Market concentration
Of the three major rating agencies, Standard & Poor's and Moody's are the most prominent, with S&P being the more quantitative and rule-bound institution and Moody's occupying the other extreme. Two reasons for the strong concentration are the difficulties involved in building up the aforementioned reputational capital and the SEC, which has thwarted the formation of new competitors through various regulatory actions. Although a slow process, things are expected to change through the continued adoption of the Credit Rating Agency Reform Act of 2006.Consistency
Moody's claims to have the same expected loss targets for structured finance securities and corporate bonds. Contrary to such statements, popular media often brings up the fact that only 27 sovereign nations have a AAA-rating, while over 8,000 RMBS hold the same AAA-rating. This argument is slightly misleading, as there are only 100 rated countries and over 96,000 structured finance products. More critically, Charles Calomiris and Joseph Mason (two financial economists) have found that the five year cumulative default rate on corporate bonds rated Baa between 1983 and 2005 was only 2.2%, while the five year cumulative default rate between 1993 and 2005 for CDOs with the same Baa rating was 24%. The numbers have been contested by Moody's, but their argument only concerns the amplitude.Misaligned incentives
One valid complaint is the obvious conflict of interests. Since the early 70s, the rating agencies have commercial relationships with issuers. Up until then ratings were paid by subscribers - an approach suffering from other serious shortcomings, with low if any profitability and free-riding problems being the most critical. Aggravating the problem is the composition and opacity of the rating process for structured finance assets. There is nothing stopping originators from enhancing the rating through continuous negotiations with their rating agency. If still unhappy, otherwise returning clients might shop around until they receive the desired rating. Moody's reported that after the July downgrades, its market share in mortgage-backed securitisations dropped from 75% to 25%, which is quite noteworthy considering the share of profits derived from structured finance.As in many other areas, these conflicts of interests have their roots in misaligned incentives: the upside of timely downgrading is limited to achieving hard-to-value goodwill in the long run, while the potential downside includes upsetting important issuers, investment banks and large institutional investors. In fact, there is no substantial incentive for any participant to choose objective information over sanguinely partial assessment. The same asymmetry afflicts the much discussed staleness of ratings: the downside to overreacting on something that could be short term noise widely outweighs the benefits of being right a few months earlier than your competitors.
Modeling tribulations
A somewhat more technical topic is the problems related to modeling securitised mortgages. The most obvious difficulty concerns statistics - there is no trustworthy data for subprime mortgage securities. Some pundits claim that the agencies have calibrated their models on prime mortgage data and then simply extrapolated these results to subprime. The cynical reader might conclude that Moody's talk about "a combination of building models and using algorithmic expertise, utilising both quantitative and qualitative analysis" is an innovative synonym for guessing.A fundamental flaw in the rating agencies' models was the assumption regarding delinquency correlation. Historically, 100% LTV ratios and no-documentation loans have been extreme outliers. As we all know, this changed considerably from 2006 and onwards (close to 50% of the subprime loans made in 2006 were low or no documentation loans), leading to a severe decoupling from traditional correlation patterns. Put differently, fraud was not sufficiently incorporated in the models. Another thing missing was some sort of model for the clear link between increased issuance and exploding default rates (evident from several other asset classes).
Taking the other side
In defense of the rating agencies, the loss estimates of Moody's from 2004 to 2006 increased by approximately 30% (this occurred during relentless spread tightening). Regarding the delay in rating changes, it can be said that it takes time to see how collateral pools of mortgages are performing. Ratings are supposed to address credit, not marketability and a majority of current events is not overly related to credit, but rather driven by liquidity and valuation.Going forward
Some observers have discussed the probability of litigation. Mimicking rating agencies by looking at history, one finds that to date, no rating agency has been held liable for any questionable action. The main obstacle to such an outcome is that ratings allegedly constitute First Amendment protected speech. One might also note that the quantity of rated securities numbers in the thousands and even a small wave of litigations could mean the elimination of one or more of the three major players, something which could potentially disrupt financial markets for a long time.As the mortgage securitisation process has played an ever-growing role in providing housing opportunities, its relevance to society in general has led to calls for increased regulation. Apart from the free market counter-argument, it is interesting to note that SEC has already looked twice at the transparency of the mortgage securitisation business and deemed it satisfying. A better way to prevent repeated crises is to punish speculating investors/borrowers for taking supposedly calculated risks.
An anonymous bond trader points out that substantial enhancement could most easily be made in the performance modeling of loans. Just allowing dynamic correlation and recovery rates would improve the results and predictive power significantly. Incorporating information from other assets should also be able to increase the accuracy of the modeled default rate.
To conclude, there is room for improvement in the ratings business, but close to impossible to legally prove any wrongdoing.
