Credit Rating Firms

Posted by Avinash Persaud on 26 March 2009

From Avinash Persaud, Jakob Vestergaard & Jean-Louis Warnholz
On April 2, expect G20 policy makers to let loose the dogs of war on credit rating firms. Back in the summer of 2007, the collapse of confidence in the credit ratings of the once $1trn asset-backed commercial paper market triggered the global financial crisis. Separately, investors lament that they were lured into dodgy assets by credit ratings that were upwardly biased by the conflicted business model of the rating firms. Even without lofty bonuses and Lear jets, rating firms have become everyone’s favourate punching bag; and policy makers are under pressure to do some punching. Unfortunately, the solutions to be offered on April 2 will do no good and could cause harm. There are more worthy proposals to consider.
Nobel Laureate George Akerlof’s tale of the market for lemons aptly describes the credit markets. The owner of a second hand car knows whether it is good or bad; whether gizmos live in the electrics or the gearbox is temperamental. The potential buyer does not. Buyers try to hedge their bets. They offer prices between a good second hand car and a “lemon”. Sellers of good second hand cars pull them from the market because they know they are under-priced. Buyers are left with the “lemons” and the market develops a reputation for being untrustworthy. Buyers and sellers stay away. To deepen markets, second hand car sellers have introduced “approved used car schemes” and borrowers have invested in credit ratings. Without a rating, Ghana would not have been able to launch its first international bond issue in 2006.
No surprise then that the loss of credibility in ratings has had dire consequences. To restore confidence in credit ratings there must be such adverse consequences for poor credit ratings that it spurs innovation in credit research that in turn leads to more accurate ratings. Some hope that this will be achieved by switching the business model from ratings being paid by borrowers to investors. This seductively simple idea is flawed. In today’s, information-free, equal-disclosure world, the value of a rating is that everyone knows it. But if everybody already knows it they will not pay for it. This is the tragedy of the commons.
EU Commissioner McCreevy has suggested that rating firms must “disclose their models, methodologies and assumptions”. This seemingly innocuous proposal does not create consequences to spur innovation in credit research. If anything it discourages innovation. In the experience of one of the authors who was a research manager for 20 years, if changing the forecasting framework is a big deal then it will not get changed until it is too late. Yet rating firms need to be quick to adapt to changing circumstances, such as the quick evolution of the sub-prime market from being under-banked to over-banked. Government oversight of rating firms could also introduce a fresh conflict of interest as they are the largest single borrowers, and their credit quality is deteriorating.
We have tried disclosure of methodology before with unhappy results. Since 2004, US rules requiring disclosure of rating methodologies helped banks arrange credit structures so as to maximize their credit rating. But this destroyed the statistical independence that underpinned the ratings and made the breakdown of structured finance ratings inevitable. While the issuer pays business model is common across all ratings, rating failures are concentrated on structured finance. According to Standard & Poor’s, the likelihood that a structured finance product held on to a “BBB” rating throughout 2008 was a desultory 58%. The likelihood that a single-issue borrower  - where it is almost impossible to build to rating –  held on to a BBB rating last year, a year of recession, was an impressive 88%. Detailed disclosure of methodology can lead to a “build to rating” behaviour that undermines ratings quality.
Governments must instead require standardized rating definitions so there can be better comparison between firms and no investor can claim to be rating-confused. However, improving the transparency of ratings may not deepen the consequences of rating-failure. Many investment rules require investors to use all the major rating firms, neutering market discipline. Governments may need to respond to this market structure problem by raising the fear of ratings failure. The FSA’s Hector Sants has already said that he wants to be less soft and cuddly to the touch and more fearsome.
The trick is to devise a system that does not incentivise firms to become overly conservative – developing countries and small companies already feel their ratings are too low. A symmetrical measure of ratings performance is a Gini-coefficient, which measures the ordering of defaults relative to the order of ratings. A high co-efficient would indicate that if you ordered defaults by their ratings, the highest rating would have the lowest defaults and the lowest rating the highest. In 2006, the Gini-coefficient of defaults in instruments rated by Standard & Poors ratings was a near perfect 90%. In 2007 this remained high in sovereign and corporate credits, but slumped to 73% in structured finance. Firms with low Gini-coefficients could pay a penalty and firms with high co-efficients could receive a subsidy. A results-based, not process-based, intervention creates consequences while keeping governments out of the ratings kitchen.   
Professor Avinash Persaud is Chairman, Intelligence Capital Limited and Emeritus Professor of Gresham College. Dr. Jakob Vestergaard is a project researcher at the Danish Institute of International Studies and Dr. Jean-Louis Warnholz is an economic researcher at  Oxford University.