This statement is also available here as an Adobe PDF.

Statement No. 257
February 11, 2008

For Information Contact:
Marshall Blume
215-898-7633
Kenneth Scott
650-723-3070
Chester Spatt
412-268-8834

As a result of the subprime mortgage turmoil there has been considerable focus in the last year on the role of credit rating agencies in determining the risk or quality ratings on mortgage-backed securities. Many investors rely almost exclusively upon the investment ratings of third-party rating agencies, thereby effectively outsourcing their due diligence. All of the major agencies use relatively similar risk models and criteria. This limits the extent of independent information and analysis in assigning a quality rating. Furthermore, research evidence suggests that ratings changes lag changes in the underlying risk. The use of common models is a key source of systemic risk as they are likely to err in the same direction. Indeed, critics of the work of the credit rating agencies suggest that the agencies misevaluated almost the entire asset class of structured mortgage products.

Another important component of the problem is the rating agencies' lack of a financial stake in their ratings, potentially limiting the reliability of their assessments. Of course, lack of "skin in the game" is not necessarily inherent in due diligence. For example, the underwriters of securities offerings and the accounting group auditing the firm may incur liability should their representations turn out to be negligently inaccurate. Analogously, it would be sensible to consider the possibility of imposing a similar contingent liability on credit rating services. The shortcomings of the rating services in evaluating mortgage-backed securities is illustrated recently by both the numerous downgrades of AAA structured products once the market began questioning the ratings and announced changes in ratings methodology by the major rating agencies.

Because various aspects of investor preferences and the regulatory process require that banks and some other institutions invest in investment quality securities (the top four ratings), the regulatory process created substantial demand for high ratings. In this regard, issuers were able to issue large quantities of highly-rated securities as the buyers obtained a false sense of security and reliability in the quality of these securities. This misplaced reliance on ratings backfired as the past history upon which the ratings were based did not adequately reflect the disruption of a major housing downturn.

The combination of "ratings shopping" and the input of the ratings framework into the structuring of these mortgage-backed products also influence the validity of ratings. An issuer's decision to acquire a rating reflects a conscious choice by issuers of which ratings to acquire (e.g., the originator of a structure would presumably select a firm that it anticipated would offer a higher rating rather than a lower one). Furthermore, the ratings do not have the traditional interpretation, because those defining the structure can endogenously adapt the structure of their offerings to reflect just the minimum requirement being satisfied. This is inherent in a system in which there is pooling of a range of types to a small number of distinct ratings. In addition, rating agencies generate a large portion of their income from consultation services to issuers helping them to tailor their offering to acquire the rating they seek.

Many of these issues arise in a variety of other forms. For example, the FDIC partially ties its risk-based insurance premium to credit ratings, Basel II risk capital standards are linked to credit ratings, and the standard credit scoring model played an important role in the sub-prime mortgage turmoil. These examples illustrate the dangers of reliance upon regulatory mandates that focus on outside certification rather than on a process of market competition and pricing.