Berating agencies

Rating agencies are under attack. A sharp rise in delinquencies in US subprime mortgage loans has forced them to embark on a plethora of rating actions since July. Hundreds of residential mortgage-backed securities (RMBSs) and collateralised debt obligations of asset-backed securities (CDOs of ABSs) - including some AAA-rated tranches - have been downgraded so far, with more set to follow. The actions have severely shaken confidence in the reliability of credit ratings, and have caused regulators to take a close look at the role of rating agencies.

"Certainly, we've been concerned by the apparent slowness of the credit rating agencies to react to the deterioration of the US subprime mortgage market," says an official at the European Commission in Brussels.

On August 16, the Commission announced it was planning an intensive review of rating agencies in response to widespread questions about the accuracy of credit ratings. In particular, it will focus on the timing of recent subprime RMBS rating actions, given evidence of a deterioration in the subprime market since mid-2006, and will investigate the resources available to rating agencies, their governance and management of potential conflicts of interest. This last issue is periodically raised by critics, who note that rating agencies earn their income from investment banks, who pay for ratings on the CDO tranches they structure. The Commission expects its review to be completed as part of a Committee of European Securities Regulators (CESR) report on the rating industry by April 2008.

Another battle front has also opened up across the Atlantic. The US House Financial Services Committee, which oversees the financial services industry in the US, commenced hearings on the role of rating agencies in September. Its chairman, Democratic congressman Barney Frank of Massachusetts, has previously criticised the agencies' reactions to the US subprime mortgage rout. "We want to see what, if anything, can be addressed," says a spokesman for the committee. A panel of the committee was set to examine the role of credit ratings in the structured finance market during late September.

In response to the announcements, Fitch Ratings and Moody's Investors Service stressed they remained committed to a continuing and constructive dialogue with regulators. Standard & Poor's (S&P) said it would participate in the CESR review, but would not comment further.

Since July, the main rating agencies have been forced to downgrade hundreds of first- and second-lien US subprime RMBSs and review their ratings on hundreds of CDO of ABS tranches. By August 21, Moody's had downgraded 417 first- and 721 second-lien US subprime RMBSs dating from 2005 and 2006, and had 43 first- and 58 second-lien US subprime RMBSs from 2005 and 2006 on watch. The agency also had 203 synthetic and cash CDOs of ABS tranches - including nine with Aaa ratings - on watch for a possible downgrade.

S&P, meanwhile, had downgraded 612 first- and 418 second-lien US subprime RMBSs and, by August 21, had downgraded 171 tranches spanning 95 CDOs of ABSs, including 18 tranches previously rated AAA. By August 21, it continued to have 26 first- and second-lien US subprime RMBSs on watch, as well as 137 tranches from 45 CDOs of ABSs.

Agencies under fire

This slew of rating actions has left the agencies fending off broadsides from banks, investors and regulators. It is not the first time they have endured heavy fire. In the wake of the bankruptcies of Texas-based energy firm Enron and Mississippi-based telecoms provider WorldCom, rating agencies were criticised for failing to downgrade the companies' bonds quickly enough.

And, earlier this year, Moody's was forced to back-pedal after incorporating its new joint default analysis - which takes into account how likely it is that governments will prevent their troubled banks from defaulting - into its rating methodology. The approach, which led to a number of Icelandic banks being upgraded to Aaa, was swiftly revised in March after a torrent of criticism from dealers and analysts (Risk April 2007, pages 66-68).

But is the current criticism of rating agencies entirely fair? Given the losses racked up by dealers and hedge funds, they are clearly not the only ones to have had their models severely tested by the poor underwriting standards employed by originators in the US subprime mortgage market. The agencies are having to adjust their assumptions to deal with the poor quality of data received on loans dating from late 2005 and 2006, in addition to losses on US subprime RMBSs that will significantly exceed historical precedent. On July 11, S&P put total aggregate losses on all subprime transactions issued since the fourth quarter of 2005 at 29 basis points, compared with 7bp for similar transactions issued in 2000 - previously the worst-performing year of the past decade.

"No matter how clever they are, no matter how well they understand the structures, they can only base their analysis on historical precedent. To be fair, how can you ask them to do more than that?" says Terri Duhon, London-based managing partner at structured finance consultancy B&B Structured Finance.

Nonetheless, critics claim the wider spreads available on AAA-rated CDO of ABS tranches versus AAA-rated corporate bonds prove ratings on structured finance transactions are unreliable. In other words, the fact the market is willing to pay more for a CDO of ABS tranche than an equivalently rated corporate bond proves there is more risk in the former. Indeed, even before the shakeout in the US subprime mortgage market, AAA-rated CDO of ABS tranches offered average spreads of roughly 45bp - more than 10 times the average spread on AAA-rated US corporate bonds, at around 4bp.

"It's the free lunch thing. If you're seemingly being offered some free basis points, you've got to stop to consider why that is. The excess spread is often in exchange for a bit more risk or uncertainty," says Michael Hampden-Turner, European head of CDO research at Citi in London.

CDO tranches

Hampden-Turner recently looked at the ratings volatility of BBB and AAA-rated CDO tranches versus plain vanilla corporate credit. Using data from all of S&P's publicly rated deals from 1985-2006, the weighted average annual rating transitions were compared for tranches of CDOs and corporate credit carrying the same rating.

In general, the findings show that CDO tranches are more prone to severe movements, while being less prone to smaller ones. "In normal market conditions, CDOs and structured products tend to be more stable than bonds and other types of fixed income. But, in very extreme market conditions, they tend to deteriorate faster," says Hampden-Turner.

The findings show that BBB-rated corporate bonds are more than twice as likely to get upgraded by one notch to A as a CDO tranche carrying the same rating (see figure 1). BBB-rated corporate bonds are also 1.65 times as likely as a BBB-rated CDO tranche to get downgraded by one notch to BB. However, BBB-rated CDO tranches have a greater tendency towards larger movements along the ratings scale, with the data indicating they are 1.25 times as likely as BBB-rated corporate bonds to get downgraded by two notches to B, and 3.5 times as likely as BBB-rated corporate bonds to get downgraded three notches to CCC.

"In distribution terms, the CDO tranches have fatter tails. The probability of a small loss is smaller and the probability of a big movement is greater," asserts Hampden-Turner.

The results for AAA-rated tranches of CDOs and bonds were less conclusive (see figure 2). During the period 1985-2006, AAA-rated tranches of CDOs have ostensibly shown greater ratings stability than bonds carrying the same rating. According to the findings, AAA-rated CDO tranches have a higher propensity to retain their AAA rating than AAA-rated vanilla bonds, by 6.18 percentage points. AAA-rated corporate credit is 4.84 times as likely as an AAA-rated CDO tranche to get downgraded by one notch to AA, while being 1.43 times as likely to get downgraded by two notches to A.

Hampden-Turner says the murkier results for AAA-rated tranches reflect the fact that, over the course of the period examined, there has not been a market move large enough to shake AAA-rated CDO tranches. "The theory remains that the AAA-rated bonds and tranches should share the same characteristics as the BBBs," he says. "If you look at the size of the tail, the pattern is almost the same. The chance of a small move is reduced, but if you look at the BBB column in figure 2, the chance of moving from AAA to BBB is actually slightly higher for CDOs."

Indeed, a close look at the results show that AAA-rated CDO tranches indicate an increased probability of getting downgraded by three notches to BBB, at 0.02% versus 0.01% for AAA-rated bonds.

But there are reasons other than ratings volatility as to why the spreads on CDO tranches are wider than bonds with the same rating. For instance, CDO structurers are able to extract value from the correlation of the underlying assets, while part of the additional spread could be attributed to a liquidity premium.

Rating agencies are quick to point out that their analysis is based on the statistical analysis of historical loss rates and is not intended to assess the level of market liquidity in a particular instrument. "We're finding that there's a lot of misunderstanding about what a rating addresses and what it's meant to address. Our rating covers the expected loss on these securities," says Yuri Yoshizawa, New York-based managing director of the US derivatives group at Moody's.

For Moody's, expected loss means the probability of default and the expected loss when default occurs. It does not take into account factors such as liquidity, mark-to-market volatility nor the possibility that a rating itself may change. So, while AAA-rated tranches of CDOs of ABSs might share a rating with AAA-rated corporate bonds, they may well display a higher level of ratings volatility.

"People believe that if something is investment grade it should always be investment grade, and that's not necessarily what our rating addresses. It addresses what the expected loss is - and that expected loss could change over time," continues Yoshizawa.

While ratings are a useful tool for gauging default probabilities and expected losses, they are no substitute for hard-nosed fundamental credit analysis and an assessment of the other risks involved. In other words, investors should not be basing their investment decisions in CDOs on ratings. "To the extent (investors) are using them to measure things other than expected loss, they are using them incorrectly," Yoshizawa adds.

Nonetheless, there is anecdotal evidence that plenty of investors are buying CDO tranches based on ratings - and that is worrying regulators. Even for those investors that have conducted the underlying credit analysis, the downgrade of investment-grade CDO of ABS tranches could cause serious pain. Many institutional investors are required by internal mandates to hold investment-grade assets, with some required to hold AAA only. So far, many of these investors have held off selling AAA-rated CDO tranches. But, as more of these assets are downgraded, institutions will be forced by internal mandates to sell - at a time when a complete absence of liquidity in the secondary market means they will have to offload at fire-sale prices.

Given the serious implications of any downgrades, critics brush off arguments that the ratings are based on statistical analysis and argue that the rating agencies should have taken notice of the loose underwriting standards of US subprime mortgage lenders - clearly signposted since mid-2006 - and been more conservative in their rating of CDOs of ABSs backed by subprime loans.

However, Yoshizawa believes that Moody's original analysis was sound. "These securitisations are living, breathing animals. I don't think it's a matter of misjudgement at the beginning - it's a matter of keeping track of these deals and tracking which of the scenarios we've looked at is starting to take shape."

Regardless of the rights and wrongs of assigning AAA ratings to these instruments, recent events have undoubtedly shaken confidence in the reliability of ratings. And, with Basel II set for full implementation in Europe and parts of the Middle East and Asia from next January, this is a real concern for some regulators and bankers. "Basel II is fundamental to the issue. A lot of the assumptions are based on ratings, so you have to make sure the ratings make sense," says one senior London-based banker.

Risk weighting

Under the most straightforward standardised approach, risk weightings are determined by ratings from external credit agencies. In theory, the issue of whether a bank using the standardised approach has sufficient regulatory capital will therefore be largely determined by the accuracy of ratings.

In reality, it is slightly more complex than that. Under Pillar II (the part of the Basel Accord covering supervisory review), national regulators have the power to insist banks hold more regulatory capital. And the UK Financial Services Authority (FSA) has repeatedly said it will force banks to hold additional capital if it feels it is necessary. "Referenced credit ratings are one way, but are not the only way to measure risk under Basel II. We would be quite concerned if banks were only to use a referenced credit rating for the basis of its decisions," says Thomas Huertas, director of wholesale firms at the FSA.

Pillar II also deals with various risks outside of credit, including liquidity risk, interest rate risk in the banking book and legal risk. And Huertas says the FSA will be mindful of liquidity risk in assessing whether banks are adequately capitalised.

Nonetheless, the ratings on illiquid, structured finance transactions, and the question of whether these ratings can be used as a basis for determining the amount of capital a bank should hold against those assets, is worrying some supervisors.

"It's definitely something that supervisors are talking about," says Stefan Walter, secretary-general of the Basel Committee on Banking Supervision. "There is room for more dialogue among supervisors and the industry around how to come up with sound valuations for structured, potentially less liquid, products."

In the meantime, many in the market believe the recent rating actions, regulators' probes - and the hammering suffered by rating agencies in the media - will cause investors to place less faith in credit ratings.

"I think more people will start to focus on mark-to-market volatility, ratings volatility and this sort of measure rather than just considering expected loss," says Hampden-Turner at Citi.

One thing is for sure - the problems in the subprime market aren't likely to end anytime soon. "Fundamental corrections in structured credit pools, such as mortgages, evolve over relatively long horizons. This is very different from corporate credit, where bankruptcy allows for a clean, swift break," says Anthony Thompson, head of US ABS and CDO research at Deutsche Bank in New York. These continuing woes will mean more forthcoming downgrades of US subprime RMBSs, as well as tranches of CDOs of ABSs. For the rating agencies, this could prove a lasting headache.