Jeon, Y., & Vernengo, M. (2008). Puzzles, Paradoxes, and Regularities Cyclical and Structural Productivity in the USA (1950–2005). RRPE.pdf

Review of Radical PoliticalEconomics http://rrp.sagepub.com/ Puzzles, Paradoxes, and Regularities: Cyclical and Structural Productivity in the United States (1950−2005) Yongbok Jeon and Matías Vernengo Review of Radical Political Economics 2008 40: 237 originally published online 20 June 2008 DOI: 10.1177/0486613408320002 The online version of this article can be found at: http://rrp.sagepub.com/content/40/3/237 Published by: http://www.sagepublications.com On behalf of: Union for Radical Political Economics Additional services and information for Review of Radical Political Economics can be found at: Email Alerts: http://rrp.sagepub.com/cgi/alerts Subscriptions: http://rrp.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://rrp.sagepub.com/content/40/3/237.refs.html >> Version of Record - Aug 12, 2008 OnlineFirst Version of Record - Jun 20, 2008 What is This? Downloaded from rrp.sagepub.com at U.N.E.D. Hemeroteca on November 11, 2014 Also. University of Utah. Review of Radical Political Economics. We would also like to thank Al Campbell.co. Seoul. Laurie Johnson. Korea.com at U. Hemeroteca on November 11. No.E. The labor productivity slowdown of the early 1970s is well known as the “productivity slowdown puzzle. and accepted a greater degree of variability than is actually guaranteed by the empirical evidence. The revival in U. Hyundai Research Institute. and there has been no significant decrease in the value of the parameter since the 1960s. our estimation method and our results suggest that conventional measures of Okun’s Law have overestimated the value of the Okun coefficient. Scott Carter. Okun’s Law has been relatively stable through time.utah.kr MATÍAS VERNENGO Department of Economics. Okun’s and Verdoorn’s laws explain 87 percent of all the variations in labor productivity. Introduction According to conventional wisdom. 2014 . and Ramaa Vasudevan for their comments. and Susan Schroeder for organizing the session. labor productivity growth has been a source of puzzlement and paradoxical results. 3. The 1990s intensified the sense of bewilderment. Philip O’Hara. Summer 2008.N. James Galbraith. who was instrumental in the development of the most consequential ideas in the paper. Jongno-gu. Volume 40. 0510 Keywords: productivity. Paradoxes.S. structural change 1. Responsibility for any mistakes is entirely ours.D.1177/0486613408320002 © 2008 Union for Radical Political Economics 237 Downloaded from rrp. and Regularities: Cyclical and Structural Productivity in the United States (1950–2005) YONGBOK JEON Industrial Strategy Division. 1645 Central Campus Drive. productivity growth after 1995 has led many authors Authors’ note: We would like to thank Per Berglund. 0490. Salt Lake City.Puzzles. e-mail: vernengo@economics. Hyundai Building 11th Floor 140-2 Gye-dong. UT 84112-9300. cycle.edu Abstract Changes in labor productivity have been a source of puzzlement and paradoxical results for economists. Edward Nell.sagepub.” and several authors suggest that its causes have not been completely identified. Room 326. JEL Codes: E320. 237-243 DOI: 10. We suggest that puzzles and paradoxes vanish once two simple regularities are properly acknowledged. e-mail: ybjeon@hri. the omission of structural effects leads to an overestimation of the Okun coefficient. more rapid productivity growth leads to increasing rather than decreasing employment in manufacturing. we presume that it is the long-term potential growth of GDP that puts pressure on the economy.D. once cyclical and structural effects are taken into account there is little role for puzzles and paradoxes in explaining productivity growth. See Arthur Okun (1962). in this version.E. This paper argues that there are two significant problems with conventional wisdom on productivity growth. 2.sagepub. The idea is that the productivity payoff of the IT investments was delayed because of lags in adopting software and business practices associated with the rapid improvements in IT hardware capabilities in the previous decades. One exception is Robert Gordon (2000) who emphasizes the cyclical component of productivity generally associated with Okun’s Law. Okun’s Law is usually represented as a time series relation in a given economy. 2014 . Okun’s and Verdoorn’s Laws The traditional representation of Okun’s Law2 says that if the economy grows above the potential growth rate. β is the autonomous increase in productivity. We have: (g – gp) = –αΔu (1) where g and gp are the real rate of GDP growth and potential GDP growth. are also part of the story.1 We believe that cyclical and longterm structural effects of output growth on productivity.238 Review of Radical Political Economics / Summer 2008 to suggest that there is a “new economy” based on information technology (IT). William Nordhaus (2005) finds that the productivity rebound since 1995 has been pervasive. respectively. Note that. and forces the 1.com at U. or in Adam Smith’s famous dictum: “the division of labor depends on the extent of the market. See Johannes P. For example. Δu is the change in the unemployment level. at least in Nordhaus’s view. Verdoorn’s Law suggests that productivity growth is a function of the rate of growth of manufacturing production.” Paradoxically. The next section presents our main results.3 A more encompassing version would suggest that productivity growth results from total output growth. 3.” We have: p = β + γgp (2) where p is the rate of labor productivity growth. with approximately 40 percent of it taking place in sectors related to the “new economy. and α is the Okun coefficient. which we suggest are associated with Okun’s and Verdoon’s laws. Downloaded from rrp.N. In fact. and γ is the Verdoorn coefficient. unemployment will fall. Hemeroteca on November 11. Verdoorn (1949) and Nicholas Kaldor (1966). First. but less than proportionally since productivity is pro-cyclical. regularities explain a staggering 87 percent of the changes in productivity. and the last section summarizes the implications. 2. Second. Almost all discussions tend to suggest that explanations for the collapse or surge in productivity growth must be related to the supply side. ruling out the possibility of a spurious regression. and that once we obtain the Verdoorn coefficient we can derive the Okun one. Hence. and averaging output growth over long periods eliminates the effects of cyclical variations.D. we do a partitioned regression and obtain a residual Δu series that is not correlated with the rate of output growth. if productivity fell in the 1970s and recovered in the mid-1990s it must be because of changes in output growth. a cyclical effect. eliminating the bias in conventional estimates of Okun’s Law. In other words.N. The conventional interpretation of Okun’s Law is that there is a supply capacity limit. and more interestingly are quite stable even for different sub-periods. and not changing parameters. The econometric results are presented in Table 1.5 As expected. except the constant. which show a Verdoorn coefficient of approximately 0. Substituting (1) into (2) we obtain: p = β + γ(g + αΔu) (3) p = β + γg + γαΔu) (4) Rearranging we get: To solve the omitted variable problem. while the latter should be seen as a structural or trend effect. All the coefficients. in contrast to conventional claims that suggest a reduction from around 3 to close to 2. are significant at the 1 percent level.4 This paper tries to estimate both effects. The reason for that is that Verdoorn’s Law is usually measured in a cross section of countries. and inflationary pressures will develop. which we call Δu*. 4.com at U. which would lead to overstating the value of the Okun parameter. For the Okun coefficient we divide the coefficient for Δu* (γα) by the Verdoorn coefficient (γ).sagepub. Downloaded from rrp. which separates the cyclical and trend effects of output growth on unemployment. The parameters are robust to different specifications. our results suggest that the Okun coefficient has been remarkably constant. we believe that estimates of Okun’s Law have been biased because of the tendency of omitting Verdoorn’s Law. Vernengo / Productivity in the United States 239 process of “creative-destruction” associated with productivity growth.E. 2014 .Jeon. It is important to note that the unobservable potential growth rate is eliminated in this specification. The simple explanation is that part of what is usually assumed to be explained by Okun’s Law. This formulation of Okun’s and Verdoorn’s laws suggest that the former should be interpreted as a cyclical effect of output on productivity. Hemeroteca on November 11. is actually the result of trend or structural effects that are caused by Verdoorn’s Law. Verdoorn’s Law measures have not suffered the problem since they often average long periods. if the economy grows significantly above its potential. in turn. the conventional view suggests that supply changes affect the labor market equilibrium. In other words. Further. and this. hence eliminating cyclical variation.63 and an Okun coefficient of around 1. An exception is the study by Ho-Chuan Huang and Su-Chin Lin (2006). the Okun coefficient is smaller than what is usually found in the literature (closer to 2). This view of a supply constraint is the basis of the natural rate of unemployment or the non-accelerating inflation rate of unemployment (NAIRU). Augmented Dickey-Fuller tests show that labor productivity and output growth are not integrated. determines the non-inflationary limit to economic expansion. In fact. then unemployment will eventually fall below the natural rate. 5.69. 7.0000 0. the Ramsey RESET test verifies no mis-specification.117267 Mean dependent var S. the equation is estimated with an AR(1).223076 0. In the same vein. As the old dictum states: necessity is the mother of invention. the estimated coefficients are not significantly different. that greater labor productivity resulted from output growth. Hemeroteca on November 11.216860 119.48320 7.D. Demand expansion does not lead directly to inflation.sagepub. Error t-Statistic Prob. following Adam Smith.868404 0.08445 –52.356941 0.D. Downloaded from rrp. that is. 6.0000 0. since it has a positive cyclical effect on productivity.240 Review of Radical Political Economics / Summer 2008 Table 1 Regression Results6 Dependent Variable: P Method: Least Squares Sample (adjusted): 1951 2005 Included observations: 55 after adjustments Convergence achieved after 6 iterations Variable C G ΔU* AR(1) R-squared Adjusted R-squared S. In other words. Kaldor’s interpretation of Verdoorn’s Law reinforces the demand-oriented interpretation of Okun’s Law. It must be noted that lower unemployment may increase labor bargaining power and indirectly lead to wage pressures beyond the increase of productivity. Also.554642 0.069260 0.070872 2. growth should be above its trend by a certain magnitude.813996 2. causality goes from output variations to the labor market. 0. and that this reflected a process of circular and cumulative causation or increasing returns to scale.032122 0. and then to productivity.7 Further.072422 .729753 0.690311 1.E. of regression Sum squared resid Log likelihood Durbin-Watson stat Inverted AR Roots Coefficient –0.000000 A more Keynesian reading.7820 0. Kaldor (1966) suggested. The Keynesian interpretation suggests that both Okun’s and Verdoorn’s laws imply causality from output growth to labor productivity. 2014 .0000 1. The Jarque-Bera goodness of fit test shows no departure from normality. since the Durbin-Watson test is in the indeterminate region. and as a result to higher inflation.658049 22.N. Finally.55 Std. After adjusting to AR(1) the model passes all diagnostic tests and. more importantly. In other words. Okun’s Law can be seen as an empirical regularity that can be exploited for policy purposes. Also. with imperceptible changes in their sizes. Verdoorn’s Law implies that a system that is growing fast is forced to innovate in order to maintain its pace.94898 2.145355 0.875715 0. dependent var Akaike info criterion Schwarz criterion F-statistic Prob (F-statistic) –1. in order to achieve a certain unemployment target. while conventional wisdom would suggest reverse causation. and probably closer to Okun’s (1962) intentions.625842 1. while the White’s test does not reject the null hypothesis of no heterocedasticity at the 5 percent level.com at U. Okun’s and Verdoorn’s laws interpreted in this way imply that the notion of a relatively stable supply limit to the economy is at least questionable. higher labor productivity increases the space for non-inflationary wage increases. would suggest that.600089 19.1158 0. Okun’s Law results from labor hoarding.E.356206 4. see Berglund (2006). the misperception about the causes of inflation in the 1970s led to contractionary policies that had negative cyclical and trend effects on output and. as a result of Okun’s and Verdoorn’s laws. and the willingness of the Federal Reserve not to increase interest rates when unemployment fell below what was accepted as the correct level of the natural rate. associated with the oil shocks and the depreciation of the dollar. were central for labor productivity recovery. Downloaded from rrp.8 However.D. it is important to note that contrary to conventional wisdom that suggests that the Great Inflation of the 1970s was caused by demand pull pressures.Jeon. to a great extent. higher levels of unemployment was the slowdown of labor productivity. In that respect. 11. For a complete model. They are essentially explained by the same phenomena. on average. see Robert Eisner (1997). In this light. were more important. This has been underscored by the increasing acceptance of a timevarying NAIRU (TV-NAIRU). 9.com at U. First.” Galbraith argues that business cycle peaks and valleys are endogenous to policy.” In other words. If a reduction in the current rate of unemployment reduces the natural rate of unemployment.sagepub. Both are related to the expansion of output growth. Vernengo / Productivity in the United States 241 Causality is difficult to ascertain. then the economy can grow without the risk of accelerating inflation. on the “dot. through Okun’s and Verdoorn’s laws.10 These policies were maintained in spite of the significant increase in unemployment. The same can be suggested about the “new economy. our econometric results have serious implications for the interpretation of the “productivity slowdown puzzle” of the early 1970s. the expansion that started during the Clinton administration.11 There is nothing particularly different between the increase in productivity growth in the 1960s. Although not central to our argument. see Robert Pollin (2003). referring to the macroeconomic policies that followed the Great Inflation. 2014 . we would add the trend as well. For a critique of the Clinton years.N. In other words. The side effect of an economy with. and it was based. if the expansion of the economy—both its cyclical and structural components—leads to increasing productivity. The justification for accepting higher levels of unemployment was that the NAIRU had increased steadily from slightly above 3 to more than 6 percent. we believe that cost push forces. had a positive impact on labor productivity. Hemeroteca on November 11. We do not provide a fully consistent stock-flow model in which Okun’s and Verdoorn’s laws are integrated in this paper. then the supply limit is affected by the expansion of demand. and the “new economy” of the mid-1990s.9 The variability of the NAIRU—and the related idea of an exogenous potential GDP—undermines the supply interpretation of Okun’s Law. there is little evidence for a constant or even relatively stable NAIRU. This does not imply that the Clinton boom was devoid of problems. For a devastating theoretical and empirical critique of the stable NAIRU. and our econometric analysis does not intend to shed light on the issue. The slowdown was associated with the contractionary macroeconomic policies implemented as a result of the acceleration of inflation. coherent with our discussion of Okun’s and Verdoorn’s laws.com” bubble. both. contractionary demand was not central for stabilization as much as a reversal of cost conditions.” and the “new economics” of the 1990s. 10. that “erring policymakers have in the past reacted imprudently to ‘supply shocks’ in ways that prematurely and systematically curtailed economics expansion. In other words. associated with the Keynesian “new economy. on labor productivity. James Galbraith (1997: 99) correctly points out. 8.E. and whereas the two expansions have certainly had different effects on income distribution. the recovery did little to improve income distribution. because output growth was slow. P. in fact. before the 1990s computers were visible everywhere but in productivity statistics. and accepted a greater degree of variability than is actually guaranteed by the empirical evidence. Our estimates cannot determine what optimal monetary policy would look like. New York: Routledge. We suggest that puzzles and paradoxes vanish once two simple regularities are properly acknowledged. our estimation method and our results suggest that conventional measures of Okun’s Law have overestimated the value of the Okun coefficient. our results suggest that the cyclical variation and the trend of labor productivity are to some degree endogenous to macroeconomic policy. and increase the risks of demand pull inflation.N. Hemeroteca on November 11.D. ed. that as the recovery from the last recession has been relatively weak when compared to the other postwar recoveries. Hence. In fact. G. Okun’s and Verdoorn’s laws explain 87 percent of all the variations in labor productivity. Okun’s Law has been relatively stable through time. This is. 2006. but we believe that the recognition of the importance of Okun’s and Verdoorn’s laws should reinforce the increasing skepticism about the NAIRU and the very notion of a rigidly fixed supply limit to the economy. Also. and we have no intention of arguing one way or another. this has serious implications for conducting macroeconomic policy. Berglund and M. given the established regularities. an antiinflationary stance that induces lower levels of activity would have a negative impact on labor productivity. In terms of monetary policy Okun’s and Verdoorn’s laws suggest that the trade-offs confronted by the Federal Reserve are not as clear-cut as often argued. and there has been no significant decrease in the value of the parameter since the 1960s. G. the technologies that are used (and improved) are those that are already available. Concluding Remarks Changes in labor productivity have been a source of puzzlement and paradoxical results for economists. Paradox of thrift and budget in a simple Keynesian growth model. It is clear that when productivity picks up. First. In addition. but as soon as it expanded in the mid-1990s they became quite ubiquitous. In other words. then eventually labor productivity will slow down. Last but not least. it seems reasonable to assume. without necessarily affecting cost pressures. 2014 . Okun’s and Verdoorn’s laws imply that the supply limits to the economy are to some degree endogenous and this should be introduced into policy discussions. However. 3. References Berglund. Downloaded from rrp. our discussion of Okun’s and Verdoorn’s laws has important implications for current macroeconomic policies.E. as Robert Solow (1987) famously put it. A thorough discussion of the implications goes beyond the scope of our paper. IT is considerably less relevant than usually assumed.sagepub. In The means to prosperity. Solow’s paradox is also explained by two simple regularities. In our view. more so if the economy enters a recession or accommodates to lower rates of growth on average.242 Review of Radical Political Economics / Summer 2008 Also. Vernengo. something that is already visible in the recent evolution of labor productivity. P.com at U. the cyclical and structural effects of the variations in output growth account for almost all the change in labor productivity. vol. Lin. the Chinese economy. Korea. His research interestsare macroeconomics. R. Downloaded from rrp. Okun. Kaldor. On the factors determining the growth of labor productivity. Gordon. Time to ditch the NAIRU. applied econometrics. N. ed. 2014 . L. Kregel. Hemeroteca on November 11. ed. Proceedings of the business and economics section. R. 11354. Oxford: Oxford University Press.com at U. and S.and history of economic thought. Davidson and J. A flexible nonlinear inference to Okun’s relationship. and history of economic thought. New York: Verso. [1949] 1993. J. P. P. II. 1962. 2006. New York Times Book Review (July 12): 36. Galbraith. Does the “new economy” measure up to the great inventions of the past? Journal of Economic Perspectives 14 (4): 49–74. H. His areas of interest include economic growth and development. Journal of Economic Perspectives 11 (1): 93–108.sagepub. Verdoorn. econometrics and the Chinese economy. 1997.E. 1987. Thirlwall.D. We’d better watch out. Applied Economics Letters 13 (5): 325–31. 2000.Jeon. He acted as an external consultant for theInternational Labor Office (ILO) and the United Nations Development Program (UNDP). In American Statistical Association. NBER Working Paper No. 2003.economic development. Yongbok Jeon has been conferred a PhD in economics at the University of Utah and is currently a research fellow at the Hyundai Research Institute. R. A new view of the NAIRU. DC: American Statistical Association. J. and his research interests are in macroeconomics. Nordhaus. London: Duckworth. A. In The essential Kaldor.N. ed. Contours of descent. In Italian economic papers. Cheltenham: Edward Elgar. 2005. F.. Seoul. In Improving the global economy. Targetti and A. Causes of the slow rate of economic growth in the United Kingdom. [1966] 1989. W. Potential output: Its measurement and significance.-C. Pollin. Pasinetti. Matías Vernengo is a research fellow at the Hyundai Research Institute. Vernengo / Productivity in the United States 243 Eisner.-C. Washington. 1997. Solow. The sources of the productivity rebound and the manufacturing employment puzzle. P. Huang. R.