Both the Dodd-Frank Act and the new Financial CHOICE Act have sought to address the causes of the 2008 financial crisis, but neither measure deals adequately with one set of Great Recession culprits: the “big three” credit rating agencies.

While those agencies -- Moody’s, Fitch, and Standard & Poors -- were widely criticized for their role in the financial crisis, the credit rating business has only changed marginally since the enactment of Dodd-Frank. Unfortunately, the current version of the Financial CHOICE Act eases some rating agency regulations mandated by Dodd-Frank without addressing the fundamental question: whether the federal government should be licensing and regulating rating agencies in the first place.

The Origins of the “Big Three”

As with many regulatory frameworks, today’s rating agency governance regime is the unintended consequence of decades-old policy choices. Today’s “big three” rating agencies originated as private publishers in the early 20^th Century. In 1909, John Moody began publishing manuals containing information about railroads and the securities they issued. The manuals also included rating symbols for each security. Moody based his symbols on those used by credit reporting agencies R.G. Dun & Company and the J.M. Bradstreet Company, which had been reporting on the creditworthiness of merchants since the mid-19^th Century.

Demand for Moody’s railroad manuals was strong, and he soon added securities guides for industrial, utility and government bonds. Competitors also began marketing their own manuals: by the late 1920s, four firms were publishing bond ratings (two of which, Standard Statistics and Poor’s Publishing later merged to form a single company).

Ratings in Government Regulation

Ratings first attracted regulatory attention during the Great Depression. In 1931, Comptroller of the Currency J. W. Pole announced that Nationally Chartered Banks would not have to write down bonds in their portfolios if they were rated Baa/BBB or above. Banking regulators in several states followed this approach, while the Federal Reserve made more informal use of rating agency manuals.

In August 1935, President Roosevelt signed a Banking Act that made federal deposit insurance permanent and limited the liability of bank owners and managers. To guard against resulting moral hazards, the new law restricted banks from buying securities that were not investment grade, as defined by the Comptroller of the Currency. The following year, Comptroller J.F.T. O’Connor furnished a definition that once again relied on rating agency symbols. The precise regulatory language (including an all-important footnote) was as follows:

By virtue of the authority vested in the Comptroller of the Currency by . . . Paragraph Seventh of Section 5136 of the Revised Statutes, the following regulation is promulgated as to further limitations and restrictions on the purchase and sale of investment securities for the bank’s own account, supplemental to the specific limitations and restrictions of the statute. . . . The purchase of “investment securities” in which the investment characteristics are distinctly and predominantly speculative, or “investment securities” of a lower designated standard than those which are distinctly and predominantly speculative is prohibited.*

*The terms employed herein may be found in recognized rating manuals, and where there is doubt as to the eligibility of a security for purchase, such eligibility must be supported by not less than two rating manuals.

By defining “investment securities” so, the federal government implicitly delegated an important regulatory function to a group of relatively small firms.

Although few of the consequences of the Comptroller’s language could have been predicted at the time, that language was still met with some contemporary objections. In a January 1938 journal article, Melchior Palyi sharply criticized rating agency performance, reporting that 70% of defaulting railroad bonds in 1924 carried investment grade ratings from Moody’s. With respect to the company’s 1937 bond manuals, Palyi concluded:

The staggering cost of detailed study of some 23,000 issuing units, or even of the almost 9,000 rated issues, is prohibitive. Accordingly, the responsible agencies advise the customer not to rely upon the ratings alone but to use them together with the text of the manual and even to buy special investment advisory services which they are ready to supply. The candid observer cannot help wondering whether it would not be a still more responsible attitude to stop the publication of ratings altogether in the best interest of all concerned. (emphasis added)

Palyi’s criticism of OCC reliance on credit ratings was equally sharp:

The meaning of the ratings device, as enforced, is the provision of an "objective" standard of discrimination between good and bad investments… The use of ratings, however, merely shifts the burden of the problem. Instead of providing objective standards, the ratings introduce a new arbitrary factor, namely, the unknown and undefined philosophy of the rating agencies.

Regulatory Reliance on Ratings and the Decline of Innovation

Despite such concerns, regulatory reliance on credit ratings did not result in any crises for several decades, perhaps owing to the benign credit conditions that prevailed in the post-World War II economy. Regulators thus felt comfortable increasing their use of ratings. In 1975, the SEC adopted Rule 15c3-1, which established net capital requirements for broker-dealers. The capital calculation required a haircut for non-investment grade assets held by these entities, and relied on credit ratings to determine which investments were non-investment grade. At the same time, the SEC granted the newly-created status of “Nationally Recognized Statistical Rating Organization (NRSRO)” to Moody’s, S&P and Fitch. From that point on, dozens of new laws and regulations incorporated references to NRSROs.

The big three cashed-in on their new officially conferred powers by migrating to the “issuer pays” model. Instead of charging investors for ratings, they instead began charging bond issuers -- thereby introducing an incentive toward bias. Meanwhile, the incumbent agencies failed to keep up with evolving best practices in the credit assessment space, most notably the use of default probability models such as Edward Altman’s Z score (1968) and Robert Merton’s option-based corporate bankruptcy model (1974). Rather than leverage computer technologies to calculate quantitative outputs -- like default probability and expected loss -- they continued to use the same imprecisely-defined letter grades from the 19^th century. And instead of adopting advanced modeling techniques like simulation and logistic regression, agencies continued to rely on an outmoded credit committee process under which rating actions were verbally debated by credit analysts.

Meanwhile, emerging non-NRSRO credit assessment providers employed newer techniques and generally marketed their products to investors. New entrants often struggled to gain penetration given the availability of free ratings from the NRSROs; while those that succeeded were often purchased by NRSROs, thereby eliminating them as a competitive threat. For example, Moody’s purchased KMV -- a firm that successfully commercialized Merton’s corporate bankruptcy model -- for $210 million in 2002.

Still, several independent non-NRSRO firms provide credit scores or tools that enable clients to calculate them. For example, Kamakura Corporation calculates estimated default probabilities on several thousand public firms each day using market capitalization and other inputs. Rapid Ratings International uses financial statement data to score public and private companies on a 0-100 scale. Non-NRSRO evaluators and analytics providers in the structured finance sector include Intex Solutions, Trepp, Credit Spectrum, and PF2 Securities. Municipal bond market providers include Lumesis and Munitrend.

When rating agencies ultimately adopted credit modeling for structured finance securities in the late 1990s and early 2000s, they did not use state of the art methods and employed biased modeling assumptions that resulted in higher-than-appropriate ratings.

Consequences of Mis-Rating

Although the inadequacies of credit ratings had become well-recognized within the financial industry by the 1990s, the issue first gained widespread attention in 2001 when rating agencies failed to downgrade Enron until just before its bankruptcy. Congress responded with the Credit Rating Reform Act of 2006. But the additional SEC oversight of NRSROs mandated by this Act failed to prevent rating agency errors in assessing RMBS (residential mortgage-backed securities), CDOs (collateralized debt obligations), SIVs (structured investment vehicles), and municipal bond insurance during the run-up to the financial crisis.

The systematic mis-rating of these investment vehicles resulted in widespread asset mispricing. When the actual risk of the mispriced credit assets became apparent, the repricing process produced shocks to the financial system that destabilized the overall economy.

A Better Alternative

The 2010 Dodd-Frank Act mandated the removal of NRSRO ratings from regulations while further increasing SEC oversight of the agencies. But if rating agency symbols no longer have financial regulatory implications, what need is there for continued oversight? Can we not return to the pre-Depression business model in which ratings were available for sale to investors, who could then choose to use or ignore them?

While financial markets are admittedly much more complicated now than they were in the 1920s, the amount of research available to bond investors is also much greater and -- thanks to the internet -- much more easily accessible. Further, as NYU’s Lawrence J. White reminds us, institutional investors have developed or can develop the in-house expertise needed to assess bond offerings, and thus should not require federal protection from bad credit ratings.

As of today, the big three credit rating agencies have maintained high market share due to delays in the implementation of regulatory changes, inertia, and the high cost of obtaining and keeping NRSRO status. Even though NRSRO ratings are supposed to be written-out of regulations, NRSRO status continues to place a federal imprimatur on credit assessments that are of questionable value. A better alternative would be for the government to completely deregulate the credit assessment industry and allow both legacy NRSROs and new entrants to compete on a truly level playing field.