Nomura Fixed Income ResearchThe Evolution of Commercial Real Estate (CRE) CDOs I. Introduction January 4, 2006 Recent months have seen a surge in issuance of CDOs backed by commercial real estate assets. A commercial real estate CDO ("CRE CDO") is a type of collateralized debt obligation (CDO) that is backed by commercial real estate assets, such as CMBS, commercial mortgage loans, REIT debt, etc. In our view, CRE CDOs require more expertise in the individual commercial real estate sectors than for regular structured finance (SF) CDOs. Some CRE CDOs use debt instruments that are low down in the capital structure. Moreover, compared to other SF CDOs, CRE CDOs tend to have a relatively small number of underlying assets that makes their portfolio "lumpy." In this paper, we highlight some of the important issues for an investor who may be considering investing in a CDO backed by various CRE assets. II. A Brief History The CRE CDO is not an entirely new product. The first CDO backed by commercial real estate 1 assets appeared in 1999. These earlier deals typically involved a static collateral pool of subordinate CMBS and REIT debt. The CDO structure allowed slightly more flexibility than the traditional repackaging of CMBS securities (i.e., re-REMICs), while providing a long-term, non-mark-to-market financing for CMBS B-piece buyers. The collateral managers of early CRE CDOs were allowed to sell only defaulted or impaired securities. Also, principal repayments were used to sequentially pay down CDO liabilities, rather than to purchase new assets. By the end of 2003, about forty CRE CDO deals amounting to nearly $20 billion were outstanding. CRE CDO entered a new era in 2004, with the emergence of transactions that were backed by 2 actively managed pools of commercial real estate assets. At the same time, new types of unsecuritized collateral assets, such as B-notes and whole loans, began to appear as collateral assets. Many of these new assets are floating-rate instruments and can be prepaid after 12-18 months. In the past, buyers of B-notes and mezzanine loans often used short-term repo to finance these assets. Instead, by putting them into a revolving-pool CRE CDO, these investors are able to obtain matched-term funding at a lower cost. A total of three CRE CDO deals were issued in 1999. Two of them were: Diversified REIT 1999-1 and Fortress CBO I. Also, so-called "multi-sector" CDOs of structured products often included a small portion of CRE securities. There were a few managed deals as early as 2002. One example is Newcastle CDO I, where the collateral manager, Newcastle Investment, was allowed to trade up to 20% of the underlying portfolio per year during the 5year reinvestment period. 2 1 Contacts: Michiko Whetten (212) 667-2338
[email protected] James Manzi, CFA (212) 667-2231
[email protected] Nomura Securities International, Inc. Two World Financial Center New York, NY 10281-1198 www.nomura.com/research/s16 This report and others are available online at Nomura's research website. To obtain a user id and password, please contact Diana Berezina at
[email protected]. The web address is http://www.nomura.com/research/s16 Please read the important disclosures and analyst certifications appearing on the second to last page. Nomura Fixed Income Research Issuance of CRE CDOs further accelerated in 2005. The volume approached $14 billion at the end of 3 2005, marking more than 50% growth from $8.4 billion in 2004. Today, about two thirds of new transactions have a revolving structure. At the end of the third quarter of 2005, there were over 50 CRE CDOs and nearly 500 tranches outstanding. From a CDO investor's perspective, this new breed of CRE CDOs provides access to traditionally unrated, unsecuritized commercial real estate assets that do not go into CMBS. On the other hand, with the booming CMBS market and benign credit fundamentals, sourcing cheap, high-quality CMBS collateral has become challenging and deal economics has deteriorated for the traditional, CMBSbacked CRE CDOs. The downside of this evolution, however, is the added complexity in analyzing exposures to the commercial real estate sector that involve multiple layers of pooling and tranching. III. CRE CDOs: Old vs. New The evolution of the CRE CDO market has been gradual. Nevertheless, a comparison between deals from a few years ago with more recent deals reveals stark differences. Graph 1 below compares the average collateral mix of CRE CDOs by issuance year. As we can see, the portion of commercial real estate loans ("CREL") has soared in 2005. Graph 1: CRE CDO Asset Mix by Vintage 100% 90% 80% Underlying Assets 70% 60% 50% 40% 30% 20% 10% 0% Pre-2002 2002 2003 Vintagel Source: Fitch 2004 2005 CREL Other CMBS REIT debt It is interesting to compare old vs. new CRE CDOs (see Table 1 in the next page). One example of older CRE CDOs is Anthracite CDO I, issued in May 2002. The deal was backed by CMBS (79%) and REIT debt (21%). The portfolio was completely ramped up at issuance, and reinvestment of principal payments was not allowed. However, the collateral "administrator" was able to sell defaulted or impaired securities and securities with significant credit risk. In contrast, one deal priced in April 2005, Guggenheim Structured Real Estate Funding 2005-1, featured a revolving pool of B-notes (49%), mezzanine loans (22%), and whole loans (11%), in addition to subordinate classes of CMBS (16%). Most of the collateral assets in the Guggenheim 3 According to Moody's. See, Philipp, T., et al., US CMBS 2Q Review: Credit Metrics Slip as Issuance Surges, Moody's special report (28 July 2005). (2) 5 CRE CDOs allow them to achieve higher leverage and greater diversity in their investments. For this type of CRE CDOs. first-loss portions of CRE loans have expanded the benefits of CDOs for both B-piece buyers and other CDO investors. the manager's ability and expertise have become ever more important. The deal's collateral portfolio was revolving and actively managed. Not surprisingly. Newer CRE CDOs that include unrated. some asset managers seek to increase their assets under management via CRE CDOs. Motivation for Creating a CDO Since the early days.0 years Collateral Fixed-rate only Liabilities Floating-rate Handling of prepayment Sequentially pay off the liabilities.0% (whole loan) 5 years 1.6% (CMBS) Reinvestment period None Weighted average maturity 7. Subordinate lenders often exercise great influence on the fortune of troubled CRE loans. CRE CDOs of this type tend to have higher collateral quality than the first type. Furthermore. Diversity / lumpiness Moody's D-score = 5 Triple-A subordination 59% Triple-A spread 45 bps Triple-B-minus subordination 28% Triple-B spread 350 bps Comment High servicer concentration Source: Moody's and S&P IV. In other words. unlike CMBS. where they typically retain the below-investment-grade portion of a CRE CDO. therefore. many collateral managers are first timers with no previous experience in managing a CDO. For B-piece buyers. New CRE CDOs at a Glance Deal Issue date Portfolio Anthracite CDO I May 2002 Static CMBS (79%) REIT debt (21%) Guggenheim Structured Real Estate Funding I April 2005 Revolving B-Notes (49%) Mezzanine loans (22%) CMBS (16%) Whole loans (11%) Baa3 to NR (below Caa3) B3 (3464) 8.Nomura Fixed Income Research deal were first loss pieces. and the involvement of commercial real estate experts also benefits other CDO investors. the fact that a cash CDO does not require mark-to-market also makes it an attractive alternative to other means of financing. as underlying portfolios consist primarily of investment grade collateral. who have extensive experience in the commercial real estate market. a CDO allows the term of financing match that of underlying assets. Table1 : Old vs. in some of the recent CRE CDOs. the recent tight spread environment in the CMBS market has caused a decline in issuance of these asset 4 However. the primary motivation of CRE CDOs has been the financing needs of B-piece buyers and special servicers. More importantly. these assets were mostly floating-rate and had relatively short 4 maturities. B-piece buyers can "condense" risk and reward of lowly quality CRE assets by putting them into a CDO portfolio. the existence of "arbitrage" between collateral spreads and liability spreads is essential. Given the low collateral quality and the revolving structure. which are generally longer-term fixedrate bonds with strong call protection. These deals open the door for investors who otherwise could not invest in these types of assets. Also. 5 (3) .8 years Floating-rate only Floating-rate Reinvested Herfindahl score = 20 64% 28 bps 18% 180 bps 32% of collateral pool Caa1 or lower Underlying asset Collateral rating A3 to Caa1 Average rating (Moody's WARF) Ba3/B1 (1831) Max single exposure 12. Furthermore. giving CMBS originators some advantage. Bloomberg Table 2 above shows the economics of a sample CRE CDO. CRE CDOs. and whole loans (13%). even non-CMBS collateral can be difficult to source in the current environment. although some deals included small amounts of whole loans. CMBS. debt 20% 475 AA 15% 55 Whole loans 10% 130 A 20% 100 CMBS 18% 250 BBB 12% 225 REIT debt 2% 113 NR 18% Average spread 293 bps to Treasury 68 bps (excluding NR) to 1mL Average rating B2/B3 WARF 3500 * All spreads are indicative levels as of January 2006. we began to see many deals solely consisting of non-CMBS 7 assets. firstloss tranche. and REIT debt. The Data Appendix at the end of this report lists CRE CDO deals since 2001. The deal's underlying portfolio was comprised of mezzanine loans (60%). which the deal's manager often retains. synthetics. In the second half of 2005. As we can see. One example is a deal called Carbon Capital II. etc. B-notes (27%). Table 2: Deal Economics of a Sample CRE CDO Assets Liabilities Type Portion Spread* Tranche Size Spread B-Notes 50% 275 AAA 35% 35 Mezz. The available return on the equity portion. Table 3: Collateral Composition of Recent CRE CDOs (H1 2005) Collateral type B-Notes CMBS Mezzanine loans Whole loans Credit tenant leases Preferred equity REIT debt Other (corporate mezzanine loans. Moody's special report (28 July 2005). excluding the non-rated. Source: Nomura Securities International. priced in August 2005.Nomura Fixed Income Research management CRE CDOs. is the important driver of a transaction.) Total Source: Moody’s Share 28% 25% 20% 14% 6% 4% 2% 2% 100% 6 7 US CMBS 2Q 2005 Review: Credit Metrics Slip As Issuance Surges. the older CRE CDOs were predominantly comprised of CMBS bonds and REIT debt. (4) . CRE CDO Ingredients – What They Are As mentioned above. to Libor. The triple-A tranche and the tripleB tranche priced at 31 bps and 165 bps. Mezzanine debt. which is backed by B-Notes. Table 3 below shows the average deal composition for the first half of 2005. the average portion of B-notes in CRE CDOs issued during the first half of 2005 was greater than that of 6 CMBS. The average spread of the underlying assets is about 300 bps. while the average spread on the liability side is only about 70 bps. respectively. the portion of CMBS and REIT in a CRE CDO has declined significantly over the past year or so. whole loans. However. V. reported by Moody's. this would represent the entire $100 million. on the capital structure. are typically compensated with higher yields (see Table 4 in the next page). as they offer better execution to the issuer. B-Notes and Mezzanine Debt B-notes represent a subordinate portion of a commercial mortgage in the so-called A/B structure. better pricing to the 8 property owner. The following are brief descriptions of various commercial real estate assets included in many of the recent CRE CDOs. In the sample structure on the following page. it is very difficult for a borrower to take equity out of a property. A. and thus. Within a REMIC trust. Sample Capital Structure AAA AA $55 mm $70 mm Loan (A-Note) CMBS Transaction BBBUnrated Class $100 mm Property $10 mm B-Note $5 mm C-Note CRE CDO $15 mm Mezzanine Loan $30 mm Equity $5 mm Pref. CDO structures. it is a pledge of equity interest of the first-mortgage borrowing entity that is subordinate to the first-mortgage. and as such. because prepayment options. Investors who choose to purchase B-notes and/or Mezzanine debt expose themselves to greater risk than those buying investment grade bonds. can be very costly.Nomura Fixed Income Research While all these assets represent risk exposures to the commercial real estate sector. (5) . 8 With conduit/fusion deals increasing in size. which may offer substitution of collateral. Whole loans are more commonly included in REMIC deals. like defeasance. but above the hard equity holder. borrowers are looking to CDO structures more and more because of the greater flexibility. their risk profiles vary significantly. A mezzanine loan is not secured by a lien on the property in question. However. Whole Loans A whole loan is a un-tranched loan secured by a commercial mortgage. Equity $10 mm Common Equity Source: Nomura B. may not involve such prohibitive prepayment constraints. issuers were able to include the entire loan without hurting the diversity of the deal. However. and how much real estate experience does he/she/they have? What amount of assets. More specifically. what the triggers are. How the local economy is forecasted to perform going forward. • For an office. All parties involved are concerned about an exit strategy in order to protect their principal investment. what kind of leverage is associated with the loan? (Note that floating rate B-Notes tend to be more leveraged because the asset is typically more transitional in nature) What is the reputation of the sponsor of the project. if the loan contains any lockbox provisions. Typically. if not exclusive. Along the same lines. he will have to exercise the option usually granted the B-Note holder of buying out the A-Note holder’s participation (6) .CMBS Moderate BB CMBS Moderate-High B CMBS High Source: Nomura Securities International The fundamental concerns facing these investors (and thus. the B-Note holder’s right to principal and interest payments will be subordinate to the holder of the A-Note. especially liquid assets.Nomura Fixed Income Research Table 4: Risk and Return Characteristics of Selected Commercial Real Estate Assets Position on Capital Structure A-Note (senior) B-Note (junior) Mezzanine Equity Perceived Relative Risk Low Moderate Moderate-High High Likely Range of Spreads/ Expected Returns Treasury + 75-150 Treasury + 200-350 Treasury + 400-550 Mid Single Digits – 15% + Swaps + 60 Swaps + 125 Swaps + 185 Swaps + 240 area Swaps + 650 area LTV 0 – 55% 55 – 65% 65 – 90% A CMBS Low BBB CMBS Low-Moderate BBB. the A-note holder will generally have greater. and if so. control over any bankruptcy proceedings dealing with the workout of a troubled loan. on the underwriting side: • How much equity does the borrower have at risk in the property? It is important to assess the borrower’s capacity to back a project should it get into trouble. Also. or tertiary market? More specifically. etc. for specific property types. In the event of default. this will deter future investors. secondary. overall and in your market? Is your hotel easily accessible? How far is the nearest airport from your property? Is it located on or near a main thoroughfare? • As for structural matters. B-note /Mezzanine debt investors look at whether the borrower may incur additional debt. • • Investors also analyze the health/value of the property. for the B-note holder to obtain greater control over the workout of a troubled loan. paying close attention to: • • • Historical operating performance Competing properties in the area Typical incomes of people living near the property (demographics). investors prefer amortizing loans to IO loans. the potential CDO investors) deal with the underwriting and structural features of the loans. types of financial covenants involved with the property. Is there any amortization in the capital stack through the life of the loan? What is the exit strategy? Due to the higher leverage involved in B-notes. what do your rent rolls look like? Are there any above market rent-payers coming to the end of their lease? Below market-payers? What is the financial health / ratings of your tenants? Are you heavily dependent on any one tenant/industry? For hotel properties: How is the tourism industry performing. Is the property located in a primary. do they own or manage? If a sponsor has a reputation of bailing out when its projects experience problems. between the two. Rake bonds Rake bonds were common on older floating rate deals. D. 2. at 2:25:1 ratio (9 upgrades. REIT Debt Real Estate Investment Trusts. but have become less popular in recent days. and property managers The mezzanine lender is generally afforded certain rights as well. but are not limited to: • • The right to hire and fire the special servicer The right to cure defaults in order to keep the senior lender from foreclosing (usually. or REITs. or government securities. In order to qualify for REIT status. and gains from securities sales. According to Moody's. and yield spreads have historically been very similar to that of BBB rated conduit/fusion CMBS classes. REIT upgrades outpaced downgrades during 2005. at par plus accrued interest. The REIT must have at least 100 shareholders and must have less than 50% of the outstanding shares concentrated in the hands of five (5) or fewer shareholders. as dividends to its shareholders. cash. Other important rights that the B-note holder may have include. shares in other REITs. since both positions are above him on the capital structure. 4. to purchase in the event of default) . interest. are companies that purchase and manage real estate or real estate loans using capital invested by shareholders. or gains from the sale of property. at a certain point in time. excluding capital gains. See. Thus. and are tied to the performance of a specific loan. At least 95% of gross income must come from the aforementioned sources. according to NAREIT. Derive at least 75% of its gross income from rents. Distribute at least 90% of its annual taxable income. This reversed the trend from 2004. there is a limit to the number of times this right can be exercised) Approval rights associated with the property budgets. by definition. mortgage loans. leases. 9 This means that the mezzanine lender would have to buy out both the A-note and B-note holder. which the mezzanine lender 9 normally has the right.Nomura Fixed Income Research in the loan. together with dividends. when there were 9 downgrades and only 6 upgrades. As of October 2005. As might be expected. 3. REIT unsecured debt ratings are typically clustered around the BBB range. including. portfolio managers often view the investments as alternatives for one another. These bonds are typically subordinate on the capital structure. the company must: 1. Moody's Credit Opinions (November 2005). but not limited to: • • The right to cure defaults in the first mortgage loan or loans The ability to foreclose on the borrower’s ownership interest and take control of the underlying property (subject to the lien of the first-mortgage. 4 downgrades). total unsecured REIT debt outstanding was about $98 billion. and make sure they are comfortable with the loan/asset. mortgage interest. CDO investors should be aware of exactly which property in a deal their “rake” bond is exposed to. 10 (7) . the lack of diversity inherent in this type of bond leaves it more open to downgrade and default risk relative to a bond backed by a pool of mortgages. Invest at least 75% of its assets in real estate. with the investment decision 10 generally based on the relative spreads. but not the obligation. C. Nomura Fixed Income Research Certain REITs are specialized. Inc. Others. Graph 2: Spreads of Triple-A CRE CDOs vs. and Apartment REITs. (8) . The types of REITs that are often involved in the CDO market are classified as “commercial financing” REITs and include names such as Gramercy. which have generally been moving in sync. such as Vornado Realty Trust. CMBS has been one of the best performing sectors within the structured finance market. Arbor Realty Trust. Graph 2 below shows the triple-A spreads for both SF CDOs and CRE CDOs. and CRIIMI MAE. VI. CMBS U. Inc.. commercial real estate sector over the last few years.S. usually diversifying by geography (e. Spreads and Credit Performance of CRE CDOs vs. In general. CDO spreads have tightened significantly over the past couple of years. Anthracite Capital. that is. The most common are Office. Triple-A SF CDOs 80 70 60 Basis Points 50 40 30 20 2Q01 3Q01 4Q01 1Q02 2Q02 3Q02 4Q02 1Q03 2Q03 3Q03 4Q03 1Q04 2Q04 3Q04 4Q04 1Q05 2Q05 3Q05 4Q05 AAA CRE CDO Sources: Nomura Securities International AAA SF CDO Credit performance has been extremely positive for the U. though there are also a substantial number of Industrial and Lodging REITs. Retail. are more general and diversify across property type and location. Equity Office Properties Trust). only invest in a particular property type.S.g. CRE CDO spreads have moved in line with structured finance (SF) CDO spreads. Graph 3 shows the numbers of upgrades and downgrades of North American CMBS over the past quarters. . Both of the two deals were issued prior to 2002. On the other hand. U. Rezak. and more than 190 tranches were upgraded to triple-A. Commercial Real Estate CDO 2005 Performance Review.Nomura Fixed Income Research Graph 3: CMBS Quarterly Rating Changes 400 Number of ratings changed 300 200 100 0 1Q03 2Q03 3Q03 4Q03 1Q04 2Q04 3Q04 4Q04 1Q05 2Q05 3Q05 11 Upgrades Sources: S&P Downgrades In 2005 (through the third quarter). just five tranches from two deals have been downgraded 12 to date. However.S. below-investment-grade tranches dominated downgrades. non-CRE factors. Fitch predicts that rating performance is likely to be different for newer deals that have a revolving structure and different underlying assets. S&P research (27 October 2005). S&P upgraded 600 CMBS tranches while downgrading just 80. only a handful of the deals experienced negative rating migrations. Not surprisingly. J. representing more than 70% of all downgrades over the same period. Investment-grade tranches dominated upgrades. Of the CRE CDOs issued to date. According to Fitch. CRE CDOs have exhibited particularly good performance as well. of the 52 CRE CDO deals it rated. Table 5: Credit Performance of CRE CDOs Issued Deals Tranche Prior to 2002 11 62 2002 12 83 2003 9 76 2004 9 93 2005 YTD 11 84 Total 52 398 Source: Fitch Vintage Downgraded (Tranche) 5 Upgraded (Tranche) 19 29 16 5 69 5 11 12 CMBS Quarterly Insights. Third-Quarter 2005. and K. Kendra. Fitch special report (6 October 2005). In these deals. contributed to the downgrades (see Table 5 below). such as exposures to the manufacturing housing (MH) ABS sector. (9) . vintage. these synthetic CRE CDOs involve a static portfolio of highly rated CMBS. subordination levels can significantly vary depending on the underlying portfolio's characteristics. rather than subordinate 16 pieces. in order to assess the portfolio's diversity. the rating agencies assign their CMBS groups to rate CDOs that are backed predominantly by commercial real estate assets. The approach is reasonable. The risk profile of an asset could vary significantly. The junior triple-A tranches of the two deals priced at 1mL+55 and 1mL+60. because CMBS and REIT debt are treated as more highly correlated with each other than other assets. 13 According to the presentation given at Moody's 5th Annual U. 14 The rating agencies now use a simulation-based approach in analyzing CRE CDOs. Some market participants believe that CRE CDOs are punished for their lower diversity than other types of CDOs. N.. such as CW Capital Cobalt I. Bailey. New York (7 September 2005).S. Moody's rating methodology (17 June 2005). such as corporate credits. Some cash CRE CDOs. and geographic location. depending on its position in the capital structure and other structural features. information is aggregated to assess losses at the CDO level.. See. the first step is to "shadow rate" them using a similar approach used in rating CMBS. et al. At the portfolio level. In general. 14 Moody's admits it now "relies exclusively" on the newer simulation-based model. After the loan-level analysis. the rating agencies have been applying similar rating methodologies for CMBS and CRE CDOs. For the time being. property type. where the CDO's reference pool consists of triple-A CMBS securities. Also. Also. Levidy. servicer. there have been several 100% synthetic CMBS CDOs in 2005. were such deals. and P is the aggregate principal balance. Halcyon 1 and Abacus 2005-4. rather than on its binomial expansion model.Nomura Fixed Income Research VII. respectively. CDO Investor Briefing. If a CDO includes unrated assets. B.e. a senior CRE CDO tranche tends to require larger subordination than a similarly rated tranche of a high-yield corporate CBO. pi is the principal balance of each asset. VIII. Synthetic Overview for CMBS Investors. How CRE CDOs Are Analyzed Since the early days of CRE CDOs. However. the rating agencies examine the loan-to-value ratio (LTV) and debt service coverage ratio (DSCR) to assess the credit risk of each loan. Synthetic CRE CDO Over the second half of 2005. 15 Moody's Herfindarl score measures a pool's lumpiness and is calculated as Herfindahl _ score = 1 ∑ (p i =1 N i / P )2 where n is the number of assets. such as B-notes and mezzanine loans. At Moody's. its CMBS group rates a CDO if the transaction contains 75% or more CRE assets. began to include a small portion (about 6%) of synthetic CMBS along with cash CRE assets. because the source of risk is the same: underlying commercial real estate loans. However. issuance of synthetic CDOs referencing CMBS picked up.. others think that there is little room for additional benefit from pooling CMBS and REIT debt that are already diversified. such 15 as diversity scores or the Herfindahl score. 16 17 See. Fitch special report (26 September 2005). issued in July 2005. (10) . CMBS: Moody's Approach to Rating Static CDOs Backed by Commercial Real Estate Securities. 13 for instance. "chunkiness") of individual deals or loans.. et al. both referencing 30 17 triple-A rated CMBS. Simulation allows incorporating correlation of underlying assets more easily. At the loan level. simulation can show the implications of having a "lumpy" portfolio more precisely than arithmetically derived measures. they would focus on the concentrations (i. when analyzing a CRE CDO. Naturally. we started to see the application of leveraged super senior (LSS) technology to CRE 18 CDOs. assessing the likelihood of spread triggers can be challenging. CMBS portfolio tends to be "chunky" with several large loans making up a large portion of the portfolio. these structures tend to come with additional triggers. and this feature drives a large portion of risk in LSS. Like inner CDOs of a CDO-squared. resulting in a high probability of tranche wipeout and thus a higher severity rate. trading of CMBS CDS represented as much as 30% of all structured finance (SF) CDS volume. . As of the third quarter of 2005. Unfortunately. where investors can take on a leveraged exposure to the super senior portion of the capital structure. see. a CRE CDO comprised of 30 CMBS tranches can be viewed as a CDO-squared comprised of 30 "inner" CDO tranches. one at the asset (loans or CMBS) level and the other at the CDO level. According to the report. Leveraged super-senior became a popular structure in corporate-backed synthetic CDOs in 2005. CMBS tranches tend to be relatively thin. The risk profile of the CDO-squared tranche is greatly affected by characteristics of the inner CDOs such as portfolio quality. the "resecuritization" of existing CDOs. These triggers are included to compensate a sponsoring dealer for the so-called gap risk. Moreover. Second. there are several important factors to consider. each of the underlying CMBS tranches represents a portion of a diversified portfolio. such a structure is subject to similar sensitivity and volatility to CDOs-squared. (11) . one can benefit from pooling a large number of these assets. synthetic corporate indices. a consortium of dealers is working on creating synthetic CMBS index. First. which priced in October 2005. rather than a pool of CMBS. In addition. some CMBS have overlapping exposures to the same mortgage loans in the form of "pari passu" bonds.Nomura Fixed Income Research Issuance of synthetic CMBS is also gaining some traction. while the benefit of diversification is likely to be much smaller for CMBS that are already diversified. Third. CDOs-squared are known to be very sensitive to the level of 20 correlation and "overlaps" among underlying assets. Like CDOs of structured finance products (i. In our view. Nomura Fixed Income Research (4 February 2005). Breaching of a trigger forces the investor to reduce leverage or take mark-to-market loss. On the other hand. In our view. These assets generally represent lower portions of the capital structure and hence are by themselves more risky than CMBS. Given the relatively short history of structured finance products. this particular transaction is built on a synthetic CRE CDO deal called Halcyon 2. ABS CDOs). a CRE CDO that includes tranches of existing securitizations involves two layers of tranching. CRE loans that are increasingly included in CRE CDOs are likely to have significantly different risk characteristics from CMBS. such as the Dow Jones CDX. 19 such as loss triggers or spread triggers. however. For example. tranche size and subordination. If we look at a CRE CDO as a CDO-squared consisting of many inner CDOs (CMBS). Hypo Leverages Super Senior CMBS Tranche. CDOs-Squared Demystified. Synthetic CMBS is a credit default swap (CDS) referencing a CMBS tranche.e. In late 2005. The rapid expansion owes largely to the introduction of the standardized documentation for SF CDS in June 2005. or CDO-squared (CDO^2). 18 See.. 20 19 For detailed discussion of CDOs-squared. One Final Thought: a CRE CDO has Some Characteristics of a CDO-Squared! Between 2004 and 2005. Securitization News (26 December 2005). investors should distinguish between the underlying CRE risk (at the loan level) and the structural implications when analyzing a CRE CDO. it is not clear at present if the rating agencies and market participants fully appreciate the implications of structural characteristics in different CRE assets. along SM the lines of the publicly traded. As in the corporate version of LSS. IX. However. or the risk that the reference portfolio suffers a large amount of loss and the amount invested is not adequate to cover the loss. has attracted much attention for its complexity and high leverage. Conclusion In this paper. we discussed the evolution of commercial real estate (CRE) CDOs. Asset Securitization Report (7 November 2005). However. The newer. While there would be additional risk associated with including assets from the lower part of the capital structure and in allowing reinvestment in the collateral portfolio.Nomura Fixed Income Research X. In our view. see: Who Let the CRE CDO Issuers Out?. Such an approach is likely to be able to grasp how the risk of the underlying commercial real estate filters through complex structures of each loan or CMBS deal. the analysis of a CRE CDO should involve a similar approach to the drill down process used for a CDO-squared. more flexible structure of CRE CDOs allows more efficient risk transfer by expanding the investor base beyond the traditional CRE players. some market participants have expressed concerns about the rapid expansion of 21 the CRE CDO sector. 21 For example. (12) . we think that significant risk lies in their structures. One aspect of the evolution is that CRE CDOs now include unrated. non-securitized CRE assets that were previously held only by a handful of commercial real estate players. with stress added for issuers and vintages. Moody's calculate a "bottom-dollar" LTV. REIT. one for loan pool losses and the other for 22 bond-level losses. et al. O. Moody's shadow rating of an unrated asset is based on the so-called "bottom-dollar" default probability. To determine it. On the other hand. The table below shows Moody's correlation assumptions for CMBS. which is unadjusted for loss severity. Correlation for a CRE loan is generally assumed to be lower than CMBS. Moody's At Moody's. Naturally. derived from the historical ratings transitions over the past 20 years. Moody's Revisits Its Assumptions Regarding Structured Finance Default (and Asset) Correlations for CDOs. which is a measure of diversity of a CDO portfolio. a bond loss simulation generates correlated defaults for each CMBS security based on its credit rating. location. In a more general CDO context. Moody's has recently revised its assumptions for correlation among TM 23 structured finance products used in its CDOROM model. and RE CDO sectors.. correlation is adjusted depending on characteristics such as property type. Appendix: CRE CDO Analytical Approach by Rating Agency A. Moody's special report (27 June 2005). whole loans are analyzed much in the same way as loans included in a CMBS deal. as well as bonds' ratings. issuers.Nomura Fixed Income Research XI. For B-notes and mezzanine debt that are behind securitized A-notes. In the loan pool loss simulation. loss severity rate depends on property type and thickness of an asset in the loan's capital structure. On the other hand. Correlation between non-CMBS CRE assets is assessed based on the pool's diversity. Moody's also reports a Herfindahl score. Correlation is determined based on sectors. This approach is similar to the simulation commonly used in synthetic CDOs of corporate credits. the CMBS group rates CDOs when more than 75% of the collateral pool consist of CRE assets. Moody's assigns "shadow ratings" to unrated CRE assets using a CMBS-like loan analysis. It would arguably yield better results than treating these CMBS tranches as separate assets in the same industry. Moody's assigns pair-wise correlations between asset classes. Correlation depends on the deal type. as a LTV of the last dollar of debt on the property. Once the loss amount for the deal is simulated. the corresponding tranche loss can be determined and aggregated for the CDO portfolio losses. On the other hand. Simulating pool losses for each underlying CMBS deal is advantageous when multiple tranches of the same CMBS deal are included in a CRE CDO.. the loss curve of an individual CMBS deal is derived from the initial subordination levels of the deal. The table below summarizes the general assumptions Moody's currently uses. vintage and issuer. and vintages. Under the new approach. the analysis conducted for the A-note is used to assess the risk of the subordinate assets. 23 Toutain. which are viewed as already diversified. Moody's CRE Analysis Asset Default probability B-notes Mezzanine debt Based on bottom-dollar LTV Whole loan Source: Moody's Loss severity rate Lower than CMBS Lower than CMBS but higher than B-notes About 40% (much lower than CMBS) Correlation Lower than CMBS Moody's employs two types of Monte Carlo simulations. Loss severity for a CRE loan is generally lower than that for a thinly tranched CMBS. 22 (13) . Moody's has published detailed criteria for revolving pools both at the pool level as well as the asset level. et al. Moody's expressed its concern about reinvestment flexibility in CRE CDOs. CDO Evaluator. and REITs (and REOCs). In order to control the uncertainty of changing portfolio composition. to analyzing CRE CDOs. 24 B. diversified CMBS (conduit) and CTLs. while loss severity 24 Philipp. According to the presentation at the S&P 25 26 Annual Global CDO Conference. Moody's stresses the important role played by subjective judgments in assessing credit quality of certain unrated CRE assets. Fitch conducts a loan-level review that is similar to the one used for analyzing conduit CMBS deals. S&P Recovery Rate Assumptions Senior B notes First-loss B notes Whole loans Mezzanine loans / preferred equity Source: S&P 35% 30% 50% 25% C. S&P released the latest version of its model. on December 19. as 26 listed below. office and retail properties. managed) pools. T.e. CDO Evaluator 3. 2005. unsecuritized assets.0. For CMBS and REIT debt. Also. Apparently. the most significant change in the CRE CDO market is the shift from static pools to revolving (i. risk tends to be much greater for a low-diversity pool with "chunky" assets. single borrower. New York (19 September 2005).. S&P 25 S&P applies its CDO simulation model..Nomura Fixed Income Research Pair-wise Correlation for CRE Sectors (Assuming North American assets with full vintage effect but no key agent stress) CMBSCMBS-Conduit CMBS-CTL REIT-Hotel REIT-Office Large loans CMBS-Conduit 35% 7% 7% 6% 6% CMBS-CTL 35% 7% 6% 6% CMBS-Large loans 35% 6% 6% REIT-Hotel 45% 15% REIT-Office 45% RE CDO Source: Moody's RE CDO 6% 6% 6% 6% 6% 34% As mentioned above. some issuers started to demand a mechanism where new collateral assets can be purchased without a detailed review by the rating agency. An "out-of-the box" analysis can be particularly important for management/capital intensive properties. mezzanine loans and whole loans. Fitch For analyzing CRE CDOs backed by CRE loans. For unrated. S&P treat them as one sector and assigns a single correlation number of 15%. S&P uses three sector categories for rated CRE assets. For B-notes.. the portfolio switches to static) once credit quality of the pool deteriorates and certain triggers are breached.. US CMBS 3Q 2005 Review: Conduit Leverage Levels Off in Moody's Rated Deals. in order to take advantage of changing market conditions. Correlation is assumed to be 30% within the same sector and 10% across different sectors. Default probability and loss severity are assessed for individual loans and then aggregated to arrive at expected losses for the entire portfolio. such as hotels and healthcare facilities. (14) . S&P also assigns "shadow ratings" before putting them into the model.e. Recently. Moody's special report (31 October 2005). compared to multifamily. default probability is determined from a debt service coverage ratio (DSCR). other CMBS (large loan. and single property). Recovery rate is assumed to vary based on the type of CRE loans. A compliance mechanism may be installed where the revolver is shut down or suspended (i. For each loan. To the extent that the deal includes rated securities. see. 27 For Fitch's approach to CRE CDOs. etc.Nomura Fixed Income Research is assessed from a loan-to-value (LTV) ratio. Rating Methodology for U. (15) . Interestingly. J. such as property type. Revolving Commercial Real Estate Loan CDOs. Both numbers are adjusted to take into account characteristics of the underlying property. the deal's reinvestment criteria would be evaluated.. Fitch criteria report (28 September 2005). At the portfolio level. cash flow volatility. Lee. et al. Fitch does NOT use its Vector simulation model for a CDO that includes only unrated CRE assets. For a revolving deal. Fitch reviews originators and servicers of the underlying loans.. the rating agency combines simulation outputs and the results from the 27 loan-level analysis. position in the capital structure.S. 40 31 28 32 60 55 33 73.1% (B notes) 12.4% (REIT) 14% (REIT) 6 (RE CDO) 64.5% 37.00 $350.65% 22.00 $500.75 $415.26% 16.2% (B notes) 60.4% (mezz notes) 2.7% (mezz loans) 8.4%/23.00 $92.1% (B notes) 23.5% (B notes) 19.000.6% 26.56% 2.3% (mezz loans) 34.21% 33% Gramercy Real Estate CDO 2005-1 N-Star REL CDO IV 6/21/2005 $1.6% (preferred equity) Arbor Realty Mortgage Securities 2005-1 ARCap 2005-1 Newcastle VII Sorin RE CDO II* Brascan RE CDO 2 Acacia CRE CDO I JER CRE CDO 2005-1 Halcyon 2* Wrightwood CDO 2005-1 N-Star CDO V 12/20/2005 12/19/2005 12/8/2005 12/8/2005 11/23/05 11/23/2005 10/25/2005 10/5/2005 10/19/2005 9/13/2005 $475.4% (B notes) 21.75% 15.80% 61.31% 3.85% 74.5% (whole loans) % Other RE 36.00 $337.00% Guggenheim RE CDO 4/29/2005 $500.3% (16) .65% % of CMBS 0% 100% 67.8% (REIT) 5% (CRE CDO) 2.Nomura Fixed Income Research XII.40 $525.7% (mezz loans) 3% (whole loans) 24% (B notes and rake bonds) 5.35% 60.75% 40.84% 35.00 26 25.2% (mezz loans) Sorin Real Estate CDO I 6/24/2005 $403.9% Carbon Capital II LNR CDO III Guggenheim Structured RE 2005-2 Abacus 2005-4* Halcyon 2005-1* FMC Real Estate CDO 2005-1 8/18/2005 8/8/2005 8/4/2005 7/21/2005 7/20/2005 7/1/2005 $455.00 $82.14% 65.2% (whole loans) 26.00 35 53.3% (sub-prime RMBS) 16.7% 100% 100% 0% 30 AAA CMBS 30 First mortgage loans 19.6%(whole loans) 15.00 $650.65% 39.7% (whole loans) 16 (B notes) 2 (mezzanine loans) 42.00 $261.86% 16% CW Capital Cobalt I Newcastle CDO VI N-Star REL CDO III Capital Trust RE CDO 2005-1 CapLease CDO 2005-1 4/18/2005 4/19/2005 3/10/2005 3/2/2005 2/3/2005 $450.100.9% (B notes) 2.5% (REIT bank facility) 42% (whole loans) 41% (B notes) 12% (mezzanine loans) 57.00% 0% 5/27/2005 $400.00 $300.30% 26.00 28 63. triple-A subordination 57.0% (mezz loans) 30 AAA-rated CMBS 30 AAA-rated CMBS 59.2% (mezz loans) 27.00 $6.1% 92.4% (STL) 1.0% Prima Capital 2005-1 5/10/2005 $376.80 $300.00 32 43.00 $400.63 ($/€) $439. Data Appendix: CRE CDO Deal List Deal Issue Date Amount Sr.00 $468.5% 27.8% (CRE CDO) B notes mezz loans whole loans 49% (B notes) 22% (mezz loans) 11% (whole loans) 21% (B notes) 23% (whole loans) 6% (synthetic CMBS) 2% (CRE CDOs) 8.5 Jr.00 28 30/40 28 29 42 42. triple-A spread (bps) 37 50 28 37 33 32 32 45 32 26.5% 0% 16.18% 31.00% 0% 82% 17.00 $1.00 100% 100% 0% 75.1% (Mezzanine loans) 4.13% 2.2% (B notes) 37.00 $377.8% (B notes) 31.00% 49% 66% 73% 16.000.5% (REIT debt) 59.2% REIT 7.26% 80.00 30 18.00 $568. 0% 83.47% 40.00 $450.8% (B note) 2.5% (REIT) 92.00% 18.0% (REIT) 16% (REIT) 3% (whole loans) 29.18 $515.99% 76.60% 22.63% 20.00 $500. Moody's.2% (B notes) 20% (REIT.7% (REIT) 5.00 $323.00% 26.5% 65. Fitch.00 $400.5% 100% 83% 79% 65.55% 58.40% 44.1% 35. whole loan.00 $801.1% (REIT) 26.7% 63.00 $435.00 $660.8% (REIT) 10.10 $325.50 $416.9% (REIT) 56.00% 63.00 $324.7% 52. IFR * denotes synthetic — END — (17) .097.50% 69.9% (REIT) 32.95 $500.7% (mezz loans) 35.49% 24.00 $350.0% 54.3% 81% 55.9% 71.4% 63.3% 50.49% 26.3% (whole loan) 12% (REIT/CTL loan) 47% (REIT) 19.81 $500.4% (RE CDO) 25. RE CDOs) 20.9% (mezz loans) 23% (REIT) 5 % (RECDO) 11.00 $480.13 $397.50 $1.18% 21.00 $300. triple-A spread (bps) 45 33 34 33 35 36 40 50 48 45 25 (MM) 42 70 53 55 58 57/90 57 65 47 48 45 45 47 55 55 49 50 47 50 Jr.60% 24.3% (REIT) 2.0% whole loans 66.Nomura Fixed Income Research Sr.0% 22.4% 88.75% 51.00 $307.12% 27.7% (REIT) 49.00% 41.0% (REIT) 58.89% 26.2% (REIT) 4.2% 42.0% CMBS CDO Crest 2001-1 2/1/2001 $500.9% 87. Bloomberg.1% (REIT) 6.00 $472.9% 66.2% (REIT) 17% (whole loans) 21 (REIT) 33.3% 100% 37.5% (whole loans) 2.4% Deal Arbor Realty Mortgage Crest 2004-1 Newcastle CDO V Capital Trust RE CDO 2004-1 N-Star REL CDO II Anthracite CDO III Newcastle CDO IV G-Force 2003-1 Crest 2003-2 TIAA Real Estate CDO 2003-1 Newcastle CDO III N-Star Real Estate CDO I Newcastle CDO II LNR 2003-1 Crest Dartmouth Street 2003-1 Crest 2003-I G-STAR 2003-3 Anthracite CDO II G-STAR 2002-2 LNR 2002-1 G-Force 2002-1 Anthracite CDO I TIAA Real Estate CDO 2002-1 Crest 2002-I G-STAR 2002-1 Newcastle CDO I Crest G-Star 2001-2 Putnam Structured Product CDO 2001-1 Crest G-Star 2001-1 Issue Date 12/16/2004 10/11/2004 9/22/2004 6/30/2004 6/29/2004 3/16/2004 3/11/2004 12/5/2003 11/26/2003 10/20/2003 9/8/2003 8/8/2003 8/5/2003 6/25/2003 3/20/2003 3/5/2003 2/27/2003 11/26/2002 11/5/2002 6/28/2002 5/31/2002 5/14/2002 5/6/2002 4/24/2002 4/10/2002 3/28/2002 11/16/2001 11/13/2001 8/7/2001 Amount $305.7% (REIT) 60.1% 55.9 48. Junior A note.00 $540.00 $500.3% (REIT) % of CMBS 0% 80.00 $600.00 Source: S&P.4% (whole loans) 28.8% (REIT) 15.0% 52% 70.00% 22.4% (RE CDO) 1.1% (B notes) 7.00% 24.00% 56% 22.00 $351.00 % Other RE 56.70 $350.3% 0% 72.9% 33.49% 25.00% 28.00 $402. triple-A subordination 61.00 $300.00 $450.67%/13.4% 85.9% (REIT) 16.1% 39.0% 21.28% 18. and IRR (21 Mar 2005) • CDOs-Squared Demystified (4 Feb 2005) • Student Loan ABS 101 (26 Jan 2005) MBS • • • • Overview of the ARMs Market (23 Nov 2005) GNMA Project Loan Prepayment Report (12 Oct 2005) FHA/VA Monthly (12 Apr 2005) Some Investment Characteristics of GNMA Project Loan Securities (4 Apr 2005) Strategy • Flatter Treasury Curve Leads to Steeper Agency Curve (6 Dec 2005) • RMBS Expect Fewer home Sales in 2006 (3 Dec 2005) • New OFHEO Report: Home Price Gains Remain Strong (2 Dec 2005) • Freddie Mac Introduces New Loan-Level Disclosures (1 Dec 2005) • Coldwell Banker's: Regional Home Price Survey (30 Nov 2005) • RMBS: Jumbo-A Credit Bonds-Problems Ahead? (29 Nov 2005) • GNMA/FNMA Swaps: Can the Rich Get Richer? (22 Nov 2005) • CME Introduces Home Price Futures Contract (18 Nov 2005) • Model Portfolio Asset Allocations (18 Nov 2005) • Gold/Oil Ratio and the Bond Market (15 Nov 2005) • MBS Market Check-up: November Update (10 Nov 2005) • Update on Bank MBS Holdings (9 Nov 2005) • U.S. GMAC – Post Delphi Bankruptcy (12 Oct 2005) • CMBS: The Hurricanes and Business Interruption Insurance (11 Oct 2005) • MBS/ABS: US Consumer FICO Score Update (11 Oct 2005) (18) . Fixed Income 2006 Outlook/ 2005 Review (15 Dec 2005) • CMBS Credit Migrations 2005 Update (30 Nov 2005) • Condition of the U.S. Fixed Income Market Conditions (Sept. and Collateral Performance – The Big Picture (27 Jun 2005) • Bond Portfolio Credit Analysis … The CDO Approach (24 May 2005) ABS/CDO • Model Risk Update – Margins of Error and Scenario Analysis (29 Nov 2005) • CDO/CDS Update (28 Nov 2005) • Correlation Redux (17 Oct 2005) • Report from Boca Raton 2005: Coverage of Selected Sessions of ABS East 2005 (20 Sep 2005) • Anatomy of a Credit CPPI (8 Sep 2005) • Constant Maturity CDS (CMCDS) – A Guide (20 May 2005) • CDO/Credit Derivatives 2005 Conference Notes (11 Apr 2005) • CDO Equity. GMAC Sale Looms (17 Oct 2005) • Bank Holdings Update – October 2005 (14 Oct 2005) • CDS: Ford Motor Credit vs. National Averages (3 Nov 2005) • Swap Spread Strategy: November Trade Ideas (2 Nov 2005) • Treasury Strategy: Value in "Off-the-Run" Bonds (1 Nov 2005) • CDS: Corporate Event Risk on the Rise-CSC & CD (1 Nov 2005) • Synthetic CMOs: Value FHLB "APLS" (1 Nov 2005) • Mortgage REITs: Priced for a Bubble Burst? (27 Oct 2005) • Summary of Bush's Tax Reform Proposals (25 Oct 2005) • Treasury Strategy: Underweight 2-Year Treasuries (24 Oct 2005) • CMM-The MBS Market's New Hedging Tool (20 Oct 2005) • MBS Credit: PMI Housing Market Risk Rankings-Fall 2005 (20 Oct 2005) • Supreme Court Decision Favors Tobacco Credits (19 Oct 2005) • Stagflation – Waiting in the Wings (18 Oct 2005) • GM Strikes Deal with UAW.S.S. Credit Support.S.Nomura Fixed Income Research Recent Nomura Fixed Income Research Fixed Income General Topics • U. 2005) (6 Sept 2005) • U. Correlation. Housing Market (3 Nov 2005) • Update on U. Fixed Income 2005 Mid-Year Outlook/Review (30 Jun 2005) • Structured Finance Trends – Yield Spreads. Housing Market Trends: Lower Real Estate Commissions (8 Nov 2005) • RMBS: Value in Non-Agency "NAS" Bonds (8 Nov 2005) • ABS/MBS Impact of New Bankruptcy Law (4 Nov 2005) • California Housing vs. S.. its officers. Singapore. is. Foreign currency-denominated securities are subject to fluctuations in exchange rates that could have an adverse effect on the value or price of. and it should not be relied upon as such. ("NSI") and Nomura Research Institute America. Neither NIplc nor NIHK hold an Australian financial services licence as both are exempt from the requirement to hold this license in respect of the financial services either provides. Nomura International plc and other Nomura Group entities manage conflicts identified through the following: their Chinese Wall. Nomura International plc or by any other Nomura Group company or affiliate thereof. The Nomura Group. client feedback and the analyst’s seniority.. directors and employees. Disclosure information is available at http://www. Nomura International plc and Nomura Research Institute Europe. United Kingdom. Opinions expressed are current as of the date appearing on this material only and the information.. Seoul Branch.Nomura Fixed Income Research I Michiko Whetten. This material is: (i) for your private information. to the extent it relates to non-US issuers and is permitted by applicable law. In addition.S.com/research/disclosures/public/search. and may not. Bhd. Australia.. directly or indirectly related to the specific recommendations or views that I have expressed in this research report. prior to or immediately following its publication. if applicable. (19) . Further information on any of the securities mentioned herein may be obtained upon request. companies mentioned herein.. firm’s overall performance and revenue (including the firm’s fixed income department). investors in securities such as ADRs. Korea. the securities. and buy or sell (or make a market in). maintenance of a Stop List and a Watch List. personal account dealing rules. Limited. a research analyst employed by Nomura Securities International. This publication has also been approved for distribution in Hong Kong by Nomura International (Hong Kong) Ltd. effectively assume currency risk. Inc. Inc. If verification is required. of companies mentioned herein. In addition. The sender therefore does not accept liability for any errors or omissions in the contents of this publication.. Inc. which may arise as a result of electronic transmission. Taiwan.. Securities Act of 1933. The Nomura Group. may not be offered or sold in the United States or to U. the values of which are influenced by foreign currencies. or Nomura International (Hong Kong) Ltd. (ii) not to be construed asa bid or an offer to buy or sell any security in any jurisdiction where such bid or offer would be illegal. Inc. I hereby certify that no part of my compensation was. corrupted. New York. or contain viruses. Inc. but we do not represent that it is accurate or complete. (collectively. Malaysia. Japan. Nomura Advisory Services (Malaysia) Sdn. nor is it tied to any specific investment banking transactions performed by Nomura Securities International. have acted upon or used this material. be redistributed to other classes of investors.. directors and employees may. hereby certify that all of the views expressed in this research report accurately reflect my personal views about any and all of the subject securities or issuers discussed herein. including the opinions contained herein are subject to change without notice. Nomura Securities International. Non-US members of the Nomura Group. the “Nomura Group”): Nomura Securities Co. ("NSC") and Nomura Research Institute. Nomura International (Hong Kong) Ltd. If this publication has been distributed by electronic transmission. Nomura Singapore Ltd. This publication contains material prepared by one or more of the following affiliates or subsidiaries of Nomura Holdings.. with the contributions of one or more Nomura entities whose employees and their respective affiliations are specified on page 1 herein. which is authorised and regulated by the UK Financial Services Authority (“FSA”) and is a member of the London Stock Exchange. or Nomura International (Hong Kong) Ltd. It is intended only for investors who are “market counterparties” or “intermediate customers” as defined by FSA. NSI accepts responsibility for the contents of this material when distributed in the United States. confidentiality and independence policies. or solicit investment banking or other business from. © Copyright 2006 Nomura Securities International. arrive late or incomplete. Taipei Branch. Inc.. such as e-mail. Indonesia.T. or related securities or derivatives. or will be.. their officers. Unless governing law permits otherwise. and (iii) is based upon information that we consider reliable. therefore. Ltd. Nomura International (Hong Kong) Ltd.nomura. you must contact a Nomura entity in your home jurisdiction if you want to use our services in effecting a transaction in the securities mentioned in this material.. or derivatives (including options) thereof. may perform investment banking or other services (including acting as advisor. please request a hard-copy version. policies and procedures for managing conflicts of interest arising from the allocation and pricing of securities and impartial investment research and disclosure to clients via client documentation. Hong Kong. which is regulated by the Hong Kong Securities and Futures Commission (“SFC”) under Hong Kong laws and in Singapore by Nomura Singapore Limited. destroyed. P. This publication has been approved for distribution in the United Kingdom and European Union by Nomura International plc. Tokyo. NY. (“NIHK”). including quality and accuracy of research. and we are not soliciting any action based upon it. including persons involved in the preparation or issuance of this material may have long or short positions in. manager or lender) for. then such transmission cannot be guaranteed to be secure or error-free as information could be intercepted. Nomura Indonesia. persons except in compliance with an exemption from the registration requirements of such Act. No part of this material may be copied or distributed without the prior written consent of the Nomura Group member identified in the banner on page 1 of this report. Fixed income research analysts receive compensation based on various factors. This publication contains material that has been prepared by the Nomura entity identified on the banner at the top of page 1 herein and. or income derived from the investment.asp. The securities described herein may not have been registered under the U. lost. and. Nomura Australia Ltd. reputation and experience. Ltd. in such case. 667.667. Japan 100-8130 81 3 3211 1811 LONDON Nomura International PLC Nomura House 1 St Martin's-le-grand London EC1A 4NP 44 207 521 2000 David P.2255 International Head of Research Nomura U. Otemachi.667.2132 212.667.NEW YORK Nomura Securities International 2 World Financial Center.9298 212.667. Fixed Income Research David Resler Mark Adelson John Dunlevy Arthur Q.2231 Head of U.667.S.667. Jacob 212. Economic Research Securitization/ABS Research Cross Market Strategist MBS Research Quantitative Research Quantitative Credit Analyst CMBS Research/Strategy Deputy Chief Economist Quantitative Analyst Quantitative Analyst Analyst Analyst Analyst Analyst Analyst Translator Translator 212.667.667.9679 212.667.2417 212. Chiyoda-Ku Tokyo.9894 .2338 212.1314 212.S. Frank Weimin Jin Michiko Whetten James Manzi Gerald Zukowski Xiang Long Cristian Pasarica Diana Berezina Benjamin Cheng Jeremy Garfield Edward Santevecchi Pui See Wong Tomoko Nago-Kern Kyoko Teratani 212.2337 212.667.667.1477 212.9054 212.667.2158 212. NY 10281 (212) 667-9300 TOKYO Nomura Securities Company 2-2-2.667. Building B New York.2415 212.