Credit Rating Agencies and the Credit Crisis: How the Issuer Pays Conflict Contributed and What Regulators Might Do about It CREDIT RATING AGENCIES AND THE CREDIT CRISIS: HOW THE "ISSUER PAYS" CONFLICT CONTRIBUTED AND WHAT REGULATORS MIGHT DO ABOUT IT Deryn Darcy* I. Introduction ............................. 607 II. Background: The Agencies, the Ratings, and the C risis ........................................................................... 6 08 A. The Credit Rating Agencies and Credit R atin gs ................................................................. 608 B . The Credit Crisis ................................................. 613 1. Subprime Mortgage Origination .................... 613 2. Securitization of Subprime Mortgages into RM BS and CDOs ............................................ 615 3. The Role of CRAs in Securitization ............... 616 4. The Burst of the Housing Bubble and Why Fingers are Pointing at the CRAs ................. 619 III. The "Issuer Pays" Conflict ......................................... 622 A. Historical Characteristics Making CRAs Vulnerable to the "Issuer Pays" Conflict ............ 624 1. Ratings-Based Regulation .............................. 624 2. Lack of Transparency in Ratings Process ..... 628 3. Lack of Accountability .................................... 629 a. Regulatory Accountability ........................ 630 b. Private Legal Liability ............................. 631 B. Pre-"Credit Crisis" Consensus: CRAs Not Compromised by Conflicts of Interest ................ 633 * J.D. Candidate 2010, Columbia University School of Law; B.S. Economics 2004, The Wharton School, University of Pennsylvania. The author offers her sincerest thanks to Professor Avery Katz for his insightful guidance, feedback, and general mentorship. She would also like to thank Professors Merritt Fox, Victor Goldberg, and Ronald Mann for their helpful research suggestions, and Kevin Coco and the survey team for their valuable comments. Last, but certainly not least, she is grateful to her husband and parents for their endless support. COLUMBIA BUSINESS LAW REVIEW C. Structured Finance Ratings are More Likely Compromised by the "Issuer Pays" Conflict ....... 636 1. The Growth of Structured Finance Products 638 2. The Concentration of the Customer Base ..... 639 3. The Structured Finance Ratings Process ...... 640 4. A Summary, and Where Do We Go From H ere? .................................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 642 IV. Countering the "Issuer Pays" Conflict: Is Disclosure Suffi cient? ........................... . . . . . . . . . . . . . . . . . . . . . . 645 A. Current and Proposed Disclosure-Based R egulations .......................................................... 646 1. The Original Regulatory Requirements ........ 646 2. Rules Amendments Adopted in December 2 008 ................................................................. 64 8 3. Proposed Amendments ................................... 650 B. The Theory Underlying Disclosure-Based R egulations .......................................................... 651 C. Why Disclosure is Insufficient for Deterring CRAs from Compromising Ratings ..................... 653 1. Failure to Perceive or Anticipate Relevant Differences Between Structured Finance and Traditional Products ............................... 654 2. Failure to Understand the Rated Instruments Themselves ................................ 655 3. Failure to Question the Accuracy of Ratings. 657 4. Ratings-Based Regulation ............................... 658 V. Performance-Based Sanctions as a Solution? ...... . . . . . 658 A. An Overview of Two Proposals ............................ 660 1. The Coffee Proposal ........................................ 660 2. The Hunt Proposal ......................................... 661 3. A Comparison of the Attributes ..................... 662 B. Shared Positives and Negatives ......................... 663 C. A Comparative Critique ...................................... 665 V I. C onclusion ...................................................................... 666 [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 607 I. INTRODUCTION "The story of the credit rating agencies is a story of colossal failure."' All good--or at least interesting-stories have a villain, and as the Credit Crisis started to unfold in 2007, credit rating agencies ("CRAs") emerged as obvious targets for fingerpointing by regulators, scholars, and commentators alike.' In short, many observers have accused the agencies of doing a poor job assessing the risks inherent in securities backed by subprime mortgages. I These securities- particularly residential mortgage-backed securities ("RMBS") and collateralized debt obligations ("CDOs") backed by subprime mortgages-have been at the center of the Credit Crisis, and many assert that these instruments would not have been as marketable without the CRAs' perceived seal of approval.4 Why did the agencies perform so poorly? There are likely many contributing factors, but a popular focal point has been the conflict of interest embedded in the CRAs' issuer pays business model.5 Indeed, several features of the structured 1 Credit Rating Agencies and the Financial Crisis: Hearing Before the H. Comm. on Oversight and Gov't Reform, 110th Cong. (Oct. 22, 2008) (Opening statement of Rep. Henry A. Waxman, Chairman). ' See, e.g., Jennifer E. Bethel, Allen Ferrell & Gang Hu, Legal and Economic Issues in Litigation Arising from the 2007-2008 Credit Crisis, John M. Olin Center for Law, Economics, and Business Discussion Paper at 58, available at http://ssrn.com/abstract=1096582 ("Many commentators have blamed the rating agencies, principally Moody's, Standard & Poor's, and Fitch, for investor losses."); John C. Coffee, Jr., Grade Inflation, NAT'L L.J., Sept. 10, 2007 ("This time, the obvious target will be the major credit- rating agencies, whose belated ratings downgrades of structured finance products precipitated a major sell-off in the debt securities market."). ' John Patrick Hunt, Credit Rating Agencies and the Worldwide "Credit Crisis": The Limits of Reputation, the Insufficiency of Reform, and a Proposal for Improvement, 2009 COLUM. Bus. L. REV. 109, 121-22 (citation omitted) (discussing common thread in criticism of CRAs). See infra Part II.B.3. See, e.g., Coffee, supra note 2 ("Rating agencies catered to investment banks."). finance ratings market, including the growth in structured finance products, the concentration of the CRAs' customer base for these securities, and the ratings process itself, suggest that the conflict may have been exacerbated. This Note explores these characteristics and argues that the current regulatory regime, with its emphasis on disclosure, is likely insufficient to protect against a similar problem in the future. This Note proceeds as follows: Part II provides background information on credit rating agencies and on the Credit Crisis. Part III explores the prevailing view that the rating agencies were not compromised by the issuer pays conflict prior to the Credit Crisis and explains what changed in the context of rating structured products. Part IV argues that the present disclosure-based regulatory regime is inadequate for guarding against compromised ratings. Part V suggests that a possible solution may lie in performance- based sanctions: penalties tied to poor quality. Part VI concludes. II. BACKGROUND: THE AGENCIES, THE RATINGS, AND THE CRISIS A. The Credit Rating Agencies and Credit Ratings Credit rating agencies produce ratings, in the form of a letter grade, that provide a relative indicator of a financial instrument's creditworthiness.' Thus, CRAs are theoretically important to investors because they distill "highly nuanced information" into a standardized symbol, allowing "markets to respond quickly and, more or less, uniformly to changes in ratings."8 CRAs are considered gatekeepers to the financial 6 See infra Part III.C. 7 JOHN C. COFFEE, JR., GATEKEEPERS: THE PROFESSIONS AND CORPORATE GOVERNANCE 284 (2006); Frank Partnoy, The Siskel and Ebert of Financial Markets? Two Thumbs Down for the Credit Rating Agencies, 77 WASH. U. L.Q. 619, 620 n.3 (1999); Claire A. Hill, Regulating the Rating Agencies, 82 WASH. U. L.Q. 43, 43 (2004). 8 COFFEE, supra note 7, at 284. COL UMBIA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 609 markets.' Along these lines, they are important to issuers because their ratings impact the issuer's cost of capital.1" In the United States, three CRAs dominate the market: Standard & Poor's Ratings Services 11 CS&P"), Moody's Investor Service, Inc. 1i ("Moody's"), and Fitch, Inc. 13 ("Fitch").14 Generally, the highest available rating is AAA, followed by AA, A, BBB, B, CCC, CC, C, and D (meaning defaulted). Each gradation may be further refined by a +/- ' See John C. Coffee, Jr., Op-Ed., The Mortgage Meltdown and Gatekeeper Failure, N.Y.L.J., Sept. 20, 2007 [hereinafter Coffee, Mortgage Meltdown] ("Gatekeepers are the 'reputational intermediaries' on whom investors depend."). This group also includes auditors, securities analysts, investment bankers, and sometimes attorneys. See also John C. Coffee, Jr., Gatekeeper Failure and Reform: The Challenge of Fashioning Relevant Reforms, 84 B.U. L. REV. 301, 308 (2004) [hereinafter Coffee, Gatekeeper Failure] (defining a gatekeeper as a party that 1) pledges its reputational capital to assure the accuracy of statements or representations that it either makes or verifies and 2) is easier to deter than the principal due to its receipt of a smaller benefit from the transaction that the gatekeeper facilitates or enables through its role in approving, certifying, or verifying information). 10 COFFEE, supra note 7, at 284. "Even within the universe of investment-grade securities, a rating downgrade or upgrade is likely to dramatically affect the issuer's cost of capital." Id. "1 S&P is a unit of The McGraw-Hill Cos., Inc. Standard and Poor's Company History, http://www.standardandpoors.com (last visited Apr. 26, 2009) (follow "About S&P" menu to "Company History" hyperlink). 12 Moody's Corporation, a publicly traded company on the NYSE, is the parent company of Moody's. Moody's Corporation Shareholder Relations, http://ir.moodys.com (last visited Apr. 26, 2009). "3 Fitch is owned by a French conglomerate, FIMALAC. It is composed of several smaller agencies including Fitch, IBCA, Duff & Phelps, and Thomson BankWatch that merged. The History of Fitch, http://www.fitchratings.com (last visited Apr. 26, 2009) (follow "About Us" hyperlink; then follow "History" hyperlink). 14 Hunt, supra note 3, at 114-15 (describing these companies as the major CRAs in the United States); U.S. SECURITIES AND EXCHANGE COMMISSION, ANNUAL REPORT ON NATIONALLY RECOGNIZED STATISTICAL RATING ORGANIZATIONS AS REQUIRED BY SECTION 6 OF THE CREDIT RATING AGENCY REFORM ACT OF 2006 35 (2008) [hereinafter SEC JUNE 2008 REPORT] (noting that Fitch, Moody's, and S&P issued over 99% of all outstanding ratings). Given the dominance of these three agencies, this Note's discussion of CRAs refers to them. suffix.15 CRAs also monitor their ratings on an ongoing basis and may upgrade or downgrade a rating as warranted. 16 They assert that their ratings on structured finance products are meant to be consistent with their ratings on traditional corporate bonds. Accordingly, a AAA-rating on a corporate bond and a AAA-rating on a structured finance product should mean approximately the same thing.17 In developing their ratings, CRAs rely on both quantitative analytics and qualitative assessments. 8 They rely on both public and non- public information; however, while some CRAs might receive private information from issuers during the ratings process, they do not verify this information. 9 " S&P and Fitch use the scale described. STANDARD & POOR'S, GUIDE TO CREDIT RATING ESSENTIALS 10 [hereinafter S&P's GUIDE], available at http://www2.standardandpoors.com/spf/pdf/fixedincome/SPCreditRatings Guide.pdf; Fitch Ratings Definitions, http://www.fitchratings.com (last visited Apr. 26, 2009) (follow "Ratings Definitions" hyperlink). Moody's uses a slightly modified scale: Aaa, followed by Aa, A, Baa, Ba, B, Caa, Ca, C, and D. Each rating may be modified by a suffix of 1, 2, or 3. About Moody's Rating Definitions, http://www.moodys.com [hereinafter Moody's Ratings Definitions] (last visited Apr. 26, 2009) (follow "About Moody's" hyperlink, followed by "Rating Definitions" hyperlink). 16 Hill, supra note 7, at 48. A CRA may put a given rating on "credit watch" if it believes a rating change may soon be appropriate. Id. at 49. 17 See, e.g., Moody's Ratings Definitions, supra note 15 ("Moody's ratings on long-term structured finance obligations . . . are calibrated to Moody's Corporate Scale."); CALVIN R. WONG, THOMAS G. GILLIS & FABIENNE MICHAUX, STANDARD & POOR'S, GENERAL: PRINCIPLES-BASED RATING METHODOLOGY FOR GLOBAL STRUCTURED FINANCE SECURITIES (2007), available at http://www.standardandpoors.com (under "Criteria & Methodologies") ("[A]n 'AAA' rated corporate bond should exhibit the same degree of credit quality as an 'AAA' rated securitized debt issue."). Section 3(a)(62) of the Securities Exchange Act of 1934 identifies five classes of ratings: financial institutions, brokers, or dealers; insurance companies; corporate issuers; asset-backed securities; and government, municipal, and foreign government securities. 15 U.S.C. § 78c(a)(62)(B)(i)-(vi) (2009). This Note will focus on ratings of corporate issuances and asset-backed securities, specifically subprime RMBS and CDOs containing subprime RMBS in their asset pools. See infra Part II.B for more on these instruments. 's SEC JUNE 2008 REPORT, supra note 14, at 37. '9 TECHNICAL COMMITTEE OF THE INTERNATIONAL ORGANIZATION OF SECURITIES COMMISSIONS, THE ROLE OF CREDIT RATING AGENCIES IN COL UMIA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 611 Ratings are intended to measure and predict the probability of default, or loss given default, of an issuer or issuanee. Instruments rated BBB- or above are considered "investment grade," meaning that the CRA believes the issuer has at least a "good" capacity to repay principal and interest as scheduled.21 Obligations with lower ratings are considered "speculative" or "junk."" CRAs stress that their ratings are simply "opinions" regarding the relative credit risk of the rated instrument or issuer.23 Thus, a higher rating reflects the CRA's belief that an instrument is of a higher credit quality than a lower rated instrument, but even a AAA rating is not intended as a guarantee that the instrument will not default; the rating only reflects that the CRA believes it is less likely to default than an instrument with a lower rating.24 Despite this intended limitation, a popular perception has developed that AAA-rated obligations STRUCTURED FINANCE MARKETS 3 (2008), available at http://www.iosco.org/ library/pubdocs/pdfIOSCOPD270.pdf [hereinafter IOSCO REPORT]; Hill, supra note 7, at 79 ("While [CRAs] do in-depth research on companies, they do not purport to go beyond what company officials tell them."). 20 SEC JUNE 2008 REPORT, supra note 14, at 37. 21 Hill, supra note 7, at 48. 22 Id. 23 See, e.g., S&P's GUIDE, supra note 15, at 3 ("[S&P's] ratings opinions are based on analysis by experienced professionals who evaluate and interpret information received from issuers and other available sources to form a considered opinion .... [Tihese ratings opinions are not intended to be a prognosis or recommendation."); Jerome S. Fons, Richard Cantor & Christopher Mahoney, Moody's Investor Service, UNDERSTANDING MOODY'S CORPORATE BOND RATINGS AND RATING PROCESS 5 (2002), available at http://www.moodys.com/moodys/cust/research/MDCdocs/06/200140000038 9218.pdf ("Moody's rating is an opinion forecast of an issuer's future relative creditworthiness."). 24 See, e.g., S&P's GUIDE, supra note 15, at 4 ("[A] corporate bond that is rated 'AA' is viewed by the rating agency as having a higher credit quality than a corporate bond with a 'BBB' rating. But the 'AA' rating isn't a guarantee that it will not default, only that, in the agency's opinion, it is less likely to default than the 'BBB' bond."). COLUMBIA BUSINESS LAW REVIEW essentially never default, perhaps given the historical performance of such highly rated instruments." A CRA may apply to the Securities and Exchange Commission ("SEC") for registration as a "Nationally Recognized Statistical Rating Organization ('NRSRO')."'26 By so doing, the CRA agrees to regulatory oversight by the SEC.27 In return, issuers and purchasers of securities rated by the CRA can receive regulatory benefits related to various SEC regulations that incorporate NRSRO ratings.2 As of June 2008, ten companies were registered with the SEC as NRSROs. 29 Despite this, the ratings market remains concentrated, with Fitch, Moody's, and S&P having issued almost 99% of all outstanding ratings. 30 While some evidence suggests that Fitch has significantly increased its market share in the last decade and has thus become a more important player in the CRA market, commentators have traditionally considered Moody's and S&P as the most 2 See, e.g., Hill, supra note 7, at 48 ("[An investor can rest almost completely assured that an instrument rated AAA will pay principal and interest as scheduled."). Indeed, based on Moody's historical data from the period 1970-2007, the probability that a AAA-rated corporate bond will have defaulted one year later is approximately 0.0%. Jennifer Tennant, Kenneth Emery & Richard Cantor, Moody's Investor Service, MOODY'S CREDIT POLICY: CORPORATE ONE-TO-FIVE-YEAR RATING TRANSITION RATES 2 (2008), available at http://www.moodys.com (follow "Rating Methodologies and Performance" hyperlink to "Historical Performance" hyperlink for list of research). 26 Registration of Nationally Recognized Statistical Rating Organizations, 15 U.S.C. § 78o-7 (2009). This statute is popularly referred to as the Credit Rating Agency Reform Act of 2006. 27 Id. 2' The SEC is not the only administrative body to incorporate NRSRO ratings into its regulatory framework. See infra Part III.A. 1. 29 SEC JUNE 2008 REPORT, supra note 14, at 34. '0 Id. at 35. The Herfindahl-Hirschmann Index (HHI), the measure generally used by economists to measure market concentration, of NRSROs is 3778 points. Id. at 35-36. An HHI above 1800 points is considered concentrated. The Herfindahl-Hirschmann Index, http://www. usdoj.gov/atr/public/testimony/hhi.htm. [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 613 prominent agencies, ranking Fitch a distant third. 1 Indeed, scholars note that a two rating norm has historically existed where issuers usually try to obtain ratings from Moody's andS&P.32 B. The Credit Crisis The Credit Crisis that began in 2007 involved a complex amalgamation of macro- and microeconomic factors, a full analysis of which is beyond the scope of this Note. However, in order to understand why regulators and commentators have identified CRAs as a villain of the Credit Crisis, an understanding of subprime mortgage origination, the securities that those mortgages were later packaged into, and the role of NRSROs in rating those securities is helpful. 1. Subprime Mortgage Origination The story of CRA involvement in the ongoing credit crisis appears to begin with the increase in subprime mortgage origination that occurred during the first half of the current 31 COFFEE, supra note 7, at 284 (noting that Fitch is most active in specialized and international markets). Cf John C. Coffee, SEC's New Rules Fail To Address Key Rating Issues, N.Y.L.J., Jan. 15, 2009 (citing Bo Becker and Todd Milbourn, Reputation and Competition: Evidence from the Credit Rating Industry, Harvard Business School Working Paper 09- 051, at 4, available at http://ssrn.com/abstract=1278150) (describing a recent study that found that Fitch increased its market share from ten percent to approximately one-third since Fitch was acquired by FIMALAC in 1997). 32 Hill, supra note 7, at 59-60 (noting that a two-rating norm has developed, where issuers usually try to obtain ratings from Moody's and S&P, and only occasionally use Fitch). 31 "The so-called 'worldwide credit crisis' that began in summer 2007 is a sprawling, interrelated series of events that continues to unfold. ... The U.S. officially went into recession at the end of 2007 as the credit crisis intensified, and as of early 2009 the weight of opinion appeared to be that problems in the credit market contributed significantly to problems in the real economy." Hunt, supra note 3, at 120 (citations omitted). For further discussion of the events and causes of the Credit Crisis, see Bethel et al., supra note 2, at 6. 614 COLUMBIA BUSINESS LAW REVIEW [Vol. 2009 decade. 4 As housing boomed in the early 2000s, peaking in 2005, the rate of subprime mortgage origination increased faster than the rate of mortgage origination overall." Bethel et al. report that of the more than $2.2 trillion of mortgages originated in 2001, $190 billion, or 8.6%, of those mortgages were subprime.3 6 In contrast, subprime mortgages accounted for 20%, or more than $600 billion, of all mortgages originated in 2005.11 Subprime origination remained nearly as high in 2006.8 While the volume and rate of subprime mortgage origination increased, the risk profile of the mortgages granted also increased from the perspective of the lender in at least three ways. First, the structure of subprime mortgages became increasingly vulnerable to interest rate risk. For example, many loans extended were "2/28" or "3/27" hybrid adjustable rate mortgages ("ARMs") 9 Under the terms of a typical 2/28 ARM, the borrower pays a low fixed interest rate and mortgage payment during the first two years that the loan is outstanding. After this period, the interest rate resets every six months for twenty-eight years based on an interest rate benchmark, often resulting in much higher payments.4 ° Second, the risk profile of the subprime borrower increased. In 2001, 28.5% of subprime borrowers could not verify information about employment, "' A subprime mortgage is a mortgage extended to a borrower with problems in her credit history, such as missed or late payments, that permit a lender to charge a higher interest rate to compensate for the additional risk the borrower represents. National Community Reinvestment Coalition, Glossary, http://www.ncrc.org/fairlending/gloss ary.htm. " Justin Lahart, Ahead of the Tape, WALL ST. J., Dec. 24, 2007, at C1 ("Housing peaked in 2005. By early 2006 it was widely recognized that the boom was likely over, and by mid-2006 it was beyond question."). " See Bethel et al., supra note 2, at 6. 37 Id. 38 Lahart, supra note 35 (reporting subprime mortgage origination of $625 and $600 billion in 2005 and 2006, respectively). " See Bethel et al., supra note 2, at 6. 40 Id. No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 615 income, or other credit-related data.4' This figure increased to nearly 51% in 2006.42 Third, mortgage brokers and bankers allegedly engaged in fraudulent and/or lax practices by submitting false information to qualify borrowers or by failing to document or verify relevant information.43 Such practices created the risk that otherwise unqualified borrowers would obtain subprime mortgages. Thus, by 2006, the volume of subprime mortgages outstanding as well as the risk of those mortgages had increased greatly. 2. Securitization of Subprime Mortgages into RMBS and CDOs As subprime mortgage origination increased, securitization became an important tool that allowed originators to continue increasing their lending activity. In a securitization, a sponsor or originator sells financial assets that generate cash flow to a special purpose vehicle ("SPV") that, in turn, issues securities to investors.44 The SPV uses income received from these pools to fund its principal and interest payments.45 In the context of the Credit Crisis, subprime mortgages on residential property were sold to 41 Id. at 83 tbl.4 (citing Freddie Mac and the International Monetary Fund). 42 Id. '* Id. at 16 ("Evidence is now mounting that at least some mortgage bankers and brokers may have submitted false appraisals and financial information to qualify otherwise unqualified households for subprime mortgage loans. Others purportedly did not document or verify subprime mortgagers' incomes, net worths, and credit histories."). See id. at 7 (noting that the sponsor or originator may be a different party than the party that originated the loans being sold to the SPV); Hunt, supra note 3, at 117 (citing Kenneth C. Kettering, Securitization and Its Discontents: The Dynamics of Financial Product Development, 29 CARDOZO L. REV. 1553, 1564-80 (2008) for a more detailed description of securitization). The financial assets that are sold may include pools of mortgages, corporate loans, auto loans, or credit card receivables. Id. 41 See Bethel et al., supra note 2, at 5. The issuer is 'bankruptcy remote,' meaning that if an originator goes bankrupt, the assets of the SPV cannot be distributed to the originator's creditors. Id. at 7. SPVs and packaged into subprime RMBS. 6 In addition, some RMBS were then structured into CDOs.47 Subprime mortgage originators used securitization to monetize the mortgages immediately rather than over the life of the mortgage, providing cash flow to fund additional loans. In 2005, $508 billion in subprime mortgages were securitized into bonds, as compared to $56 billion in 2000.4" 3. The Role of CRAs in Securitization When an SPV issues a RMBS or a CDO, it typically divides the securities into senior, mezzanine, and subordinated/equity tranches according to credit risk to provide securities that match the varying risk preferences of different investors.49 Typically, the issuer pays a CRA to rate the senior and mezzanine tranches. 0 Consistent with credit 46 Id. at 5-7 ("Mortgage-backed securities are debt obligations whose cash flows are backed by the principal and interest payments of pools of mortgage loans, most commonly on residential property."). 47 U.S. SECURITIES AND EXCHANGE COMMISSION, SUMMARY OF ISSUES IDENTIFIED IN THE COMMISSION STAFF'S EXAMINATION OF SELECT CREDIT RATING AGENCIES 7 (2008) [hereinafter SEC JULY 2008 REPORT]. CDOs are created by a similar process to MBS and are comprised of approximately 200 debt securities. A key difference between RMBS and CDOs is that a CDO may be actively managed. Id. Subprime RMBS appeared to comprise an increasing percentage of CDO collateral pools; data suggests that the average percentage of CDO collateral pools comprised of subprime RMBS grew from 43.3% in 2003 to 71.3% in 2006. Id. (reporting data from an NRSRO). 48 Michael Hudson, Debt Bomb-Lending a Hand: How Wall Street Stoked the Mortgage Meltdown-Lehman and Others Transformed the Market for Riskiest Borrowers, WALL. ST. J., June 27, 2007, at Al ("At the sector's peak in 2005, with the housing market booming, loan defaults remained low. Wall Street pooled a record $508 billion in subprime mortgages in bonds, up from $56 billion in 2000, according to trade publication Inside Mortgage Finance."). " Bethel et al., supra note 2, at 7 (describing process of using tranches); IOSCO REPORT, supra note 19, at 4-5 (explaining how structuring the offering into different tranches allows the issuer to take advantage of differing risk preferences). 50 SEC JULY 2008 REPORT, supra note 47, at 31. "Typically, the rating agency is paid only if the credit rating is issued, though sometimes it COL UAMIA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 617 ratings on other instruments, the rating reflects the CRA's view regarding the likelihood that the issuer will meet its principal and interest obligations.51 Structurers use forms of credit enhancement to lower the credit risk of the asset pool. By absorbing all or a portion of credit losses, these tools are used to make it more likely investors will receive their contractual cash flows, thereby increasing credit ratings. 2 Senior tranches usually receive ratings somewhere between AAA and AA, while mezzanine tranches receive ratings of A to BBB."8 Since senior tiers carry the lowest default risk to investors, they provide the lowest interest rates.54 When determining ratings on RMBS or CDOs, CRAs consider both quantitative models and qualitative analysis, and the ultimate decision is made by a ratings committee comprised of analytic personnel.5 The agencies take into account the credit quality of the underlying collateral and cash flow protection stemming from subordinated tranches.56 While in some cases, CRAs review reports from due diligence firms on the underlying collateral-i.e., the subprime mortgages-CRAs do not perform their own due diligence. 7 receives a breakup fee for the analytic work undertaken even if the credit rating is not issued." Id. at 9. 51 Id. at 7. 52 Bethel et al., supra note 2, at 12. Forms of credit enhancement include subordination, "which creates a hierarchy of loss absorption among the tranche securities;" over-collateralization, "which is the amount that the principal balance on the mortgage pool exceeds the principal balance of the tranche securities issued by the trust;" and excess spread, "which is the amount that the trust's monthly interest income exceeds its monthly liabilities." SEC JULY 2008 REPORT, supra note 47, at 6. " Bethel et al., supra note 2, at 7-8. 54 Hunt, supra note 3, at 118 (citing Ratul Roy & Glen McDermott, ABS CDOs, in THE STRUCTURED CREDIT HANDBOOK 335, 335 (Arvind Rajan, Glen McDermott & Ratul Roy eds., 2007)). 55 SEC JULY 2008 REPORT, supra note 47, at 7. 56 Bethel et al., supra note 2, at 8. 51 John C. Coffee, Jr., Achieving Transparency, N.Y.L.J., Dec. 10, 2007 ("[Riating agencies do not perform meaningful due diligence."); Bethel et al., supra note 2, at 13 ("In at least some instances, credit-rating agencies review due-diligence firms' reports or summaries of reports when evaluating credit risk."). However, in some instances, CRAs were not COLUMBIA BUSINESS LAW REVIEW CRAs assert that they have no obligation to do so, but their decision not to may also stem from a lack of sufficient resources given the volume of issuances the CRAs rate or concern over legal liability.58 Commentators have emphasized that credit ratings were critical in convincing investors to accept novel instruments like subprime RMBS and CDOs linked to subprime mortgages, although it is not clear why.59 One possible reason is that these products were typically purchased by institutional investors such as hedge funds, banks, life insurance companies, pension funds, and mutual funds. These buyers may be subject to legal restrictions that force them to purchase investment grade debt."° In addition, nearly all SPVs are structured to qualify for an exemption under the Investment Company Act of 1940.61 Issuers may rely on Rule 3a-7, which exempts an SPV that issues fixed income securities with one of the highest four investment grade ratings from an NRSRO at the time of sale. s 6 provided with due diligence reports, even upon request. Vikas Bajaj & Jenny Anderson, Inquiry Focuses on Withholding of Data on Loans, N.Y. TIMEs, Jan. 12, 2008, at Al. " Coffee, supra note 57 (noting that Moody's Code of Professional Conduct states that Moody's does not have an obligation to perform due diligence) (citing MOODY'S INVESTOR SERVICE, CODE OF PROFESSIONAL CONDUCT 6 (June 2005)). 59 See, e.g., Hunt, supra note 3, at 119 ("Agency ratings apparently have been crucial for investor acceptance of these new instruments, although commenters do not agree on the reason for this."). 60 Bethel et al., supra note 2, at 14 (noting that in certain circumstances, institutional investors are subject to legal rules that require them to purchase only investment grade or AAA-rated debt). While RMBS are often registered, CDOs are usually not, so offerees must be qualified institutional investors such as pension plans, hedge funds, investment banks, and municipalities. Hunt, supra note 3, at 119; 17 C.F.R. § 230.144A-(d)(1) (2009). 61 Bethel et al., supra note 2, at 8. 62 17 C.F.R. § 270.3a-7 (2009). Bethel et al., supra note 2, at 8 (referring to Rule 3a-7 as the "primary exemption"). In contrast, the SEC has stated most structured finance issuances do not depend on Rule 3a-7. References to Ratings of Nationally Recognized Statistical Rating [Vol, 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 619 Moreover, for ERISA purposes, investment grade ratings can be important to banks that wish to act as sponsors of SPVs.6 3 4. The Burst of the Housing Bubble and Why Fingers are Pointing at the CRAs After housing peaked in 2005, housing prices began to decline nationally in 2006.64 Home sales fell, and interest rates increased.65 By February 2008, over two million homeowners incurred interest rate resets on their mortgages.66 Data show that mortgage payments increased by up to 30%.67 Already as of June 2007, roughly 13% of subprime loans stood at or near foreclosure."6 Market participants began reassessing the credit risk inherent in subprime RMBS and CDOs as default and delinquency rates exceeded expectations.69 Organizations, Investment Company Act Release No. 28,327, 2008 SEC LEXIS 1519, at *27 (July 1, 2008). 6 Bethel et al., supra note 2, at 14. ERISA refers to the Employee Retirement Income Security Act. U.S. Department of Labor, http://www. dol.gov/dol/topic/health-plans/erisa.htm (last visited Apr. 26, 2009). " See Lahart, supra note 35 ("Housing peaked in 2005. By early 2006 it was widely recognized the boom was likely over, and by mid-2006 it was beyond question. In June 2006, sales of existing single-family homes were 9% below their year-earlier level, sales of new homes were down 15% and framing lumber prices were down 19%."); Bethel et al., supra note 2, at 18 (stating that housing prices declined about 1.5% between 2006 and 2007). 66 Bethel et al., supra note 2, at 18. 66 Id. (citing C. Cagan, Mortgage Payment Reset: The Issue and the Impact, First American Core-Logic, 2007, available at www.facorelogic. comluploadedFiles/Newsroom/Studies-andBriefs/Studies/20070048Mortg agePaymentResetStudyFINAL.pdf). 67 Id. Hudson, supra note 48. 66 Bethel et al., supra note 2, at 20 ("Over the summer of 2007, unanticipated delinquency and default rates on subprime residential mortgages caused market participants to re-evaluate the credit risk inherent in subprime RMBS and CDOs."). COLUMBIA BUSINESS LAW REVIEW Criticism of CRAs began to emerge. 70 Generally, commentators accused the agencies of failing to adequately estimate the risk of securitized products containing subprime mortgages in their collateral pools and asserted that CRAs induced investors to purchase these products by giving them undeservedly high ratings. 71 As investors realized that the securities were riskier than their ratings indicated, they lost confidence in the products generally, resulting in systemic consequences.72 Commentators also suggested that CRAs permitted their ratings to become compromised by conflicts of interest.7 3 " See Hunt, supra note 3, at 120-21 ("The crisis likely has a diverse set of interrelated causes, but observers have criticized rating agencies sharply-sometimes in quite colorful terms. Official bodies that have reported on the crisis, including the President's Working Group on Financial Markets, the International Organization of Securities Commissions, the Financial Stability Forum, and the staff of the Securities and Exchange Commission .. . have been less colorful but equally critical." (citations omitted)). 71 See IOSCO REPORT, supra note 19, at 2 ("[Clredit ratings have had an inordinate impact on the valuation and liquidity of subprime RMBSs and RMBS-backed CDOs."); FINANCIAL STABILITY FORUM, REPORT OF THE FINANCIAL STABILITY FORUM ON ENHANCING MARKET AND INSTITUTIONAL RESILIENCE 33 (2008) [hereinafter FSF REPORT] (noting the "severe underestimation" by CRAs of the credit risks of securities backed by subprime mortgages). 72 FSF REPORT, supra note 71, at 32 ("CRAs assigned high ratings to complex structured subprime debt based on inadequate historical data and in some cases flawed models. As investors realised this, they lost confidence in ratings and securitised products more generally."); IOSCO REPORT, supra note 19, at 2 ("[Als the number of delinquencies on subprime mortgages in the United States suddenly increased, some investors began to question the accuracy of many CDO and RMBS ratings, fueling a growing reluctance to invest in these products by increasingly risk-averse investors .... [T]hese questions about the quality of CRA ratings and the integrity of the rating process arguably added to the liquidity crisis that occurred on many markets beginning in August 2007."). A full examination of the systemic consequences is beyond the scope of this Note. " See, e.g., Coffee, Mortgage Meltdown, supra note 9 ("Today, as the repercussions of the mortgage meltdown continue to be felt on a worldwide basis, attention is shifting to the credit rating agencies. The finger pointing has only just begun, and predictably the SEC will study the [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 621 Indeed, CRAs began downgrading RMBS and CDOs at high rates. By September 2007, Moody's downgraded approximately 5% of subprime RMBS it rated in 2006. This contrasts starkly with Moody's historical downgrade rates: From 2002-2006, Moody's downgraded only 2.1% by dollar volume of subprime RMBS and 1% by dollar volume of all RMBS.74 From another perspective, Morgan Stanley found that of the 6431 subprime RMBS pools rated by Moody's in 2006, Moody's had downgraded 50.1% of them by December 2007, despite historical downgrade risk on RMBS of 0.3%. Similarly, S&P reported that its global structured securities one-year downgrade rate rose to 10.81% in 2007 from 1.97% in 2006 based on the number of ratings outstanding in 2007.76 S&P downgraded 18.85% of subprime RMBS in 2007, resulting in an average rating of A.77 In addition to the rate of downgrades, comparing the rates of structured product defaults to the rate of corporate bond defaults is also illuminating. Bloomberg Markets reported in July 2007 that, according to Moody's, corporate bonds with Moody's lowest investment grade rating Baa had an average default rate of 2.2% over five year increments spanning 1983-2005, problem and report that conflicts of interest were a major factor. Quelle surprise!"). For an extensive discussion of this issue, see infra Part III.C. " The Role and Impact of Credit Rating Agencies on Subprime Credit Markets: Hearing Before the S. Comm. on Banking, Housing, and Urban Affairs, 110th Cong. (Sept. 26, 2007) (statement of Michael Kanef, Group Managing Director, Moody's Investors Service); Bethel et al., supra note 2, at 20. " Coffee, supra note 57 (citing MORGAN STANLEY FIXED INCOME RESEARCH, STANDING ON THE RAZOR'S EDGE (Nov. 2007)). Morgan Stanley also found that 97% of the RMBS issuances rated A or below by Moody's in 2006 had been downgraded. Id. 76 STANDARD & POOR'S, TRANSITION STUDY: IN 2007, U.S. HOUSING MARKET WEAKNESS PROMPTED RATING VOLATILITY IN GLOBAL STRUCTURED SECURITIES 4 (Feb. 2008), available at http://www2.standardandpoors.com/ spf/pdf/media/subprime housing-market_020108.pdf [hereinafter S&P TRANSITION STUDY]. 77 Id. at 16. while CDOs with the same ratings had an average default rate of 24% over five-year periods spanning 1993-2005.78 III. THE "ISSUER PAYS" CONFLICT7 9 The business model of the big three CRAs-Moody's, S&P, and Fitch-presents each with an inherent conflict of interest: these agencies are paid by the issuers subject to their ratings.80 Today, CRAs receive approximately 90 to 95% of their annual revenues from issuer fees.8 While these CRAs originally charged subscription fees, they switched to the issuer pays model in the mid-1970s.82 The decision to do so stemmed largely from the nature of their business-CRAs produce information, and information is a public good.83 The 78 Charles Calomiris & Joseph Mason, We Need a Better Way to Judge Risk, FIN. TIMES, Aug. 23, 2007. 7' The issuer pays conflict is not the only conflict of interest that may compromise the ratings process. Individual rating analysts may be compromised by their own personal financial interests. For example, they might issue inflated ratings in hopes of obtaining a more lucrative job from the issuer. Hunt, supra note 3, at 152. In addition, the agencies have apparently taken steps to ensure that an analyst does not participate in rating issuers with whom that analyst has business or economic ties and to restrict and monitor analyst trading. SEC JULY 2008 REPORT, supra note 47, at 28. Furthermore, executive compensation tied to stock options may compromise executive decision-making. See Hunt, supra note 3, at 152. Note that a subscriber pays model would create its own conflicts of interest. See SEC JUNE 2008 REPORT, supra note 14, at 46 (discussing these potential conflicts). 80 Partnoy, supra note 7, at 652 ("Perhaps the most important change in the credit rating agencies' approach since the mid-1970s has been their means of generating revenue. Today, issuers, not investors, pay fees to the rating agencies."). 81 Id. Using 2003 data, Professor Hill estimates Moody's received 90% of its revenues from issuer fees. See Hill, supra note 7, at 50. 82 Partnoy, supra note 7, at 652. "" Id. at 653 ("One reason the payment to rating agencies may have shifted from investors to issuers is that the information rating agencies generate has the characteristics of a public good. For example, an agency publishing a rating to one or more individuals, for a fee, will find it difficult to exclude other non-paying individuals from access to that rating."). COL UMBIA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 623 subscription-based model created a "free rider" problem because the agencies could not feasibly stop paying subscribers from sharing the information with non- subscribers. 4 Through the issuer pays model, the agencies can effectively charge all users of their product since the issuer can pass the cost of the rating on to investors in the form of a slightly reduced interest rate.8 5 The adoption of the issuer pays model created the risk that a CRA will compromise the integrity of its ratings to appease the issuer-its paying customer-by issuing an undeservedly high rating or by failing to downgrade a security when changes warrant it. Despite this risk, and even though several historical characteristics of CRAs would seem to exacerbate the risk inherent in the issuer pays model, the general consensus appears to be that any deficiency in CRA ratings prior to the Credit Crisis (and thus relating primarily to securities other than structured finance products) did not result from the issuer pays conflict.86 The economics of rating structured finance products and the nature of the relationship between a CRA and an issuer of a structured product suggest, however, that the issuer pays conflict is intensified when CRAs rate securitized products. This may account in part for the seemingly low quality of ratings as evidenced by the high rate of downgrades on structured finance products beginning in 2007.7 ' The free rider problem increased as photocopying and other technology progressed that permitted users to share or re-sell information more cheaply. Hill, supra note 7, at 50 (citations omitted). In addition, around the time that the agencies switched to the issuer pays model, users were demanding "more comprehensive and resource-intensive analysis of issuers." U.S. SECURITIES AND EXCHANGE COMMISSION, REPORT ON THE ROLE AND FUNCTION OF CREDIT RATING AGENCIES IN THE OPERATION OF SECURITIES MARKETS AS REQUIRED BY SECTION 702(B) OF THE SARBANES- OXLEY ACT OF 2002 41 (2003). 85 COFFEE, supra note 7, at 296; Partnoy, supra note 7, at 653. 86 See infra Part III.B. s See supra Part II.B.4. A. Historical Characteristics Making CRAs Vulnerable to the "Issuer Pays" Conflict 1. Ratings-Based Regulation NRSROs are unique in that regulators across a range of substantive areas have incorporated their ratings into regulatory frameworks. This practice began in 1930, when the Federal Reserve System developed a method for reviewing a bank's entire portfolio based on the credit ratings of the bonds in the portfolio.88 While additional regulations incorporating credit ratings followed closely thereafter, the most significant development occurred in 1973, when the SEC revised its "net capital" rule for broker dealers.89 The first securities rule to formally incorporate credit ratings, Rule 15c3-1 requires a broker dealer to take a "hair cut" when valuing securities for its balance sheet and ties the amount of the "hair cut" to credit ratings. Fearing that the entry of "start ups and fly-by-night small firms" lacking reputational capital would result in a race to the bottom, the SEC decided that only those ratings issued by established CRAs-what it defined as "Nationally Recognized Statistical Rating Organizations"-would count 88 COFFEE, supra note 7, at 288-89; Partnoy, supra note 7, at 687 ("The Federal Reserve Banks first began using bond ratings in their examination of the portfolios of member banks in 1930.") (citing GILBERT HAROLD, BOND RATINGS AS AN INVESTMENT GUIDE: AN APPRAISAL OF THEIR EFFECTIVENESS 25 (1938)). 89 17 C.F.R. § 240.15c3-1 (2009); see also COFFEE, supra note 7, at 289 (discussing additional regulations implemented in the 1930s); Partnoy, supra note 7, at 686-90 (discussing the same). Much of Professor Partnoy's discussion cites to GILBERT HAROLD, BOND RATINGS AS AN INVESTMENT GUIDE: AN APPRAISAL OF THEIR EFFECTIVENESS (1938). - 17 C.F.R. § 240.15c3-1 (2009); Partnoy, supra note 7, at 690 (describing Rule 15c3-1 as the first securities rule to formally incorporate credit ratings). Professor Coffee explains that "hair cuts" in the context of Rule 15c3-1 are "mandatory write-downs... on the broker's balance sheet for securities that it owned, which were deemed risky or speculative." COFFEE, supra note 7, at 289. COLUAMA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 625 for purposes of the rule." Thus, the NRSRO designation was born. From there, the incorporation of NRSRO ratings in substantive regulation increased. Now, many important rules use NRSRO ratings as benchmarks.92 According to the SEC, at least forty-four of its rules and forms incorporate references to credit ratings.9' The practice of incorporating 91 Notice of Revision of Proposed Amendments to Rule 15c3-1 Under the Securities Exchange Act of 1934, Exchange Act Release No. 10,525, 1973 SEC LEXIS 2309 (Nov. 29, 1973) ("The Commission to a limited extent has also recognized the usefulness of the nationally recognized statistical rating organizations as a basis for establishing a dividing line for securities with a greater or lesser degree of market volatility."); COFFEE, supra note 7, at 289 ("Fearing the proverbial 'race to the bottom,' the SEC decided to recognize only the credit ratings issued by the major credit-rating agencies that pre-existed its new rule. To this end, it defined the term 'Nationally Recognized Statistical Ratings Organizations' . . . so as deliberately to exclude start-ups and fly-by-night small firms that lacked reputational capital."); Partnoy, supra note 7, at 690 n.344 ("As the initial source of the term NRSRO, Rule 15c3-1 effectively froze the then- approved credit rating agencies (e.g., S&P, Moody's, Duff & Phelps, and Fitch) as acceptable for rating purposes, and severely limited the possibilities for new entrants."). Duff & Phelps was since acquired by Fitch in 2000. The History of Fitch, http://www.fitchratings.com (last visited Apr. 26, 2009) (follow "About Us" hyperlink; then follow "History" hyperlink). For example, when the SEC began regulating money market funds in the early 1980s, it revised Rule 2a-7 of the Investment Company Act of 1940 and limited money market funds to investments in "Eligible Securities:" those securities with a high rating from at least two NRSROs or from the only NRSRO rating the security. 17 C.F.R. § 270.2a-7 (2009); COFFEE, supra note 7, at 290. The SEC revised the rule again in 1991, essentially prohibiting a money market fund from investing more than five percent of its assets in unrated commercial paper or more than one percent in the commercial paper of a single unrated issuer. Revisions to Rules Regulating Money Market Funds, Securities Act Release No. 6882, 1991 SEC LEXIS 268 (Feb. 20, 1991); COFFEE, supra note 7, at 290. " Christopher Cox, Chairman, SEC, Statement on Proposal to Increase Investor Protection by Reducing Reliance on Credit Ratings (June 25, 2008), available at http://www.see.gov/news/speech/2008/spch062508 cccredit.htm (discussing SEC proposal to eliminate eleven references to credit ratings in SEC rules and forms, to make changes to twenty-seven references, and to leave six unchanged). credit ratings into substantive regulation has spread far beyond the SEC and the Federal Reserve. Credit ratings are now used in pension, banking, real estate, and insurance regulation. 14 The new Basel II banking regulations incorporate credit ratings, taking ratings-based regulation global. 5 The use of ratings in substantive regulation is so pervasive that one scholar has developed the "regulatory license" view of CRAs.9 6 Professor Frank Partnoy argues that the SEC and other regulators have conveyed a meaningful degree of market power on NRSROs by endowing them with the ability to grant regulatory licenses, that is, to convey regulatory advantages to an issuer through the provision of a good rating. 7 Professor Partnoy asserts that NRSROs are in the "business of selling regulatory licenses" and that ratings are essentially only valuable because they reduce the issuer's regulatory costs, not because they are credible or accurate. 8 While it is arguable whether NRSROs only provide value through conferring favorable regulatory treatment, the extensive use of credit ratings in substantive regulations and the practical necessity those regulations may impose on issuers and investors to rely on credit ratings could very well skew CRA incentives to provide high-quality ratings. 11 ' See Partnoy, supra note 7, at 690 (stating that the number of rules, releases, and regulations that rely on credit ratings is in the hundreds); COFFEE, supra note 7, at 290 (noting that regulations incorporating credit ratings affect nearly all types of institutional investors). " See COFFEE, supra note 7, at 291 (noting that the Basel II Accords "place considerable weight on the credit ratings applicable to each bank's investments"). Under Basel II regulations, banks holding AAA assets may keep less capital relative to how much they are lending. AAAsking for Trouble, ECONOMIST, July 14, 2007. 96 See generally Partnoy, supra note 7, at 623, 681-703 (outlining his "regulatory license" view of rating agencies). 97 Id. at 623, 681. 98 Id. ' Hill, supra note 7, at 64-66 argues that Professor Partnoy overstates the case that ratings serve only to provide favorable regulatory treatment. She does, however, note that there is one specific exception where ratings may only serve to provide regulatory benefits: "a structured finance deal [Vol. 2009COL UMIA B USINESS LA W RE VIE W No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 627 Because NRSRO ratings (and in particular, investment grade ratings) may reduce regulatory transaction costs for issuers and may be necessary for certain investors to purchase a given security, those CRAs with the NRSRO designation might choose to sacrifice quality (in the form of an accurate rating) in favor of producing inflated ratings since investors and issuers alike need ratings from NRSROs and have few sources from whom they can purchase these ratings. 100 Moreover, investors and issuers might actually prefer a higher rating to a more conservative one because they can only reap regulatory benefits if the credit rating is high enough to qualify for preferred regulatory treatment."1 While the SEC has proposed new rules to reduce reliance on credit ratings in its rules and forms, it appears unlikely that the systematic use of credit ratings in substantive regulation will ameliorate in the near-term since credit ratings are incorporated into many rules and regulations extending beyond the SEC.0 2 structured for a particular investor who is studying the deal closely. The investor may not need any information about the deal, and may not even be subject to guidelines that require him to buy highly rated instruments; he may just be buying the favorable regulatory treatment." Id. at 66 & n.114. 100 See, e.g., Part II.B.3, supra (discussing how ratings were critical to the growth of subprime RMBS and subprime-backed CDOs); see also supra Part II.A (discussing historical dominance of the big three CRAs). 101 Some academics and commentators have noted that no constituency actually prefers conservative ratings, particularly ratings downgrades. See, e.g., Coffee, supra note 2 ("In truth, no one likes a tough, objective rater-not the issuer that pays for it, nor the institutions that hold portfolios of investment grade debt that will be devastated by a ratings downgrade."). 112 References to Ratings of Nationally Recognized Statistical Rating Organizations, Exchange Act Release No. 58,070, 2008 SEC LEXIS 1518 (July 1, 2008); Security Ratings, Securities Act Release No. 8940, Exchange Act Release No. 58,071, 2008 SEC LEXIS 1510 (July 1, 2008); References to Ratings of Nationally Recognized Statistical Rating Organizations, Investment Company Act Release No. 28,327, 2008 SEC LEXIS 1519 (July 1, 2008) [hereinafter SEC July 2008 Proposed Amendments] (all outlining proposed amendments to SEC regulations intended to reduce reliance on NRSRO ratings in SEC rules and forms); see also Hunt, supra note 3, at 151 (explaining that credit ratings are also COLUMBIA BUSINESS LA W REVIEW 2. Lack of Transparency in Ratings Process Regulators and commentators have also criticized the CRA market for a lack of transparency. 10 3 One scholar has categorized transparency issues as arising in two distinct areas. First, the industry suffers from insufficient methodological transparency, meaning that the CRAs do not provide enough information about their rating procedures and methodologies for an outsider to know just how an agency determined a given rating.' Second, CRAs provide insufficient information on performance, making it difficult for an outsider to clearly evaluate the quality of ratings.?° While the CRAs voluntarily provided information on both of these counts prior to any formal regulatory disclosure requirements, the disclosures provided were apparently inadequately specific or incomplete.' For example, in a study of S&P, Moody's, and Fitch that began in August 2007 and culminated in a report issued in July 2008, the SEC found that the agencies did not always fully disclose significant components of the ratings process and methodologies for rating RMBS and CDOs, despite claims from the CRAs that they disclosed their ratings processes." 7 Specifically, the agencies did not always disclose criteria and frequently incorporated into our system of private contracting: "[O]ther private contracts, including bond indentures, contain 'rating triggers,' under which a rating downgrade below a particular level is a technical default"). This could also call into question the efficacy of the SEC's proposal to reduce reliance on credit ratings. '03 See Registration of Nationally Recognized Statistical Rating Organizations, 15 U.S.C. § 78o-7 (2009) (defining the act, in part, as an act to foster transparency in the credit rating industry); Hunt, supra note 3, at 129 (citing absence of transparency as a characteristic of the rating market that may "cause it to perform worse than it could"). 104 Id. at 138 (defining methodological transparency as "an outsider's ability to tell just how the agencies reach the ratings they award"). 'o' Id. (defining performance transparency as "the ability to discern how well the ratings perform"). '06 On methodological transparency, see id. at 139 (stating that the SEC's first set of methodological disclosure rules contributed little to what was already available). 107 SEC JULY 2008 REPORT, supra note 47, at 1, 13. [Vol. 2009 No. 2:605] CREDIT RA TING AGENCIES AND THE CREDIT CRISIS 629 deviated from their quantitative models.1 °8 Moreover, the SEC's proposed rule amendments suggest that the agencies' current disclosures on performance do not provide for comparison across rating agencies and permit data manipulation. 109 Both methodological and performance transparency are important to constraining the issuer pays conflict because they assist investors in evaluating both how an issuer may exert pressure in the ratings process and whether poor quality ratings suggest a CRA is compromised.110 3. Lack of Accountability Historically, the CRA market has been characterized by a lack of accountability both to regulators and through private legal liability. This historical lack of accountability could '0' Id. at 13-14. Indeed, the staff of the SEC concluded that "[nione of the rating agencies examined had specific written procedures for rating RMBS and CDOs." Id. at 16. In addition, one scholar has testified that "[w]hile the statistical techniques used by the NRSROs are transparent, the ratings criteria (the variables incorporated into the statistical techniques) are not disclosed up to a level of replicability." The Role of Credit Rating Agencies in the Structured Finance Market: Hearing Before the Subcomm. on Capital Markets, Insurance, and Government Sponsored Enterprises of the H. Comm. on Fin. Servs., 110th Cong. (Sept. 27, 2007) (testimony of Joseph R. Mason, Associate Professor, Drexel University), available at http://financialservices.house.gov/hearingll/mason.pdf. " Proposed Rules for Nationally Recognized Statistical Rating Organizations, Exchange Act Release No. 57,967, 2008 SEC LEXIS 1613 (June 16, 2008) (proposing that NRSROs be required to disclose performance statistics separately for each class of ratings so that various ratings users could more readily identify the statistics relevant to them, in a manner that facilitates comparison across agencies, and in a form that lessens the potential for data manipulation). 110 Hunt, supra note 3, at 140 n.100 ("One area in which methodological transparency does seem desirable is in relationships with structured-product arrangers."); id. at 142 ("If investors cannot determine how an agency's ratings have performed, they cannot develop informed views about the quality of those ratings."); see also SEC JULY 2008 REPoRT, supra note 47, at 1 (discussing how staff reviewed processes for rating RMBS and CDOs to gain insight into conflicts of interest in the ratings process). COLUMBIA BUSINESS LA W REVIEW arguably have rendered CRAs more vulnerable to the issuer pays conflict. a.Regulatory Accountability Prior to 2006, the CRA market was only regulated to the extent that the SEC controlled the NRSRO designation process. The SEC conveyed NRSRO status through a "no- action letter" mechanism; a CRA would apply to the SEC and, if accepted, the SEC would send a no-action letter indicating the SEC would accept the agency's ratings as those of an NRSRO. "' This process apparently lacked transparency, leaving applicants unclear on the criteria for designation and the status of their applications. 12 In 2006, Congress, in an effort partly aimed at increasing accountability of CRAs, passed the Credit Rating Agency Reform Act of 2006.11 Through this Act, Congress clarified the SEC's authority over NRSROs and established a clear process for achieving NRSRO designation.'14 It is unclear whether the Act will successfully increase accountability. On one hand, the Act gives the SEC exclusive authority to enforce provisions of the Act if an NRSRO issues credit ratings in contravention of the procedures it discloses in its application or in other required disclosures."' On the other hand, the Act prohibits the SEC from regulating "the "' Hunt, supra note 3, at 133. 112 See id. at 133 (citing Examining the Role of the Capital Markets: Hearing Before the S. Comm. on Banking, Housing & Urban Affairs, 109th Cong. 197 (2005) (statement of Kent Wideman, Executive Vice President, Dominion Bond Rating Service) (describing the process as unclear and resulting in untimely decisions); H.R. 2990, 109th Cong. (2005); The Credit Rating Agency Duopoly Relief Act: Hearing Before the H. Comm. on Fin. Servs., 109th Cong. 21 (2005) (testimony of Sean Egan, Managing Director, Egan-Jones Rating Co.)). 113 Registration of Nationally Recognized Statistical Rating Organizations, 15 U.S.C. § 78o-7 (2009) ("An [a]ct [t]o improve ratings quality for the protection of investors and in the public interest by fostering accountability, transparency, and competition in the credit rating agency industry."). 114 See generally id. 5 15 U.S.C. § 78o-7(c)(1) (2009). [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 631 substance of credit ratings or the procedures and methodologies" employed by NRSROs and provides that rules and regulations "shall be narrowly tailored" to meet the Act's requirements." 6 While the effectiveness of the Act remains an open question, the historic lack of regulatory oversight that existed prior to the Act could have created an accountability void that permitted CRAs to compromise their ratings to benefit issuers. At a minimum, CRAs did not face any regulatory consequences for inflating ratings to appease their customers. b.Private Legal Liability In addition to a historical lack of regulatory accountability, CRAs have remained essentially insulated from liability in private litigation."7 The securities laws seem to provide a natural hook for holding CRAs liable, since CRAs issue ratings on bonds and other publicly traded notes."' To date, however, no CRA has apparently been held liable in a securities class action. "I Section 11 of the Securities Act of 1933 exempts CRAs from liability. 120 Moreover, while an argument exists for holding CRAs liable under Section 10(b) of the Securities Exchange Act of 1934 and Rule 10b-5 promulgated thereunder, private plaintiffs, including issuers and other market participants, have faced 116 15 U.S.C. § 78o-7(c)(2) (2009). 117 See, e.g., COFFEE, supra note 7, at 302 (noting that CRAs enjoy "virtual immunity" from private litigation); Jeffrey Manns, Rating Risk after the Subprime Mortgage Crisis: A User Fee Approach for Rating Agency Accountability N.C. L. REV. (forthcoming) (manuscript at 42, available at http://ssrn.com/abstract=1199622) ("What is clear is that rating agencies face little or no accountability to issuers and generally have no liability exposure to creditors or other financial market participants."). 11 COFFEE, supra note 7, at 302 (noting that bonds and other publicly- traded notes are "securities") (citing Rating Agencies and the Use of Credit Ratings under the Federal Securities Laws, Securities Act Release No. 8236 (June 12, 2003)). 119 Coffee, supra note 31 (noting in a footnote that it does not appear a CRA has been held liable in a securities class action). 120 Hill, supra note 7, at 56-57; 17 C.F.R. § 230.436(g)(1) (2009). at least two hurdles when attempting to hold CRAs liable under any cause of action."' First, and primarily, courts have proven somewhat receptive to the agencies' arguments that their ratings are opinions protected under the First Amendment. 122 In some situations, CRAs have invoked press shield laws to avoid having to respond to subpoenas. 123 121 15 U.S.C. § 78j(b) (2000); 17 C.F.R. § 240.10b-5 (2009). As Professor Coffee put it, "statements made by any person 'in connection with the purchase or sale of a security' are potentially within the reach of SEC Rule 10b-5. Thus, a particular rating (or its upgrade or downgrade) could constitute a material misstatement and be actionable under Rule 10b-5; similarly, the failure to update ratings in light of new information could be a material omission." COFFEE, supra note 7, at 302 (citing Rating Agencies and the Use of Credit Ratings under the Federal Securities Laws, Securities Act Release No. 8236 (June 12, 2003)). 122 Manns, supra note 117, at 42 ("[The] First Amendment hurdle has made it extraordinarily difficult to establish that rating agencies engaged in libel and has left issuers without legal recourse except in outlier cases."); Hunt, supra note 3, at 186 (noting that CRAs have won some high profile victories on its First Amendment defense in lower courts). The First Amendment defense may be invoked both in the context of securities laws and against other claims. See, e.g., Newby v. Enron Corp., 511 F. Supp. 2d 742, 817 (S.D. Tex. 2005) ("[Wlhile there is no automatic, blanket, absolute First Amendment protection for reports from the credit rating agencies based on their status as credit rating agencies, the courts generally have shielded them from liability for allegedly negligent ratings for various reasons."); id. at 825 (noting that "there is no blanket First Amendment protection for published credit ratings," but concluding that the plaintiff failed to meet the enhanced pleading requirements necessary to overcome First Amendment protection for a claim of negligent misrepresentation against a CRA). But see Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 753, 761-62 (1985) (affirming decision of lower court not to grant heightened First Amendment protections to CRA that distributed a false report to five subscribers, with plurality emphasizing that the report did not address a matter of public concern). 123 See, e.g., In re Pan Am Corp., 161 B.R. 577, 586 (S.D.N.Y. 1993) (holding that CRA gathered information sought by subpoena with newsgathering intent and that CRA is thereby entitled to invoke journalist's privilege); In re Scott Paper Co. Sec. Litig., 145 F.R.D. 366, 370 (1992) (holding that CRA qualified as member of press for First Amendment purposes, allowing it to invoke qualified journalist's privilege). But see In re Fitch, Inc., 330 F.3d 104, 111 (2d Cir. 2003) ("We simply conclude that on these facts, the district court did not abuse its COLUMBIA BUSINESS LA W RE VIEW [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 633 Second, at least one court has held that a plaintiffs reliance on the defendant CRA's rating was unreasonable.'24 This seems illogical given the prominence of credit ratings in regulatory frameworks. The Credit Rating Agency Reform Act of 2006 likely did little to dismantle this insulation from legal accountability, as the Act states that it does not create a private right of action.'25 While a full review of the case law involving CRAs is beyond the scope of this Note, it is worth noting that the trend may be moving away from this protective legal regime. 16 The bottom line, however, is that CRAs have historically enjoyed immunity from private liability. This absence of legal accountability to market participants would have made it easier for CRAs to compromise their ratings in favor of issuers than if CRAs faced a real risk of incurring damages for inflated ratings. B. Pre-"Credit Crisis" Consensus: CRAs Not Compromised by Conflicts of Interest Prior to the current Credit Crisis, scholars and commentators seemed to agree that CRAs had not allowed the integrity of their ratings to become compromised by the issuer pays model, despite the aforementioned characteristics of the industry that could have exacerbated discretion by concluding that Fitch had not sufficiently shown that the information it sought to protect was gathered pursuant to the newsgathering activities of a professional journalist."). 124 Quinn v. McGraw Hill Cos., Inc., 168 F.3d 331, 336 (7th Cir. 1999) (concluding that, on the facts of the case, no reasonable jury could find that plaintiff reasonably relied on defendant CRA's rating). 125 15 U.S.C. § 78o-7(m)(2) (2009). 126 See, e.g., COFFEE, supra note 7, at 313 n.71 (noting that the view of ratings as editorials seems increasingly dated); it re Nat'l. Century Fin. Enter., Inc., 580 F. Supp. 2d 630, 639-40, 653 (S.D. Ohio 2008) (rejecting CRAs' First Amendment defenses on the grounds that the complaint did not allege that the ratings were disseminated to the general public); Coffee, supra note 31 (discussing Nat'l Century case and noting insulation from liability may be changing). COLUMBIA BUSINESS LA W REVIEW this conflict.127 This notion is consistent with the prevailing view of CRAs: the "reputational capital" model.128 In contrast with the regulatory license view, adherents of this model emphasize the informational value of ratings and believe that the need to maintain reputations for accuracy and reliability will give CRAs optimal incentives for high- quality ratings.129 Accordingly, S&P and Moody's have maintained their market shares by producing high-quality ratings, thus building and maintaining reputational capital, and agencies that did not produce accurate ratings may have been forced out of the market. 3 ° Under this view, a CRA 127 See, e.g., Hill, supra note 7, at 51 ("Markets apparently do believe- correctly I think-that rating agencies are independent, notwithstanding that rating agencies are paid by their issuers."); Hunt, supra note 3, at 154 (noting that the SEC's rules promulgated under the 2006 Act "reflect the view that no serious regulatory intervention is needed to deal with agency conflicts of interest"); see also COFFEE, supra note 7, at 287-88 (discussing factors that provide for checks against issuer pays conflict). But see Hill, supra note 7, at 74 (noting the risk of Arthur Andersen-style conflicts should rating agencies increase their offerings of ancillary services). 128 Hunt, supra note 3, at 113 ("The dominant view of rating quality in the legal literature and among policyrnakers comes from the 'reputational capital' model .... "); id. at 15-16 (elaborating on the reputational capital view); Steven L. Schwarcz, Private Ordering of Public Markets: The Rating Agency Paradox, 2002 U. ILL. L. REV. 1, 26 (2002) ("Rating agencies are already motivated to provide accurate and efficient ratings because their profitability is directly tied to reputation."); Partnoy, supra note 7, at 633 (noting that law journal commentary seems to accept arguments that "collectively present the credit rating business as competitive and reputation-driven"). 129 Hunt, supra note 3, at 113 (describing how the reputational capital view holds that a well-functioning reputation mechanism will provide CRAs with optimal incentives to produce ratings of high quality); id. at 116 ("[Tlhe mainstream view is that investors value the agencies' assessments of credit quality."); Partnoy, supra note 7, at 631 ("Financial economists maintain that the value of a bond rating lies in its certification of the debt issue's credit quality."). See supra Part III.A. 1 for discussion of the regulatory license view. "' See Partnoy, supra note 7, at 634 (elaborating on this view). [Vol. 2009 No. 2:605] CREDIT RATING A GENCIES AND THE CREDIT CRISIS 635 would withstand pressure from issuers to inflate ratings out of concern for its reputation."' To the extent observers have complained about the quality of ratings, the complaints have centered around shirking rather than conflicts of interest.13 In a series of high profile events, such as the Enron and Worldcom bankruptcies and the 1997 Asian Financial Crisis, CRAs have been notoriously slow to downgrade ratings.33 Some commentators suggest that the absence of competition in the market permits the agencies to devote fewer resources to research and that the agencies are reactive, downgrading securities after negative public information has been released, rather than anticipating these developments.13 4 CRAs counter that investment grade securities rarely default, but one might question how meaningful this assertion is. 35 Academics point to examples where the agencies have downgraded securities from investment grade within days of default. 136 Despite these examples, investment grade ratings-particularly initial ratings--of corporate securities are accepted by academics as reliable indicators of relative credit risk.13 7 131 Professor Hill also notes that the two-rating norm would help Moody's and S&P to withstand this pressure, since issuers could not credibly threaten to take their business elsewhere. Hill, supra note 7, at 74. 132 See, e.g., Claire A. Hill, Rating Agencies Behaving Badly: The Case of Enron, 35 CONN. L. REV. 1145, 1152 n.58 ("Interestingly, the criticism has largely centered around 'shirking' rather than conflicts of interest.. "). "' Hill, supra note 7, at 67 ("[Wlhere [CRAs] have been more vociferously criticized, is in their ongoing ratings-their upgradings and downgradings. Consider, Enron, Worldcom, the Asian Flu, and the many other examples where, it seems, downgrades came much too late."). 134 See COFFEE, supra note 7, at 285. 131 See, e.g., id at 297 (noting that the industry points out that few investment grade securities default); Schwarcz, supra note 128, at 13 (discussing CRA's successful track record (citations omitted)). 136 See COFFEE, supra note 7, at 297 ("Default statistics are thus as cooked as Enron's books."). 137 See, e.g., Coffee, supra note 31 ("In truth, the credit rating agencies were and remain relatively accurate raters of corporate debt."); Hill, supra C. Structured Finance Ratings are More Likely Compromised by the "Issuer Pays" Conflict While CRA ratings on corporate securities may be reliable, Part II.B provides data illustrating that the rate of downgrades on structured finance products increased in 2007 relative to both historic levels and the relative rate of downgrades on corporate securities. 13 1 If a rating on a structured finance product is supposed to mean roughly the same thing as a rating on a corporate bond, what explains the disparity in rating performance? 39 One answer is that rating structured finance products is harder than rating traditional products, but this explanation seems unsatisfactory to elucidate the difference in rating performance between these asset classes, notwithstanding the greater complexity of the securities.4 ° Commentators have pointed to a number of more meaningful factors to help explain the downturn in performance. They have complained that models used by the rating agencies to rate these products contained fundamental defects. 141 For example, to the extent that quantitative models rely on historical data to estimate default risk, subprime mortgage loans performed strongly between 2001 and 2005. This may have contributed to overly optimistic ratings on later vintage RMBS. 142 More generally, the market may have had note 7, at 67 ("Rating agencies do particularly well when it comes to initial ratings."); Hill, supra note 132, at 1151 ("[Tlhere is evidence that whether or not the agencies generally know how good a debt security is, they know how good it is relative to other securities . . . "' See supra Part II.B.4. 139 See supra Part II.A. 140 Coffee, supra note 57 (noting that rating structured finance products is much more difficult than rating bonds of public companies); SEC July 2008 REPORT, supra note 47, at 11 (noting that the structured products themselves became increasingly complex over time). 141 See Hunt, supra note 3, at 123 (stating that observers have criticized CRAs for fundamental defects in their methodologies for rating structured finance products backed by subprime mortgages). 142 Bethel et al., supra note 2, at 25. See generally id., at 23-33 (discussing "what went wrong" in the credit crisis). COL UMBIA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 637 inadequate experience understanding the credit risks associated with new types of subprime loans, such as hybrid ARMs. 143 Some commentators have also noted the prevalence of herding behavior among CDO structurers eager to achieve investment grade ratings. 144 Once a structure obtained a desired rating from an agency, other issues copied that structure.'45 If the rating agency failed to capture a risk inherent in the CDO structure, that risk would be inadequately captured in the rating of several CDOs. 1'46 From a different perspective, government officials are inquiring whether sponsors of MBS and CDOs did not disclose information about high risk loans to CRAs. 147 In addition to these explanations, CRAs have come under fire for allowing the issuer pays conflict to improperly influence structured finance ratings.' This section explores 143 Id. at 23-24 ("In part the answer will likely involve the experience or lack thereof of market participants with this particular asset class .... The credit risks of the pools of mortgages that included subprime loans, especially hybrid ARMs, were different than the credit risks of mortgage pools previously securitized."). 14 See id. at 27-28 (explaining that the practice of copying a novel CDO structure that has obtained an investment grade rating can result in subsequently correlated downgrades if a rating agency did not receive full disclosure regarding risk inherent in the structure or if the agency failed to fully appreciate the risk). 141 Id. at 27 ("Once a relatively novel CDO's senior tranche is structured to receive an investment-grade rating, other CDOs tend to mimic that structure." (citing Peter Tufano, Financial Innovation and First Mover Advantages, 25 J. FIN. ECON. 213, 213-40 (1989))). 146 Id. at 27-28. 147 Id. at 17-18. '" See, e.g., Coffee, supra note 2 ("Striking new evidence supports ... that structured finance products have been the subject of rampant 'grade inflation."'); Arthur Levitt, Jr., Conflicts and the Credit Crunch, WALL ST. J., Sept. 7, 2007, at A15 (stating that the issuer pays conflict deepened with the increase in structured finance products and that the lucrative business of rating these products calls into question the objectivity of the ratings); Aaron Lucchetti, Ratings Firms' Practices Get Rated-SEC Probes if Conflicts Fueled Subprime Troubles, WALL ST. J., Sept. 7, 2007, at C1 ("In the wake of mortgage-market turmoil, regulators plan to probe how the big credit-rating companies are paid and whether they are independent enough of the Wall Street firms that issue bonds."). COLUMBIA BUSINESS LAW REVIEW three key reasons why CRAs may have been subject to greater pressure from issuers to inflate ratings on these products than on corporate securities. First, the market for structured finance ratings began to increase dramatically by 2003, becoming an important source of growth for CRAs. Second, the economics of the structured finance ratings market differ from that of the corporate securities rating market in simple, but critical ways. Third, the ratings process for products like RMBS and CDOs presents a greater opportunity for an issuer to exert pressure on a CRA. Combined, these differences indicate that the issuer pays conflict may have contributed to the relatively poor performance of recent structured finance ratings. 1. The Growth of Structured Finance Products The early-to-mid 2000s witnessed incredible growth in structured finance product issuance. 141 Indeed, between 2004 and 2006, CDO issuance alone increased from $157 billion to $552 billion.15 1 In turn, the volume of RMBS and CDO issuances rated by the CRAs increased substantially between 2002 and 2006, and structured finance became vital to the growth of the agencies' business.' For example, Moody's reported that its structured finance ratings revenue grew 87% between 2003 and 2006 and accounted for more than half of its ratings revenue in 2006.152 Perhaps more 149 Coffee, supra note 2 (describing growth of structured finance as explosive). 150 SECURITIES INDUSTRY AND FINANCIAL MARKETS ASSOCIATION, GLOBAL CDO MARKET ISSUANCE DATA, available at http://www.sifma.org/ research/pdffSIFMACDOIssuanceData2008.pdf. '5' SEC JULY 2008 REPORT, supra note 47, at 10; Hunt, supra note 3, at 172 (citing Joseph R. Mason & Josh Rosner, Where Did the Risk Go? How Misapplied Bond Ratings Cause Mortgage Backed Securities and Collateralized Debt Obligation Market Disruptions 17 (May 3, 2007) (unpublished working paper) (noting that structured finance ratings have been critical to the growth of the CRAs' businesses)). 152 Hunt, supra note 3, at 173 (citing Moody's Corp., Annual Report (Form 10-K), at 76 (Feb. 29, 2008); Moody's Corp., Annual Report (Form 10-K), at 39 (Mar. 22, 2002)). [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 639 importantly, structured finance produced two thirds of Moody's ratings revenue growth during this time period." 3 Structured finance was not only critical to revenue growth; some sources suggest that for the CRAs, rating structured finance products may have been highly profitable." While the growth of structured finance does not alone explain why the CRAs may have been more susceptible to the issuer pays conflict when rating these products, it does suggest that the success of the agencies during the relevant time period was quite dependent on the generation of business from structured finance issuers. 2. The Concentration of the Customer Base An important difference between the markets for rating structured products and for rating corporate bonds is the greater importance a structured finance issuer may play, compared to a corporate issuer, to a CRA's revenue base. The market for rating corporate bonds consists of thousands of small issuers, each accessing the debt markets on an occasional basis."55 For traditional products, CRAs typically charge a fixed fee of two to three basis points of the amount of the offering. 156 Given the infrequency with which corporate issuers purchase ratings and the relatively low level of fees generated for a corporate debt issuance, each corporate issuer pays an insignificant amount of fees.'57 In contrast, the client base in the structured finance market is far smaller. Based on a review of 642 sample deals, the SEC found that twenty-two different arrangers 153 Id. 154 See, e.g., Coffee, supra note 57 ("[Tlhe credit-rating business was destabilized by the extraordinary growth and profitability of structured finance."). 155 COFFEE, supra note 7, at 305 (noting that CRAs cover thousands of corporate issuers); Coffee, supra note 31 (explaining that large public corporations only access the debt markets periodically). 116 COFFEE, supra note 7, at 286. 7 Id. at 305 (noting that no corporate issuer pays "fees approaching the same order of magnitude as the annual multi-million dollar fee paid by Enron to Arthur Andersen"). performed the underwriting function on RMBS deals, but twelve of these arrangers accounted for 80% of the deals by number and by dollar volume." 8 The SEC also found that for a sample of 368 CDOs of RMBS deals, twenty-six arrangers performed the underwriting function, with eleven accounting for 92% of the deals and 80% of the dollar volume.19 Moreover, the twelve largest CDO underwriters were also twelve of the largest thirteen RMBS underwriters. 60 Thus, several arrangers are frequent players in the structured finance market.16 On top of this concentration, CRAs typically charge higher fees on asset securitizations given their complexity. While Moody's reportedly charges $10,000 to $25,000 on a traditional issuance, depending on its size, a complex deal may result in a fee of $90,000.162 The result is that certain clients may contribute materially to a CRA's structured finance revenue base. 163 Importantly, the underwriter of the deal has significant influence over which CRAs rate their deals. 66 A CRA may logically choose to inflate its ratings to appease a client that provides the agency with a material amount of its structured finance revenues, particularly as structured finance becomes a more important component of the company's overall revenue stream. 3. The Structured Finance Ratings Process Many commentators have noted the close interaction that occurs between the arranger and the CRA hired to rate the 158 SEC JULY 2008 REPORT, supra note 47, at 32. 159 Id. 160 Id. 161 See Coffee, supra note 31 (commenting on the frequency with which investment banks brought asset securitizations to market after 2000). 162 COFFEE, supra note 7, at 296-97 (noting that agency fees on asset securitizations are typically higher); id. at 309 n.15 (stating that fee of $90,000 not uncommon for complex deals). "8 Coffee, supra note 2 ("[Tlhe investment bank is a repeat player that accounts for a material share of the rating agency's structured finance revenues and that could take that business elsewhere.'). 164 SEC JULY 2008 REPORT, supra note 47, at 31. COL UMBIA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RA TING AGENCIES AND THE CREDIT CRISIS 641 deal.'65 Securitizations are ratings driven transactions; if the structurer cannot achieve the desired ratings, the transaction does not go forward. 16 Thus, a process has apparently developed where the rating agency and arranger have a back-and-forth dialogue which allows the arranger to adjust the transaction in order to achieve the necessary rating.67 It does not appear that there is another type of asset class for which CRAs may be so severe a bottleneck for an issuer seeking to accomplish a transaction. Reports suggest that ratings shopping has been prevalent; if an underwriter does not get the ratings it wants from one CRA, it will try to have another CRA rate the security.'68 This is particularly problematic because a CRA is usually only paid if it actually issues the rating, although it may receive a fee if it completed work but did not release a rating. 169 Data suggests underwriters are sensitive to ratings criteria when choosing a CRA, as Moody's reported that its market share declined from 75% to 25% when it increased its rating standards. 170 For the arranger, switching rating agencies is low visibility and appears nearly costless.' 7 ' Thus, arrangers can credibly threaten to remove 165 Coffee, supra note 2 ("[TIhe relationship between the ratings agency and the investment bank becomes close to the point of incestuousness."). 16 Hill, supra note 7, at 49. 167 Coffee, supra note 2 ("If the ratings agency will not deliver the desired investment grade rating on the first request, the investment bank will ask what more must it do. Gradually, the securitized pool of assets is improved and 'sweetened' until it just makes it over the 'investment grade' line."). 16 Lucchetti, supra note 148 ("Issuers often work with the rating firms to restructure the securities that are deemed high-risk, or even attempt to get another firm to rate the bonds."). 169 SEC JuLY 2008 REPORT, supra note 47, at 9. 170 Lucchetti, supra note 148. 171 Coffee, Mortgage Meltdown, supra note 9 ("[F]iring an auditor is difficult because SEC rules require full disclosure and permit the auditor to comment. Precisely because issuers usually hire multiple rating agencies, they can drop one with less visibility or adverse consequences .... In short, business in the market for ratings is mobile, retaliation is COLUMBIA BUSINESS LAW REVIEW business from a CRA that does not provide the customer with the high credit ratings it desires. 4. A Summary, and Where Do We Go From Here? To summarize the above discussion, the CRAs had much to gain from inflating their ratings on structured finance products when issuance of these securities increased in the early-to-mid-2000s. Each CRA had the opportunity to realize substantial growth in revenue and profitability from rating these securities. The price of this growth, however, may have been to sacrifice some of the integrity in the ratings process. With each potential customer bringing the possibility of frequent repeat business of a profitable nature, a rational CRA may have chosen to inflate ratings to lock in business. Indeed, as the structured finance market continued to develop, a CRA may have feared losing its opportunity to establish a dominant position in rating these products if it failed to appease arrangers from the beginning. An interactive ratings process that provided the customer with significant opportunity to exert pressure may have only exacerbated the problem. Moreover, while the agencies had the chance to reap substantial rewards from inflating their ratings, the costs of doing so did not simultaneously increase.172 As recounted earlier, the agencies have faced no meaningful legal or regulatory sanctions for issuing ratings of poor quality."7 3 The primary cost that a CRA might incur from releasing a rating that later proves too high is a reputational cost. This cost should not be overlooked. For this cost to deter a CRA from inflating its structured finance ratings, however, the relatively costless, and hence the gatekeeper becomes compromised, particularly with regard to structured finance products."). 172 While Congress enacted the Credit Rating Agency Reform Act of 2006, 15 U.S.C. § 78o-7 (2009), rating agencies did not become subject to the provisions of the Act and the SEC's rules promulgated thereunder until September 2007. SEC JULY 2008 REPORT, supra note 47, at 22. Further, it is arguable whether the act and the associated rules increased accountability in any meaningful way. ... See supra Part III.A.3. [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 643 CRA would need to anticipate that the inflated ratings would impact its reputation for rating other types of products; since many of the securities at issue in the Credit Crisis were novel, the CRAs did not have reputations for rating them and thus had no reputation to damage. 17 4 At least one scholar has noted that the "[r]evenues from ratings on novel products seem likely to have been large enough to cause agencies to risk negative spillover."175 At a minimum, it is at least plausible that some CRAs, rather than being shut out of the market for rating structured finance products, opted instead to inflate their ratings to some degree. Thus, one explanation for why CRAs may have inflated their ratings on structured finance products is that the gains of doing so increased while the costs of compromising their ratings essentially stayed the same. Indeed, this is consistent with Professor Coffee's "deterrence explanation" of gatekeeper failure. 176 As Professor Coffee explains, "[e]conomics 101 teaches us that when the costs go down while the benefits associated with an activity go up, the output of the activity will increase."177 At least in the short- term, it may have been economically rational for CRAs to inflate ratings on subprime RMBS and subprime-backed CDOs. 17 If this is the case, then a possible regulatory 174 Hunt, supra note 3, at 114 ("It is not plausible to argue that rating agencies have a valuable reputation for rating instruments they have never rated before."). See Bethel et al., supra note 2, at 23 (noting the lack of experience many market participants had with the asset class at the center of the Credit Crisis controversy). 17 Hunt, supra note 3, at 172. 176 See generally COFFEE, supra note 7, at 60-61. See also Coffee, Gatekeeper Failure, supra note 9, at 318-23 (describing the deterrence explanation as a possible reason for the failure of auditors to detect Enron and other accounting frauds in 2001-2002). 177 See Coffee, Gatekeeper Failure, supra note 9, at 318. 171 See Frank Partnoy, Barbarians at the Gatekeepers: A Proposal for a Modified Strict Liability Regime, 79 WASH. U. L.Q. 491, 497-98 (2001) ("[A]bsent from much of this scholarly debate is a focus on the argument that economically rational gatekeepers, like economically rational issuers, will balance the benefits and costs of accurate (or inaccurate) certification in an effort to maximize profits. On the one hand, gatekeepers may achieve short-term gains by providing inaccurate certification or by response should focus on imposing a cost on CRAs of acquiescing to the issuer pays conflict. An optimal response would only impose such a cost to a degree that it would neutralize any benefit to the CRA of engaging in ratings inflation. In theory, the regulatory response could apply to all classes of ratings, not only structured finance products, so long as it corresponded accurately to any ill-derived gains. A second possible explanation may flow from the bubble nature of the housing and related structured finance market that developed after 2000.179 Professor Coffee has elaborated a "market bubble" account of gatekeeper failure.18 ° This theory posits that in an environment of market euphoria, investors are less concerned with gatekeepers' certifications, so those purchasing the gatekeeper's service regard the services as more of a formality.'81 The gatekeeper loses leverage over its client and experiences a reduction in the value of its reputational capital. 18 2 As a result, the rational best response of the gatekeeper is to acquiesce to the client's wishes and minimize the costs associated with its service. 183 Under this hypothesis, CRAs may have recognized the bubble quality of the structured finance market and realized that investors were less concerned with obtaining accurate ratings on structured products than with capitalizing on the higher returns the securities offered. Without investors overstating the value of securities. On the other hand, gatekeepers could suffer long-term losses from a decline in reputation if investors realize that the gatekeeper originally overstated the assessment of value. Although investors will discount the securities of issuers who use disreputable intermediaries, gatekeepers should be willing to incur losses in reputational capital so long as the gains from inaccurate certifications exceed the costs."). "' Coffee, supra note 31 (characterizing the market as bubbly). 180 Coffee, Gatekeeper Failure, supra note 9, at 323-26 (elaborating on the theory in the context of auditors). 181 Id. 182 Id. at 323. " Id. at 324 ("Thus, if we assume that the auditor will be largely ignored by euphoric investors, the rational auditor's best competitive strategy (at least for the short term) was to become as acquiescent and low cost as possible."). COL UMIA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS demanding accurate ratings, issuers had no reason to demand a high-quality product from the CRAs. To maximize its business in this market, therefore, the CRAs may have chosen to compete by appeasing issuers. It is not necessary to determine whether the deterrence explanation or the market bubble explanation better accounts for the CRAs' seemingly compromised ratings. The two theories are complementary rather than contradictory."' Under both theories, the CRAs seemingly inflated their ratings to reap the gains presented by the growth of the structured finance market, and a possible regulatory response may flow from increasing the costs to the CRAs of engaging in this undesirable conduct. The remainder of this Note seeks to evaluate whether the current regulatory regime, which is largely reliant on disclosure, will adequately increase the costs to the CRAs of engaging in ratings inflation, thereby optimally deterring this behavior. Concluding that the disclosure-based regime is not sufficient, the Note explores and critiques two proposals that explicitly impose penalties, referred to as performance-based sanctions, on CRAs for low-quality ratings. IV. COUNTERING THE "ISSUER PAYS" CONFLICT: IS DISCLOSURE SUFFICIENT? Congress enacted the Credit Rating Agency Reform Act in 2006, and the original rules promulgated by the SEC became effective in September 2007.1 Thus, by the time the CRAs became subject to this regulation, the low quality of the ratings on many subprime RMBS and CDOs had already been revealed."8 6 This section seeks to evaluate whether disclosure offers a promising solution for deterring CRAs " Id. at 330 (noting that the explanations are complementary). Professor Coffee does note that if one accepts the bubble explanation as the cause of gatekeeper failure, the problem may be self-correcting. Id. at 330. 185 Registration of Nationally Recognized Statistical Rating Organizations, 15 U.S.C. § 78o-7 (2009); SEC JuLY 2008 REPORT, supra note 47, at 15. "' See supra Part II.B.4. from acquiescing to the issuer pays conflict. It begins by describing the disclosure presently required under the Act and the SEC's rules. Recently adopted amendments to these rules as well as proposals currently under consideration will also be discussed."8 7 The section then explores the theory underlying disclosure as a regulatory tool in this context and suggests reasons why disclosure may not be sufficient to correct against a similar conflict of interest problem in the future. A. Current and Proposed Disclosure-Based Regulations 1. The Original Regulatory Requirements The 2006 Act requires that an applicant for the NRSRO designation disclose any "conflict of interest relating to the issuance of credit ratings by the applicant" on its Form NRSRO, the application document that an applicant must submit to the SEC. 8' Once an applicant is granted NRSRO status, it must make the information provided in its application available on its website or through another "comparable, readily accessible means."189 While Congress gave the SEC authority to prohibit conflicts of interest, including those stemming from the issuer pays model, the SEC opted only to require that NRSROs establish, maintain, and enforce written policies and procedures to manage the 187 On December 3, 2008, the SEC adopted amendments to the rules it originally promulgated under the Act and proposed additional amendments presently under consideration. Press Release, SEC, SEC Approves Measures to Strengthen Oversight of Credit Rating Agencies (Dec. 3, 2008) [hereinafter SEC Dec. Press Release], available at http://www.sec.gov/news/ press/2008/2008-284.htm; Amendments to Rules for Nationally Recognized Statistical Rating Organizations, Exchange Act Release No. 59,342, 2009 SEC LEXIS 239 (Feb. 2, 2009); Re-proposed Rules for Nationally Recognized Statistical Rating Organizations, Exchange Act Release No. 59,343, 2009 SEC LEXIS 240 (Feb. 2, 2009). 188 15 U.S.C. § 78o-7(a)(1)(B)(vi); 17 C.F.R. § 249b.300 (2009). 189 15 U.S.C. § 78o-7(a)(3). COL UMBIA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 647 conflict. 10 An agency has to disclose these procedures as an exhibit to its Form NRSRO.15 ' On the one hand, Congress gave the SEC exclusive authority to enforce the provisions of the Act if an NRSRO issues ratings in "material contravention" of the procedures it has developed and disclosed for managing conflicts, and the SEC may revoke the registration of an NRSRO that does so.192 On the other hand, this would necessarily result in a fact-intensive inquiry that would require a substantial commitment of investigative resources on the part of the SEC. Thus, policymakers are primarily relying on disclosure of the conflicts and of the procedures an NRSRO has in place for managing those conflicts to protect investors. 93 190 15 U.S.C. § 78o-7(h)(2)(A) (granting the SEC authority to prohibit conflicts of interest, including conflicts related to the way an obligor compensates the CRA); 17 C.F.R. § 240.17g-5(a)(1)-(2) (prohibiting an NRSRO from having a conflict of interest unless it has disclosed and taken the necessary steps to manage it). 17 C.F.R. § 240.17g-5(c) does in fact prohibit NRSROs from having specific types of conflicts. An NRSRO may not issue or maintain a credit rating 1) solicited by a person that provided the NRSRO with 10% or more of the NRSRO's total net revenue in the most recently ended fiscal year; 2) where the NRSRO, a credit analyst that participated in determining the rating, or a person responsible for approving the rating directly owns securities of, or has any other direct ownership interest in, the person subject to the rating; 3) where the person rated is associated with the NRSRO; and 4) where a credit analyst who participated in determining the rating, or a person responsible for approving the rating, is an officer or director or the person subject to the rating. 1 17 C.F.R. § 249b.300 (required as Exhibit 7). 192 15 U.S.C. § 78o-7(c)(1)(A)-(B); see also Kara Scannell & Aaron Lucchetti, Credit Crunch: SEC to Seek Rule on Added Disclosure by Bond- Rating Firms-Proposals Unlikely to Quell Criticism That More Be Done, WALL ST. J, June 11, 2008, at C2 (paraphrasing Chairman Christopher Cox that an NRSRO that violates its own rating policies and procedures may have its registration revoked). 193 See Hunt, supra note 3, at 154 ("The SEC's rules seem to reflect the view that no serious regulatory intervention is needed to deal with agency conflicts of interest."). COLUMBIA BUSINESS LA W REVIEW 2. Rules Amendments Adopted in December 200819 On December 3, 2008, the SEC adopted final rule amendments to supplement their existing rules. These amendments were originally proposed to "address concerns about the integrity of [the] credit rating procedures and methodologies in the light of the role they played in determining credit ratings for" subprime RMBS and CDOs.195 Specifically, an NRSRO must now provide transition statistics for each asset class for which it is registered or seeking registration over one-year, three-year, and ten-year periods. 19' While this amendment specifies with greater particularity what performance statistics an NRSRO needs to disclose, it is unclear how much this requirement adds to the performance statistics that CRAs already disclose voluntarily. Moreover, a registered agency must provide enhanced disclosure regarding the due diligence it undertakes when rating structured finance products and how this due diligence is incorporated into the rating process. 197 In addition, an NRSRO must provide on its 194 SEC Dec. Press Release, supra note 187; Amendments to Rules for Nationally Recognized Statistical Rating Organizations, Exchange Act Release No. 59,342, 2009 SEC LEXIS 239 (Feb. 2, 2009) (containing final text of adopted amendments). The adopted amendments were originally proposed in June 2008. Proposed Rules for Nationally Recognized Statistical Rating Organizations, Exchange Act Release No. 57,967, 2008 SEC LEXIS 1613 (June 16, 2008). Again, while the amendments adopted by the SEC are aimed at correcting problems beyond the conflicts issue, this section summarizes those disclosure-based regulations intended to address the issuer pays conflict. '9 SEC Dec. Press Release, supra note 187; Proposed Rules for Nationally Recognized Statistical Rating Organizations, Securities Exchange Act Release No. 57,967, 2008 SEC LEXIS 1613, at *1 (June 16, 2008) (explaining purpose of the then-proposed amendments). 196 Amendments to Rules for Nationally Recognized Statistical Rating Organizations, Exchange Act Release No. 59,342, 2009 SEC LEXIS 239, at *183-85 (Feb. 2, 2009). 197 Id. at *185-86 (requiring that NRSROs provide, for example, information on "whether, and if so, how information about verification performed on assets underlying or referenced by a security or money [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 649 website a random sample of 10% of its issuer-paid ratings and the histories of those ratings for each ratings class for which the NRSRO is registered and has issued 500 or more issuer-paid ratings.198 An NRSRO, however, has six months to include this information on its website from the time the rating action is taken. 99 The SEC did adopt one noteworthy amendment that might strike a blow to a part of the structured finance ratings process that may be problematic. The SEC has amended Rule 17g-5 to prohibit an NRSRO from issuing a rating where the NRSRO made recommendations about the "corporate or legal structure, assets, liabilities, or activities" of the issuer.2 °0 How far reaching this amendment will be remains unclear. The agencies maintain that they refrain from making these recommendations.2 "1 More importantly, while the prohibition seems to prohibit the iterative, interactive ratings process that many have criticized, the Commission indicated that the amendment is not intended to stop issuers from engaging in a feedback process with a CRA. °2 Rather, the Commission is primarily concerned with CRAs "rating their own work."203 market instrument issued by an asset pool or as part of any asset-backed or mortgage-backed securities transaction is relied on in determining credit ratings"). 198 Id. at *178-79. '9 Id. at *179. 200 Id. at *181-82. 201 See Hunt, supra note 3, at 154 n.155 (questioning the practical effect of this rule). 202 Open Meeting of the U.S. Securities and Exchange Commission (Dec. 3, 2008), http://www.connectlive.com/events/secopenmeetings/2008 index.html. This limitation became clear during a dialog between Chairman Christopher Cox and Erik Sirri, Director, Division of Trading and Markets. 203 Id. COLUMBIA BUSINESS LA W REVIEW 3. Proposed Amendments 0 4 The SEC is presently considering two additional disclosure-based rules that are relevant to the issuer pays conflict. First, the SEC is considering extending the newly adopted requirement that an NRSRO disclose a random sample of 10% of its issuer-paid ratings to require the NRSRO to disclose 100% of these ratings.2 5 This rule would, however, provide the agencies with a twelve-month time lag to disclose the information and would apply only to ratings issued after June 26, 2007.206 Perhaps more meaningfully, a second proposal would prohibit an NRSRO from providing a rating on a structured finance product that is paid for by the issuer, sponsor, or underwriter unless the information provided to the NRSRO to rate and perform ongoing monitoring on the product is also made available to other NRSROs. ° NRSROs seeking to access this information would need to certify to the SEC that they intend to use the information only to provide credit ratings and that they will produce a minimum number of ratings based on the information. 08 The idea behind the proposal appears to be that CRAs would be able to issue unsolicited ratings on instruments, thereby uncovering any inflated ratings. In theory, this requirement would provide an NRSRO with increased leverage over an underwriter that 204 Again, this summary emphasizes amendments geared at mitigating the issuer pays conflict through disclosure. The SEC's present proposals regarding NRSROs are far broader. Re-proposed Rules for Nationally Recognized Statistical Rating Organizations, Exchange Act Release No. 59,343, 2009 SEC LEXIS 240 (Feb. 2, 2009); SEC July 2008 Proposed Amendments, supra note 102. 205 Re-proposed Rules for Nationally Recognized Statistical Rating Organizations, Exchange Act Release No. 59,343, 2009 SEC LEXIS 240, at *149 (Feb. 2, 2009). 206 Id. 207 Id. at *150-54 (outlining process for implementing this proposal). 208 Id. at *153-54. An NRSRO must certify that it will rate at least 10% of the securities for which it accesses information if the NRSRO chooses to access information for ten or more securities in the calendar year covered by the certification. [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 651 is pressuring it to inflate its ratings. Whether this proposal would rectify the problem, however, is uncertain. On the one hand, an NRSRO may have an incentive to issue these unsolicited ratings as a competitive strategy to increase business. On the other hand, a truly widespread practice of issuing unsolicited ratings is costly for any one NRSRO, and the practice would need to be fairly widespread to uncover systematic inflation. In addition, an issuer may resist switching its business to the more conservative CRA that gave it a lower rating absent some functioning market mechanism that penalizes the issuer for failing to switch. Beyond that, an NRSRO could abuse the practice of issuing unsolicited ratings to strong arm issuers into purchasing its ratings. In the past, Moody's was heavily criticized for doing this and largely retreated from the practice of issuing unsolicited ratings.2°9 Consistent with the existing regulatory framework, the proposed amendments largely rely on disclosure requirements to protect investors from the problems that the issuer pays conflict creates. Indeed, SEC Chairman Christopher Cox acknowledged this when he stated that the SEC "want[s] investors to understand what they're looking at.210 B. The Theory Underlying Disclosure-Based Regulations In order to evaluate whether a disclosure-based regulatory regime can effectively deter CRAs from favoring the interests of their clients over investors, it is helpful to explore the theory underlying why disclosure is favored as a regulatory tool. In the United States, disclosure is a 209 Hill, supra note 7, at 52; Partnoy, supra note 7, at 651 n.154. 210 Scannell & Lucchetti, supra note 192 ("The bulk of the agency's proposals will focus on disclosure aimed at holding the bond-rating firms more accountable and encouraging them to change their behavior."). COLUMBIA BUSINESS LA W REVIEW touchstone of the federal securities laws.21' So long as investors receive full and fair disclosure of all relevant information regarding a security on the market, investors are considered adequately protected because they can evaluate the merits of the security for themselves.212 One problem underlying this notion is that investors are heterogeneous and differ in their abilities to understand and evaluate information.213 The basic solution to this is known as the efficient markets hypothesis. In simple terms, this hypothesis posits that in an efficient market, stock prices will reflect all publicly available information. 214 Furthermore, they will do so almost instantaneously because of the large number of sophisticated investors trolling the market for informational advantage.215 Therefore, it is not 211 See THOMAS LEE HAZEN, TREATISE ON THE LAW OF SECURITIES REGULATION 740 (4th ed. 2002) ("[Fjederal securities law's exclusive focus is on full disclosure."). 212 See id. at 28 ("The theory behind the federal regulatory framework is that investors are adequately protected if all relevant aspects of the securities being marketed are fully and fairly disclosed. The reasoning is that full disclosure provides investors with sufficient opportunity to evaluate the merits of an investment and fend for themselves. It is a basic tenet of federal securities regulation that investors' ability to make their own evaluations of available investments obviates any need that some observers may perceive for the more costly and time-consuming governmental merit analysis of the securities being offered."). 211 See Steven L. Schwarcz, Rethinking the Disclosure Paradigm in a World of Complexity, 2004 U. ILL. L. REV. 1, 17 ("There is, after all, a wide variation of expertise among investors, and no doubt some investors are incapable of fully understanding at least some of the information disclosed."). 214 Id. (citing CHARLES R. T. O'KELLEY & ROBERT B. THOMPSON, CORPORATIONS AND OTHER BUSINESS ASSOCIATIONS 170-171 (3d ed. 1999) (referring to this belief as the semi-strong form of the efficient market hypothesis)). 211 Id. at 17-18 ("The explanation for how markets can assimilate new information so rapidly stems from the existence of a large number of sophisticated market investors who trade for their own account or for investors they represent. These professionals track all sources of information affecting the value of stock and act instantaneously to capture the 'profit' available by buying stock at the 'old' fair price." (citing O'KELLEY & THOMPSON, supra note 214, at 171) (emphasis added)). [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 653 necessary for all investors to understand every disclosure so long as a sufficient number do.216 Drawing a parallel to the regulatory regime in place for NRSROs, the SEC appears to believe that if an NRSRO fully and fairly discloses a particular conflict of interest, its procedures for managing those conflicts, and its ratings methodologies, an investor should be able to adequately evaluate the risk that a given rating is compromised by the disclosed conflict. Armed with this information, the investor will discount the value she places on the NRSRO's rating by the risk the conflict presents. For investors to make this calculation, the goal must be complete transparency; as Chairman Cox said, "[w]e want investors to understand what they're looking at."217 C. Why Disclosure is Insufficient for Deterring CRAs from Compromising Ratings While policymakers hope that a disclosure-based regulatory regime will increase accountability and incentivize CRAs to modify undesirable behavior, the emphasis on disclosure places the onus on investors to hold CRAs accountable.218 Even if CRAs are fully and fairly making mandatory disclosures on the issuer pays conflict, investors must adequately perceive and evaluate that information and must penalize CRAs via the market mechanism if disclosure is to adequately deter the agencies from reaping the gains of compromised ratings. This section explores some possible reasons why investors may not adequately perceive and evaluate the risks posed by the issuer pays conflict in rating structured finance products and why they may not convey, through the market mechanism, costs on CRAs for compromising their ratings. 216 Id. 217 Scannell & Lucchetti, supra note 192. 211 Id. ("The bulk of the agency's proposals will focus on disclosure aimed at holding the bond-rating firms more accountable and encouraging them to change their behavior."). COLUMBIA BUSINESS LAW REVIEW 1. Failure to Perceive or Anticipate Relevant Differences Between Structured Finance and Traditional Products Investors may have failed to adequately discount ratings on structured finance products issued by CRAs because they did not fully appreciate some of the relevant differences between ratings on these products as compared to traditional products. For example, even if CRAs fully disclosed the level of interaction that goes on between the CRA and an arranger when rating a structured product, investors may not have fully appreciated that this practice could result in and allow for ratings shopping by arrangers.219 Some commentators have noted that an interactive process is not necessarily inherently problematic.22 ° In addition, arrangers can move business from one rating agency to another without attracting much, if any, attention.221 Since the practice may appear innocuous and the detrimental result of ratings shopping may not be visible to investors, full disclosure of the ratings process itself may not alert them to the risk that the process can exacerbate the issuer pays conflict. The failure to perceive these differences between rating structured products and traditional products may have been exacerbated by the CRAs' claims that structured and corporate securities are essentially rated on the same scale. 2 If investors cannot adequately perceive differences in the ratings processes that exacerbate the issuer pays conflict, they may not adequately appreciate the amount of 219 See infra Part III.C.3; Lucchetti, supra note 148 ("Issuers often work with the rating firms to restructure the securities that are deemed high-risk, or even attempt to get another firm to rate the bonds."). 220 See Hunt, supra note 3, at 140 n.100 ("In principle, arrangers ought to be able to understand the methodologies in determining how to structure their offerings. Concerns about the fact that agencies and originators engage in a dialogue that allows the originator to optimize the offering with knowledge of the rating outcomes of various structures ... seem misplaced unless coupled with some specific argument that other market defects make the practice undesirable."). 221 Coffee, Mortgage Meltdown, supra note 9. 222 See supra Part II.A. [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 655 risk resulting from the conflict that is inherent in the rating. To the extent investors can use a market mechanism to penalize CRAs, they will underutilize it if they do not perceive the extent to which ratings have been compromised. 2. Failure to Understand the Rated Instruments Themselves For an investor to adequately discount the value of a rating based on risks posed by conflicts, investors must also make an assessment of the rating's informational value before taking the conflict into account. To do this, an investor must understand the transaction that the agency is rating--even if the CRA discloses its ratings methodologies and procedures in painstaking detail, this information is meaningless to someone who does not understand the transaction itself. But structured transactions are far more complex than traditional corporate securities, and structured finance products have become even more complex over time. 223 Therefore, perhaps part of the problem lies in investors' ability to understand the underlying transactions, rather than wholly in their failure to appreciate the full risk of the issuer pays conflict. 224 Indeed, one scholar has noted 223 Coffee, supra note 57 ("[Rlating structured-finance products is far more difficult than rating the bonds of public corporations."); SEC JULY 2008 REPORT, supra note 47, at 11 (noting the increasing complexity of structured products). 224 This seems even more plausible since some commentators have noted that the CRAs themselves likely did not have enough information to fully appreciate the risks inherent in the transactions they were rating. For example, sometimes CRAs did not have information on the collateral underlying CDOs. See Credit Rating Agencies and the Financial Crisis: Hearing Before the H. Comm. on Oversight and Gov't Reform, 110th Cong. (2008) (opening statement of Rep. Henry A. Waxman, Chairman) (reporting that one former S&P analyst was told that a request for loan level information when rating a CDO was "totally unreasonable"); SEC JULY 2008 REPORT, supra note 47, at 36 (describing how CRAs relied on credit ratings on the RMBS in the asset pool); Bethel et al., supra note 2, at 25 (noting how reliance on data regarding performance of subprime loans between 2001-2005, given strong performance during that time, could have resulted in ratings being too high when rating later vintages). COLUMBIA BUSINESS LAW REVIEW that "most investors do not have the ability to evaluate structured transactions."22 If we were dealing with actively traded equity securities, the efficient market hypothesis would correct for this problem.226 Under that hypothesis, not all investors would need to comprehend the underlying transaction because the price of the security would reflect the most sophisticated investors' understanding of the deal as well as their evaluation as to whether the security's rating was compromised by conflicts. However, although ultimately an empirical question, the efficient markets hypothesis likely does not hold in markets for RMBS and CDOs. Public and private debt markets are often not efficient like markets for equity securities. 27 True, purchasers of these securities tend to be among more sophisticated investors, such as hedge funds, banks, pension funds, corporations, mutual funds, and life insurers.22 ' That being said, it seems likely given the growth 225 Schwarcz, supra note 213, at 12. 226 See supra Part IV.B. 227 John C. Coffee, Jr., The Securitization Bubble, N.Y.L.J., Mar. 17, 2008 ("Debt markets are seldom efficient in the way that equity securities markets are, and the 'fraud on the market' doctrine will likely be inapplicable."); Schwarcz, supra note 213, at 19 ("Furthermore, the efficient market hypothesis might not even apply to public or private bond markets, yet investors in those markets suffer from the same information asymmetry discussed in this article."); Morey W. McDaniel, Bondholders and Stockholders, 13 J. CORP. L. 205, 242 (1988) ("There is evidence that the market for corporate bonds is not very efficient. For many bond issues, it is not unusual to find infrequent trading activity and large spreads between bid and asked prices." (citations omitted)); Yedidia Z. Stern, A General Model for Corporate Acquisition Law, 26 J. CORP. L. 675, 709 (2001) ("[Sltudies show that the bond market is not efficient; and therefore, one cannot expect the market prices to compensate bondholders for the risks to which they are exposed."); see also Schwarcz, supra note 213, at 18 (noting that complexity can undermine the efficient markets hypothesis because "fewer than the necessary critical mass of investors understands the disclosure"). 22 Bethel et al., supra note 2, at 14 (listing the types of investors who typically purchase RMBS and CDOs). See, e.g., Schwarcz, supra note 213, at 4-6 & n.36 ("Disclosure of a complex and convoluted structure to investors of the originator-the audience on which this article primarily [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 657 in issuances of these complex securities over recent years that at least some purchasers were not among the most sophisticated investors. If these investors did not understand the transactions, they may have underestimated the risk that conflicts compromised the ratings. 3. Failure to Question the Accuracy of Ratings Another reason why investors may have failed to discount the value of ratings despite the disclosure of conflicts is that investors failed to scrutinize the informational value of ratings, period. At least one scholar has noted that it is not clear whether some investors do anything besides rely on the rating as a simple heuristic.229 Thus, investors may not be exploiting the disclosure available to them to conduct their own analytic assessments regarding the risk of the rated securities and checking those assessments against the assigned rating. If investors are failing to scrutinize the value of a rating via an independent assessment, the use by the CRAs of the same ratings scale for structured and traditional products could have exacerbated the problem, since a structured product with an equivalent rating to a corporate security often paid a higher yield.Y0 Investors may have simply preferred to buy the security that provided the highest rate of return within a particular rating class. focuses-may well be either too detailed for many of those investors, even institutional investors, to understand and assimilate, or too superficial to allow investors to fully assess the transaction and its ramifications.... Those investors are, of course, the originator's shareholders and bondholders.... They should be distinguished from the very narrow and highly specialized class of sophisticated investors in securities issued by the SPEs that are parties to the originators structured transactions."). 22" See Steven L. Schwarcz, Complexity as a Catalyst of Market Failure 13 (Duke Law Sch. Pub. Law & Legal Theory Paper, Paper No. 217, 2009), available at http://ssrn.com/abstract=1240863 (discussing how investment analysts may rely on ratings as a simple heuristic). 230 See Aaron Lucchetti, Kara Scannell & Craig Karmin, SEC Aims to Rein In the Role of Ratings, WALL ST. J., June 24, 2008, at Cl; Lahart, supra note 35 (noting that MBS and mortgage-backed CDOs were popular because they offered higher rates of return than similarly rated traditional bonds). 4. Ratings-Based Regulation Finally, even if CRAs fully and fairly disclosed all information necessary for investors to adequately perceive how and to what extent the issuer pays conflict impacted the agencies' ratings, disclosure can only deter CRAs to the extent investors can penalize them for producing low integrity ratings. The ability to punish CRAs in this manner requires a functioning market mechanism, and the prevalence of incorporating credit ratings in regulatory frameworks may impair this market mechanism.231 Unless investors can force issuers to withdraw business from rating agencies that compromise their ratings, disclosure will not produce optimal incentives for CRAs to maintain integrity in their ratings processes. V. PERFORMANCE-BASED SANCTIONS AS A SOLUTION? If it is true that the current regulatory focus on disclosure is unlikely to convey adequate costs on CRAs to deter them from inflating ratings on structured finance products, what can regulators do to fix this problem? That is, what regulatory mechanism could they adopt to impose costs on the agencies that will, in a sense, cancel out any gains that result from compromising the integrity of their ratings process? One possible solution is to subject CRAs to some sort of sanction if it is shown on an ex post basis that an agency's ratings were too high. This Note refers to these types of penalties as performance-based sanctions. Indeed, two scholars, Professor John Coffee and John Patrick Hunt, have proposed solutions since the onset of the Credit Crisis that would impose this sort of penalty on CRAs if their ratings 21 Coffee, supra note 2 ("[Blecause the ratings agency has a de facto licensing power, the issuer needs its rating even if the market distrusts the rating's informational value."). See also supra Part III.A. 1. COL UMBIA B USINESS LA W RE VIE W [Vol. 2009 No. 2:605] CREDIT RA TING AGENCIES AND THE CREDIT CRISIS 659 were shown to be of low quality over time.232 The remainder of this Note will review Professor Coffee's proposal (the "Coffee proposal") and Mr. Hunt's proposal (the "Hunt proposal") to explore generally the desirability of a performance-based sanction and whether either suggested solution might convey an appropriate cost on the CRAs for issuing inflated ratings. Before reviewing these proposals, it is worth stopping to consider whether policymakers should make any regulatory changes to (try to) prevent CRAs from issuing low integrity ratings in the future. The question is whether what happened with structured finance ratings during the Credit Crisis was a relatively short-term problem that will essentially correct itself without government intervention. Along these lines, the problem will naturally abate for at least some period since the market for structured products has contracted. 3 Indeed, if the market bubble explanation is the dominant reason why CRAs failed investors during this structured finance boom, regulatory intervention may be unnecessary. 34 In addition, if CRAs fear either market or regulatory repercussions now that public interest is focused on them (and particularly focused on the conflicts issue), the firms may correct the problem on their own.2 Therefore, it 232 Coffee, supra note 2 (outlining proposal); Coffee, Mortgage Meltdown, supra note 9 (outlining proposal); Hunt, supra note 3, at 182, 200-09 (outlining proposal and discussing implementation). 233 See Aaron Lucchetti & Kara Scannell, Moody's, S&P Answer Critics Over Bond Calls, WALL ST. J., Sept. 26, 2007, at C1 (describing how former Federal Reserve Chairman Alan Greenspan has said that more regulations for NRSROs are unnecessary because the issuance of structured finance products will decline due to concerns over CRAs' ability to rate the products). 21 Coffee, Gatekeeper Failure, supra note 9, at 330 (noting that gatekeeper failure may be self-correcting if the market bubble explanation is the dominant reason for gatekeeper failure in a particular context). See supra Part III.C.4 for an overview of the market bubble explanation. 235 Hunt, supra note 3, at 154 (noting that there have been actions in the face of regulatory pressure to address the issuer pays conflict); SEC JULY 2008 REPORT, supra note 47, at 25 (stating that the big three CRAs enhanced their procedures for addressing conflicts when they sought NRSRO registration). COLUMBIA BUSINESS LA W REVIEW may be premature to reinvent the regulatory wheel when CRAs have generally produced fairly accurate ratings over the years and have appeared, at least until now, to withstand the pressure of the issuer pays conflict: 23 Regulatory action may still be preferable, however, for two reasons. First, securitization is a beneficial financial tool when it is not abused. Regulatory action may be necessary to restore investor confidence in structured finance if ratings remain practically necessary for these products to be marketable. Second, as long as players in the capital markets innovate, there remains a risk that the next new security that comes along will create similar risks, especially since CRAs will likely compete intensely to capture the market for rating any new products. A. An Overview of Two Proposals 1. The Coffee Proposal Professor Coffee has suggested that the SEC tie NRSRO- status to the default rates experienced by an agency's credit ratings.237 According to Professor Coffee, the SEC could establish a maximum default rate for each letter grade and suspend the NRSRO status of an agency whose five-year default rate exceeded that maximum. 238 The agency's NRSRO-status would be suspended until its five-year default rate returned below the established parameter."9 While it would not be necessary, the SEC could limit the sanction imposed on a CRA to a particular asset class and a particular rating.24 ° For example, the SEC could determine that an NRSRO must maintain a five-year default rate below 5% for all asset-backed securities rated BBB; if a CRA's five-year default rate on asset-backed securities rated BBB exceeded 236 See supra Part III.B. 237 Coffee, supra note 2; Coffee, Mortgage Meltdown, supra note 9. " Coffee, Mortgage Meltdown, supra note 9. 239 Id. 240 Coffee, supra note 2. [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 661 this limit, the CRA's BBB rating on this class of assets would no longer be recognized by the SEC until the agency's five- year default rate declined below 5%.241 Professor Coffee recommended this proposal specifically in light of the issuer pays conflict problem. The idea is that the threat of losing NRSRO status would "make ratings agencies more conservative... and thus create a countervailing pressure to the current incentives that produce grade inflation."42 2. The Hunt Proposal While his proposal is intended to address the problem of low-quality ratings in general and not merely the issuer pays conflict, Mr. Hunt advocates for requiring a CRA to disgorge profits received from issuing ratings on new products that fall below a predetermined quality level unless the CRA reveals in advance that the ratings are of low quality.24 3 Mr. Hunt explains that the proposal could be implemented in a number of ways, including through private liability, an administrative system under the SEC, or private "' "The rating agency could continue to issue ratings during this interim, but they would be useful only for their informative value, not their legal impact." Coffee, Mortgage Meltdown, supra note 9. 242 Coffee, supra note 2. The idea of tying NRSRO status to ratings performance is not an entirely new idea. After the Enron debacle, two commentators, Fidelity and the Investment Company Institute, suggested implementing a process analogous to that used for broadcast licenses. Essentially, the SEC would hold public hearings where parties could comment on an agency's performance, and the SEC would take these comments into consideration when deciding whether to renew a CRA's NRSRO-status. This idea was suggested prior to the adoption of the Credit Rating Agency Reform Act of 2006. See Hill, supra note 7, at 88 (citations omitted). 24 3 Hunt, supra note 3, at 181. In his article, Mr. Hunt argues that reform efforts consistent with the reputational capital model of ratings agencies are incomplete because a well-functioning reputation mechanism will still not guarantee high-quality ratings of new products. When rating novel products, rating agencies have no reputation for rating the product. While low-quality ratings on new products could harm the reputation of the agency for rating other products, ratings agencies will still rate the novel product if the volume of that product is high enough. See id. COLUMBIA BUSINESS LAW REVIEW contracting. 244 He also provides several suggestions for measuring quality, such as the "accuracy ratio,"245 comparing the default rates of products across classes (i.e., by using the default rate of traditional bond ratings as a benchmark),246 or by setting the minimum quality level at the point below which the market for a product collapses. 47 In order to more easily compare Mr. Hunt's proposal to Professor Coffee's, the author assumes the Hunt's proposal would be implemented by a regulatory agency (most naturally the SEC). The disgorgement penalty would be applied for a given product type and for a given vintage.248 3. A Comparison of the Attributes Both the Coffee proposal and the Hunt proposal would impose a penalty on a CRA whose ratings underperformed relative to a predetermined quality metric. Each system could be implemented by the SEC, and both have a quasi strict liability element since a CRA would be sanctioned if its ratings missed the quality metric, irregardless of the standard of care employed by the CRA. In addition, while Mr. Hunt suggested several quality metrics that could be used to implement his proposal, we might assume for purposes of comparison that the proposals would be undertaken using the same quality metric. The proposals differ in four primary respects. First, Professor Coffee's proposal would sanction a CRA by removing its NRSRO designation for a particular asset class and letter category. This penalty would primarily impact the NRSRO's future revenues and profitability, since investors would not be able to rely on the CRA's rating for the effected asset class and letter grade until the agency's ratings began performing above the required quality threshold. In 244 Id. at 207. 245 Id. at 201 (noting that the accuracy ratio is a "way of expressing the performance of an ordinal rating system in terms of a single number"). 246 Id. at 203, 205. 247 Id. at 205. 248 Id. at 206. [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 663 contrast, the Hunt proposal would force an underperforming CRA to disgorge profits that the CRA had already earned. Second, a CRA sanctioned under the Coffee proposal would lose its NRSRO designation for a particular asset class and letter grade, while an agency penalized under the Hunt proposal would lose the profits it earned over a set time period for a product type and vintage. Thus, Mr. Hunt's proposal would review a CRA's performance in more narrow increments because a product type may not be construed as broadly as an asset class and because the performance of each product type would be reviewed by year of issuance. Third, while both proposals do not inquire into a CRA's standard of care, Mr. Hunt would grant a reprieve if the CRA disclosed ahead of time that its ratings were of low quality. The Coffee proposal does not provide such leeway to the agencies. Fourth, Mr. Hunt's proposal is concerned only with ratings on new products, whereas the Coffee proposal does not distinguish between novel and established securities. B. Shared Positives and Negatives Both the Coffee and Hunt proposals have several shared positive attributes that may make them effective proposals for countering the issuer pays conflict in the context of rating structured finance products. Most obviously, both proposals impose a tangible cost on the CRAs in the form of lost profits that would counterbalance any gains derived from issuing inflated ratings. Moreover, several benefits flow from the quasi strict liability character of the proposals. First, each seems relatively straightforward and low cost to administer. Assuming that both proposals were implemented using default rates, the SEC would need to set the appropriate threshold default rate for each letter grade and then monitor defaults by asset class or product type. While the SEC would need to expend a meaningful level of resources to administer the program, it would probably be less costly to run than a system that actually inquired whether particular ratings were truly compromised by the issuer pays conflict. Second, whereas a disclosure-based system places the onus on the COLUMBIA BUSINESS LAW REVIEW investor to evaluate whether ratings are compromised, a quasi strict liability system shifts the burden to the CRAs to perform up to the required quality level. Third, by refraining from inquiring into the standard of care used by a CRA or into the particular processes employed, the proposals seem consistent with statutory limits on the SEC's authority to regulate the substance of credit ratings.249 In addition to these benefits, these proposals both contain relatively high visibility sanctions. If a CRA either loses part of its NRSRO status or is forced to disgorge profits, it would probably be subject to negative publicity. This may provide an added incentive to CRAs to produce accurate ratings. Last, the proposals appear to provide for measured sanctions and would likely not result in astronomic liability. These proposals also contain possible downsides. One potential problem, which is inherent in any proposal with a strict liability flavor, is the risk of overdeterrence. The goal in creating a performance-based sanction to counteract the issuer pays conflict would be to subject CRAs to the optimal level of legal threat, but default rates would not be a perfect proxy for compromised ratings. 250 A CRA might underperform relative to the predetermined quality metric if correlated macroeconomic events cause many securities to default for reasons completely unrelated to conflicts of interest. However, the SEC could choose to forego sanctions if it appeared the underperformance resulted from something truly extraordinary. A second possible problem may be the impact on smaller rating agencies that are trying to establish a foothold in the industry. A small new entrant may not as easily be able to withstand a sanction as an established CRA. From a different angle, a CRA will only be subject to these penalties if it cannot perform up to the necessary quality standard, so the forced exit of a low-quality entrant may not necessarily be a bad thing. 249 15 U.S.C. § 78o-7(c)(2) (2009). 25 See Coffee, Gatekeeper Failure, supra note 9, at 345. [Vol. 2009 No. 2:605] CREDIT RA TING AGENCIES AND THE CREDIT CRISIS 665 C. A Comparative Critique In general, both the Coffee proposal and the Hunt proposal offer many benefits that, on balance, appear to outweigh their potential negative consequences. The proposals are not identical, though, so they do offer competing positive and negative attributes. In particular, a benefit that may be unique to the Coffee proposal is the particular stigma that might attach to a CRA when it loses its NRSRO status. While both proposals offer relatively high visibility sanctions, the true threat of NRSRO recognition may motivate a CRA in a way that the possibility of profit disgorgement may not. On the other hand, however, issuers and investors may strongly resist a proposed regulation that might actually withdraw NRSRO status from the big three CRAs. If ratings remain ingrained in the regulatory framework, issuers and investors may prefer to retain the status quo, even at risk of low-quality ratings, than to have to work with new CRAs with whom they are unfamiliar.2 51 A second and seemingly more severe problem inherent in the Coffee proposal is the way that it relies on restricting future profits as the penalty for inflating current ratings. The cost imposed on the CRA is the future profit the CRA loses while its NRSRO status is suspended. This lost profit may have little correlation to the gains the CRA made from inflating its ratings in the past. For example, assume that innovators develop a new type of security. Seeking to reap short-term gains, a CRA may inflate its ratings to capture as much of the market for rating that product in the near-term as possible. Once the inflated nature of the rating is revealed, issuance of that new security may slow as investors lose confidence in the product (as happened with RMBS and CDOs). While the CRA may lose its NRSRO status for rating that class of securities, the CRA forgoes less profit than it made in the short-term because the market for rating the security has contracted. Essentially, the proposal still permits opportunistic behavior by CRAs at the expense of 251 See supra Part III.A.1. COLUMBIA BUSINESS LAW REVIEW investors. The cost imposed by the Hunt proposal, on the other hand, better matches the gains that the CRA makes from issuing compromised ratings. While the Hunt proposal seems to more closely impose the optimal cost on the agencies for issuing inflated ratings, it is unclear how one aspect of the Hunt proposal might be implemented. The Hunt proposal was advocated to address the problem of low-quality ratings in general, not only to address ratings deficiency flowing from the issuer pays conflict. Given his objective, Mr. Hunt would not require disgorgement for low-quality ratings if the CRA disclosed the quality deficiency ahead of time. It is unclear how this would translate if one wanted to view the proposal with respect to the issuer pays conflict only, since it does not seem logical to permit a CRA to escape penalty if it disclosed ahead of time that its ratings are or might be compromised by a conflict of interest. While both proposals share many positive attributes, the Hunt proposal seems to provide a better chance of imposing the optimal cost of the CRAs for deterring them from issuing inflated ratings. By limiting disgorgement to the CRA's profits that it earned on the low quality ratings, the Hunt proposal subjects CRAs to a penalty that is probably the best proxy for the gains received from acquiescing to the conflict. However, if implemented to counteract the pressure that results from the issuer pays conflict, the proposal might be more effective if the agencies could not avoid being penalized by disclosing either that their ratings may be of low quality or that their ratings could be subject to a conflict. VI. CONCLUSION While the Credit Crisis that began in 2007 has many complicated and interrelated causes, several commentators have criticized CRAs for providing undeservedly high ratings on subprime RMBS and CDOs containing subprime mortgages in their collateral pools. Although the issuer pays conflict inherent in the dominant CRAs' business model cannot fully explain why the agencies failed to provide high- quality ratings on these securities, several features of the [Vol. 2009 No. 2:605] CREDIT RATING AGENCIES AND THE CREDIT CRISIS 667 structured finance ratings market suggest that conflicts of interest are partly responsible. Specifically, the concentration of the customer base for these ratings and the iterative, interactive ratings process employed by the CRAs to rate structured finance products may have combined to make it more difficult for CRAs to withstand pressure to inflate ratings. Moreover, the growing importance of structured finance ratings to CRAs' revenue growth may have compounded the problem. The current SEC-administered regulatory regime and proposed rule amendments rely largely on disclosure to protect investors from risks created by the issuer pays conflict. Disclosure alone, however, may not protect against a similar problem in the future. Even if CRAs make the required disclosures fully and fairly, investors must adequately perceive and evaluate that information and must penalize CRAs via the market mechanism for risks associated with the issuer pays conflict. Otherwise, disclosure will not impose adequate costs on agencies to deter them from reaping the gains of compromised ratings. Several factors, including investors' failure to appreciate relevant differences between structured finance ratings and ratings on traditional products, inability to understand structured finance products, and failure to scrutinize the informational value of ratings, may prevent investors from fully appreciating mandatory disclosures on structured finance ratings. In addition, ratings-based regulation may further undermine the market mechanism, since issuers and investors alike might depend on ratings for their regulatory consequences and not solely for the information they convey. Instead of relying exclusively on disclosure, regulators should consider implementing a rule that would penalize CRAs for unacceptably poor performance. A performance- based sanction could fill the current regulatory gap by imposing costs on CRAs to offset any gains reaped by issuing compromised ratings. Such a measure could deter CRAs from capitalizing on short-term gains that they can realize by appeasing issuers at the expense of investors. One difficulty that policymakers would encounter in crafting a 668 COLUMBIA B USINESS LAW RE VIE W [Vol. 2009 performance-based sanction lies in formulating a rule that would convey the appropriate costs on rating agencies to avoid over- or under-deterrence. The Hunt proposal may provide a solid foundation for formulating a regulatory response.