In this Article, we evaluate the viability of credit default swap (CDS) spreads as substitutes for credit ratings. We focus on CDS spreads based on the obligations of financial institutions, particularly fifteen large financial institutions that were prominently involved in the recent financial crisis. Our data from 2006 through 2009 show that CDS spreads incorporate new information about as quickly as equity prices and significantly more quickly than credit ratings. Although CDS spreads did not identify accumulating risk exposures before 2007, they quickly reflected disclosures and developments beginning in the summer of 2007 at the latest. Thus, CDS spreads are a promising marketbased tool for regulatory and private purposes, and they may serve as a viable substitute for credit ratings.
INTRODUCTION I. THE PROBLEM OF CREDIT RATINGS II. CDS SPREADS VERSUS CREDIT RATINGS A. Aggregate Analysis B. Firm-Level Analysis C. Event-Specific Analysis III. CDS SPREADS VERSUS EQUITY PRICES CONCLUSION APPENDIX INTRODUCTION
In response to the recent financial crisis, commentators have criticized certain credit rating agencies, known as Nationally Recognized Statistical Rating Organizations (NRSROs), (1) and credit default swaps (CDSs). (2) Regulators have proposed a range of reforms in both areas but have not considered the links between them. This Article considers one potential link: whether CDS markets could play a role in financial reform related to NRSROs.
More specifically, one set of proposals suggests that legislators and regulators reduce the dysfunctional incentives associated with overreliance on NRSRO credit ratings. (3) Although there is support for this proposal in theory, (4) one objection forestalling reform has been that, notwithstanding the problems associated with credit rating agencies, there do not appear to be viable substitutes for credit ratings. Indeed, ongoing skepticism and criticism about CDS markets has reinforced the objection. In simple terms, the objection is that regulators and investors should not replace one broken system (credit ratings) with another broken system (CDSs). (5)
In a typical CDS transaction, one counterparty (the buyer of protection) agrees to pay a periodic premium to the other counterparty (the seller of protection). (6) In return, the seller of protection agrees to compensate the buyer of protection if a reference entity specified in the CDS contract experiences a default or similar "credit event." (7) For simple CDSs, the reference entity might be a corporation or government entity. (8) For more complex CDSs, the reference entity might be a portfolio of structured financial instruments. Parties usually document the various CDS terms through a standard form agreement created by the International Swaps and Derivatives Association (ISDA). (9)
One of the most important terms in a CDS agreement is the definition of a "credit event," which has become largely standardized. The most common credit event is the failure to pay by the reference entity. Bankruptcy or restructuring credit events can vary depending on how much interest reduction or maturity extension the parties wish to specify in the CDS agreement. (10)
CDS "prices," as measured in the market, represent the size of the premium paid by the buyer of protection and are generally known as CDS "spreads." CDS spreads change over time based on supply and demand for particular CDS contracts. CDS spreads are analogous to insurance premiums and similarly reflect market participants' assessment of the risk of a default or credit event associated with the underlying obligation.
In general, CDSs are widely and deeply traded, and they help to reflect market information about the credit risk of underlying financial obligations. Several studies have shown that CDS markets generally reflect valuable information. (11) Broad market participation suggests that CDS prices should convey information about counterparties' assessments of this risk. Notwithstanding the evidence that CDS markets generally reflect valuation information, regulators and market participants have resisted moving away from NRSRO credit ratings toward CDS spreads.
In this Article, we present recent evidence that CDSs based on financial institutions' obligations are potentially useful for regulatory purposes and private investors. Overall, the data show that changes in CDS spreads reflect information more promptly than changes in credit ratings, even during a period of intense market discord. CDS spreads increased during 2007 and 2008 as information became available showing that the probability of defaults by financial institutions was increasing. During this same period, credit ratings nevertheless remained relatively unchanged. We explore the implications of this evidence for the debate about whether markets or institutions are better for regulatory purposes. We argue that CDS spreads are a viable alternative to credit ratings in reflecting information because of their market-based nature. In other words, markets (CDSs) responded to new information more promptly and responsively than institutions (credit rating agencies) did.
In Part I, we summarize some of the problems associated with reliance on credit ratings and discuss the need for a viable substitute. Part II presents our analysis of our CDS-spread data. We examine the relationship between CDS spreads and ratings on an aggregate basis and provide firm-specific examples showing the utility of CDS spreads relative to credit ratings. We also investigate how the CDS spreads for selected firms in the sample respond to specific, high-profile events. In Part III, we test the efficiency of CDS spreads compared to equity prices and find that the information in CDS spreads, like the information in equity prices, is timely and accurate. In other words, CDS markets, though not perfect, were generally efficient before and during the recent financial crisis. We conclude by suggesting some promising areas of future research and discussing the implications of these findings for monitoring purposes.
THE PROBLEM OF CREDIT RATINGS
In theory, credit ratings are a potentially valuable source of information. During the early 1900s, John Moody intuited that investors would pay him if he could synthesize the complex data in various reports on the railroad industry into a single letter rating. (12) By the 1920s, Moody and his competitors were rating most new bond issues, including government bonds. These private companies, which came to be known as rating "agencies," acted as information intermediaries, and their letter ratings reflected valuable information. Investors--not issuers--paid for ratings. (13) Indeed, bond issuers complained about credit rating agencies and opposed the concept of credit ratings as intrusive. (14)
Over time, rating agencies shifted from selling information to selling "regulatory licenses," (15) the "keys that unlock the financial markets." (16) By regulatory licenses, we mean the property rights associated with the "ability of a private entity, rather than a regulator, to determine the substantive effect of legal rules." (17) The regulatory-license model of information intermediaries differs substantially from the traditional reputational-capital model. According to the reputational-capital model, ratings providers survive and prosper primarily because of the continuing value of the information their ratings incorporate and reflect. In contrast, raters who benefit from regulatory licenses can continue to earn economic rents even if their ratings do not reflect valuable information. The regulatory-license view seeks to explain why market participants might continue to rely on particular rating agencies and why those rating agencies might continue to earn high margins even after a sustained period during which their ratings did not reflect valuation information.
The leading rating agencies began to shift from the reputational-intermediary role to the regulatory-license role after the 1929 crash, when regulators turned to particular rating agencies, notably Moody's and Standard & Poor's, for measures of bond quality to be used in banking and insurance guidelines.
Federal Reserve examiners proposed a system for weighting the value of a bank's portfolio based on credit ratings. Bank and insurance regulators expressed the "safety" or "desirability" of portfolios in letter ratings, and used such ratings in bank capital requirements and bank and insurance company investment guidelines. States relied on rating agencies to determine which bonds were "legal" for insurance companies to hold. The Comptroller of the Currency made similar determinations for federally chartered banks. (18) A second wave of regulatory reliance began in the mid-1970s, when the Securities and Exchange Commission (SEC) introduced the concept of NRSROs and thus further encouraged regulators to rely on their ratings. (19) Not coincidentally, NRSROs shifted the focus of their business model from investors to issuers during this time: specifically, they began charging the issuers of the debt they rated.
The issuer-pay model introduced significant new conflicts of interest--chiefly, the challenge for credit raters to impartially rate securities of companies that generate their revenues. But the rating agencies believed that they could manage these conflicts internally. Regulators now mandate that institutions of all types pay heed to NRSRO credit ratings as a necessary step for regulatory compliance. Some rules require that certain investors can only buy bonds with high ratings. Other rules reduce capital requirements for institutions that purchase highly rated bonds. Without high ratings, bond issuers cannot access certain markets, because they do not have a "regulatory license" from the NRSROs to comply with NRSRO-dependent regulations. (20) Tying regulations to ratings has created more demand for rating agencies' services and...