Rewarding Failure

The collapse in 2001 of Enron's long-running pyramid scheme was, at heart, an accounting scandal. Enron's own accountants were captive to corruption at the top, and external auditors -- whose whole raison d'etre is to produce a true record of financial activity without relying on the honesty of the in-house staff -- failed to detect any problem. Accountancy, as a whole, could not deliver what people expected of it.

The legislative response to Enron's fraud was driven by an irresistible imperative to take some action, and constrained by refusal to reevaluate the premises of the existing system. The result, predictably enough, was to double down on a failure: Congress, especially with the Sarbanes-Oxley act, has sought to address this problem by massively increasing the power of accountants and auditors, and increasing each corporation's reliance on them.

Consider the 2007 credit market turmoil in this light. This problem is, at heart, a rating-agency failure; it is worth understanding this in some detail.

Ratings are important to most institutional investors because of regulatory restrictions on ownership: mutual funds, for example, cannot own "junk" (BB/Ba or lower rated) bonds. But investors tend to seek the highest yield, or equivalently, the lowest price for a bond paying a given coupon. The business of structured credit finance, in the presence of these regulation-induced incentives, becomes a search for products which are attractive to the rating agencies but unattractive to buyers, thus combining high ratings with high yields. Investment banks create such products by bundling low-priced bonds together, then selling claims on part of the resulting bundle. This can degenerate into a contest to best game the rating agencies' models; the existence in early 2006 of AAA-rated paper paying 500 basis points above Libor indicates that such gaming, if it was undertaken, succeeded.

The rating-agency model has proved inadequate for credit risk management, and it cannot be rescued. If ratings agencies are involved in trading, they cannot retain the trust of investors or of regulators; without trading revenue, they cannot keep pace with the modelling capabilities of investment banks or of hedge funds.

Now we turn to the government response. Again impelled to do something, and again unwilling to reexamine their premises, our government is preparing "reforms" which will increase the power of the rating agencies, and the financial system's dependence on them. The government is preparing to force the banks to bail out the monoline insurers, so that debt of questionable quality but insured can continue to be rated AAA, and thus can continue to be used to meet government-imposed regulatory capital requirements. This will still further entrench the ratings-driven system, just when it has manifestly failed.

[Cross-posted to Chequer-Board.]