the junk bond market in the 1980s. He took that knowledge “all the way to the bank” until price and default were “in sync” . Consequently.which . the market factors in the inefficiency. More earnings retained at the same time that LTD/CAP is decreasing will give the firm a more favorable interest rate in addition to providing internally generated funds that will be less expensive from an accounting perspective. and the firm will try to take advantage of its equity position as long as demand for the stock is being uplifted by more earnings. 527 occurred as soon as the investment became popular. Similarly, Harry Markowitz did convertible bond arbitrage before investors realized that such an opportunity existed. Again, popularity led to its demise. Since these leverage states involve changing dynamics between leverage and the market, they are not exploiting an inefficiency per se, only the definition of the factors leading to the highest combination of return and risk. They need to be redefined whenever the market reaches a turning point, usually because of an imminent downturn, but sometimes because of technical innovation (popularization of the Internet) and even because of catastrophic events. Eight such market-changing factors are as follows: • • • • • • • • 1. Interest Rate Levels and Changes 2. Yield Curve Changes 3. Liquidity Preference 4. Legislation - New tax laws, Sarbanes Oxley, etc. 5. Foreign Market Risk 6, Exchange Rate Risk 7. Inflation Risks: Both current and Expected 8. Confidence in Equities As an example, after a downturn, more debt for a company with a strong balance sheet might lower capital costs because interest rates would have been lowered by the Federal Reserve. In that scenario, a leverage state that increases the proportion of debt and yet increases operating momentum as well, might be better than one that increases the proportion of equity. Stock would not be in high demand at this point, nor would retained earnings be high either. In fact the market might even downgrade a firm’s prospects for not taking on enough risk. Where does that leave companies that must finance with all equity? Because of their higher betas, such firms must depend more on earnings which would be relatively low immediately after a downturn. These firms must do all they can to eliminate higher operating risk by expanding customer bases so that there is less singular 528 dependence on a small group of customers or vendors. Such actions will lower risk without changing the fundamental nature of their businesses - usually a high profit albeit high risk and low debt scenario. MATCHING THE LEVERAGE STATE TO THE BUSINESS CYCLE The following table indicates tendencies and not absolutes. There are firms who take contrary actions to the business cycle and also do well. Table 21-2 ECONOMIC STATE RECOVERY CAPM Variables Market - Low Risk Free Rate - Low Beta - Low EXPANSION CAPM Variables Market - Medium, Increasing Risk Free Rate - Medium, Increasing Beta - Increasing MARKET PEAK CAPM Variables Market - High Risk Free Rate - High Beta - High and Decreasing Debt - High and Decreasing Favored Industries Capital Goods Transportation Operating Risk - High, Decreasing Financial Leverage Medium to Low and Decreasing Debt - Medium Debt - High and Increasing Favored Industries Secondary Goods Favored Industries Interest Sensitive Consumer Durables Operating Risk - Low Operating Risk - Medium, Increasing Financial Leverage Medium to High and Stabilizing Financial Leverage High Increasing The capital asset pricing model (CAPM) offers a simplistic but sometimes valid explanation for the conformity of some firms’ capital structures to the business cycle. If we 529 can accept that beta increases with the proportion of debt in a firm’s structure, and that the Federal Reserve raises interest rates as the economy improves, we can surmise that the greatest amount of debt should be incurred when both interest rates and beta are at low points. When the firm does the greatest amount of debt financing at the beginning of an expansion, it is incurring the least amount of risk for the greatest amount of potential return. As beta rises with more debt, the firm can take advantage of rising stock prices. At this point, earnings will be accelerating faster than the cost of equity and capital will be flowing into the company. As the Fed raises rates, debt financing becomes less and less attractive and the firm builds equity through higher retained earnings, and sometimes new stock issues, which have become more attractive with increased earnings per share. Decreasing the proportion of debt in the capital structure will lower beta, which is exactly what the company needs to do: by this time, the economy has peaked and begins to plateau, and earnings have begun to decelerate. Although the strategically astute firm’s beta has decreased, both the risk- free rate and the market rate have risen, creating more risk. Eventually, the cost of capital rises faster than the velocity of collective earnings and the economy goes into a downturn. From a CAPM perspective, the market risk premium will begin to fall, as more money flows into bonds; stocks will begin to offer both more risk and less return, although dividend yields will rise. The market acts like a barometer for earnings expectations as the rate of change of earnings becomes less than the rate of change for the cost of equity, and stocks become over priced. It is as difficult for any company to fall into this strategic imperative, as it is for investors to profit from it. The idea of moving “fund “ money into utilities and consumer staples during a downturn, is directly reflective of those sector's characteristically low betas. These firms have low operating leverage and steady demand even in a bad economy, and offer a return that exceeds that of treasuries. Acting, like a microcosm of the larger company, the investor profits by shifting assets into areas of lower risk. While that 530 technique offers both entities (the company and the investor) an immediate reward, it is not as conducive to lowering risk as is diversification. PROBABILITY AND DIVERSIFICATION Most academics have a healthy skepticism about beta. In fact, this book offers several alternative calculations for the cost of equity simply for the purpose of analytical diversification; if all methods point us in the right direction, the probability of being wrong is much less. One can never forget the Black, Jensen and Scholes study of the late fifties, early sixties that found high beta stocks to be negatively correlated with return, while lower beta stocks compensated investors much more than their inherent risks warranted. Since the theoretical ideal encompasses a “risk return tradeoff”, beta was dropped by many who were on the CAPM bandwagon. Mathematically, each company has a different correlation coefficient to the market index, which makes beta an issue of individual response and not collective dynamics - although the CAPM sometimes makes perfect predictions about the market. When the correlation coefficient is smaller, beta will take less preeminence than “Alpha”, which will explain such events as lawsuits, favorable legislation, and even industry collusion. The downward bias of high beta stocks should be viewed as both a warning and an opportunity for investors and corporations alike. Most of the stocks in the tech bubble were high beta stocks and yet the low betas stocks rose as well during that period. When a large company is growing by acquisitions, it can minimize its market volatility, but increase the “Alpha” component and respond with large price jumps at the time of the purchases. Such actions will keep beta to a minimum and yet potentially increase the stock price at a rate greater than the market. Secondly, when a company diversifies its operations into other related areas, it can expand its sales and lower its “unlevered” beta, creating a low probability of default. In fact, acquisitions are often implemented for this very reason: The reduction in risk is derived from a lower covariance between sales and the market - which is a theoretical determinant of beta. 531 SALES AND BETA A higher mean increase coupled with a lower standard deviation should be the sales goal of any corporation. Steadier sales will increase profits allowing more debt to be incurred when interest coverage increases. For all equity companies, more retained earnings will be produced, and funds flowing into the stock will follow a steadier pattern. In effect, a company can sacrifice some profit margin for asset turnover and produce the same return on assets while lowering risk. Of course, most companies do not want to give up the inherent pricing power that comes from higher profit margins and so they settle themselves into a “niche” where there is little competition. Since technology is a great “equalizer” and often removes niches, the importance of sales diversification is self-evident. The cornerstone of Mandelker and Rhees’s decomposition of beta was the strong relationship of a company’s sales to the stock market. To review that connection, recall that three main factors were involved: • • 1. The historical return on equity, ROE. 2. The current period Total Leverage, determined by multiplying the pure form operating leverage (operating momentum) by the pure form financial leverage: ( % ∆Sales) x (% ∆Ν Income / % ∆ ∆Νet ∆Operating Income) ∆Operating Income / % ∆ • 3. The most recent covariance of the return on the market with company sales, divided by the variance of the market: COV( Sales, Market) / Market σ 2 The final equation is merely ROE x Total Leverage x COV(Sales, Market) / Market Variance. The last factor (3) is the principle component of both sector rotation in the greater market and diversification within the company itself. For example, in the late 1990s, tech sales were prized more than natural resource sales and produced both a greater covariance with the market, and large amounts of collective operating leverage. During the Y2K fallout, the economy soared because of tech “upgrades”, but became more risky because the more volatile tech sector became a larger proportion of total GDP. One look at the basic 532 structure of covariance can tell us as much as most journalistic histories: All three components of covariance rose at the same time. • 1. “R”: Pearson’s Correlation Coefficient:: The market correlation with company sales soared as tech became more prominent and the Internet became popular. • 2. The standard deviation of the market return: As interest rates rose, the far greater return on the market produced a high risk premium. Money flowed into equities, but was often invested in a “promise” rather than real assets. IPOs, start ups, and biotechs that had no earnings history received large amounts of capital. • 3. The standard deviation of the change in company sales. Most firms enjoyed steady sales throughout the 1990s; their risk remained low. Unfortunately, the prominence of tech sales induced more operating risk into the economy and a collectively higher standard deviation for sales. The covariance is thus composed of “R” multiplied by the standard deviation of the change in company sales, multiplied by the standard deviation of the market return. The strategic implications of this equation are vast. Although few companies can time the market, financial management does try to take advantage of low interest rates or an above average demand for company stock. During an expansion, there is less risk from raising both leverage components, and beta will rise with the market. As the market “overheats”, companies can expand customer bases, and simultaneously lower both operating leverage and the standard deviation for sales. At the same time, they can lower their proportion of debt in the capital structure. Beta will naturally drop with less risk, but we need to remember that it will always be within the bounds of the industry. A high beta biotech will never drop down to 0.5, if the inherent beta is 1.5. Thus, unless the company wants to change its basic nature and become a conglomerate, it will “always be at the mercy of the market”. One other strategy that is more esoteric, would be to expand the alpha component in the regression equation and minimize the correlation coefficient, “R”. For example, a 533 fertilizer company that is registered on the New York Stock Exchange, but primarily mines bat guano in Africa, and sells it to the Middle East, might be somewhat buffered from the shocks of the American market. Although markets are more globally connected, unusual circumstances such as protective legislation, will immunize a company from beta. Another example of low correlation occurs when commodity prices in certain industries rise contrary to the economy; executives will juggle the strategic advantages of these associations to insulate their companies from market volatility. The connection between the individual investor and corporate finance is cemented with beta. The market favors a sector because sales generate income at a more rapid rate than the overall market can produce. Individual investors direct the flow of capital to these companies because they offer the highest return for the least amount of risk. On a corporate level, the industries in the favored sectors are optimizing their collective capital structures by changing the risk characteristics of leverage to lower capital costs and raise profits. Beta in these industries may be quite low, but tends to “over react” to positive changes in the market. Ultimately, sales accretion through investment in projects that balance risk and diversify operations, will expand the customer base, and create a favorable beta. PROBABILITY AND ANTICIPATION “Beta” and the CAPM have a theoretical base that is not always applicable. The market can change in unanticipated ways that can leave analysts scratching their heads. With more volatility comes increased uncertainty and functions that measure past performance are simply inadequate as forecasting tools. Even the marginal benefits equation and EVA are ex - post measurements of movement toward the optimal without reference to the future. While we can create an ongoing compendium of corporate results and check to see if the company is doing what it needs to do, data is received after the fact. Undeniably, a system that anticipates large capital movements would be superior to “chasing earnings”. If analysis were as easy as sighting large capital inflows and 534 calculating pay off times, both the risk and return in the market would be minimized; stocks would be like treasuries, and profits would be like clockwork. Although leverage states can point us in the right direction, the market will value each state differently at different times, and we need to gauge the larger economy as well as the individual company. By checking the leverage states of the most favored sector in the economy, we can gain some valuable information, especially if the sector falls into a pattern. For example, if my company is taking on equity while the most favored sector is taking on debt, I need to look at my company’s response to interest rates. If my company has an all-equity capital structure and eschews debt, I observe the operating momentum of the favored sector and ask questions such as, “Is operating momentum decreasing on the basis of more sales while operating margins increase? Can I anticipate a higher EVA because equity gains are made from retained earnings and not from stock issues? In some markets, leverage states in the top sector may form no discernible patterns: companies that incur large amounts of debt may be increasing their stock prices just like those who issue large amounts of equity. In such a chaotic sector, it would be best to avoid investment unless one intends to “get out fast”, because these firms will be “out of sync” with the capital markets, and a surge in profits may be followed by a case of classic “buyer’s remorse”. PRINCIPLE COMPONENTS ANALYSIS The discovery of leverage states was based upon pattern recognition from several sets of data. In the same way that a gambler checks off horse racing charts, leverage data was filtered through a series of pluses and minuses before it was submitted to more sophisticated testing. The data set exhibited a balanced array of patterns through what statisticians call “principle components analysis”. This technique establishes which linear combinations of data will create the greatest amount of variation and identifies components using an Eigen system. Each component is given a weight as a percentage of the whole. The greatest weight is not the most “important” component (although it often is), but the 535 element causing the most variation in the data. For example, the following chart applies to the data in the correlation studies: Table 21-3 VARIABLE Operating Income Assets Capital Sales EIGENVALUE 2.4335 1.0555 0.39864 0.1124 PERCENTAGE 60.8 % 26.4 % 9.97 % 2.81 % We know that no operating income can exist without a sale, so the paltry 2.81 % contributed to the variation by sales does not diminish its importance. We merely claim that any categorization including the operating income variable will be heavily influenced by it. In fact, with most combinations of three to four variables, one variable will be preponderant, and the others will not contribute an equitable share of the variation. However, the following table applies to the leverage state variables: Table 21-4 VARIABLE LTD/CAP Financial Leverage Ratio Asset / Capital Operating Momentum EIGENVALUE 1.1687 0.99438 0.95529 0.88154 PERCENTAGE 29.2 % 24.9 % 23.9 % 22.0 % These ratios are very close to each other and add similar amounts of information. They work well together as a “team” because both numerator and denominator contain interconnected pieces of data that contribute to the whole. Like the ROC / ROE Du Pont equations from which they were derived, the leverage factors are flexible and easy to manipulate, containing common dependencies such as fixed costs, interest expense and operating margins. While equality in effective strength does not imply that these ratios can 536 categorize leverage, they do form a measured interface between profitability and investment, albeit an imperfect one. STATIC VS FORWARD LOOKING RATIOS Making investment decisions based on past data is almost always doomed to failure. In fact, the efficient markets hypothesis evolved for this very reason - enough mistakes have been made while observing sales and earnings histories to decipher a causation. We term ex-post measurements to be “static”, not as a negative connotation, but to declare that more information is needed. For example, that net income increased twenty percent last quarter is a “static” measurement. Now if the analyst proclaimed that net income is anticipated to change at a rate of three times the market in the next quarter, that would be a forward-looking measurement. Unfortunately, such proclamations must be accompanied by a caveat, and are sometimes outrageously wrong as well. The leverage ratios, particularly operating momentum and the financial leverage ratio, are inherently forward-looking. The derivation of operating momentum increases is based on three years of data, and implies a high probability of a consistent change in operating margin, especially when momentum increases. However, operating momentum by itself is too volatile to use as a predictive tool. Both increases and decreases can lead to more operating risk and the measurement can increase ten fold when a company turns around from a bad year. Nevertheless, operating momentum has a strong tendency to revert to its mean simply because it is indicative of the type of cost structure inherent in the industry. If it decreases well below its mean for two years in a row, the probability of an increase is fairly high. Naturally, if analysts will give the investor a sales forecast, then operating income has at least some predictability. In fact, as mentioned in previous chapters, operating momentum will mimic true operating leverage in the next year, as long as costs are stable; that is, variable costs remain a stable percentage of sales and fixed costs do not change. Analogously, the financial leverage ratio will mimic % ∆ Net Income / % ∆ Operating Income as long as interest expense remains constant. On a theoretical basis, as 537 long as costs and interest expense were stable, net income would be totally predictable from changes in sales. We know that a precise number is not determinable because of volatility in both costs and leverage, but that the flow of change tends to revert to its mean. Thus, if we view total leverage as the product of operating momentum, and theoretical financial leverage, our equation would be: (% ∆ Operating Income / % ∆ Sales) x (% ∆ Net Income / % ∆ Operating Income ) = % ∆ Net Income / % ∆ Sales. Inevitably, a highly volatile electronics company with no debt can have the same total leverage as a staid chemical company that chooses to finance with too much debt; the total risk will be implicit in each company’s stock price. The connection between total leverage and the firm’s capital structure will be reflected by three aspects of the ratio, % ∆ Net Income / % ∆ Sales: • • • 1. The size (level) of the ratio 2. The magnitude of its component parts (net income and sales) 3. The volatility of net income and sales as measured by their respective standard deviations. To examine reversion to the mean, we will use Enesco as an example: This branded gift company had a stormy economic relationship with the 1990s, and was not amenable to prediction. To begin, we obtain a five year geometric average of both sales and net income using the “growth trend” technique. This is our target “guesstimate”, which does not have to be precise; in fact, we can even drop a year that we sight as “not in the trend” and add another year in its place. The figures in the table are absolute values from the respective year and not percentage changes because the geometric average is implicitly defined by the change between periods. 538 Table 21-5 ENESCO NET INCOME SALES 1996 38.4 845 1997 10.5 476 1998 -22.4 451 1999 26.9 384 2000 15.1 325 2001 1.1 269 Rather than choose the recession year of 2001 as typical, we opt for 1996 to 2000 because it is more in trend. Separate spans can even be chosen for each ratio as long as we keep them in trend. The geometric growth rate will be the ratio of the last year in the trend divided by the first year and then "exponentiated" by the inverse of the number between the years - four. For net income the method yields, (15.1 / 38.4) .0.25 = 0.791883. Note that the inverse of four, the number of years between the five-year interim, is 1 /4 or 0.25 which is the exponent. Finally, to obtain a percentage rate of growth, we subtract one from 0.791663 to obtain -0.2081 or - 20.81 %. For sales, the ratio is (325 / 845) 0.25 = 0.787511. Subtracting one yields - 21.25 %. The final total leverage ratio is - 20.81 / -21.25 which is equal to 0.97929. It is the “nature” of Enesco to move toward this number from either above it or below it, mostly through changes in operating momentum. Although Enesco was an extremely volatile company, we will use an earlier period to check our calculations, and additionally, we will use the financial leverage ratio (EBIT/EBIT-I), and not the theoretical ratio to derive total leverage. came close to it.59 -1.32 %.9424 Multiplying that figure by 4. 1992.1058 1.0607 1. we will use 1992 as an example.11 -1.4154 2. the imperative movement is reversion to the mean .0232 1.0227 1. what operating momentum would achieve the mean level of total leverage? To answer this question.375 -1.406 2.79 % equals 4.035 1.67 1.539 Table 21-6 Years Financial Leverage Operating Momentum Total Leverage 1990 1991 1992 1993 1994 1995 1996 1997 1.0845 1.0388 = 0.11 with an operating increase of 5. Part of the risk/return structure encourages innovative management to take chances such as increasing financial leverage when total leverage is well above its average.642 -5.6865 -1. The question for the investor and analyst is: given a level of financial leverage. The actual operating income rose by 5. in normal risk environments.568 1.2305 -1. In 1992.979/1. had analysts given a perfect sales forecast of 4.1549 -1. each year saw a move toward the mean of .0614 1.79 %. The company achieved a ratio of 1.05 -0.8658 -5. our best extrapolation would have been . although only one year.7679 1.51 %.692 Except for 1991.075 -0.0388 1.32 % versus a sales increase of 4.979. However.79 %. The entire basis of capital structure strategy hypothesizes that the stock could have moved up on these paltry figures alone providing that they were consistent enough to allow incremental increases in financial leverage. 540 simply because this measurement defines the nature of the company. Average analysts will not pick up on the movement and will dismiss the firm as “debt ridden” . and a “strong” company who uses debt as a strategic imperative to fund projects and optimize capital structure. and is hoping for a switch to a leverage state in which the company is taking on equity from higher earnings. when an investor is putting capital into a firm that is incurring debt. he or she is taking a higher risk. each market is different. and even sales to long-term debt ratios. THE QUICK PAY OFF Certainly. While we can measure the history of capital turnover (Sales/Capital). At other times. downside risk will be reduced. Thus. the more quickly investors get paid. the investor differentiates between a “weak” company who must take on debt to maintain basic production. Inevitably. the more rapidly capital is turned into profit. Unfortunately. In essence. Since vendors provide a specific cost structure and customers expect a specific price structure. the best run companies will try to structure their leverage to stay as close to the mean as possible. Often. the best prospect will be a company that is well above the industry average in LTD/CAP. Besides examining marginal benefits and default probabilities. an increase of all four leverage ratios will provide a high risk leverage state that is worthy of investment. and uses default probabilities to evaluate a firm. but moving back towards it. The debt issue should be large because it is less efficient to go to the credit markets frequently. which lends an air of predictability to this volatile calculation. Moreover. a small amount of debt is more indicative of a firm that is . companies must work within limitations and the result will be an attempt to keep this ratio stable. but if the investor operates in the capacity of a quasi.ratings agency. simply because the odds of “losing one’s shirt” are greater. Sales need to be consistent. more return is often culled from these “debt states”.which it is . several other tenets need to be adhered to: • • • Operating history should be strong enough to get a preferred interest rate. there is no magic formula that will guarantee a quick pay off. In the latter case.on paper. the investor has one of two choices: check analysts’ opinions about the acquisition and coordinate them with the firm’s present position in the business cycle. Therefore “long-term projects” are often funded by longterm debt. Barra. These “transition states” can be profitable but should be avoided if a large interest expense will be incurred in a few years. • Increasing operating momentum will lead to a higher operating margin . Such an increase can be a sign of greater vendor activity. This risk adverse approach allows steady inflows of income to be balanced by steady outflows of expenses . such a “run-up” may be accompanied by a sell-off. • Increases in the assets/capital ratio are generally a positive indication when long-term debt is incurred. Synergies in operations may take a few years to accomplish. or leave the stock entirely if it appreciates wildly in the first year after a major purchase. BARR ROSENBERG AND RESPONSE COEFFICIENTS Risk management was revolutionized by Barr Rosenberg. subsequent years may be lean. Usually. a large purchase will be accompanied by enough “hullabaloo” to increase speculation in the stock. small amounts of debt can indicate cost over runs. However. but the investor does not want to wait. At this point. trade credit. Additionally.which is needed to offset higher interest expenses. and a greater return on capital. while the infusion of debt gives an immediate boost to tax benefits.including sinking fund payments and interest expenses. with under performance of the market. The founder of the firm. Rosenberg formulated what he termed “relative response coefficients” which were . • Taking on debt when the LTD/CAP ratio is rising while the financial leverage ratio is decreasing leads to risky imbalances that must be closely examined. The student will remember from earlier sections that most firms will match expected cash flows with the type of funding. alternative financing.541 looking to fund existing assets and not purchase new assets. and that may be a euphemism for an acquisition that is not expected to pay off immediately. and the firm does not want to invest in obsolete machinery. a “fundamental beta” was derived. Rosenberg’s derived betas improved the predictive power of the measurement to forty-five percent with no market information 2 .3). even psychological factors may intertwine with capital funding to determine a target structure. Changes in interest rates and the potential for credit default weigh heavily on most companies.changes in beta. the theory of fundamental beta entails changing capital structure as a response to changing events in the economy. or an “energy crunch”. While the average individual stock is about thirty percent correlated with the market (R = 0. As an example. the company can either under fund projects or under perform the market because it is not minimizing capital costs. In the case of default. Thus. but the information and data requirements were onerous: over sixty ratios in seven different categories were used. to Rosenberg. their response coefficient was “five”. As an example. technology may be rapidly replacing the standards in the industry. The relative response coefficient of the individual security to the event When the coefficients were weighted by the overall impact of each major event. Rosenberg’s fundamentally derived betas did outperform historically derived ones in prediction models. The proportional contributions of major economic events to market variance 2. By not financing with lower cost debt. Therefore. which further reflect global conditions.542 weighted valuations between an individual stock and numerous macro variables such as inflation. extraneous factors count a great deal. if General Motors responds to an oil embargo five times as much as the S & P index. Barr Rosenberg theorized that beta was essentially composed of two elements: • • 1. The measure to which the economic impact of events can be explained is the same measure to which relative response coefficients can be interpreted . a company may not take part in financing with lower interest rates after a recovery because it observes ominous trends in fixed asset acquisition in the near-term. the response of the leverage states is more dependent on capital markets and is more derivative. while “fundamental betas” are comprehensive. the investor needs to be aware that the interaction between variables also changes. Rosenberg’s “fundamental betas” are similar to “leverage states” because they are driven by a probabilistic response to world events. each taking on a level of low. Additionally.543 whatsoever. the number of variables used in forming a leverage state should be kept to a minimum. medium or high. the average investor should not have to resort to such extensive cross referencing. leverage states are characterized by their economy. However. leverage states would yield myriad combinations Since changes in the CAPM and the over all economy are somewhat implicit in the leverage states. while defining a broader trend in capital structure. the predictive power went up eighty-six percent. As in the capital asset pricing model (CAPM). Analyzed in the context of the three CAPM variables. Only with a few key variables can a leverage state be both flexible enough to cover a variety of conditions. but economical enough to specify a capital structure type. (Back to Table of Contents) . Leverage states attempt to establish a category that is strict enough to keep that change to a minimum. When market information was added. Since correlations in any market change frequently. consider the fact that Wall Street employs many technical analysts. a method of rationalizing a probability that is already heavily in the market’s favor. stocks will follow a smooth trend without much deviation and investors choose a sell point that yields a twenty percent profit before a moving average begins descending. but we will not know the exact location of where the marble will end up. technical analysts will claim that with enough information about the various sub vectors and velocity of the marble. the stock market can take on an eerie coherence. Nevertheless. On the other hand. The gist of the argument revolves around whether charting and technical indicators merely give a false interpretation to the random variation that occurs when a large force propels an object in a specific direction. For example. both arguments have merit simply because “the market’ makes room to accommodate them. Just as financial markets begin to “make sense”. The stock breaks through its “support line” and keeps moving downward.a priori. If this seems like a religious debate where we need to choose one side or the other. The volatility indicator that “worked so well before” is no longer operable. In fact. .544 22 TECHNICAL ANALYSIS AND CAPITAL STRUCTURE Academics have pilloried technical analysis for many years. claiming that it was the financial version of “snake oil”. we can throw a marble across a hard surface and we can control the speed and direction of the toss. Academics will counter that such information is concurrent with the movement and is impossible to obtain . At times. Wall Street has continued to shell out millions to research firms that proclaim the ability to chart and predict the future direction of a stock. They portray it as mathematical illusion. and that solid financial decisions have been made with their recommendations. an ugly thing happens: the Dow drops by two or three hundred points. a mathematical conclusion can be drawn about the specific end location. Several questions need to be addressed. If we toss a coin hundreds of times. it rapidly rises. Capital in the form of a demand for equity will flow into the company as long as the company maintains its comparative earnings advantage with the market. Technically. and it makes esoteric patterns like “breakouts” and “triangles”. they believe that the market factors in all information concerning a security as it moves up and down. how can my back-testing of a moving average be so successful? How can Wall Street chartists make the right decisions with millions of dollars on the line? If stock picking is merely following trends between “resistance” and “support” lines. That is a mathematical fact. a company will move toward its optimal capital structure at this time. computer scientists were showing that repeated patterns could not predict the future direction of prices. If the academics are correct. a competitive advantage exists that is usually sector wide. Those who believe in the “efficient markets hypothesis” believe that prediction from past performance is impossible and allude to the “random walk” of this coin toss. If the . Ordinarily. and comparatively higher earnings are generated from six months to two or more years.545 Academics like to compare random volatility to a coin toss. MAJOR FORCES When a combination of high product demand coalesces with lower capital costs. and the financial forces that move a stock in one particular direction or the other. It trends. there is credence to this theory because even as far back as the late 1950s and early 60s. we will create a picture that looks like a facsimile of a stock chart. why wouldn’t it be standard practice for everyone? The answers to these questions involve a combined knowledge of the reasons for statistical variation. In fact. What is uncertain is the length of the moving averages that detects the stock movement. and put an up tick mark on a chart to indicate “heads” and a down tick mark for “tails”. the rise in a stock price will be mirrored by a rise in a short-term moving average above a long-term average. That technical indicators will work in these periods of low volatility. and mistakenly believe that a wider distance will give us immunity from a quick descent. and are infusing the firm with capital. When this “noise” trends in a smooth pattern. “smart money”. the various runs amount to statistical “noise”. No! What happens is that the upwards pressure caused by earnings creates a mathematical illusion: we begin to measure the performance indicator. most stocks will make several “runs” well above the moving average. less debt etc. can we predict how long it will stay above it? The answer is a resounding. and it appears that the trader is “out thinking” the market. If the upwards pressure is great. and taking advantage . earnings. traders believe they can make money by “getting out” while volatility is low. The technical equivalent of this rise is illustrated in a series of moving average cross-overs. the distance between moving averages.546 stock movement is slight. Obviously.) are great enough to propel the shorter term average above the longer term for a sustained period of six months or more. a fifty day moving average would be the short-term indicator and it would move above a longer moving average such as two hundred days. trying to predict short-term volatility. the real money is made when the stock is below its long-term moving average and is beginning to rise upwards. are anticipating an investment “pay off’ in greater product demand and higher earnings. there will still be a lot of money to be made if the financial forces (sales. it will be detected by a short interval -even by a few hours. and then make another quick ascent upwards As long as earnings out pace the market. However. and we detect a stock whose fifty day moving average is rising above its two hundred day moving average. those who have an insider’s knowledge of the industry. At this juncture. the imperative for the investor is to detect the forces that move these averages. On the other hand. Another uncertainty lies in how long these forces can continue If the upwards pressure is great enough. adds to the illusion of predictability. In fact. At this time. starting out very short-term and gradually gaining strength to where the trend is sustained by longer period lengths such as fifty and two hundred days. decline to a point right above it. try to predict volatility using an exponentially weighted moving average .547 of price run-ups on the basis of historical patterns is fruitless. the stock did indeed go up. and the stock can drop ten points a few days later. Although technical indicators can correctly reveal pricing pressure in either direction. such techniques are applied to asset value and not individual stocks which are inherently more volatile. The “new” indicator will now display a downward bias. In the interim (those few days). at major financial institutions. THE BANE OF VOLATILITY If researchers could predict the changes in the standard deviation for a security. and then quickly selling it before any uncertainty arises. bank stocks went up and down by as much as ten to twenty percent in one day. In fact. the bane of volatility is the cause of uncertainty. the difference between longterm rates and short-term rates. he or she is satisfied. Those investors who made money during a brief two week up trend. an investor can buy one hundred shares of IBM based on a short-term “relative strength” indicator. Wall Street developed an adaptive. When the curve inverted. would subsequently lose it in a one-day slide. The objective is to capture one day’s gain. value at risk calculations. “survivalist-type” technique to circumvent volatility: enter the day-trader. the market would be fully predictable. Although the day-trader may miss out on any further appreciation of the stock. The reason? Online trading sites allow traders to exchange huge blocks of stocks for a small commission. they fail miserably in telling the trader how long it will last. the day trader tries to eliminate volatility by catching a stock on the upswing.usually for one day. but the trader wasn’t “quick enough on the draw” to make a profit. Banks need to borrow short-term and lend long in order to make money. However. For example. For stock traders. As a reaction to the vicissitudes of technical trading. Consider the banks that suffered in the “credit crunch” of 2007: these banks had low betas because they generally made conservative movements based on the “yield curve”. attempting to profit from . along with software such as RiskMetrics. With a framework of only one day. most avid traders have knowledge of a specific portfolio of stocks that are especially configured for large movements upwards or downwards and conducive to technical trading.S. such that a “melt down” in the Chinese stock market will more immediately affect the U. several variables were weighed before a financial decision was made. In fact. and the question needs to be asked “so why isn’t everybody doing it?”. technical traders become obsessed with finding the one “trade tool” that will give an . Thus. SELF-FULFILLING PROPHECY The advent of the “information age” has created more random variation because response to world events is instantaneous. When all financial managers act in concert. and investments were firmly connected to their foundation . however. On the other hand. rumors about specific stocks and companies have nearly destroyed some institutions (Bear Sterns). if traders have enough stock. On a microcosmic level. For the same reason that people don’t take out a mortgage on their house and go play blackjack in Las Vegas: the risk for most people is far greater than the reward. they can trade frequently and not tie up money in commissions. and it is no coincidence that online technical analysis can move a stock when enough traders follow it. the loss is relatively small. it would be considered a sickness. At one time.548 the “asset turnover”. but has created more risk by inducing massive movements of capital into and out of certain sectors. economic variables have become more interconnected on a global level. market. That interconnectivity has demanded not only more immediate reaction to events. and so thus we have the “tech boom” of the late 1990s or the “housing boom” of the early millennium. an investment becomes a gamble. If he or she loses. In recent years. When the chance of loss (risk) is greater than the average return. there is less diversification of assets. Like golfers searching for the “perfect” driver that puts twenty more yards on their shots. This type of enterprise can be both profitable and nerveracking. if middle class America gambled with the size of the portfolio needed to make money in day trading. A wealthy trader can take a small bit of the portfolio and go day trading.a set of fundamentals that determined earnings potential. the stock did not trend long enough to make money. paging through historical data and not realizing that each market is defined by a different set of parameters. Low inflation rates will affect those indicators differently from high inflation. how does it react to the transition of one state from another? The simple “stochastic” indicator is familiar to most technical traders. most technical indicators are excellent tools for mapping out the past performance of a stock. The objective becomes to find a cohesive “exit strategy” that enables an investor to get out of an investment before “others see the light” However. Choosing one technical indicator. EX-POST PERFORMANCE On a historical basis. An investor who makes a large three-day profit and leaves the stock. Thus. The same dynamic that makes economies of scale an advantage to the largest firms in an industry. they will back test technical indicators to see how well they worked in the past. if a small investor is technical trading. On a technical basis. the acceleration of earnings can easily be coordinated with the volume flow of capital into the stock. a different yield curve is going to produce different results. In effect. the . and the leverage states of capital structure. A financial executive can match the demand schedule for a product line with the support and resistance lines for a security and view any correlation. and although there are many complex adaptations of this tool. gives large traders an absolute advantage in technical trading because of their potentially shorter time frame. Since most stocks track earnings.549 edge over the competition. “Stochastic” is a word from statistics that signifies random variation. but those who spotted the trend in a short time frame end up making a “killing”. Small investors simply must wait longer to make a profit because they have so little invested. he or she may see a rapid rise and fall before any money was made. will have an advantage over the “odd lotter” who waits four days and sees a big volume movement on that third day. This same historical correlation can be done between technical indicators. the simplest version was chosen to display. that territory is mapped out by those investors who have the most money and the shortest time frame. Thus. In capital structure terms. the indicator has some mathematical substance to it. its price will fall and the stochastic along with it. why would it lead to a higher stochastic? The answer has to do with two concepts: concurrent performance and momentum. and reduce beta and long-term debt to capital simultaneously. while a figure of 20 represents a “buy” signal because it has been “oversold” and will rise again. a fiscal year was chosen as the time period. Such an event often occurs when product demand is low. and the firm is also incurring more debt. Companies who were earning more would increase their amount of retained earnings. Thus the . However. The paradox should be apparent: if reducing risk by lowering beta is so important. when a firm moves away from its target capital structure. and the leverage state for that year was matched to the stochastic that was present at the beginning of the period. The reader will be reminded that the type of market and level of interest rates will determine the leverage states with the most potential profit. investing in a stock that out performed the others will pay off because that stock will rise again. any stock that out performs the market for any length of time is also at a greater risk of dropping.Low Price)) equates range with price level and can be done for any time period. as well as the stochastic that was present at the end of the period. Values of 80 and over signify that the stock may be a “risky buy” because it is “overbought”. When a market has momentum. It was our hypothesis that in this market (one with rising interest rates) companies that incurred more debt would have a lower stochastic because they would close lower on December 31 when the measurement was made. these companies would close higher on December 31.550 indicator is a ratio that shows how close a stock is to its highest level: ((Closing Price . If there is momentum in a market. Many recent markets have mimicked that “winner takes all’ scenario. and also have a higher stochastic. For our purposes.Low Price) / (High Price . stocks with high stochastics will continue to out pace stocks which are mired in a lower stochastic. Since the range (distance between high and low) can give an estimate of the variance when it is squared and multiplied by 1/16. 29 6. and to match analysts’ estimates of earnings with the leverage state. investing in a lower stochastic “debt” state would also be a prelude to a higher pay off and a reduced debt ratio when earnings accelerated and raised the price of the stock.35 Although we have alluded to the importance of raising the asset / capital ratio when debt is incurred. Lev. The stochastic and percentages are skewed even though medians and not means were used to average them. it is not a categorical separator. The following tables show the result of this brief study: Table 22-1 DEBT STATE 1 2 3 4 5 6 STATES Fin.18 -17.67 N/A 45. The burden remains on the investor to know which sector will be favored in the business cycle. The data is also market dependent and is not meant to guide the investor into a decision because it can change as the economy changes. while number 6 shows the worst.42 N/A 84. Start Stoc.64 % Change 3. In the above data. However. or raise the financial leverage ratio. N/A 66.21 -4.551 stochastic is as much a measure of performance as it is of risk. .23 74. In effect.47 55. a lower stochastic would preface “getting in early” before a stock took off. + + + + + + + + + + + - Asset/Cap. It does not delineate between those firms who will continue to perform well and those who are about to drop.39 End Stoc. In order to profit from the transition to an equity based leverage state. Nevertheless.29 N/A 54.98 45.83 74. number 2 and number 5 show the most promise as an investment vehicle. LTD/CAP Op Lev.65 -37.8 N/A 74. 47. or both. the lower stochastic may imply that the leverage state has more long-term potential as an investment.54 N/A 35. since these are states where the firm either raised the proportion of debt to equity.51 36.96 -14. recognize that this data was from a specific era (1990s) and that this state was actually a “favored” state three years later in 2002.73 80. when a shift in capital proportion is caused by asset sales.7 81.93 -0.99 17. and the low stock keeps rising. there is no guarantee that a firm with a debt laden state will not take on even more debt and wallow in a low stochastic for three or four years.32 From the above data. Table 22-2 EQUITY STATES (PROFIT) Leverage Fin.81 77. However. or 10 seem to offer the most momentum and stability while state 7 loses its price strength. At this juncture. A savvy investor will see the potential in a transition from a debt-oriented state to an equity-oriented one and invest when the stochastic is low enough to make a large profit. the company may not be growing and generating sufficient operating income. Similarly. if a debt state transitions into an equity oriented state without sufficient demand for the company’s products. states 8.79 78.78 72. but it can measure the earnings power of a firm in the middle of a downsizing or divestiture. beta may decrease.78 % Change -54. Thus. Lev. + + 80. Not only can EVA reveal the mismanagement of equity. 9. LTD/CAP Op Lev. the range of the stock .24 End Stoc.17 -0. the stock can fall precipitously. Before the reader begins selling this leverage state. 7 8 9 10 + + Asset/Cap Start Stoc. Leverage states are never set in stone because the economy changes around them. 26. One sign of financial strength is a rising low stock over a number of years. For example. They respond to shifts in the way that risk is being priced by the market.552 earnings must be accelerating faster than the cost of equity.22 77. the EVA/capital dynamic remains the best measurement of true investment worth. When the high stock is relatively stable. EVA will rise and the stochastic should rise as well. the opportunities presented to each are similar. It seems that we measured the stock as it hit a speculative high and then descended.5 %. a decrease of 25 %. Nucor. the turnaround time depends on :technology.Low =70 . Now if the high price ran up to $100. Two notable exceptions exist to stochastic conformity. stocks tend to rise and fall as a sector. and the resources with which the projects are endowed.Low / High .5 or 50 %.14 % gain and yet the stochastic would be going down to 37. management. and that is true for stochastics as well. On the other hand.375 or 37. the risk of this stock going even lower is very high in spite of the lower stochastic. the beta of the stock is lowered relative to its industry which will be actualized by a lower covariance with the market. but still had a profit “attached” to it.553 narrows and it becomes less risky. Thus. The stock would have increased to 75 which is a 7. It would not behoove a company to take on masses of debt when the top competitor is “raking in the dough” and paying off its latest bond issue. for example. targeted customers etc. Notice the mechanics in the following example. Obviously. The first exception is that many companies will occupy a unique product “niche”. While it is not sacrosanct that corporations must act in unison to be successful.which is indicated by their rapidly falling stochastics. STOCHASTIC CONFORMANCE Competition in an industry takes place within the framework dictated by such variables as demand. may be on its way down because a speculative high may have been reached. a stock that has risen with a lower stochastic. earnings potential and the cost of capital. The two most risky leverage states in our stochastic review both had price run ups before falling . specializes in steel alloys that have made them .5%.00: 75 .60 = . They have diversified and then specialized in an area in which they have particular competence that gives them a competitive advantage. The process is similar in each firm in the sector with slight differences between service and manufacturing orientations.60 / 100 . Close . Once capital is infused into a firm.60 / 80-60 = 0. In effect. the standard deviation of the stock declines. and the firm must work to lower its probability of default. and the stock will continue to rise. These companies will generally see a declining stochastic as initial speculative run-ups begin to falter.TECHNICAL ANALYSIS? Years ago when the CAPM was first proffered by academia. Many corporations have become so large that their only hope for growth lies in buying other companies. The second exception is in the area of acquisitions. Since it is difficult to integrate a large amount of assets rapidly. variable costs are not as affected as they would be with other steel producers. and with conflict comes (hopefully) profitable resolution. however is probabilistic and not deterministic. and there is always room for another legitimate idea. we see monetarists squaring off with Keynesians. . higher bankruptcy costs can be avoided. growth stock investors versus value stock investors. Thus. The search to make finance more objective and scientific has progressed into a linearity of thought where a specific postulate is staked out. but their core industries may be growing at the same pace as the economy . ad infinitum. turnaround times for capital will usually dictate a slightly higher bankruptcy cost in the next period depending on the competence of the integration. operating income will begin rising almost immediately. CAPITAL STRUCTURALISM: QUASI . supply-siders versus rationalists. and marginal bankruptcy costs must decrease in the next period. free traders and protectionists. beta was accused of being another version of technical analysis. tax benefits are rewarded in the year of the acquisition. Adopting opposite sides of an issue allows debate to occur. or the stochastic will be much lower.four percent. except that it had a university pedigree and came with a set of “white papers”. When commodity prices go down. The politics of opposition are especially apt for the rift between Wall Street and academia because each panders to extremist positions.554 less dependent on steel as a commodity. If the integration is smooth. Wall Street dictates that the firm needs to expand faster than the industry itself. From a capital structure perspective. Finance. they need to grow at twenty percent. promulgated and then defended. While technical analysis can be adapted to capital structuralism. Over many years there has been an upward bias in the market of approximately 9. FIGHTING WORDS: “THE EFFICIENT MARKETS HYPOTHESIS” The specter of the “efficient markets hypothesis “has haunted Wall Street for decades. information travels so rapidly that no individual can gain a competitive advantage. In another. Thus. and since technical analysts possess the same library of indicators. diversifying operations and lowering operating leverage (and beta) is the key. Consequently. it is no wonder that most mutual funds under perform the market despite access to millions of dollars worth of research. and each company being researched extensively.8 percent per year (2008). and remain flexible.555 Capital structure can be technical tool when it is matched with an indicator like a moving average. industry and company. Another product of academics. but in isolation. Since most fundamentals are published after the fact. income and default. but technical analysis attempts to measure and predict performance while capital structuralism tries to supply the impetus behind what is being measured . it is a legitimate method of optimizing the wealth of a company.the economy. it briefly states that with millions of buyers and sellers in the market. In a third economy. some of the touted success of both fundamental research and technical . in one economy more debt maximizes the stock price. will ultimately optimize wealth. Capital structuralism recognizes that only strategies that adapt to the confluence of risks for economy. sales. It professes not to predict the price of a stock but to predict the risk of a firm in the domain of several variables . each field makes extensive use of the different adaptations of the standard deviation. and three out of four stocks rise at this time. only mergers and acquisitions will work. Even rumors of earnings deficiencies become factored into the market. taxes. optimizing capital structure is under the aegis of risk management. All of these strategies will have the one commonality of movement toward an optimal target capital structure. The upward surge of a bull market has been far more frequent than any downward shifts. Because the market keeps changing. the stock would have sky rocketed anyway. While an efficient markets purist would discount pricing pressure as random variation. . Burton Malkiel. it does. An additional “chink” in the armor of efficient markets includes momentum. even those who strongly believe in the efficient markets hypothesis will claim that this technique has been substantiated. However. Many technical indicators measure the force of momentum which can logically be applied to those select stocks that keep rising. Stocks that out perform the market in price or earnings for a six-month period tend to do better than the market for another six months. What about capital structure? Wouldn’t an efficient market factor in such movement? In fact. Even when an initial investment is made. Several researchers have found a momentum variable that can be exploited in some stocks. Indeed. If the investor observes a state that is conducive to growth.556 analysis has developed from the illusion that the acumen of the analyst helped pick the “right security at the right time” . neither the market nor the investor has any certainty as to the length of time the investment will begin to produce more earnings. the tax benefits may cause some upward movement in the stock. and because the mathematical variety of trends.when in reality. it would create its own demise because the market would factor it in. This inside information is a great opportunity for the capital structuralist. the market may have changed and the company may no longer respond to it. The advantage of corporate insiders is that they have knowledge of project timeframes when investments are expected to pay off. In fact.. However. there is always one indicator available to exploit a specific company’s situation. in his classic book A Random Walk Down Wall Street . by the time this “needle in the haystack” is found . there is a high probability that the firm is moving toward an optimal capital structure. and company executives have been putting up their own cash. because the leverage state can be matched with public knowledge about insider sales. emphasized that if any specific technique (such as stochastics) worked for any length of time. volatility and correlation are so great. Even great fundamentals will elicit downward pressure in a bear market. EVA will decline and the stock price will be diminished. the association is not nearly as strong as it is over the long term. Specific . However. A trend can be exploited for as long as it continues. Analysts have taken this need one step further: by extrapolating earnings from sales and then observing the correlated movement of the stock price over many years. Sector domination and increased market share would be observed in a surge upwards. in the near term. low operating leverage products cannot be forecasted correctly one hundred percent of the time. interest rates and inflation. By correlating performance to actionable policy. would be mirrored in a downturn. for example. There is no predicting how long that trend will continue. Although prediction of stock prices is forever suspect. Since many stocks move in large irregular “jumps”. technical analysis is an instructive tool. THE ART AND SCIENCE OF FORECASTING The chain of logic behind forecasting begins with the fact that all companies need a sales forecast to coordinate the various activities of a firm. it can be matched with the forces that affected stock price. If earnings are extraordinary but too much equity is incurred at too high a price. and 2.557 two points appear to be valid in technical analysis: 1. Losing out to competitors. That high cost of equity is one of those “outside forces” to consider because it is partially imposed by the relationship between earnings. Our own analysis of EVA confirms this near-term inability to correlate earnings with stock price. Students of finance should be especially familiar with forecasting techniques with the caveat that even steady. a demand forecast morphs into a stock forecast. However. as a management tool that gives detailed description of past performance. the application of technical analysis to exploit market inefficiencies appears to be limited. planning is essential to any firm. predicting the forces that move a stock and combining them into a consensus that smoothes out inaccuracies in any specific prediction is standard for financial executives. because random variation from outside forces has such a strong pull on the stock. specific industry and personal wealth that stock selection entails. When the short-term average is above the long-term. . these trends are indicative of the first and second derivatives in calculus. “Upwards pressure” does not necessarily mean that sales are accelerating. Starting in November. and finally drop to where it is equal to the long-term average in March. No one “needs” to buy a stock .moving averages-can let the student observe the connection between day to day activities and the larger financial imperative of maximizing wealth. sales are accelerating. and above all .number of vehicles on the road. business or government. capital structure begins with the sales forecast because too much or too little capital raised will cause a deficiency in the relationship between earnings and the target proportion of debt to equity. Selling tires has a degree of predictability . Holt-Winter exponential smoothing. When that number is rising. the short term moving average for coffee will rise faster than the long-term average . you may do the bookkeeping for a “mom and pop” ski lodge in Vermont. but that there is more tendency to do so. The first derivative (upwards pressure) will show whether a number is increasing or decreasing. MOVING AVERAGES TO USE WISELY Some tools of technical analysis can be very helpful to anyone who needs to spot an accelerating trend. season. the bookkeeper subtracts the long-term average from the short-term average. Indeed. While tire forecasting can be a sophisticated science. In December or January it will be above the long-term average. psychology.consumer. One major reason for discrepancies between sales forecasts and stock predictions is that sales are generally based on the needs of an economic entity . You want to contrast the amount of coffee sold to skiers in the winter versus the amount sold in summer. structural changes in rubber etc. To measure the acceleration. Box-Jenkins analysis. A coffee sales moving average will show the difference. For example. it does not approach the random variations in economy.although some people may think they do. there is upwards pressure on sales.558 techniques like Arima. In fact. To illustrate the advantage of creating a smoothing constant. Like a human being. understanding the concept of contrasting moving averages will give even the small business person a usable form of calculus. (2/(N+1)) where N is equal to the period that you specify.559 while the second derivative (acceleration) will show if the number is increasing at an increasing or decreasing rate. THE EXPONENTIAL MOVING AVERAGE This moving average is a very flexible tool.2. we create two smoothing constants based on the length of the average. . consider a spreadsheet where a moving average requires a specific number of lead cells. and yet simple to implement. we repeat the same procedure for both the short and long-term averages. and it is adaptable to any time period you specify. this simple spreadsheet function offers the student an introduction. Unlike many spreadsheet-type moving averages. or self-correcting behavior of the average. it does not need any lead cells before it begins to work. For example. and is discussed in the next section. The Wall Street Journal uses these to display trends and many other businesses use them to track sales and inventory. The exponential form is self-correcting (recursive). if you want the short-term average to be nine days. “neural networks” are supposed to make comparisons based on learned adaptation. the smoothing constant would be: (2/10) or 0. To circumvent the need for a lead time. Any time the average is used. it must have the specific number of cells as a “lead”: This example is a simple five day moving average. Thus. The basic kernel of “artificial intelligence” is observed in the recursive. When forming a working moving average system. While the concept is more profound than the practical calculus of an exponential moving average. The smoothing constant is merely. or 50. or inventory data. one cell below.560 Table 22-3 EXCEL COLUMN E1 E2 E3 E4 E5 SPREADSHEET 25 30 35 40 30 EXCEL COLUMN F 1 (BLANK) F 2 (BLANK) F 3 (BLANK) F 4 (BLANK) F 5 = SUM (E 1:E 5) / 5 Notice that we needed (N-1) blank spots before we could begin the average. setting it up against itself: Cell I 3 = (2 / ($G$2 + 1)) * (E2 . 21. and the other for specifying the long-term length STEP 1: We will use cells G 2 and H 2 as blanks for inputting short and long periods respectively. We form the first of two equations beginning in cell I 3. It uses the answer in cell I 3 to correct itself: I 4 = (2 / ($G$2 + 1)) * (E 3 . and the beginning .E2) + E2. The first equation begins the process of smoothing. and that we could not specify a different number for period length like 10. the logic of the process is illustrated. In the step by step process I am about to give you.which of course. STEP 3. stock prices. STEP 4. We can copy data from a CSV or text file by pasting it into E 2 or we can input the data manually . The data can be sales data. Label cell G 1 “short” and label H 1 “long”. only two blank spots are needed. preferably at the top of the spreadsheet and side by side. The formula in cell F 5 would be copied down to match the remaining length of the data. In cells J 3 and J 4 you will copy cells I 3 and I 4 respectively except for changing the smoothing constant. Although the net outcome is the contents of E 2. The second equation is inputted in cell I 4. One blank spot will be for inputting the short-term period length.I3) + I 3 STEP 5. STEP 2. is tedious for very long sets. These cells will be your long-term averages. Our data will begin in E 2 and we will label cell E 1 “price”. We identify a column in which to place (copy from another file or input manually) our data. currency exchange. For stocks. many traders use a 50 and 200 day average for short and long respectively. the sub trends can continue for weeks at a time in either direction and will be punctuated by large unpredictable jumps. K 3 = I 3 . and vice versa. Both equations in I4 and J4 can be copied down for thousands of cells if you wish. From now on you can just input periods into cells G2 and H2 and then copy data into E2. he or she will be trading frequently with no guarantee of making any money. we subtract column J from column I starting in cell K 3 and copying down for the length of the data. The student should be aware that the smaller the time period.561 part (the part to change) will be (2 / ($H$2 + 1)) which will refer to cell H 2 (the long period) rather than G 2 (the short period). A word about period length. In this game. BRIEF INTERPRETATION The gist of technical analysis is to follow an upwards trend and sell before the risk of a downward trend is too great. While the investor can be aware of the overall up trend because the 50-day average is above the 200-day average. While the investor can observe acceleration very acutely with small averages (say 9 and 14). STEP 6. In fact. STEP 8. When the distance between short and long becomes greater. and the stock stays in a stable trend. the greater the sensitivity to acceleration. there is upwards pressure on the data. Some stability of trend is established with this combination. . you click “do not save” and you are ready to input another set of data. STEP 7.To establish the acceleration factor. tempting fate by depending on one measurement invites disaster. When you are finished with analysis. the brokerage (“the house”) gets to keep their commissions while the investor (“bettor”) barely breaks even. Gambler’s Anonymous is full of certified geniuses who have tried several systems and failed. use of moving averages can mimic Three Card Monte: one can “up the ante” for a big payday if luck falls the investors way. and you can save the set up as a template. Inevitably.J 3. The time to be in the stock was when the initial force of earnings was propelling it above the long-term moving average. investing on the basis of capital structure will at least give the investor time to leave the stock after making a profit because these stocks will be rising for a period of six months to two years or more.562 PRIMARY TRENDS AND SECONDARY TRENDS The combined force of higher earnings and a lower cost of capital will push a fiftyday average above a two hundred day average. Those who are still greedy after a twenty percent run-up will kick themselves if the stock begins to “tank”. (Back to Table of Contents) . once the stock has “jumped up” (short term average moving above long term). In the case of a speculative run-up. A company that is dominating a favored sector will create a large difference between short and long moving averages. In the case of a general economic expansion. and the student/investor should understand the difference. However. The secondary trends that are a combination of volume generated rises and noise can fool an investor into losing money. If one is addicted to watching stocks. if it were not included in this text. the chartists will immediately look for an exit strategy leaving slow investors “holding the bag”. speculation can create such an up-trend so it is mandatory that the investor do the research and find out whether the trend is sustainable. the more strength and stability the price rise will have. investing in the secondary trends is a bit like fast. The reason? The longer the short-term moving average remains above the long-term. an investment becomes a gamble. Indeed. many stocks will trend upwards without much impetus. but this author would be negligent. This longer term trend is the primary trend and is manifest from a firm’s pay off on its investment (either debt or equity). “Vegas action”. At times. Investors must decide when to get out by balancing the profit already made with risk aversion. When the risks far out weigh the rewards. and the investor wants to be in a stock whose company is in the process of optimizing its capital structure. much of the movement will be a combination of sustainable trend and statistical “noise” which will be difficult for even top mathematicians to decipher. The larger the amount of variation. Variation among stock price increases. taking liberties that are mathematically invalid. Thus. each advocate must prove a point. as focusing on measuring the variation itself. Unfortunately. The individual must make several decisions based on facts that he or she culled together. but nevertheless making an attempt to draw actionable intelligence from them. And variation in finance is where the root of most problems lies. For the small investor. we are in fact.it is not heavily concentrated around an average but tends to be distributed with large amount of sample points at one end or the other. Whenever two agendas collide.to point us in the “right” direction. financial data is usually highly skewed: that is .563 23 STATISTICS PRIMER There is a vast market in America for badly applied statistics. When we try to fit financial data to a “normal” curve. However. very few of the measurements are statistically valid because of small sample size and large amounts of variation. We concentrate not so much on measuring an average that is derived in the absence of variation. access to real time corporate data is simply impossible. If the individual . the harder it is to prove a premise is true. The pragmatic reality of investing is to make decisions with insufficient data. for example. finance and statistics make strange bedfellows: without adequate measurement. can approach two or three times the average which renders many price related studies inconclusive. And what better way to lend an ambiance of pseudo-science to one’s methodology than by exhibiting an average that has no real statistical significance? In fact when we speak in terms of the “average American”. it is an extreme misapplication of statistical technique because there is so much variation among the population. a sample type and size must mathematically adjust to the amount of variation or the study will be inconclusive. proper capital allocation could never be made. And that is the essence of financial statistics . We count the amount of figures in the sequence and obtain a sample size of “7”. any statistical liberties we take must be used to avoid excessive risk rather than to speculate on large returns. The sum of any sequence of numbers divided by the amount of figures in the sequence (sample size) will yield the mean. the student/investor is encouraged to “diversify” indicators and use many different types. 32. the two numbers in the middle would be divided by 2. . MODE AND MEDIAN MEAN . defines the relationship between numbers through “parameters” such as the mean. Since there are three numbers both lower and higher than 17. The following example is a sequence of numbers: 7. 3. the flexibility of what is termed “parametric statistics”. say eight instead of seven. and dictates a balanced approach.When a sample of numbers is arranged from low to high. the median is the midpoint with an equal count of data points both lower and higher. in the above sample. In that case. Throughout this book. 60. we arrange the sequence from low to high: 3. 21. Since financial data is so heavily skewed. 32. all entries are unique and there is no mode. We then sum all of the sequence and divide the sample size into the sum: (7+17+14+32+60+3+21) / 7 = 22. but the median would be an average of the two middle numbers. THE MEAN. But . If the sample were an even number. 60. MEDIAN . 14.The mode is the most frequent number encountered in a sample. 17. that is our median. In most cases.if several measurements point us in the same direction. Thus. the investor can be satisfied that a rational decision was made given the available information. the median better incorporates inherent variation. 21. However. 14. MODE . In the sample above. 17. in many cases the median is a far better indicator of distribution than the mean. 7. there would still be three numbers both lower and higher. any “statistical significance” is tested by the amount of variation.564 is to be guided correctly.The arithmetic average of a population is its mean and is the most used measurement in statistics. the sample size will be inadequate to draw firm conclusions. populations are often “skewed” and heavily weighted in one direction or the other. In fact. Median. The following distributions show the locations of mean in relation to median. and Mode Figure 23-2 LEFT OR "NEGATIVE" SKEW Mean Median . Figure 23-1 NORMAL DISTRIBUTION Mean. In finance. median and mode will be the same number.565 When a population is “normal”. data points will often take on the characteristic depiction of a “barbell” and be heavily weighted on both sides with few points in the middle. This “bimodal” distribution is difficult to interpret because its lack of central tendency eliminates the statistical significance of mean or median. the mean. it appears that two separate populations exist which makes it difficult for the analyst to determine significant relationships. Only the type of industry actually mattered.566 Figure 23-3 RIGHT OR "POSITIVE" SKEW Median Mean In a bimodal distribution. any oil related stock was doing quite well while the rest of the market lagged behind. Any relational statistic between the two sub populations such as the amount of assets or sales would have been insignificant. Median. for example. Mode THE VARIANCE AND ASSORTED ADAPTATIONS . Figure 23-4 BIMODAL DISTRIBUTION Mean. In the market of 2005 to 2007. and can be used in predictive forecasting. µ is equal to the mean of the sample.µ) / N. For example if X is a number in the sequence (1. “∑” is the Greek letter “epsilon” ∑ which designates that each operation in a sequence should be summed.3.567 THE VARIANCE The variance measures the deviation of individual observations about a central value. Thus. and “N” is equal to the sample size. Its Greek symbol is the letter “sigma”. it is the prime indicator of risk and used to compare uncertain outcomes with the ”normalcy” and implicit “certainty” of the central value. THE COVARIANCE .5). If the reader refers back to the “normal” distribution. and then divide by the sample size. examining whether the combination reduces risk or increases it. To measure the variance. The formula for the variance is ∑ (X .At times. the analyst needs to determine the risk of a combined population. and then the processed data points are summed.the probability of being on an interval between the mean and a certain number is based on the number of standard deviations away from the mean. with a mean of “3”. STANDARD DEVIATION . In this function. σ. each data point is subjected to the same process.2. It is more commonly used in finance than the variance because it has a close relationship with the mean. In the above sequence of (1.5).4.4.3. or 1. we would proceed: (1+2) + (2+2) + (3+2) = 12. each point within the distribution is some amount of standard deviations away from the center with a specific level of probability attached to it: that is . A 2 2 2 2 2 2 2 . Most data points will trend toward either greater or lesser distance from a central tendency and the variance measures this distance. In a sample population of (1.The standard deviation is merely the square root of the variance.2. The symbol for the variance is the Greek letter “sigma” squared. we found that the variance was “2”. 2. the process would be (1-3) + (2-3) + (3-3) + (4-3) + (5-3) / 5 which is 10 / 5 or “2”. we sum the squared values of distances from the mean of each data point. X is equal to the individual data point in the sequence (sample). or σ . In effect. 3) and a function calls for ∑ (X + 2). Now the standard deviation is √2.41. I know the return from product line # 1 already because it has the same mean as in the standard deviation example. while lower risk is indicated as the covariance approaches zero. but uses the entire sample .2. In effect.7. if I am an operations manager. the parenthetical “i” and “j” represent the two different samples. In fact. Each sample point in one population should have a corresponding entry in another population and it behooves the analyst to drop data points if one population is larger than the other.By examining the population in parts that are above or below a certain value.568 much applied adaptation of the variance is termed “the covariance” For example.5). a positive covariance indicates the tendency for the populations to rise together.10). I then sum each product and divide by the sample size. and then multiply it by the same operation done on corresponding numbers in the second sequence. In terms of risk. Downside risk measurements establish a risk level that is customized to the risk aversion of the practitioner. I might want less risk between the returns from product line # 1 (1.8. the analyst creates a threshold level.4. Although the numbers in the second sequence were much bigger than the first. I take each individual number in one sequence and subtract it from that sequence’s mean.µ(i)) (X(j) µ(j)) / N. The formula is ∑ (X(i) . one would want the lowest covariance possible which is achievable if some numbers rise while their corresponding number falls. and product line # 3 (6. DOWNSIDE RISK . Working this example through: ((1-3)(6-8) + (2-3)(7-8) + (3-3)(8-8) + (4-3)(9-8) + (5-3)(10-8)) / 5 =( 4 + 1 + 0 + 1 + 4) / 5 =2 The covariance is the same as the variance in this case and the two standard deviations are the same as well. I now obtain the mean for product line # 3 and determine it to be “8”. I then obtain a sum (not squared this time) of the product of the deviations from the respective means and divide by the sample size.9. (3). The semi variance sums only those data points that are less than the mean. the risk of achieving them was the same simply because the distance to the mean was similar: each number increased by “1” in the second sequence just as they did in the first. In this formula.3. the differencing position matters and the individual data point is set up to subtract the threshold amount. The function is: ∑ (X(i)T) / N. To use the semi variance. The function is simply ∑ (µ(i) . 40). since the )) µ lower partial moments function may be cubed. the lower partial moments function can be carried to the third fourth or fifth powers at the discretion of the analyst.06 The lower partial moment will yield the same results as the semi variance when the threshold level is set to the mean. if few points are summed.17 .79 + 173. Notice also that the difference is set up for the possibility of raising it to the third fourth and higher powers. we will set a threshold rate of “7” which is not part of the data point sequence from above: ((1 -7) + (2 .3) )+ (15. Usually. To differentiate the two functions.569 size for the divisor. A chosen target level µ can also be substituted for the mean. and we only sum those data points less than that number in the operation: Semi variance = ((15.11 + 0.45 + 148. cash flows. the positions of the mean and individual data points in the differencing do not matter((µ(i) .µ(i)) ) . Notice that the mean is nowhere to be found and that the threshold value.17 is the mean.0289) / 6 = 87. “T”. and “M” is a value greater or equal to “2”. 3.17 .17 .1) + (15.17 . However. the deciding factor can be their downside risk. 15.X(i)) / N. 2. stocks) and when the two populations have similar standard deviations. with X(i) < µ(i). is large. we form a function that will yield the risk of being below the mean. The lower partial moments indicator also sums those points that are below a user defined threshold level which is substituted for the mean in a variance-like function. While the semi variance is always squared. “X” are those values less than “T”. both of these functions will be compared among alternative populations (portfolios.7) +(3 . even as “N’.83. and the function is squared. 15.2) + (15. “N”. 2 2 2 2 2 M 2 2 2 2 . the sample size. the downside risk will be negligible. In the data sequence (1. 30. where “T” equals some threshold value that the user wants to “beat”.15) 2 / 6 = (200.7) )/ 6 =77 / 6 =12.X(i)) = (X(i)) . Thus. replaces it. However. little known. because it assumes a perfectly normal distribution which is rarely the case. . • 2. if I wanted to convert from monthly to millennial (1000 years) data. finding its square root.Often. which is fifty-two.0625. Naturally. By using the square root of the period one is converting to – in terms of the existing data . I would multiply my current monthly standard deviation by the square root of 12000. and then multiplying the existing standard deviation by that number. quick estimator of a standard deviation.570 ANNUALIZED VOLATILITY . I convert one year into weeks. shorter time frames will lead to more accurate estimates. Follow the logic behind these conversions: Annual to monthly: = (annual standard deviation) (√ 1 / 12)) √ Monthly to daily = ((monthly standard deviation) (√ 1 / 30) √ Daily to yearly = (daily standard deviation) ( √ 365) ESTIMATED VOLATILITY . but we need to recognize that any conversion without the actual data is a “guesstimate” at best.we can make these conversions effortlessly.This tool is a valuable. 0. we need to convert data without having the actual sample. The period to which one is converting is expressed in terms of the periodic type of data one currently possesses. For example. I would express the millennium in terms of months: 12 months = 1 yr. Square this range and then multiply it by 1 / 16 or decimally. I then take the square root of fifty-two. Thus. Thus. It needs to be used cautiously. The key to converting volatility lies in converting an existing period to the new period. however. and the longest time most analysts will convert is daily to annual data. As an extreme example. • 1. Subtract the lowest point from the highest point to determine the range of the distribution. and I need a yearly standard deviation. if I currently have weekly data. we may have monthly data but we want a yearly value for the standard deviation or vice versa. and multiply the weekly standard deviation by that number. 12000 months equal one millennium. The formula is modified as: ∑ (X . We take the square root of this number to determine the standard deviation 2 2 • 4. the sample size needs to be adjusted to increase the accuracy of the calculation.VARIANCE . In any sample.(7. 21. the student /investor can detect the discrepancy from the high skew of the population with two close numbers at the low end and a number more than twice any other at the high end.0701.µ) / (N .1) for the sample variance and the application of a square root for the sample standard deviation: √ (∑ (X .1)). If the denominator were “15”and not “14”.1 and not 12. the greater the combination of return versus risk. Example: A highly skewed population .µ) / (N . 14.σ ). SAMPLE STANDARD DEVIATION . the smaller sample size added seven percent of volatility to the calculation. we would divide 60 by 14 and get 4. we merely subtract the standard deviation from the mean and compare it to another population. 56 . thus: √ (0. THE MEAN . 56). 2.0625 (range) = variance. Thus.In small samples.625 (49) -150. Although this indicator gives a 2 2 .2857 for the sample variance. • • • 1. We simply reduce (N). if the sum of squared deviations ∑ was 60. which is the high minus the low. usually less than thirty elements. The function is ( µ .25 = Standard deviation 2 Since the actual standard deviation is 18. “safer” investments. 8. It is a great equalizer between high return investments that are volatile and lower risk.571 • 3. the sample size by “1” and proceed as normal.0625 = Variance 3. The higher the number. the variance would have been calculated as “4”. This resulting figure is the estimated variance. Note that the smaller value of “N”. and the sample size was 15. In formula terms it is: 0. √ 150.25. Obtaining the square root would yield 2.0625 = 12.7 = 49. 0.This is a major but often neglected indicator of risk. pushed up the expected volatility of the number.0625 (range) ) is the estimated standard deviation. In the above example. the analyst may want to make provisions for the statistically worst case scenario that would occur by chance one percent of the time. At times. and then subtracting that product from the mean. statistically derived number with the standard deviation. it is less accurate when comparing a small return to a very large one.24 = -4) < (6 . Stock X is a better choice with more potential return and less risk associated with it . The next indicator better clarifies the tradeoff when two populations are divergent. This indicator will assess whether the extra return is worth the risk. Stock Y has a coefficient of variation of (8 / 6) = 1. X has a mean of 20 % for a return and a standard deviation of 24 %. These standard numbers are as follows: . WORST CASE SCENARIOS . Stock Y has a mean of 6 % and a standard deviation of 8 %. the analyst will produce an estimate of a worst case scenario. The comparison is: (20 .Neither the mean-variance rule or the coefficient of variation cover all risk scenarios.572 very accurate assessment of the risk . by the coefficient of variation. By this rule.8 = -2).in combination.333. Thus. we merely divide the mean into the standard deviation.2. By risk factors alone. The following example shows this divergence: X and Y are two different stocks. THE COEFFICIENT OF VARIATION . A lower value will indicate better characteristics.return tradeoff when populations have a similar magnitude.To overcome the limitations of the mean variance. The function is ( σ / µ). Stock X has a coefficient of variation of (24 / 20 ) = 1. stock Y has better risk-return characteristics even though the investor would forego 14 % in potential returns. any population with a higher standard deviation is more risky. By multiplying a standard. The standard numbers cover specific percentages of a “normal” population and so this technique should not be applied to highly skewed data. 57 Figure 23-5 5% 90 % 1.328 = -0.1 . Thus.0. For every five out of one hundred periods.2 by 1. 0. I would multiply 0.33 2. DEV. and then subtract that product from 0. (0.328. leaving (1 . I would .228) or -22.1.64 = 0. if I have a stock whose mean yearly increase is 10 % with a standard deviation of 20 %.64 5% These standard numbers cover the curve from the middle.2 x 1. as an example. and I wanted to find out the worst case scenario that would happen only 5 % of the time.64 which is the normal deviate at 90 percent.Percentage / 2) on each side.96 2.573 Table 23-1 NORMAL DEVIATES Percentage of the Curve Covered 90% 95% 98% 99% Standard number (Normal Deviate) x STD. 1.8 percent.64 1. The function yields. These are 2 2 combined: σ (i) + σ (j) + (2)(∑ COV i j) Just as with any variance. or even not at all. Risk is reduced by the application of two principles: 1) The sample size is increased and 2) The relationship between the populations is opposed.574 expect a loss of 22. better known as “VaR”). In fact. it is used in combination with several other indicators (exponentially weighted moving averages and Monte Carlo analysis) to diversify away its major premise . and σ (j) = variance of population (j). the average investor should be wary of such forward looking statistics. a rarity in the world of finance. multiplied by the number “2”.that the population is considered “normal”. To form a combined standard deviation. Although an entire branch of risk management is dedicated to this type of analysis (Value at risk. as one moves up. The function is defined as σ 2 2 (i) = variance of population (i). Notice also that while five percent of the curve is uncovered on the lower side. this reduction forms the foundation for “modern portfolio theory” a concept that is still as valid today as when Harry Markowitz first proposed it many years ago. board feet. almost all diversification has the mathematical concept of combined standard deviations as its premise. we start by forming a variance by adding the separate variances of the populations and then adding the summation of all possible combinations of covariances.dollars. In effect. we merely apply the square root ∑ to the function to obtain a standard deviation. Like volatility estimations. ACCOUNTING FOR ADDITIONAL RISK: COMBINING STANDARD DEVIATIONS When units are similar . The covariance between the populations becomes part of the combined standard deviation and determines the amount of risk reduction.we can add standard deviations and sometimes reduce risk by combining two or more populations. widgets. but not percentages . This next example forms a portfolio . the upper side has a companion area of five percent that has a mirror probability.8 percent. the other might move down. The reader will observe that the largest amount of variance was derived from stock Z.9998 9. we will have a sum of three separate variances and covariances.605 Z 30 41 49 40 9.605) +(9.499 Formula for the Variance: σ (i) + σ (j) + (2)(∑ COV i j) ∑ = (1) + (3.5006 + 6.006.9998 + 9. 2 2 2 2 2 . in addition to the large amount of covariance among the combinations.006 = 11.5006 6. Obviously. In fact. Y and Z. using this formula will show the impact of any portfolio decision from a perspective of risk.499) = 139.Y) COV (X.79 = The same as the portfolio standard deviation.79 Table 23-3 Combinations of Covariances COV (X. In this case. when more covariances are used. = √ 139. Table 23-2 Three Populations Year 1 Year 2 Year 3 Population Means Standard Dev. Z) COV (Y.54 Portfolio 43 50 66 53 11.575 between stocks X. Z) 0. Had some of the numbers declined as others rose (as in the case of Y) the standard deviation might have been smaller. (Sample) X 1 2 3 2 1 Y 12 7 14 11 3. a spreadsheet is the best tool to calculate them.54) +(2)(0. cautiously. Multiple standard deviations can be matched with the coverage of a standard area which was the basis for the formation of “normal deviates” in the “worst case scenario” analysis. it may or may not be a good estimator of a population.5 % of the population lies within two standard deviations of the mean.crave certainty and expect them. Each standard deviation represents a specific percentage of area of that curve. there will be (1 . Thus.997/2)) = . if I have a mean of “7”. but welcome their informational content .003(decimal) in total area on the ends of the curve that are uncovered. 99. historical data will misdirect the analyst unless it is weighted with more recent data. he or she will need to make them because customers . Although the capital structuralist avoids making predictions and concentrates on detecting an environment that is conducive to earnings . or a pension fund manager . hence the evolution of the EWMA or “exponentially weighted” moving average. and a standard deviation of “2”. only if we have other corroborating measurements that diversify and validate the analysis. a fairly symmetrical curve develops with the mean in the center. In case after case. plus one standard deviation covers approximately 68 % of the normal curve. the analyst should avoid any dependence on forecasts.7 % leaves just 0.7 % of the population will be between (7-3(2)) and (7+3(2)) or the numbers “1” and “13”.a CFO. we must be cognizant that financial data is usually highly skewed Creating a “ball park” estimate is better than having none at all. The normal curve theoretically covers an entire population. Even more tenuous than the shape of the curve is the fact that we are making determinations from historical data.3% or 0. and 99. a big investor.7 % lies within three standard deviations. The mean. 95.(percent covered / 2)) or (1-(0. The 99. depending on the size of the variance and by default. the standard deviation.576 PREDICTION AND VARIANCE Statisticians have found that when a sample population is large enough. Thus on a normal curve. CONFIDENCE INTERVALS .Once a mean is determined. While we can use this “normal” curve for estimations and predictions. Moreover. On each end. Another misconception can be made about the area not covered by the standard deviations. the negative. Finally.3 % in the above example) lies at one end of the curve or the other.15 % Remember that the literature always regards standard deviations as either added to or subtracted from the mean.7 %. and the total area covered.5 standard deviations of the mean. Thus when we speak of total area covered:. We often say “within” such a distance but in actuality. 6 σ covers a total of 99. Many falsely believe that the percentage not covered (0. Some confusion exists among student / investors as to the coverage of standard deviations. Thus. the total coverage is twice as much -or five standard deviations. 0. we mean that the distance covers twice the reference because it applies to both sides of the mean.15 % 7 99.577 0.7 % 13 3 Std.5 %. Dev. Statisticians will often refer to the positive variant of a standard deviation (the area greater than the mean) without regard to its opposite. 0. if a statistician says that a certain number is “within” 2. Dev. In fact. we need to divide this area by two and realize that this percentage will exist at both ends since we expand from the .15 % at each end. 2 σ covers a total of 68 %. and 4 σ covers a total of 95. and one must realize that the coverage goes in both directions: Figure 23-6 1 3 Std. but it does not mean that the left over 4.25 percent are at both ends.33 3 Percentage of the Normal Curve Covered 90 95 95.64 for 90 %. THE NORMAL DEVIATE OR “Z” SCORE -It is often helpful to determine the probability of an event happening or not happening by examining the number of standard deviations that a certain number is away from the mean.5. We increase the accuracy of the interval by dividing the standard deviation by the square root of the sample size: The function is: µ ± σ / √ N. Thus a “confidence” interval is merely a specific number of standard deviations that are matched with a percentage on the normal curve . he or she will observe that this is merely a specific case of determining the” negative” effect of a standard deviation going below the mean. 2.5 percent is at one end of the curve. Since the number of standard deviations corresponds to a known area of the curve.5.5 or 2. We merely subtract the mean from the number and divide . Thus the 98 % confidence limits are 5. In fact. we can determine probability by how much of the curve is covered.64 1.5 percent of the curve is covered.5 and 2. If the student will recall “the worst case” analysis.33)(5 / √60) = 4 + 1. what are the confidence limits at 98 % if the sample size is “60”? This is equal to 4 ± (2.96 for 95 % etc. Table 23-4 Number of Standard Deviations (above mean) 1.7 Example: with a mean of “4” and a standard deviation of “5”.5 98 99.1. it means that 95. Thus we are making a prediction that a number will between a certain interval a specific percentage of the time.5 or 5.5 and 4 .1. 1.578 centrally located mean.96 2 2. If I am “within” two standard deviations of the mean (or covering four standard deviations in total area). two choices are available: we can find the frequency of the number in the population and divide it by the sample size of the population . To obtain a number that is greater than “33” or less than “7” in this distribution. 20 . This is the area between the mean of “20” and the data point of “7”.0047 or 0.47 percent. we find the area under the curve for the numbers that are adjacent to the number of which we want to find the probability. In this case. To interpolate.4953. The probability of obtaining a specific number under the normal curve is zero percent. suppose our mean is “20” with a standard deviation of “5”.or . The functional representation of this assumption is: (33 .579 by the standard deviation.53 % just as it is between “20” and “33”. because the area under any one unit is zero.6. I look up 2.µ) / σ . I have the decimal equivalent of the total area that would be covered on both sides of the entire interval. we can subtract the 0. In effect.we can approximate it by finding the area between that number and the closest possible number and interpolate.20) / 5 =2.6 standard deviations below the mean.6 and find that it corresponds to the number “0. and we want to find out how much area would be covered from the mean to the number “7”. the number “33” would be the same area above the mean (20 + 13 = 33. is 49. For example. On a table that is appropriately called the “normal deviate” which is found in all statistical texts and on some computers. except this number is 2. The number of standard deviations produced can be matched with a standard statistical table to determine the percentage of area covered. If I double the 0.06 %) from 100 percent (1) and then divide by two to obtain 0. We then subtract the areas to . this is the area that covers the distance between the mean and those adjacent numbers. Thus the interval between “7” and “33” would cover 0. (7-20) /5 = -2.6 and the absolute number of standard deviations is 2. both above and below the mean. Again.06 percent of the population The probability of a number in the population that is between “7” and “20”.9906 (99.4953”. the function is: (X . To find out the probability of any one number.6.4953 x 2 or 99.4953.13 =7) and would also correspond to 0. Thus. respectively. To illustrate. 32. showing the significance of the most recent data (52). depending on the sample. The methodology behind a moving average is that it is measured by dropping the last data point in the sample. Thus.580 produce the interval between the numbers.4918 from 0. we drop the very last data point (20) and add a new data point . and updated with the addition of a new data point. 52) / 6 =34. we will have six data points.20) / 5 = -2. it exhibits pressure on the data to move up or down.21. The average in this set is 29.5.20) / 5 = -2. Thus the probability of having a “7” in the population when the mean is “20” and the standard deviation is “5”. 30. using the same data as in the last example.30). it shows trend movement better than a static average. it may be an integer or it may be continuous. “7” is between “6” and “8”. this type of vectoring makes moving averages the mainstay of technical analysis but will not predict volatile jumps or reactions to random events. 0. The Z deviates are (6 .33. In effect. (20. These correspond to the decimal areas. there are an infinite number of values in the interval. If it is continuous.4974 and obtain 0.32.let’s say “52”.4918. let’s say. When we go into a new month. 21. While a “mean” gives the arithmetic average over an entire population. and the new moving average is (41.83 The moving average has increased. We subtract 0. is 0.0056. This type of value-at-risk calculation attempts to compensate for its normal distribution with a technique that gives more weight to recent data by continually updating it with a moving average. and the probability of getting any one of them is still the difference between the adjacent areas. 41. The time span between first and last data points provides the sample size. when a “short” moving average is compared to a relatively longer moving average.4974 and 0. a moving average updates it to reflect the most recent data.4. The true average (mean) of the entire population would include all data points in the .56 percent UPDATING THE MEAN AND STANDARD DEVIATION: MOVING AVERAGES Many banks will use the “worst case” scenario analysis to track the value of their assets (especially derivatives) on any given day.8 and (8 . In a six month moving average. 33. Its statistical table is based on the sample size minus one. who had to publish under the nom de plume. and obtain the number. sample standard deviation = 3. Now.90) / 2). the standard deviation (sample in this case) is divided by the square root of the sample size. REGRESSION The problem with regression . and so an upper tail area of 0.699”. the ability to forecast a trend is more significant in a less volatile population. The student/investor will remember that the upper tail represents one half of the uncovered area (the other half being in the lower tail). We match the “29” (also called the degrees of freedom) with the 0.0.05. “1. because his employer did not allow him to publish. from the chart. W. The upper tail area is also called “alpha” and has a symbol of “ α “and is equal to ((1 . Gosset. It is used in confidence intervals just as the normal deviate except that it will be multiplied by the sample standard deviation (with N . “Student”. when used in a confidence interval. 90 %).699. If our sample size is “30”. but for its inventor. we subtract “1” and obtain “29”.S. The student t distribution is very much like the normal deviate except that it applies to small sample sizes under 30.025 signifies a total uncovered area of 0. Although technical analysts use moving averages to track volatile populations like stocks and commodities.581 sample and not just the six that were part of the moving average.is the same as that of using parametric (normal curve) statistics: our analysis fits normal data but loses accuracy . This will be our number of standard deviations away from the mean. moving averages are more useful when they track relatively stable populations like month to month changes in asset turnover. named not for students. Again.05 and a coverage of 95 %. sample size = 30. we need a corresponding confidence level which we can pick arbitrarily (say. Calculation: 7 ± 1. confidence limits of 90 % = 1.0695. “N”.9305 or 6.Analogous to the normal deviate is the “student t distribution”.as applied to financial data . ACCOUNTING FOR SMALL SAMPLE SIZE .699 (3 / √30) = 7.1 weighting) and not the larger population standard deviation. Example: mean = 7. and a correspondent “upper tail” area. and the difference will be dependent on costs. Graphically. we can derive “ball park” estimates from regression but must never be dependent on it for decision making. The practitioner has to remember that only two variables are being associated and that some other variable (one that is not being measured) may be at the root of causation. while another population is represented on the horizontal axis. Since financial data is varied and anything but “normal”. called the “Y” axis. the more the data is related because it tends to rise or fall at the same rate. The closer the data comes to forming a straight line. regression is a more useful tool when there is some deterministic relationship between variables such as there is between sales and operating income. Regression has several different forms. that is . statistics based on a normal distribution can give only a “guesstimate” and not a true prediction. playing football in fall weather has nothing to do with the temperature. for example. but at the other extreme. it offers a valuable tool for comparison. the more related are the two sets of data. For example. Therefore. Used in combination with many other indicators. it uses all of the mechanics of covariance and variance but places them in an associative context. It attempts to relate one set of data to another set by fitting them to a straight line. it can mislead the analyst into thinking that highly associated variables are proof of a ‘cause”. the number of football games played is perfectly negatively correlated with the temperature in autumn. called the “X” axis. However. and measures their association by the “correlation coefficient” or “R”. although it may have a high correlation with it. In effect.more games are played as the temperature keeps dropping. On a scale from -1 to 1. Each point on the graph represents some combination of single “X” and “Y” variables. one population is represented on the vertical axis. the farther from zero is the correlation coefficient.582 when higher variability is encountered. but the type most encountered is simple linear regression. Operating income is formed when total costs are subtracted from sales. It may help the student/investor to view it as another adaptation of the standard deviation because in effect. The “Y” . Correlation must be examined closely because there is a tendency to confuse it with cause. which is determined by the amount of change in “Y” in terms of the change in variable “X”. Even though the hypothesis is true in this case. but not necessarily physically. we can infer that some action upon “X” causes a change in “Y” statistically. the “least squares” methodology could be logically adapted for use with many variables. “A” is called the Y intercept and occurs when X equals zero. Thus. we call the “Y” variable the dependent variable and “X” the independent variable Regression will attempt to fit the two populations into a straight line in the form.Y = X / 2. In effect. we can proclaim that the two variables are directly related. Before the advent of personal computers. statisticians would perform regressions with multiple variables requiring extensive calculations. regression does not provide us with the tools to make an assurance. variances and means. called the slope. regression can provide valuable information but does not determine the cause of any event. This author prefers to view the analysis as an extension of the standard deviation because it is simpler to remember and also integrates other areas of risk into it. 2 .583 population changes as we input different values for “X.even if they are perfectly related in some other way such as in the function . may have played a role in the experiment. while “B” represents the coefficient of “X”. Regression can be learned by making ratios of the techniques we have already learned: covariances. If we were to determine the amount of kilocalories provided by one barrel of oil. The following section delineates some of the relationships in linear regression. atmospheric pressure for example. standard deviations. Y = A + BX . However. closely associated variables can lead to further experimentation and identify an hypothesis. In an experimental mode. but we can not claim that more oil provides more kilocalories because the proclamation is limited to just two variables. only an actual measurement of observations in a controlled environment can point us in the direction of cause and effect. some other variable. Most statisticians will lead the student/investor through what is termed “the least squares” approach to regression. Y) / ((Standard deviation of X)(Standard deviation of Y)). because the pattern of movement between data points is similar for each. X. “beta” can refer to the change in any population in terms of the change in another population. • 4) THE SLOPE OR BETA . and there are only two variables.We defined covariance as ∑ (X(i) . if we multiply the slope by the standard deviation of X and divide by the standard deviation of Y. This measurement is used to determine the degree of variation between two populations. has previously been mentioned. If we know the correlation coefficient.In most cases. The “Y” intercept. we use the covariance in the numerator and this time we divide it by the variance of “X”: (Covariance. the more one population varies with another population when each data point moves to the next. Another way of determining “R” is through the use of the slope. In effect. In this . The two populations are closely “correlated”.µ(i)) (X(j) . we . Y) / (Variance of X). The “B” or slope is the coefficient of “X” and determines the amount of variation of “Y” in terms of “X”. we can substitute “Y” for “X(j) because it designates a different population. and we apply it to a particular stock and the market. we will determine the covariance using the standard methodology. However. is a point on the vertical axis that will occur if we input zero for “X”.1”. “A”. we often refer to this ratio as “beta” because “B” has the Greek equivalent of that description.The straight line. “R”. • 2) COVARIANCE . X. In finance. we need only to multiply the standard deviations of X and Y by it to derive the covariance: (R)(Standard deviation of X)(Standard deviation of Y) or R σxσy.584 RELATIONSHIPS IN LINEAR REGRESSION • 1) THE STRAIGHT LINE . Y = A + BX.µ(j)) / N.Again. “R” is simply the covariance of X and Y divided by the product of the standard deviations of X and Y: (Covariance. format. which allows us to determine the correlation coefficient “R”. the closer the measurement gets to “1” or “. • 3) CORRELATION COEFFICIENT . 5) The covariance of “X” and “Y”. subtracted from the mean of “Y”: (Mean of Y) ((Slope)(Mean of X). • 5) THE “Y” INTERCEPT . Thus a value of 0. the stronger is the potential relationship. However. sales can be regressed against operating income.585 again obtain the correlation coefficient. is the coefficient of determination or “R ”.(∑ (X(i) ∑ ∑ µ(i)) (Y(j) .900 for “R”. “R”.5 for “R” will be cut in half. the closer R gets to “1”. and a fairly linear pattern would result because only the variance of costs would separate them. 3) The standard deviation of “X”. but a value of 0. is at their core. and so the resilience of such a regression is only 2 2 2 2 . the standard deviation.“Alpha”.Forecasting with simple regression is most appropriate for near deterministic relationships that are stable. When we square “R”. we create a proportional indicator that establishes the ratio of the amount of explained variation to the amount of total variation. 2) The mean of “Y”. Knowing the relationships will enable the student/investor to understand theory and build even more relationships because the basic measurement of risk.Even more important than the correlation coefficient. R = (Slope)(Standard deviation of X) / (Standard deviation of Y). for “purists” who prefer terminology. MAKING PREDICTIONS . or the “Y” intercept is merely the product of the mean of “X” and the slope. costs can vary a great deal over a decade. once we square it. lower values are discounted at an increasing rate because the act of squaring “R” creates an exponential distribution. ) THE BASIC FIVE -It should be evident that regression can be performed on the basis of only five computations: 1) The mean of “X”. Like the correlation coefficient. 4) The standard deviation of “Y”. only goes down to 0.810 for “R ”. Most readers will perform these operations on a calculator or in a spreadsheet. THE COEFFICIENT OF DETERMINATION . “R”. For example. However. in an industry that is shifting margins.µ(j)) / N) / σ x)(∑ X / N)). it is (∑ Y / N) . or when the data is so stable that more computation is deemed unnecessary. In effect. regression is more amenable to short periods in which the relationship between the populations is observed to decrease or increase and creates some sort of contingent action for management. some growth rates may not be meaningful. Thus. We can combine a regression technique with standard deviations to capture both the magnitude and stability of a growth rate. this method establishes a trend line and predicts the next sequential entry. few relationships in business are deterministic enough to form a line that can be predicted for long periods. logarithmic regression. GROWTH RATES In a nutshell. GEOMETRIC APPROXIMATION . X number of periods apart. The measurement can be a fundamental like sales or income. the nearest period measurement is the numerator. We use these rates freely when the evidence they exhibit is corroborated by other indicators. and some logistics. growth rates reveal more about a relationship between two variables than any other measurement. or any measurement that can . It is the essence of calculus. Therefore. and the measurement taken X number of periods out forms the denominator. the student/investor must remember that the coefficient of variation ( σ / µ) or (Standard deviation / Mean) applies to percentage growth rates as well as static units. while others are so volatile that they defy meaningful measurement. all growth rates can be measured in terms of the same risk-return characteristics with which we measure fundamentals like net income.A very quick and easy method of calculating growth rates is to obtain the geometric mean by using the intervals between periods.586 temporary. Outside of engineering. and should not be used to meet the standards of professional analysis. However. However. it will not be as accurate as the technique described in the next section. logarithms and measured regression. The base figure for the calculation is a ratio formed from taking the same measurement. The technique is best used when a quick comparison is needed. The judgment of the analyst is paramount. and indeed fiscal performance is determined by a period length of five (December 31 to December 31) with four periods between that period length.25 = (0.which is when growth occurs. (7 / 17) and raising it to the 0. 10.18). this technique is more accurately applied to stable populations like dividends than it is to volatile measurements like stock prices.25 power. 11. in the above sequence. and (N . Thus if N is equal to the period length.41176) . In this case.25.80105 . (7. which will require a calculator: (7/17) 0. then the ratio will be raised to the power of 1 / 4 or 0. Thus. . we would throw out the “9” and use the “17” instead. even if the number is farther out in terms of the number of periods.80105 by “7” and obtain “5.1) is equal to the number between periods . Once we have a ratio of measurements taken at period length. simply because “17” is more representative of a trend. is our number between periods. we will dictate “7” as the current year.1) or four. 14.80105. then (N . The period number between the length is the determining factor in this analysis because we are measuring growth. dropping “9” from the sequence. 9. if we had a sequence of (7. We find that measurements over five periods confirm a trend better than other intervals. 14.18). if the number is not representative of a trend (it is heavily skewed in one of the measured years). 11. because he or she must decide if a particular number is part of a trend. and “17” as the measurement five periods ago. there is “negative growth” because 7 is less than the “17” from five periods ago. we raise it to the inverse of the number between years.25 0. However. I would multiply 0. The number of periods is determined by the needs of the analyst. When we subtract “1” we obtain a decimal increase of the sequence: 0. 9. 17. 10. If five is our period length. = 0. For example. For example.587 be physically counted.1 = -0. The growth factor is determined by making a ratio of the numbers. If I wanted to know the next predicted number in the sequence.19895 or 19.607”. it should be replaced by a more representative number. 17.89 percent. .14186.The current period would be designated as period “5”. but using “9” as the data from five periods ago. 10. 17. We take the logarithm of both sides of the equation: Log (2008 SALES) = Log (2004 SALES) + 5 Log (1 + Growth Rate) In this construction.19 percent growth rate. Thus.25 or 1. the squared deviations of the periods. 2008 SALES = (2004 SALES)(1 + Growth Rate) . Any current measurement is equal to the product of the measurement X periods ago and a growth rate that is raised to the power of the number of periods. The theory behind using logarithms is that they will form a straight line when applied to a growth rate. The denominator will merely be like a variance. The ratio would be (17 / 10) and the growth factor would then be (1. and if I multiply the growth factor by the current period to predict the next period. To see how this works. To determine the growth rate. we can obtain accurate growth rates for a sequence even when numbers are not “in trend”. the second. We then form a mean of the periods ((1+2+3+4+5) / 5) = 3 If we designate the logarithm of the measured fundamental as (X(i) and the period as T. 11. If I subtract “1”. again using “7” as our current data. 9. (7.588 On a more positive “growth” note. if we take the logarithms of each number in the sequence. and a period to identify it.412. and then multiply it by the deviation of the period. we merely use the rule of logarithms where multiplied products turn into added logarithms. I could use the number from five periods ago. This makes them amenable to linear techniques like regression. ACCURATE GROWTH RATES FROM LOGARITHMS As long as data is a positive number. “10”. I obtain (1. In the sequence. we will 5 . 14. The last number in the sequence (the period farthest away from the current ) is designated period 1. Each number in the sequence is matched with its logarithm.7) 0.18).14186)(17) = 19. period 2. we set up a chart that designates T = (total number of periods). I obtain a14. we can form a summation that acts as a numerator for a beta type variable. if “17” were the current year. because it is in trend sequence. we will do a five period growth rate. The actual negative growth was much less than dictated by the current trend.808 / 10 = EXP(-0.922378 .922378 = (1 + Growth rate).398 2. “T”.197 T Period 5 4 3 2 1 (X(i)) (T. and subtract “1”. and Mean T are not logarithms.Mean T)) / ∑ (T . the relationships between variables rarely form a straight line. Emphasis tends to be placed on linear solutions when the variables act in a curvilinear fashion.1 =-0.3) + (2 .TMean) 3.Mean T)2. can sometimes find correlation when regression fails. SPEARMAN RANK CORRELATION: NON PARAMETRIC STATISTICS Although parametric statistics dominates the financial world.303 0 -2.892 2. (X(i) (T – T Mean ) STEP 2: Form the denominator. STEP 3: Divide the numbers and put them in analog form (EXP). Σ (T – T Mean)2.639 2.808 / 10 which needs to be in analog form. If X rises in an exponential pattern.364 STEP 1: Form the numerator. the methodology that eschews the formalized use of mean and standard deviation.303 2.3) + (3-3) +(4-3) +(5-3) = 10. Table 23-5 Sequence Number 7 (Current) 10 11 14 9 (Beginning) X(i) Logarithm 1. 0.07762 or negative 7.946 2. 2 2 2 2 2 . -0. non parametric methods are often more appropriate. The resulting figure is the ratio 0. Sum the last column. non parametric statistics will capture the relative pattern better than parametric statistics. we sum the last column and obtain -0. The denominator is formed by (1 .0808) = 0.76 percent growth.639 -4. Non parametric statistics.808. To form the numerator. The primary difference between parametric and non parametric techniques is that the latter will usually rank data from lowest to highest before performing operations on it. Note that the period.589 refer to the period mean as (Mean T) and we can form the function: ∑ ((X(i) (T . parametric statistics will emphasize an average increase in a population without reference to the amount of data points that significantly contributed to it. In Spearman. we are actually making a ratio of the sum of squared deviations among the different populations. one would accomplish a similar objective. Alternatively. this formula applies to the unaltered data points.Mean X) . it is SS(xy) / √(SS(x))(SS(y)). Thus “R” is properly referred to as the “Pearson product-moment coefficient of correlation”. if one did linear regression by assigning “1s” and “0s” to the “Y” variable for increases and decreases. text type notation. extreme values contribute to the correlation as much as values that are right around the average. If one performs regular linear regression. For example. On the other hand. When we derive the “R”. we do the same linear regression that is done in normal parametric statistics. in the Spearman rank correlation. In effect. non parametric statistics will refer to the amount of data points above a certain level with less reference to the average. “R”. As an example.590 In this regard. The sum of squares methodology applies to Spearman. except that we perform it on ranked data. the methodology becomes ordinal and “monotonic”. Stock price changes tend to increase with earnings changes. “X” refers to the numbered rank of data from lowest to highest and “Y” is the ranking from its data from lowest to highest. and more regard to the level: high medium or low. In normal linear regression. In the Spearman rank correlation. it is: SS(xy) = ∑ (X(i) . and the fact is firmly established by Spearman rank correlation. less responsive to the amount of increase and more responsive to the incidence thereof. consider the correlation between earnings and stock prices.Mean X)(Y(j) Mean Y). SS(x) = ∑ (X(i) .Mean Y) . the association is high because there is less regard to the amount of change and more regard to its occurrence. one finds the correlation is low because the variability of the changes is so great. correlation coefficient. 2 2 . on earnings and market price changes. SS(y) = ∑ (Y(j) . it will summarize the contribution with a correlation coefficient. In mathematical notation. In extended. Thus. the smaller is the sample size. we have to test whether it is significant given the sample number and prospective confidence limits. Note that the sample size does not match the number of rankings when there are repeated entries.Y(i)). In this function. the higher “R” must be at any confidence level to be significant. 34.[(6 ∑ (D) ) / (N(N . many of the rankings may be tied when there are frequent duplicates among data points. the following is an incidence in the number “7” rank of a repeated number. The number of ranks will match the total number of elements in the sample (including repetitions). the practitioner can get by with the following equation which simplifies the process: R = 1 . To circumvent this problem. 34 7. and N represents the sample size. We use appropriate statistical tables or a computerized program to match the obtained correlation coefficient with critical values 2 2 .33 9 10 NUMBER 14 16 17 22 25 27 34 34 34 41 44 If there are very few repeated data points. For example. Once we obtain “R”. “34”. which is (X(i) .33 8 41 7. “D” represents the distance between the ranks. Table 23-6 RANKINGS FOR ONE SEQUENCE OF NUMBERS (X or Y) NORMAL RANK NUMBER SPEARMAN RANK 1 14 1 2 16 2 3 17 3 4 22 4 5 25 5 6 27 6 7 34.591 Naturally.1))].33 9 44 7. ties are divided proportionately among the rank (with a decimal) and the next rank up is eliminated in favor of the divided rank. 05 at both ends. Thus if the chart displays the “alpha” area. this simple example will suffice: Table 23-7 Rankings X 1 2 3 4 5 Rank for X 1 2 3 4 5 Y 66 65 64 67 68 Rank for Y 3 2 1 4 5 By the Pearson . SS(xy) = 6. SAMPLE SIZE Designing an experiment with a sample that reflects the characteristics of the greater population. we multiply it by two and subtract from one to determine the confidence interval to which it refers. An “R” value that lies in the middle of the curve has a high value of not being associated. and it displays 0. Like the normal deviate chart. most flawed studies have an inadequate sample at their core. R = 0. is a methodology beyond the scope of this text. SS(x) = 10. The correlation coefficient correlation of 0.900 and so the sample is not considered significantly correlated at N = 5. while points at the ends of the curve have significant values (a low probability of no association). If we obtain a very low probability of “no association”. the probability of “no association” or the “null set” lies between the critical values.05 “alpha”).Spearman “R”.6.592 given the sample size. and a sample size of N = 5. We look on the chart for 90 % confidence limits (√ (. SS(xy) / (√(SS(x))(SS(y)) = 6 / √100 = 6 / 10 = 0.6. To illustrate. This is an example of “hypothesis testing”. SS(y) = 10. 90 % confidence limits. However. we interpret it as a significant correlation.900.6 does not exceed 0. at any confidence level that we choose. A sample becomes tainted by the bias or . a confidence level of 90 % would have . While it is the objective of the financial researcher to decipher trends and glean actionable conclusions. there is often a temptation to make the sample fit a desired objective.the need to coordinate new information with what we already know. may have no correlation at all today. Obviously. The variables that affect financial data are often uncontrollable. a well run study with an adequate sample should ideally be interpreted by a dispassionate observer who may know little about the expectations and details of the research. Analogously. designing a significant sample can be a difficult task because there may be information or knowledge that the study does not consider. and disregarded Spain. consider the attempt to determine the percentage of brown eyed people on the continent of Europe. Italy. In effect. Armed with a knowledge of statistics. As an extreme example. when we interpret data. The statistician must always attempt to build the sample . or republicans poll a section of their constituency. our research would be highly skewed because so many Scandinavians are blue-eyed. the transience of any statistical conclusion is confirmed each time a probability is less than one hundred percent.593 lack of knowledge of the practitioner. the search must be tempered by the recognition that it is primarily a function of time with results that are subject to change. It is therefore difficult to establish any deterministic relationships among them. technology. This is the essence of the “blind” studies favored in the physical sciences. If we decided to take our sample from the country of Sweden. the sample size must be larger to be significant. it is often similar to sample bias because we view it from a narrow perspective . such as happens when democrats poll only democrats. A premise that was true thirty years ago. Tax legislation. demographics and foreign competition can alter the interactions between variables. Thus. and France. With the great variation among data that one encounters in finance. the observer can verify the hypothesis by coordinating numbers alone. established by a cause and effect relationship. We arrange the three pieces of information in a function ((Normal deviate at a given confidence level) (Variance)) / Tolerable error . ( 4σ / E ) σ and it is best to check all three calculations and choose those that agree. we can also obtain an estimate from the previously discussed “range estimation method”.1) (D) + σ reduces to. A normal deviate that establishes a confidence interval. N σ / (N .1) (D) + σ . the judgment of the practitioner is paramount. When the entire population is very large. “N” represents the size of the entire population from which the practitioner wants to garner a sample and “D” is equal to E / 4. Since each of these functions can yield a different answer. Always choose the largest sample size. Thus. or Z σ / E . the estimated variance is 1 /16 (range) .594 based on expected variation. say 95 percent which is equal to 1. whatever level of error we choose will have an effect on sample size and attempts to mimic this constraint For the variance. A variance that can be estimated from a smaller sample from the population with N 1 weighting. this function “begs the question” because we do not know the sample size. or 2 σ / √ N. To establish an estimate. The amount of error that will be tolerated. However. the practitioner needs three pieces of information: • 1. Obviously. suppose I want to estimate the mean return for companies with over ten billion dollars in debt and 2 2 2 2 2 2 2 2 2 2 2 2 2 . In this instance.96 • 2. • 3. The range from lowest number to highest number is usually about four standard deviations in a “normal” population. also called “bound on error” is equated with two standard deviations divided by the square root of sample size. As an example. We multiply 1 / 16 by the range and then square that number to obtain an estimated variance. The amount of tolerable error. A variation of this function exists in some statistics texts: N σ / (N . I choose the largest standard which is a sample size of “64”.101875 = 39.16 / 0.46.05) = 61. Finally.1) (D) + σ ) variant. I next do the ( 4σ / E ) calculation and obtain 0.000625) + 0.004) σ / (99)(0.26 Since the first two calculations are in agreement and are much larger.96) (.0025 = 64.04 = 4 / 0. A sample standard deviation shows a 20 % variation among returns and I set 5 % as my tolerable (bound on) error at a 95 % confidence level. I do the other variant. A quick calculation shows: (1. I conclude that the population size is too great to use the (N σ / (N .2) / (0. 2 2 2 2 2 2 2 (Back to Table of Contents) .595 my research shows that 100 such companies exist. 100 (. pdf. Quality is Free. Dictionary of Finance and Investment Terms. “A Pedagogical Examination of the Relationship Between Operating and Financial Leverage and Systematic Risk.nyu. Downes. 1974. Davidson III. “Debt and Value: Beyond Miller-Modigliani. 1980. John Wiley and Sons. “The Determination of Optimal Capital Structure: The Effect of Firm and Industry Debt Ratios on Market Value. and Wallace N. Philip B. Financial Management: Theory and Practice.org/SLC/Papers/zero_debt2006_01. 2008). 2003. Damadoran. 8(Fall 1994): 1-14 Crosby.” Journal of Financial and Strategic Decisions. “Why Don’t Some Firms Use Debt?.p. Inc. Tsirijakis.:Harcourt College Publishers. Soku. Stern.afajof. Ehrbar. Barron’s Educational Services. 1998. EVA: The Real Key to Creating Wealth.. Brigham. Edwin.” PowerPoint Presentation. Hinsdale Illinois: The Dryden Press. Edward L.pdf Cheng. (August 13.x SELECTED REFERENCES Altman. Soku. 2001 Byoun. 1968.” Journal of Finance. 9(Fall 1996): 1-8. and David Lindsey. Edwin G. 6th ed. Basic Macroeconomics: Principles and Reality. G. New York: Mentor. Hatfield. Ehrhardt. Gay B. “Financial Ratios.” http://www.” Journal of Financial and Strategic Decisions.. Al. Discriminant Analysis and the Prediction of Corporate Bankruptcy. “How and When Do Firms Adjust Their Capital Structures Toward Targets?” http://www.edu/`adamodar/pdfiles/country/levvaluehort. Inc. http://pages. Aswath.ppt..fma. John and Jordan Elliot Goodman. September.org/afa/forthcoming/3724. Dolan. Andreas Merikas and Manolis D. . Louis T.stern. n. Joel M. (April 13.W. 2008) Byoun. Eugene and Michael C.Bennett Stewart. Duett. “An Empirical Investigation of the Rank Correlation Between Different Risk Measures. Homewood. New York: Touchstone. F. A Time to be Rich.um-lj. Hagin. Robert L. . Feb 20. Illinois: Dow Jones-Irwin. Journal of Applied Corporate Finance. Andreas Pfingsten and Peter Wagner. 1996): 39-70. May 1972: 435-452. “The Capital Asset Pricing Model: Theory and Evidence.. 12(Summer 1999): 18-31 Fama.” Journal of Financial and Strategic Decisions.” Feb.ef. The EVA Revolution. “The Effect of Firm’s Capital Structure on the Systematic Risk of Common Stocks. New York: John Wiley and Sons. Techniques and Implementation. 8(Summer 1995): 4755. French. Hahn. Carsten. 1995. 4 1999. John Wiley and Sons. “The Impact of a Firm’s Characteristics on Junk Bond Default. Macmillan (Rawson Associates) 1987. Sam Ramsey and David Shimko. Hendricks. “On Accounting Flows and Systematic Risk.xi Ehrbar. William. Darryl . 2008). Eugene. Philip A. and Kenneth R. Inc. Hagstrom. Garrod. Grieder.com/sol3/papers. “Evaluation of Value-At -Risk Models using Historical Data.doc.” Journal of Finance. Hunt. R. Hamada. April.ssrn. Lacy H.” http://papers. 1976. (August 8. Bennett Stewart III. The Dow Jones-Irwin Guide to Modern Portfolio Theory.. Common Stocks and Uncommon Profits. Fisher. Al and G. Robert G. 2003. Lacy H. http://miha.si/-dokument/wp/garrod1. 2002 1-48 Hakim. CT: JAI Press. Hunt.” Federal Reserve Bank of New York: FRBNY Economic Policy Review. Inc. Neil and Dusan Mramor. Dynamics of Forecasting Financial Cycles: Theory.cfm?abstract_id=301982. Greenwich. 2004): 25-46. The Warren Buffett Way. 16(Summer. 1987.” Journal of Economic Perspectives. Secrets of the Temple: How the Federal Reserve Runs the Country. 1979. Risk and Maximal Gains from Diversification. “The Equity Risk Measurement Handbook.” Stern Stewart Research: Evaluation. Probability Models for Economic Decisions. 13(April. Modigliani. Mendenhall. Juan Golic. Levy.” Journal of Financial and Quantitative Analysis. Statistics for Management and Economics. Inc. Michael E. L. John. A Random Walk Down Wall Street. Statistics for Modern Business Decisions. Franco and Miller M. “Capital Structure Optimization: A Constrained Minimization Model. Raymond M. Lintner. Reinmuth.. Corporation Finance. William and James E. “The Cost of Capital. Haim and Marshall Sarnat.” Journal of Accounting and Finance Research. (June 1958). and S. Norton and Company. Second College Edition. “Security Prices. 1981.” Journal of Finance. 1982. Belmont. McGraw-Hill.” American Economic Review. Modigliani. Burton G. 1978. Lee Gurley III. Philiippe. Massachusetts: Duxbury Press. Jorion. Newman and A. New York: W. 2001 Lapin. Prentice -Hall International Inc. Harvard Business School Press.H. Mandelker. “The Impact of the Degrees of Operating and Financial Leverage on Systematic Risk of Common Stocks. March 1984: 45-57.W.xii Johnson. Brace. Justin. Roger B.” American Economic Review. (December. On Competition. 2005 Petit. 1982. CA: Thompson/Brooks/Cole. Jovanovich. 1965). North Scituate. Value At Risk Second Edition. Malkiel. 2005): 177-189. and April Park.. Franco and Miller M. 3rd ed.H “Taxes and the Cost of Capital. (June 1963). New York: Harcourt. 3(March 2001): 1-12 Porter. G. Rhee. Capital Investment and Financial Decisions. 1998 . 3rd ed. 2nd ed. Joseph A. Lawrence. Myerson. and the Theory of Investment. ” Journal of Financial and Strategic Decisions. Michael. Inc. 2001. William F.xiii Prasad. 2008) Sivy.” August 4. Sam. Inc. 10(Spring 1997): 47-58 Ramesh. “Predicting and Pricing the Probability of Default. Garry D. Crown Copyright. Office of Fair Trading. Lloyd B. Stovall. Banking and Economic Activity. Michael Sivy’s Rules of Investing. 1996. New Jersey: Prentice-Hall. 2nd ed. Part 1 Main Report: Empirical Indicators for Market Investigations. Huntington S.edu/alessiosaretto/default. Ram. McGraw-Hill. 1964). Publishers. 2004. Dev.” Journal of Finance. BA II Plus: Guidebook.pdf. and Robert Ed Woolsey. “Long Run Strategic Capital Structure. Thomas.ucla. Texas Instruments. Inc. Saretto. Englewood Cliffs. Standard and Poor’s Guide to Sector Investing. “Capital Asset Prices: A Theory of Market Equilibrium. Financial Analyst’s Indispensable Pocket Guide. Texas Instruments Instructional Communications. http://personal. Operations Research for Immediate Application: A Quick and Dirty Manual. Alessio A. 2004. 1995 Swanson. Bruton and Andreas Merikas.. (August 8. New York: Warner Books. (Back to Table of Contents) . New York: Harper and Row. Money. 1975. New York: McGraw-Hill. Sharpe. 1991. (September.anderson. 1982 United Kingdom.