GARP Risk Review (Global Association of Risk Professionals) January/February 2005 Issue 22
by Janet Tavakoli
Rating agencies play an important role in the evaluation of structured finance deals. But do the agencies provide consistent and reliable ratings for the credit risk community? Do they use different approaches when evaluating synthetic CDOs? And what purpose do these ratings really serve? Janet Tavakoli investigates.
Rating agencies are in business to make money by providing a formal assessment of credit risk. They face constant pressure to earn fees for rating deals, and they are not public service companies. The pressure is applied by frequent issuers or placement agents, such as investment banks, and the heat gets turned up an extra notch when rating agencies are asked to quickly rate complicated deals.
These ratings, unfortunately, are not a good barometer for investors. In addition to enjoying a cozy relationship with structured finance issuers, rating agencies have also made it clear that they do not hold themselves accountable for unearthing fraud.
What’s more, the belief held by some industry observers that rating agencies bring greater transparency to the structured finance market is a myth. Ratings are valuable evidence that steps were taken to evaluate cash flows from disclosed public documents, but investors should not rely on ratings to assess the risk of structured finance deals. Disclosure is not the same as transparency.
Taking these facts into account, what purpose do structured finance ratings serve? This article will analyze the role of rating agencies in structured finance deals. There are a number of important questions to consider including: How do structured finance deals compare with the traditional credit risk rated by rating agencies? What types of approaches do agencies use to rate synthetic collateralized debt obligations? How do the rating agencies differ when it comes to rating super senior tranches? And how does the lack of standardization among rating agencies affect cash structured finance deals?
Generally speaking, a rating is a rough indication of the probability of an investor receiving principal and interest payments on time. The rating agencies sometimes rate only the repayment of principal and receipt of interest payments are sometimes rated separately. In addition, rating agencies sometimes only rate the ultimate probability of an investor getting his or her money back – instead of the timely return of principal and interest.
Rating agencies primarily attempt to assess two types of risk: rating transition and default risk. Conventional general market risk to a portfolio is not captured by ratings. In structured transactions involving interest rate and/or currency risk, rating agencies attempt to assess the cash flow impact of these risks – and their hedges, if applicable. But this is not their core expertise, so investors need to double check this risk on their own.
Even with traditional credit risk rated by the rating agencies, there are interpretation risks. For instance, recovery rate in the event of default is debatable. The major rating agencies compile data on public and rated securities and their recovery rates, but reliable data isn’t always available. For example, in Europe, where recovery rate data is sparse, rating agencies may arbitrarily assign a percentage recovery rate to the corporate obligations of an entire country. It would seem logical that the rating agency ratings can be directly mapped onto one another, but they cannot.
The reason this can’t happen is partly because rating agencies tend to disagree. For instance, Lehman Brothers recently decided to add ratings by Fitch to those of Moody’s and S&P to its corporate bond index, allowing it to keep General Motors in the investment-grade index and avoid a panic among investment grade-funds who otherwise would have had to dump General Motors bonds. S&P is expected to downgrade General Motors to junk status, below the investment grade-rating of BBB, later this year. However, since Moody’s and Fitch will probably maintain their investment-grade ratings, General Motors can stay in the Lehman investment grade bond index for now.
Collateralized debt obligation (CDO) structuring makes the rating agency issues even more interesting. Moody’s and Fitch may supply ratings on cash flow structures for which S&P will refuse to provide a rating, because it doesn’t fit into the S&P framework. For example, S&P will not rate raw equity cash flows, no matter how robust.
One would think that rating agencies would at least be internally consistent. But that isn’t necessarily true. Even within the same rating agency, portfolio tests and restrictions may vary by deal, and some deals are better protected than others. Different structurers within a rating agency may choose different stress scenarios when evaluating cash flows for an ultimate rating.
The rating agencies use different modeling approaches for synthetic CDOs. Each of the rating agencies publishes information about their data and models, and the magnitude of the differences from the varying approaches differs by deal type and collateral.
That said, it is possible to arbitrage the rating agencies. For instance, one might tranche a synthetic corporate deal using S&P methodology because it results in less subordination to get an AAA-rated tranche than Moody’s methodology.
But if you want to find a reliable rating, it’s wise to evaluate data produced by multiple agencies. In 2004, Fitch’s model showed such unreliable results for structurers using Fitch’s frequently changing correlation matrix that they dubbed it the “Fitch Random Ratings Model.”
Some investors therefore prefer deals that are rated by at least two rating agencies. However, many synthetic CDOs have been issued with a rating by only one agency – or even with no rating whatsoever.
Moody’s, S&P, and Fitch suffer from macular degeneration in regard to rating synthetic CDOs with super senior tranches. The rating agencies do not acknowledge the existence of a super senior tranche.
The super senior tranche is a convention created by structurers, and all the rating agencies care about is whether you have solved for the AAA attachment point in a manner consistent with their methodology. That methodology will be the same whether or not a structurer later creates a super senior tranche in the deal.
So far as the rating agencies are concerned, the super senior is a retrofit that has nothing to do with them. Except that it does.
There is a compelling reason why the rating agencies might wish to take a point of view on super seniors in the context of their rating on Aaa (or AAA) tranches. I hear many arguments about rating agency models, but no arguments about AAA tranches.
This focus on models make us miss a huge problem that is right before our eyes: a synthetic portfolio can be tranched two different ways. Suppose, for example, the amount of subordination required to get an AAA rating for a tranche of a synthetic portfolio is 12%, and that the first tranching method uses only an AAA tranche. Using the second method, we solve for the 12% AAA attachment point, and subsequently create a super senior tranche.
The synthetic CDO would then be comprised of two components: a superior senior tranche that comprises 83% of the portfolio (17% subordination), and an AAA that accounts for 5% of the portfolio. Further dissecting this example, the main point is that we could have just as easily created a “correct” super senior tranche with a smaller or even a slightly larger sized AAA tranche provided the combinations of the super senior and AAA tranche totaled 88% of the deal.
Both CDOs have an AAA tranche. Are you indifferent to owning the AAA from the first deal versus the second deal with the super senior tranche? The rating says you should be happy to buy either one. But you shouldn’t be.
The second AAA is a first-loss tranche supporting the super senior tranche. Suppose you were unlucky enough to have bought one of the deals that had two defaults with zero recovery, each 1% of the deal, so that 2% of the total names in the portfolio defaulted. Chances are that other names in the portfolio were downgraded, so that the amount of subordination to maintain an AAA rating would have to increase by more than 2%. But you know for sure, that it would have to increase by at least 2%.
Obviously, then, the two AAA tranches are not equivalent in a non-static world. If everything else remained the same, but 2% of the portfolio defaulted, slightly more than two percent of the first AAA tranche would not be deemed AAA. The AAA of the second CDO presents a different picture, because 40% of the formerly AAA tranche would no longer be deemed AAA. Of course, it would take the rating agencies several months to catch up with this difference – if they caught up at all.
The AAA tranches of static synthetic arbitrage deals often trade close to levels of cash AAA tranches. Many have no added structural protections to benefit the senior noteholders.
The lack of standardization among rating agencies also affects cash deals. Some cash deals allow for payment-in-kind or PIK tranches. Instead of making a coupon payment, the principal amount of the relevant tranche is increased by the amount of the unpaid coupon amount. The PIK tranche can pay either the current coupon calculated from the new principal balance, or can pay-in-kind depending on the cash flow availability in the deal. This mechanism is meant to prevent default of a tranche payment and to avoid triggering an early termination of the deal.
Moody’s uses an expected loss approach to rate these tranches. The expected loss approach uses the probability of loss and the potential severity of loss as the criteria for mapping a rating to a specific tranche. According to these criteria, if the tranche PIKs, but eventually pays all accrued interest and interest on accrued interest (if applicable), there is no loss. The timely return of interest is not an issue for Moody’s for purposes of this rating. Moody’s will allow PIK-ing in all tranches – including the Aaa- rated tranche
S&P, in contrast, typically allows PIK-ing at the BBB level or below. The agency does allow PIK-ing for tranches rated single A – buy only if there is clear disclosure that interest can be deferred. Even so, if a bond rated A-, A or A+, PIKs for more than a year under a stress test, S&P considers it a default. This same event is not considered a default for tranches rated BBB+ or lower, and S&P doesn’t allow PIK-ing for tranches rated higher than A+. Its logic is that investors in higher rated products expect timely payment of interest.
Fitch takes what seems to be the most reasonable approach. Fitch is specific about exactly what its rating means. If a CDO tranche pays timely principal and interest under the relevant default and recovery scenarios, Fitch will rate the tranche for timely payment of principal and interest. If the tranche only accrues coupon and eventually pays it off along with interest on accrued interest, Fitch will rate the tranche for timely payment of principal, and for ultimate payment of interest.
Many investors believe a rating indicates how secure an investment is as a store of wealth. But this is a misconception. In the past, for instance, some fund managers who invested in structured financial products backed by European MBS were unpleasantly surprised to find that tranches with “double A” ratings were bid as if they were “triple B” assets when they tried to sell them.
This rating may have been a fair indication of the probability of the investor receiving all of his or her money back if he or she held the asset to maturity. Unfortunately, the rating also gave no indication of the real-world liquidity concerns of investors. All “double A’s” are not created equal when it comes to liquidating your investment; this is a fundamental weakness of ratings.
From a costs perspective, rating agencies’ charges seem to magically converge partly because fee information is easily leaked and partly because there are only three major rating agencies chasing dollars for structured finance deals. Charges are typically seven basis points with a minimum floor charge for smaller transactions. Fortunately, this initial fee is an upfront charge, and per annum maintenance is much less – at least so far.
Yet for structured financial products, investors are inadequately informed by the rating of any of the three major rating agencies and have to do their own independent evaluations.
Investors in synthetic structured finance products should strongly consider a Moody’s rating, even if it is not required. Investors in unrated deals “investment grade” deals should ask for the Moody’s tranching as the ratings benchmark because it usually provides a more conservative tranching result, meaning more subordination is required.
An efficient market will eventually begin to discount the ratings of rating agencies that consistently require less subordination, so ongoing comparisons of subordination requirements across rating agencies are always a good idea for rated transactions. Prudent investors and risk managers might also consider purchasing rating agency equity to hedge their structured products.
Endnote: Tavakoli Structured Finance, Inc. retains the copyright for this article. It was reprinted with TSF’s permission in GARP Risk Review (Global Association of Risk Professionals) January/February 2005 Issue 22 under the title: “Structured Finance: Rating the Rating Agencies.” In recent years there has been widespread manipulation, fraud, and abuse in the credit derivatives and collateralized debt obligations markets. Many of these abuses are detailed in my 2003 book, Collateralized Debt Obligations and Structured Finance, and the fully updated 2008 second edition, Structured Finance & Collateralized Debt Obligations.
See also: “Tavakoli Structured Finance Revokes the Credit Rating Agencies’ NRSRO Designation,” July 26, 2011.
More finance articles by Janet Tavakoli.