July 28, 2003
By Electronic Mail
Mr. Jonathan G. Katz
Re: File No. S7-12-03, Concept Release: Rating Agencies and the Use of Credit Ratings under the Federal Securities Laws
Dear Mr. Katz:
On June 4, 2003, the Securities and Exchange Commission ("the Commission," or "the SEC") issued its Concept Release1 ("the Release") regarding the credit ratings industry, wherein it posed and requested public comment on a series of relevant questions. In our response, Moody's Investors Service ("Moody's") will address these questions by categorizing them into the three broad areas identified by the Commission: 1) should credit ratings continue to be used for regulatory purposes under the federal securities laws; 2) if so, what should be the process for determining whose credit ratings to use; and, 3) if credit ratings continue to be used in federal securities law, what is the appropriate level of oversight for the agencies whose ratings are used?
Ratings help level the information playing field between borrowers and lenders by providing a readily understood language for credit risk. Some of the most important attributes of credit ratings as currently made available by the major rating agencies include their independence, predictive content, broad coverage, and free dissemination to the general public. These attributes are the result of decades of evolution in line with market-based needs. But they have also caused ratings to be a "public good," which in turn has led to their adoption by various authorities for certain public policy objectives.
There is an impression among market commentators that the SEC's existing Nationally Recognized Statistical Ratings Organization ("NRSRO") designation process has caused the development and success of the rating agency industry. That impression misconstrues the history of the use of credit ratings in federal securities laws, which use was appropriated because rating agencies already provided a valuable public good. It is also important to note that there has been very significant expansion of the U.S. capital market and development of new types of securities2 after the Commission commenced using the designation. Further, the widespread use of ratings in the global markets, many of which do not require ratings or have a governmental designation process similar to that of the United States, demonstrates that ratings are used because they are valued by market participants, not because they are required by governmental action. Therefore:
I. Should Ratings Continue to be Used in Securities Law?
A. Eliminate NRSRO Designation
Recent criticisms of rating agencies have included the perceived inadequacy of their performance. Some critics have asserted that national recognition is the cause of an "artificially" low number of competitors, which has reduced the incentives for accuracy. Thus, the argument continues, enlarging the pool of recognized rating agencies will create competitive pressures which will improve performance. Others argue that expanding national recognition will not improve the performance of rating agencies, but that enhancing governmental oversight or increasing their contingent liability will. Finally, it has been asserted that increased oversight will do little to improve the quality of the rating,4 and that perhaps ratings and rating agencies have outlived their utility and should be replaced by other, purportedly more advanced technologies.5
If the Commission wishes to discontinue the designation of rating agencies and the use of ratings in federal securities laws in an attempt to address any or all of these assertions, Moody's would support such discontinuance. Discontinuance of "national recognition" would also eliminate risks associated with recognizing potentially less qualified rating agencies, "rating shopping" by issuers for the highest rating for regulatory qualification purposes, and other forms of standards-lowering competition.6 Moreover, if "national recognition" results in regulation that impairs the constructive role of ratings in the capital markets, or which substantially increases regulatory or private litigation costs, then Moody's would oppose the continued use of ratings in federal securities laws or the recognition of rating agencies by the Commission.
B. Alternative Approaches
As an alternative to "national recognition," the Commission could rely directly on the regulated entities to make complete and full public disclosure with respect to the methodologies they have employed to distinguish higher from lower risk securities, and periodically approve or allow the market to determine the acceptability of such methods.
Specifically, and as indicated in the Concept Release, the Commission can rely upon broker-dealers' internal rating systems, in the case of Rule 15c3-1, and upon subjective evaluations for Rule 2a-7 purposes. A securities firm can be required to publicly disclose the framework that underlies its internal ratings system. If an institution relies on agency ratings within its internal system, it should disclose which agencies it uses, when it substitutes one rating agency for another and how it resolves split ratings. Provision of internal ratings to the SEC, for comparison with public and other private ratings, would minimize the need for direct SEC supervision and extensive firewalls. For current use of ratings in determining Form S-3 financial eligibility requirements, the SEC could substitute alternative criteria, such as issue size, number of investors, or market listing.
II. SEC Recognition of Rating Agencies
Moody's appreciates that the adoption of credit ratings for use under the federal securities laws was in response to the perceived utility of ratings in furtherance of certain public policy objectives,7 including:
(i) to help manage principal/agent risks in asset management, and
(ii) to help specify required capital to cover risks in certain regulated institutions.
We also understand that use of ratings in federal securities laws is currently viewed by many to contribute to market efficiency and investor protection. If, therefore, the Commission chooses to maintain the NRSRO system, we believe that additional regulatory measures are not inappropriate so long as they neither cause those attributes of ratings valued by market participants for credit-sensitive securities to be altered, nor materially alter rating agencies' ability to produce and maintain ratings in the manner which have resulted in their adoption for public policy purposes in the first place.
A. Transparency in the SEC's Recognition Process
Moody's has always been, and will continue to be, a proponent of transparency in the capital market. To that end, we would favor the SEC's adoption of a set of standards for recognition of NRSROs, if these criteria do not cause alteration of the core attributes of our product, are equally applicable to all designated entities, and are well understood by the market.
B. Recognition Criteria
In determining recognition criteria, the SEC must weigh: (i) the benefits of having more recognized NRSROs, with the potential for greater diversity of opinion made freely available to market participants; against (ii) the potential detriment that greater diversity of opinion could result in opinion being used in regulation which does not satisfy the objectives the regulation seeks to achieve.
We recommend that any criterion used by the Commission to designate rating agencies should be adequately broad to permit open competition, thus allowing numerous entities the ability to distinguish themselves, one from the other, based on the quality of their individually held views. At the same time the Commission may wish to balance the benefits from an increase in the number of opinions against possible lowering of the substantive standards intended by the regulation in which NRSRO ratings are used. 8
How well a rating agency performs over time is important and should ultimately be assessed by the market's acceptance and usage of its ratings. On the other hand, because there is not, nor should there be, one "sanctioned" model for rating agencies, we would suggest that precise requirements are not directly pertinent to the overall reliability of a rating agency and thus unduly raise the barrier to entry with little perceived advantage. Further, if the Commission were to adopt the following criteria as the designation requirements for recognized rating agencies, we are not convinced that the "national recognition" test would provide any added benefit.
1. Ratings must be made widely and publicly available
For ratings on public debt securities to be appropriately used in securities laws, they must be made widely available to the general public. A rating should not be construed as "publicly disseminated" if access to it is limited to subscribers, or made otherwise cumbersome or untimely. Such broad availability avoids selective disclosure of potentially material information and allows for ratings comparability. The "public-ness" of ratings helps level the information playing field among investors.9
Further, as investor protection is the ultimate aim of rules and regulations, we would suggest that the incidental benefit of public-ness of ratings is that it allows the market to test the reliability of the rating system over time.
2. Rating process and performance disclosure
Transparency in rating agencies' methodologies is important to the market's understanding of the ratings of a rating agency. Such methodologies may cover, for example, approach to broad industry divisions (such as industrial corporations, sovereigns, and financial institutions), to industry sectors within such divisions (for example within the industry sector we have published methodologies in relation to utilities), to regional sub-sectors, and where appropriate, to types of securities.10 We believe such disclosure is helpful, and provides the users of ratings with the appropriate tools to judge the reliability of the rating process. Further, all market players should be given equal opportunity to understand the procedural approach employed by an agency to arrive at its opinion. It is Moody's practice to make publicly available all rating practices, rating methodologies, and performance metrics against our stated objectives for Moody's ratings.
3. Management of the potential conflicts of interest
As we have indicated in our previous communications with Commission staff and with market participants, ratings have the same problem as all other public goods: the free-rider problem.11 Consumers of bond ratings have little incentive to pay for an opinion they can easily obtain at no cost. Thus, the rating agency model which has developed is an "issuer fee-based" model. This model has two intrinsic conflicts of interest which must be effectively managed:12 a) issuers pay rating agencies for their credit opinions; and, b) issuers are one source of input in a rating agency's formation of its opinion. Importantly, however, the issuer fee model also effectively avoids the necessity of selective disclosure of ratings, which would occur under alternative models such as the subscription service or "private sale" process models.
Issuer fees are supported by fundamental market mechanics and are critical to the public good nature of ratings - i.e., ratings offered to the public without charge. After all, investors do not need to pay - and thus will not pay - for ratings that are publicly disseminated rather than selectively disclosed. If ratings are to be made public and freely available to all interested parties, then rating agencies must generate fees by some other means. These means could include: i) fees from other services, which raise concerns about the independence of ratings that are subsidized by such services; ii) fees from sales of research, a business model that proved insufficient as early as the 1970's in funding the necessary professional staff to provide comprehensive coverage in rapidly expanding and increasingly complex financial markets; or iii) the issuer-fee model, which avoids having ratings subsidized by other activities, yet generates sufficient fees to support large professional staffs.
While it is important that rating agencies independently form their opinions, we believe that issuers can be a useful source of information in the rating process. In so far as the issuer can provide the rating agency with its perspective of its enterprise, such an exchange is both beneficial and helpful to the ultimate end user, the investor. If these discussions did not occur, it would likely lead to the dissemination of either less timely or less considered rating opinions, which would likely lead to a greater volatility and frequency of rating changes, including rating reversals.13 Rating actions can and often do have an impact on the price of debt instruments or the cost of funding for an issuer. Unnecessary volatility in ratings would therefore increase the volatility of credit spreads and increase transaction costs for both issuers and investors, without providing a corresponding benefit.
We suggest that, to protect against the inherent conflict of interest, the following criteria be satisfied:
4. Protection of non-public information
Moody's analysts can be contacted by issuers, investors, investment banks, the media, subscribers to our research service, and, in fact, the general public. Nonetheless, we have two types of primary relationships: the issuing entity, about whose debt securities we publish ratings; and the investor, for whose benefit we provide our opinions on the issuers' debt securities. As we have discussed above, we have interactions with issuers throughout the rating process. Our interaction with the investor community occurs in a number of ways. First, investors may access our ratings and press releases freely through our web-site and through media sources; second, they may subscribe to our research service; third, our rating teams hold periodic teleconferences; fourth, investors participate in investor conferences, where our analysts present their views on particular industries or sectors; and, finally, our rating analysts have individual meetings and telephone conversations with investors.
We have provided in Annex 3, some of the internal procedures we have adopted to protect confidential information in our possession. Moody's policy prohibits the disclosure of non-public information for two primary reasons:
i) our on-going relationship with the issuer is premised on the concept that the non-public information provided to us will remain out of the public domain until the issuer makes that information public; and
ii) we are fully aware that should we disseminate material, non-public information on a selective basis, we could be in violation of securities laws. While Moody's is very mindful of the necessity to safeguard confidential information, we nevertheless believe that investors gain insights from the discussions with our analysts as they are better able, for example, to understand the analytical reasoning behind a public rating.
We recommend that as part of any NRSRO designation regime, rating agencies should protect against improper communications through internal guidelines that specifically prohibit analysts from discussing non-public information concerning an issuer with anyone other than the issuer or its designated agents. We believe that such internal measures are the most effective means for safeguarding non-public information. However, should the Commission decide that more stringent rules in this area are necessary for the industry, we would support their imposition if they were applicable with equal force to all rating agencies with similar access to confidential information, including non-NRSROs.
C. Alleged Anti-Competitive, Abusive and Unfair Practices
Moody's appreciates the concern for an appropriate level of competition within the industry. We believe that vigorous competition in the rating agency industry has advantages and disadvantages in promoting primary regulatory objectives, and importantly depends on the bases upon which rating agencies are motivated to compete.
Ratings opinions reflect opinion forecasts of the future. Two rating agencies looking at the same set of facts may legitimately reach different conclusions, based on their individual points of view or understanding of market behavior. The application of the methodology will be crucial in the formulation of the rating opinion. Moreover, diversity of opinion from credit ratings indicates that these rating agencies are independent of each other.
We believe that any attempt to harmonize rating opinions and meanings would reduce the amount of information available to the market. For example, Moody's ratings are a measure of expected loss. That is to say, they forecast the likelihood of default on a bond and the severity of loss in the event of that bond's default. Not all other rating agencies intend to convey the same information with their ratings systems.17 If all opinion providers were compelled to assign the same rating symbol or adopt wholesale the same ratings as other agencies, the amount of information available to the market would be lessened, contrary to a primary objective of providing diverse opinions to promote efficient markets.
Finally, any regulation that requires the application of a prescribed methodology or approach will leave rating agencies less able to react to dynamic capital markets or future credit events. For example, the issues that have led to recent market failures - accounting fraud, ratings triggers, and corporate governance issues - have been identified, and Moody's has incorporated enhancements to our analytic processes, methodologies, and human capital to better address these issues. It is likely that the future will see different issues influencing creditworthiness, and it is imperative that credit rating agencies retain the operating flexibility necessary to meet these future demands.
In order for any opinion-based industry to produce the greatest public good, the individual opinion providers should operate independently of one-another, utilizing their own approaches and methodologies in crafting their opinions. If the SEC considers maintaining or augmenting the NRSRO system, we would ask that it provide a regime where all rating agency players are encouraged to compete vigorously on that basis. Otherwise, they would have less motivation to improve their performance as they would be discouraged from differentiating their product from other providers in the industry. If rules or regulation alter such a framework, it may change the rating product in such a way that its use will no longer be advantageous to either the market or the regulatory regime.
2. Unsolicited ratings
As a practical matter, Moody's has almost completely curtailed the assignment of unsolicited ratings in recent years. This is primarily due to hostile reactions to such ratings from many quarters, including from issuers, intermediaries, lawmakers and some investors. Beginning in January 2000, Moody's announced that we would identify in initial rating assignments those unsolicited ratings for which the issuer did not request a rating and declined Moody's invitation to participate in the rating assignment process. In order to be so designated, a rating must satisfy each of the following requirements:
In such cases, the following statement will appear in the rating assignment press release:
III. Examination and Oversight of NRSROs
Moody's will not object to continuing oversight of the designated rating agencies so long as the oversight measures are limited and focused on designation criteria similar to those recommended in this response. We believe it to be rational for the Commission to wish to reassess periodically the compatibility of a recognized rating agency with such designation criteria.
We would further suggest that the regulatory oversight and examination regime should be commensurate with that which has been evidenced as necessary by the industry practice. To that end, should the Commission decide to pursue a designation system, we urge that the continuing oversight procedures be proportionately tailored for the rating agency industry. For instance, rating agency regulation should not be modeled after the auditing, investment banking, investment advisory or brokerage industries.
Any regulatory regime that may have the ramification of inhibiting rating agencies from issuing independent rating opinions does not serve the best interests of the market. To that end, Moody's would recommend that the Commission carefully balance the public interest in oversight with the public interest in fearless and robust information gathering, analysis, and dissemination by the agencies.
A. Increased Exposure to Litigation Will Have a Chilling Effect
1. Appropriate liability standard for opinions
A credit rating is an opinion about the future,19 not a statement of fact, about the ability of a company to meet its debt obligations in a timely manner. It is based on information the agency has gathered and analyzed. It is forward-looking and predictive, resting on the agency's assessment of the information it has gathered, and in light of the agency's experience in rating other companies. Credit ratings provide only one element in support of the investment decision-making process. They are not, nor should they be treated as, recommendations, guarantees or statements of historical fact. Therefore, ratings should not be exposed to a liability standard that treats them as something they are not and have never purported to be.
A well-reasoned opinion may nevertheless be disagreed with by others, or may not in the end "come true." By way of example, no one can predict the future and similar to an insurance company which is uncertain as to exactly which holder will in fact draw on its policy, a rating agency is uncertain about exactly which issuer will default. However, insurance companies over time have learned to correlate certain factors with longevity and to assign greater or lesser weight to each factor. Thus, they know that over the next five years more 80 year-old smokers will die than 20 year-old non-smokers; and they know that over the same time period some 80 year-olds will live and some 20 year-olds will not. However, they simply cannot name names. Rating opinions are analogous to actuarial opinions in this regard.20
Further, because ratings are at their core opinions, and speech on matters of public concern, they fall squarely under the protection of the First Amendment of the U.S. Constitution. It is the fundamental premise of that Amendment that "the marketplace of ideas"21 thrives only when opinion providers have the "breathing space"22 to express their thoughts without fear of retribution. An opinion (unlike a statement of fact) cannot be "true" or "false."23 As such, the First Amendment has over time been interpreted to provide ratings with substantial protection from liability. The rating agency is entitled to decide what information to take into account, what weight to place on it, and what conclusions to draw from it.
Finally, it is the rating agency's task to make unbiased and sometimes unpopular observations regarding powerful and prestigious issuers. It is in the nature of the business that our opinions are at times neither welcomed nor applauded - not by the issuers, underwriters, governments, or current holders of the issuer's securities. We are not aware of a single instance where an issuer publicly stated that the rating of its company was too high or that a downgrade was needed. The unpopularity of some of our opinions is understandable. If, for example, an issuer's obligations are downgraded it may impact the issuer's borrowing costs, or an investment manager's ability to hold the security of that issuer. The issuer or investor may, therefore, be frustrated with our rating opinion and may contest the "rightness" of our rating.
2. Document retention and production
While Moody's does not object to the retention of documents, we would oppose measures that required NRSROs to routinely produce documents to the Commission as part of the oversight mechanism. We believe that it is part of the rating business to retain certain documents of which we, as opinion providers, repeatedly make use. Yet, the production of such documents to the Commission either in its capacity as the regulatory authority with oversight powers, or in its investigatory role, will de facto increase our exposure to civil litigation. Even assuming that a rating agency did not object to producing documents to the SEC, plaintiffs' attorneys would be certain to cite the agency's production to the SEC as a waiver of the agency's journalists' privilege. Producing documents to the SEC could thus become tantamount to making the documents public, or at least public for purposes of the plaintiffs' bar.
Such documents may include not only our work product, but also any material we may have gathered from the issuer. There are two potential unintended but quite negative outcomes:
i) Erosion of the protections for privileged and confidential information, and increased availability of rating agencies' internal documents, will likely lead to an increase in frivolous litigation over the "correctness" of our individual rating opinions. As we noted above, this increase in litigation will discourage analysts from releasing non-conforming rating opinions. It is our position that measures that limit the ability of market participants to speak freely and interact openly are detrimental to overall market transparency and ultimately to the fundamental goals of securities law.
ii) It is likely that issuers will refrain from having discussions with us about their business. Such an outcome will reduce the sources of information upon which our rating opinions are based, and may in the long run damage the quality of our ratings. In the near term, however, it may potentially increase the volatility of our ratings as we will no longer be privy to the issuer's perspective of its business. Rating agencies may, thus, be left to take action on market rumors. We would suggest that such an outcome would undermine the purpose of using ratings in the securities law framework.
3. Unintended consequence - chilling effect
It is both conceivable and likely that if regulation exposes rating agencies to increased litigation and litigation costs, rating analysts would be urged to limit dissemination of objective, insightful and at times controversial opinions. The potential of a challenge on the "rightness" of the opinion each time one is published would eventually, almost certainly, discourage rating agencies from forthright opinions. Rating agencies would moderate their commentary, and potentially distort rating results to minimize administrative sanctions and/or civil actions.
Moreover, regulation that includes routine document production or efforts to control the bases on which we reach our rating opinions might cause rating agencies to decline to provide opinions about companies in more volatile industries, and companies with less clear or potentially misleading disclosure practices. Such an outcome would reduce comprehensive coverage, thus reducing the public good attribute of ratings which offers investors a tool for broad comparability of credit quality. It would also remove opinions from the market concerning exactly the companies about which investors would most benefit.
Moody's acknowledges that use of ratings in federal securities laws is perceived by many to support existing public policy objectives. Yet ratings and rating agencies primarily seek to meet the demands of participants in the market for credit-sensitive securities. Thus, the ongoing evolution of ratings should not necessarily be expected to maximize their value for regulatory use. In order to retain their primary value to market participants, credit ratings must continue to develop and keep pace with the evolution of the capital markets more broadly, not become static objects in regulation.
We thank the Commission for this opportunity to share our thoughts and opinions on any potential regulatory initiative upon which it may choose to embark. We look forward to further discussions in the near future.
The Use of Ratings in Securities Law
Informational asymmetry is an enduring characteristic of markets. Sellers generally have more information about the good or service than buyers. In the context of the fixed income market, an issuer of a bond will, as a matter of course, have far more information about itself than is available to a potential investor. Such information can be important to enable the investor to assess the likelihood of that issuer's ability and willingness to pay. However, in the absence of some compelling force, most investors will remain ignorant of at least some important information.
This lack of transparency could lead to an environment wherein fewer transactions would be completed, and more of those transactions which are completed would likely be incorrectly priced. In the interest of market efficiency, the securities laws have sought to make transparency compulsory through a combination of information disclosure requirements and use of securities ratings and other publicly available information in regulations.25
1. Regulatory Use of Ratings - Ratings as Public Goods26
"Public goods" are a well-known economic concept. What has not been widely recognized, however, is the degree to which securities information from non-issuer sources constitutes a public good. The key characteristic of a public good is the non-excludability of users who have not paid for the good. People benefit regardless of whether they contribute to the costs of producing the good or not.
A rating is a symbol that encapsulates and conveys the agency's opinion about the likelihood of timely debt repayment. Once a rating has been publicly disseminated, one investor's use of the rating does not prevent another investor from using the same rating in exactly the same way, at exactly the same time. Therefore, publicly disseminated ratings are non-excludable.
The principal attributes of the ratings process and rating opinions as demanded by market users are:
(i) independence, including effective management of potential conflicts of interest;
(ii) predictive content;
(iii) relative stability;
(iv) comprehensive coverage to allow for comparison in the investment decision process; and
(v) broad, public and free dissemination.
Thus, from a public policy standpoint it is understandable why the private rating agency system has provided a helpful tool in support of regulatory objectives. Ratings condense a great deal of research into easy-to-use symbols that are relatively stable and accessible to all. In effect, they can be likened to other tools used by securities laws to promote transparency and stability in the capital market.27
As a result, regulatory authorities have increasingly utilized ratings as mechanisms to address dual policy goals of promoting transparency and limiting unnecessary volatility in the financial market. Over time, the regulatory use of ratings has proliferated. Securities regulators, banking regulators, insurance regulators, and legislators who oversee the activities of regulated entities use ratings. These forms of use have been similarly replicated in the private sector through adoption of ratings as proxy tools for portfolio composition and governance guidelines.
2. Market Use of Ratings
Moody's ratings provide predictive opinions on one characteristic of a corporate entity's financial enterprise - its likelihood to repay debt in a timely manner. In Moody's view, the main and proper role of credit ratings is to enhance transparency and efficiency in debt capital markets by reducing the information asymmetry between borrowers and lenders. We believe this function to be beneficial for the market as it enhances investor confidence and allows borrowers to have broader access to funds.
Yet, as discussed above, our ratings have a number of specific characteristics that have incrementally and over time encouraged their adoption across a wide range of functions. They are employed by issuers, intermediaries, counterparties to financial and commercial contracts, and large institutional investors, among others. Each group has different intended objectives in using ratings.
It is the role of ratings and the job of rating agencies to strike the appropriate balance and meet the demands of participants in the markets for credit-sensitive securities. Consequently, the attributes of the ratings process and ratings have evolved to serve the needs of these market participants. Moreover, the ongoing evolution of ratings should not necessarily be expected to maximize value for use in securities laws.
3. Competitive Landscape
As explained previously, Moody's and the other credit rating agencies operate within a framework wherein our ratings are used by numerous entities with sometimes opposing objectives and requirements. Investors associate "quality" with the most reliable ratings, while issuers associate "quality" with the highest possible plausible rating.28 Many issuers and investment bankers would welcome a greater ability to select among ratings providers for purposes of having greater control over the rating outcome, which they perceive will positively influence the marketability of their debt.
New market entrants and marginal participants historically have sought to make their products more attractive to issuers by offering more "flattering" ratings than do more established market participants.29 New entrants will understandably be inclined to compete in this manner because of the ease with which such a strategy could be implemented and the short-term benefits that might accrue to the entrant as a result.
Managing the Potential Conflict-of-Interest
Created by the Issuer-Pays Structure
Internal Procedural Safeguards
Protection of Confidential Information