Toomre Capital Markets LLC

Real-Time Capital Markets -- Analytics, Visualization, Event Processing, and Intelligence

Can Wall Street be Trusted to Value Risky CDOs?

Reporter Neil Shaw of Reuters asks whether Wall Street can be trusted to value risky CDOs? Rather cynically Toomre Capital Markets LLC ("TCM") as well as many investors with a modicum of gray hair near their temples would answer with a resounding "No!!!"

The Reuters article starts with the contention that the complex models that Wall Street uses to project cash flows and hence attempts to "analyze" risky investments such as sub-prime mortgages may be as suspect as some of the loans underlying the mortgage-backed security ("MBS") pools themselves. The recent and sharp surge in defaults on sub-prime home loans has many worried and even frightened. It has jolted both Standard & Poor's ("S&P") and Moody's into a massive downgrade of sub-prime mortgage transactions and caused a run on two Bear Stearns hedge funds that were both heavily invested in securities and derivative contracts with sub-prime exposure and leveraged.

The massive downgrade of the sub-prime MBS is just the first step though. It has been a dirty little secret on Wall Street during the past few years that there were few natural buyers for all of the intermediate credit classes from MBS structured product transactions. There were plenty of buyers for the most senior rated tranches. The excess liquidity in the financial system from the post-9/11 period had many investors desperately hunting for incremental yield and hence return in a relatively low volatility environment for fixed-income investments.

The few basis points of incremental yield were quite appealing from buying Class A-1 of "Some Unknown CDO Trust" backed by rating agency-approved collateral (including sub-prime mortgage tranches) that would result in a "AAA/Aaa" rating. Many asked, "Why not take the incremental yield over some structured mortgage product backed by the implied "AAA/Aaa" credit of FNMA or FHLMC?" And many investors reached for that incremental yield which the Wall Street deal machines were only so happy to feed.

At the other end of the spectrum, credit spreads for low-rated and "junk" corporate securities dramatically compressed. The credit markets have just completed a period on unusually low default rates. As Toomre Capital Markets has noted before, spreads of all types compressed to near all-time narrows relative to Treasury securities around the first of the 2007 year. Hence, throughout the spread tightening process, high-yield buyers in particular have been desperately searching for investment alternatives. Wall Street structured finance professionals have been only too happy to provide some of what is called MBS/CDO equity which is designed as the "first loss" tranche in the event that one or more of the collateral pools or assets defaults and/or suffers cash flow impairment of some type.

The problem for the Wall Street structured finance pros is what to do with the MBS mezzanine tranches. There were not many natural buyers for these intermediate tranches in the credit transactions. Designed by the Wall Street structuring wizards so that they just barely qualified for "BBB/Baa" ratings by the gate-keeper rating agencies, these tranches have always had one problem. They have limited upside and almost unlimited downside. A BBB mortgage-backed (or by extension CDO) tranche is very unlikely to be upgraded, but as witnessed this past week, always has the potential to be downgraded, in some cases nearly catastrophically.

What mezzanine MBS tranches could be sold to clients like insurance companies, state funds and pension plans certainly was. However, those traditional buyers simply could not absorb the surge in MBS structured product. There were simply too many billions of mezzanine tranches left in Wall Street underwriter hands.

Because most homeowners do make housing payments to provide for a roof over their heads, the vast majority of residential MBS pools are eventually repaid. Of course, there is some question about the timing of that repayment and how the investor might suffer in the event that the underlying mortgage loan is refinanced during a period of lower interest rates. As a result, across the spectrum of all outstanding residential mortgages, it can be statistically calculated that even during the worst economic times (like a depression or 1970's inflationary shock), a significant majority of a pool will be receive timely interest and ultimate principal. Hence, a "traditional" (if there even is one) structured MBS product allocates something like 65-70% of the principal to the high-grade (AAA/AA) class(es), 20-25% to the mezzanine tranche(s) and 10% to the equity class.

During 2006, S&P rated approximately $420.6 Billion in sub-prime residential mortgage-backed securities ("RMBS"), a drop of some 7% from 2005 when S&P rated only approximately $460 Billion. Using the lower 20% approximation above, this suggests that the Wall Street MBS underwriters had to place something in excess of $175 Billion in subprime mortgage mezzanine tranches somewhere. The question was where should this unwanted amount of securities go? Wall Street over the short-term certainly could use its huge balance sheet to finance some of these positions. However, eventually they needed to be "sold" at some price.

In short, the answer to the excess MBS mezzanine inventory was to dump the unwanted tranches into these new fangled thingies the Wall Street structuring pros called Collateralized Debt Obligations ("CDOs"). With everyone hunting for yield at the top end of the credit spectrum, and high yield at the low end of the credit spectrum, Wall Street and the rating agencies came up with a new idea. Aggregate a diversified pool of collateral (of course, including as much sub-prime MBS mezzanine paper as one could) and then divide the resulting cash flows from that trust into a series of credit-prioritized tranches. And Voila! There was a way to get rid of a good deal more of the unwanted MBS mezzanine inventory.

Of course, there was then the problem of what to do with CDO mezzanine tranches which also did not sell so well. Hence, Wall Street came up with a yet further enhancement called the CDO^2. The concept was simple. Take a diversified pool of the unwanted CDO mezzanine inventory and credit tranche that again resulting in high grade debt, CDO^2 mezzanine debt and more equity for the high-yield return buyers.

As the reader may recall, Lars Toomre in an earlier part of professional life has been very involved in the MBS market. He was part of the small team that put together Wall Street's first arbitrage mortgage transaction, Investors GNMA Trust 1983-1. It is so unbelievable today to hear that that $500 million Investors GNMA bond issue had more than three points of arbitrage (or more than $15 million in profits after fees for lawyers, trustee and rating agencies). That deal was bigger than all of Lehman Brother's capital was at the time. Lars went on to run the Lehman Brothers CMO New Issue and Secondary Trading desk through 1989 and experienced first-hand the deal machine at work. At that time, Lehman Brothers was a leader in the structured MBS market and achieved a number of firsts, including first whole-loan CMO issue, first floating-rate bond class, first humped-back residual and first ABS issued by Federal government with no credit support.

The deal machine is a hungry beast. It takes quite a number of highly-paid professionals at each investment bank to originate, finance, structure, sell, trade, close and administer the structured finance transactions underwritten. When there are no deals flowing through the pipeline, much of that people expense is very, very, very expensive overhead. Hence, there is almost continuous pressure to ensure that structured finance transactions continue to close on a regular basis.

Similarly, the rating agencies in recent years have had to increase staff to meet Wall Street's demand for ratings on all types of structured product. One might recall how well both the McGraw-Hill and Moody's common equity were doing last year. With each earnings report came news that the mortgage and structured finance fees were increasing and eventually surpassing the fees from rating traditional corporate and municipal debt. The rating agencies themselves then had little incentive to "toughen" or "tighten" their rating standards with so much transaction volume coming across the transom.

The Reuters article continues with the thought that the Wall Street and rating agency models might overlook swift market downturns or corrupt loan data. If that were indeed the case, there would certainly be considerably further turmoil for what are now the interconnected and global credit markets. Well-heeled hedge funds, Wall Street proprietary trading desks, rating agencies and structured finance professionals as a whole may well be too optimistic when analyzing or valuing exotic mortgage investments. As a consequence, Reuters states "future drops in market prices may be more severe and possibly trigger panic selling by sophisticated investors."

Lars Toomre loves the quote from Andrew Lo, a finance professor at Massachusetts Institute of Technology, "There models end up breaking down rather dramatically during abnormal times. And, of course, those are exactly the times that we should and need to worry about." It has been a good number of years since the fixed-income markets last went illiquid.

How many of the young people on today's Wall Street trading floors remember or even know of the lessons learned from past major events? Do they recall that the wheels very nearly came off during the 1987 stock market crash? Or how one could not sell a mortgage during the FSLIC liquidation in the recession of 1989? Do they recall what happened to David Askins and Granite Partners LP amidst the 1994 sell-off when the Federal Reserve repeatedly raised interest rates? Do they recall how illiquid markets can become when the majority of investor positions are valued at less than what was originally paid for them and that any sale must result in the booking of a realized loss? Do they recall just how frightening it was in 1998 when Federal Reserve urged the fourteen dealers to step in and "rescue" the potential disaster that a blow-up of Long-Term Capital Management might have turned into? Do recall that there simply were NO, NONE, NADDA bids for any type of fixed-income security other than the most liquid (like Treasuries)?

The balance of the Reuters article continues:

Ratings companies like Moody's Investors Service use computer models to help predict losses on thinly traded debt investments called collateralized debt obligations, or CDOs, that are often tied to pools of high-risk home loans. The models help the agencies determine what rating a security merits.

Because securities in the $1 trillion (500 billion pound) CDO market trade infrequently, it is difficult for hedge funds and other investors to mark their values to recent sale prices, called "marking to market."

Hedge funds instead use mathematical models of their own to estimate and report the value of their CDO holdings to investors -- a practice known as "marking to model."

Recent troubles at hedge funds run by Bear Stearns, Braddock Financial Corp. and United Capital Markets have highlighted the problems inherent in that approach. Even so, fund managers are resisting market views on the value of sub-prime assets and continuing to "mark to model," claiming declines represent short-term volatility.

"'Mark to model' is a joke," said Janet Tavakoli, president of Tavakoli Structured Finance, a Chicago consulting firm. "What you need to do now is vet the underlying collateral" in CDOs instead of just modeling, which wasn't done earlier, she said. "It's grubby, roll-up-your-sleeves kind of work."

Some hedge funds may now have to report losses on CDOs, while pension funds and insurance companies may dump other securities if these are dropped by raters to "junk" status.


While models may be necessary to analyze investments of such complexity and have worked well under normal conditions, they may break down quickly in times of crisis, MIT's Lo said.

Many popular hedge fund models ignore the possibility of a sudden withdrawal of liquidity, while ratings agencies may make overly abstract or unrealistic modeling assumptions and rely on the quality of the data assembled by Wall Street banks.

This week, Moody's and its rivals Standard & Poor's and Fitch Ratings slashed ratings on billions of sub-prime-related bonds, including CDOs, rattling global financial markets.

Josh Rosner, managing director at investment research firm Graham Fisher & Co., points to a recent S&P statement that the loan performance data it uses has called into question the accuracy of some of the data initially provided to them.

"I find it troubling that the rating agencies are only now publicly recognizing this," Rosner said.

The potential for self-serving pricing by hedge funds is a "serious concern," MIT's Lo said. "There needs to be more independence in pricing and valuation."

Bear Stearns on May 15 said that the riskier of its two hedge funds was down 6.5 percent for April, but then revised that figure to down almost 19 percent a few weeks later.

A logical choice for independent CDO pricing might be rating agencies, but this may be difficult in practice, Lo said.

"If the ratings agency ends up coming up with a really, really good pricing model, the individual responsible for developing those models will very quickly be hired by the hedge funds," Lo said.

The ratings agencies are themselves facing mounting complaints that they have been too slow and opaque in their tackling of the subprime crisis. Some say the agencies ignore key credit risks and cash in by doling out top-notch ratings to subprime-related CDOs.

In the wake of troubles at Bear's hedge funds, the chairman of the House Financial Services Committee said on Wednesday that he would hold a hearing on the role of credit-rating agencies in the fall.

S&P spokesman Adam Tempkin said the agency is very transparent. "We make all of our technical papers completely available to anybody that wants them. They explain every aspect of the models we use."

Noel Kirnon, senior managing director at Moody's, said, "The performance of our ratings overall suggest we're doing a pretty good job."

"The ratings aren't the output of a model," said Fitch managing director Kevin Kendra. "Ratings are the output of a credit committee."


But many say investors in lightly regulated hedge funds or risky CDO securities knew what they were doing.

"Nobody made anybody else put their money into a hedge fund," said Mark Adelson, head of structured finance research at Nomura Securities International in New York.

Fear of a broad-based sell-off in CDOs may also be overblown since some bondholders would be loathe to sell at fire-sale prices and also because any sudden sales would bring in other hedge funds hungry for bargains.

"Not all hedge funds are crying today," said Arturo Cifuentes, managing director at R.W. Pressprich and former global head of CDO research at Wachovia. "For some hedge funds, this is a great opportunity."

Toomre Capital Markets LLC has long felt that the structured finance sector was underestimating liquidity risk. With the events of the last few weeks, liquidity risk premiums certainly have increased for structured products. The interesting question has it been enough yet? Wall Street certainly has a vested interest in whether the structured product sector melts down. Do you as an investor really want to trust Wall Street to value your risky CDOs?