July 28, 2003BY ELECTRONIC MAIL
Mr. Jonathan G. Katz
This letter is submitted by Fitch, Inc. ("Fitch") in response to the request for comments of the Securities and Exchange Commission ("SEC" or the "Commission") to the concept release Rating Agencies and the Use of Credit Ratings under the Federal Securities Laws (Release Nos. 33-8236; 34-47972; IC-26066, the "Concept Release").
Fitch traces it roots to the Fitch Publishing Company established in 1913. In the 1920s, Fitch introduced the now familiar "AAA" to "D" rating scale. Fitch was one of the three rating agencies (together with Standard & Poor's ("S&P") and Moody's Investors Service ("Moody's")) first recognized as a nationally recognized statistical rating organization (a so-called "NRSRO") by the SEC in 1975.
Since 1989 when Fitch was recapitalized by a new management team, Fitch has experienced dramatic growth. Throughout the 1990's, Fitch especially grew in the new area of structured finance, by providing investors original research, clear explanations of complex credits, and more rigorous surveillance than the other rating agencies.
In 1997, Fitch merged with IBCA Limited, another NRSRO headquartered in London, significantly increasing Fitch's worldwide presence and coverage in banking, financial institutions and sovereigns. Through the merger with IBCA, Fitch became owned by Fimalac S.A., a holding company which acquired IBCA in 1992. The merger of Fitch and IBCA represented the first step in our plan to respond to investors' need for an alternative global, full service rating agency capable of successfully competing with Moody's and S&P across all products and market segments.
Our next step in building Fitch into a global competitor was our acquisition of Duff & Phelps Credit Rating Co., an NRSRO headquartered in Chicago, in April, 2000 followed by the acquisition later that year of the rating business of Thomson BankWatch. These acquisitions strengthened our coverage in the corporate, financial institution, insurance and structured finance sectors, as well as adding a significant number of international offices and affiliates.
As a result of Fitch's growth and acquisitions, it today has approximately 1,250 employees, including over 700 analysts, in over 40 offices and affiliates worldwide. Fitch currently covers 2,300 banks and financial institutions, 1,000 corporations, 70 sovereigns and 26,000 municipal offerings in the United States. In addition, we cover over 7,000 issues in structured finance, which remains our traditional strength.
B. Alternatives to the NRSRO System
Through the years, NRSRO ratings have been increasingly used in safety and soundness and eligible investment regulations for banks, insurance companies and other financial institutions. The SEC first chose to use NRSRO ratings in connection with the capital requirements for broker-dealers in 1975. While the use of ratings in regulations has not been without controversy, we believe that regulators, including the SEC, have relied on NRSRO ratings for the same reason that investors do: ease of use, wide spread availability and proven performance over time.
Although other methods can be used to assess the creditworthiness of a security, such as the use of market-based indicators and computer-based models, we believe that such methods, while valuable, lack the simplicity, stability and track record of performance to supplant ratings as the preferred method used by investors to assess creditworthiness.
The NRSRO system is designed, appropriately in our view, to assure that recognized organizations possess the competence to develop accurate and reliable ratings and protect against the establishment of rating organizations that would issue inflated ratings in an effort to achieve short-term competitive gain. Without a system to recognize rating organizations, many important capital adequacy and eligible investment rules used in financial institution regulation would be ineffective. Issuers could find in the marketplace agencies willing to inflate ratings and regulators would have no means of protecting the public from the consequences.
One criticism of the NRSRO system is that it poses a barrier to entry for new entrants. Currently in the United States, ratings by new entrants typically do not satisfy investing criteria for regulated institutional and other corporate investors unless the SEC recognizes the entrant as an NRSRO. Admittedly, the requirements to achieve NRSRO status pose some barriers to entry, but these barriers are necessary to fulfill the important purpose of ensuring that the recognized agencies demonstrate the performance necessary to create a reliable ratings system.
In addition to believing that the NRSRO system is valuable, we do not believe that abolition of the NRSRO designation would lead to a meaningful increase in competition among rating agencies given Moody's and S&P's power in the current market. Despite Fitch's own NRSRO status and the determined efforts that Fitch and its predecessors have made in the last 25 years, Fitch has encountered significant impediments in its attempt to secure meaningful shares of the corporate or municipal bonds ratings markets because of Moody's and S&P's monopoly positions in those markets. It is unlikely that the removal of the NRSRO designation would lead to greater opportunity for smaller, less established agencies. The difficulty of entering the market is a result of the competitive environment and not the NRSRO designation itself.
Over the past few years there also has been considerable debate regarding the relative predictive value of credit ratings compared to the newer credit derivative market. Credit derivatives, in addition to other market-based indicators, such as bond spreads, have gathered much publicity as corporate credit quality deteriorated. However, it has been difficult to make broad comparisons as the credit cycle remained negative. Now that the markets have begun to stabilize, and it is possible to evaluate a "credit cycle," it has become clear that despite some successes, the market-based indicators have been quite volatile and have had many "false positives," credits whose spreads widened out significantly only to rebound to prior levels.
In these cases, investors following the lead of the credit derivative market would have purchased protection at levels much higher than the credits' ultimate risk. If credit default swaps or bond spreads alone had been used to determine the credit quality of a security, an institution holding that security would have been required to significantly boost the capital related to that position or sell the relevant bond, creating even more volatility.
Several examples illustrate the volatility of market-based indicators.
In these examples, market-based indicators diverged significantly from traditional ratings for a period of time, but then reverted to levels more consistent with the relevant credit rating. While market-based indicators are helpful for gauging an issuer's liquidity and access to funding, a regime that relied only upon bond spreads or credit default swaps would have most likely required an investor to increase capital or sell an asset to maintain proper levels of capital when the market was most volatile. If many investors sold the same asset at the same time, it is possible that the resulting illiquidity would have further pressured the relevant issuer's ultimate credit quality.
We believe that the use of market-based indicators and models to address credit risk to the exclusion of credit ratings issued by NRSROs raises several other issues. These methods are not as widely available or as consistent as ratings, in particular models are not readily available in a wide variety of situations where credit ratings are used, such as for structured finance and privately held companies. In addition, we believe it would be relatively difficult for a regulatory body to approve models for use. To prove the effectiveness of the model, large amounts of resources would be required from the approving entity and the entity submitting the model for approval. In all likelihood, only the largest financial institutions would be able to obtain recognition for their models or be willing to devote the substantial resources needed to maintain the models.
The use of market-based indicators and models to the exclusion of credit ratings may create competitive disadvantage to smaller financial institutions and other investors in the market. While credit ratings are generally freely available to the public on the agencies' websites and easily obtained and understood, models are expensive and highly sophisticated and, as a result, sources of models are limited. The availability of reliable data on pricing and credit spreads is also limited and costly and models require highly qualified staff to run the model, input data, and check results. All of this makes the use of models and market-based indicators untenable to many market participants.
We believe that many smaller financial institutions would be unable to dedicate the resources necessary to develop the capability needed to use internally-developed credit ratings in calculating their capital charges, particularly in diverse and complex areas such as structured finance and leveraged finance.
In short, Fitch believes the use of credit ratings is the best and simplest system to assess the creditworthiness of securities and the NRSRO system is the most appropriate way to recognize credit rating agencies whose opinions may prudently be relied on for regulatory purposes. Regarding the entity most capable of assessing the rating agencies for the purpose of recognition, Fitch believes that the SEC itself is in fact the most logical choice. The SEC has a long history of following the rating agencies and is well aware of the facts and issues that affect them and how their behavior affects the market at large. While other agencies may be able to perform the function over time, the SEC's prior knowledge in the area makes it preferable to other choices.
C. Recognition Criteria
We believe that the SEC should formalize the process by which a rating organization is recognized. The application process, specific recognition criteria and time frames for action on all applications should be specified in appropriate regulations. We believe public comment should be solicited on applications and an appropriate appeal process should be put in place.
The criteria for recognition should include an evaluation of the organization's resources, its policies to avoid conflicts of interest and prevent insider trading and the extent to which the organization's ratings are used by market participants. Most importantly, however, recognition should be based upon the organization demonstrating the performance of their ratings over time by publication of actual default rates experienced in rating categories and transition studies showing the actual movement of ratings over time. When considering a rating organization for possible recognition, we believe the SEC should evaluate the default and transition experience of each organization's ratings against a benchmark reflecting the aggregate, historical default and transition rates of all ratings issued by rating agencies in the market2. Ultimately, we believe that recognition should be reserved for those organizations that prove the performance of their ratings over time relative to the performance of other rating systems.
We also believe that the SEC should continue the practice of limited recognition that acknowledges the special expertise of smaller organizations in selected areas of specialty or geographic regions such as the prior recognition afforded to IBCA and BankWatch for their expertise in financial institution analysis.
It should be noted that Fitch does not believe that a criteria for recognition should be adherence to generally accepted industry standards. In fact, such industry standards do not exist in the case of credit rating agencies and we believe that it would be detrimental to introduce them. Ratings are opinions, and as such ratings are based on differing criteria, qualitative and quantitative, in each agency. The market benefits from this diversity of opinion, and demands it. Requiring that a rating agency abide by strict standards would create a situation in which each agency would produce the same result on each credit, and there would be no need for competing agencies nor any benefit from competing agencies.
D. Examination and Oversight of NRSROs
Fitch acknowledges the Commission's right to revoke the recognition of any NRSRO that no longer meets the criteria for recognition. Given the importance of credit ratings in the financial markets, we believe this is an important need. As we discussed in connection with the criteria for recognition, we also believe that the examination and oversight of NRSROs should be principally focused on the performance of a rating organization's ratings over time relative to the performance of other rating systems. Accordingly, we believe that the Commission's principal oversight function should be to regularly evaluate the default and transition experience of each organization's ratings against an aggregate benchmark. Additionally, we also acknowledge the importance of our adherence to policies designed to prevent the misuse of inside information and the need of the Commission to assure compliance with these important policies.
In addition, we believe that any oversight should be narrowly tailored so as to recognize the constitutional rights of the rating agencies, which function as journalists and should be afforded the high level of protection guaranteed by the First Amendment. An excessive amount of interference with the business of rating agencies would both violate the First Amendment rights of the agencies and remove some of the flexibility in the ratings process which is critical to objective and timely ratings.
Within this framework, a narrowly tailored oversight scheme specific to rating agencies should be developed. While the rating agencies currently voluntarily file under the Investment Advisor's Act, this is not a "good fit", as agencies function as journalists providing analysis and opinion and not as investment advisers. As the Supreme Court recognized in Lowe, Congress "did not seek to regulate the press through the licensing of non-personalized publishing activities" when it enacted the Investment Advisors Act, but rather was "primarily interested in regulating the business of rendering personalized investment advice." Lowe v. SEC, 472 U.S. 181, 204 (1985). Fitch does not provide any personalized investment advice - indeed, even Fitch's non-personalized ratings do not make any recommendations to buy or sell particular securities, but rather simply analyze the creditworthiness of a security, a point that was noted by the Staff in its June 4, 2003 response to questions from Congressman Richard H. Baker3. Fitch is therefore not an "investment advisory business" within the meaning of the Investment Advisors Act and to try to make the Investment Advisors Act apply to Fitch and other rating agencies would not be productive.
In the same vein, it would be unsound to seek to impose a diligence requirement on rating agencies either for purposes of creating a private right of action or for oversight purposes. Even putting aside the significant and in our view insurmountable issues of statutory authority and constitutionality, rating agencies do not now audit or verify the information on which they rely and to impose such a requirement would duplicate the work of the various professionals (auditors, lawyers, investment bankers and fiduciaries) upon whom the law does place certain obligations of diligence and due care.
E. Conflicts of Interest
Over the years, there has been considerable discussion about the fact that the current NRSROs derive a significant portion of their revenue from the ratings fees charged to issuers of rated securities. Fitch does not believe that the fact that issuers generally pay the rating agencies' fees creates an actual conflict of interest, i.e., a conflict that impairs the objectivity of the rating agencies' judgment about creditworthiness reflected in ratings. Rather, it is more appropriately classified as a potential conflict of interest, i.e., something that should be disclosed and managed to assure that it does not become an actual conflict. We believe the measures Fitch (and, on belief, the other agencies as well) has in place to manage the potential conflict adequately prevent an actual conflict of interest from arising.
Charging a fee to the issuer for the analysis done in connection with a rating dates back to the late 1960s. It is widely known by investors, who are the ultimate consumers of the rating agency product.
By way of context, Fitch's revenue comes from two principal sources: the sale of subscriptions for our research and fees paid by issuers for the analysis we conduct with respect to ratings. In this we are similar to other members of the media which derive revenue from subscribers and advertisers that include companies that they cover. Like other journalists, we emphasize independence and objectivity because our independent, unbiased coverage of the companies and securities we rate is important to our research subscribers and the marketplace in general.
Fitch goes to great efforts to assure that our receipt of fees from issuers does not affect our editorial independence. We have a separate sales and marketing team that works independently of the analysts that cover the issuers. In corporate finance ratings, analysts generally are not involved in fee discussions. Although structured finance analysts may be involved in fee discussions, they are typically senior analysts who understand the need to manage the potential conflict of interest.
We also manage the potential conflict through our compensation philosophy. The revenue Fitch receives from issuers covered by an analyst is not a factor in that analyst's compensation. Instead, an analyst's performance, such as the quality and timeliness of research, and Fitch's overall financial performance determine an analyst's compensation. Similarly, an analyst's performance relative to his or her peers and the overall profitability of Fitch determine an analyst's bonus. The financial performance of analysts' sectors or groups do not factor into their bonuses.
Fitch does not have an advisory relationship with the companies it rates. It always maintains full independence. Unlike an investment bank, our fees are not based on the success of a bond issue or tied to the level of the rating issued. The fee charged an issuer does not go up or down depending on the ratings assigned or the successful completion of a bond offering.
Our fee is determined in advance of the determination of the rating and we do not charge a fee for a rating unless the issuer agrees in advance to pay the fee. While we do assign ratings on an unsolicited basis, we do not send bills for unsolicited ratings. Any issuer may terminate its fee arrangement with Fitch without fear that its rating will be lowered, although we do reserve the right to withdraw a rating for which we are not paid if there is insufficient investor interest in the rating to justify continuing effort to maintain it.
As noted above, Fitch believes that the disclosure of the arrangement by which an issuer pay fees to Fitch in connection with Fitch's ratings of the issuer is appropriate. Accordingly, Fitch currently discloses that it receives fees from issuers in connection with our ratings as well as the range of fees paid. This has been our practice for sometime.
While disclosing the specific fees paid by the issuer in connection with a rating or the more extensive disclosure of all amounts paid by an issuer to Fitch and its affiliates would provide the users of ratings with more information, such disclosure would create competitive issues for Fitch. We do not believe that it is necessary or appropriate to provide disclosure of the specific fees or any more extensive financial disclosure. We believe that the specific fees we charge and the revenue we derive from other sources are proprietary and if known by our competitors, both of whom possess dominant market power in certain markets, will cause us competitive injury. We believe that the far more important disclosure is that the fee arrangement exists and the range of those fees.
Another concern discussed by the SEC in the concept release is that subscribers have preferential access to rating analysts and may obtain information about a rating action before it is available to the general public. This concern is completely unwarranted in the case of Fitch. Fitch takes great efforts to ensure that all members of the public have access to our ratings and may discuss those ratings with our analysts, whether or not those interested parties are subscribers.
All public ratings and rating actions are widely disseminated through our websites and international wire services. Except for prior notification to the issuer of a rating or rating action, ratings and rating actions are never selectively disclosed to any subscriber or any other party. Fitch's ratings and related publications, including those detailing rating actions, are widely available through our public websites and wire services free-of-charge and there are no prior communications of rating actions to subscribers.
Fitch analysts do regularly conduct informal conversations with investors, other members of the financial media and interested parties discussing our analysis and commentary, but as a matter of policy those conversations can never go beyond the scope of our published analysis or communicate any nonpublic information. We believe that making our analysts available to anyone interested in discussing our analysis is a valuable service to investors and the capital markets at large. The contact information for the principal analysts and other key contact people at Fitch is included in every item we publish for the purpose of facilitating interested parties posing questions to our analysts. Anyone can call our analysts free-of-charge and discuss our analysis with them, whether or not the person is a subscriber to our subscription services.
From time to time, we also hold free telephone conferences that are available to anyone interested at which our analysts will discuss our published analysis and criteria and take questions from the participants. These telephone conferences are publicly announced in the same manner our ratings and rating actions are disseminated.
We also sponsor conferences throughout the world, as well as participate in conferences sponsored by others (which may sometimes require payment of a registration fee) at which our analysts will discuss our published analysis and criteria. These conferences are publicly advertised and all our welcome.
In addition, we firmly believe that existing antifraud remedies are sufficient to deter any inappropriate disclosures by rating agencies to subscribers or any other parties.
Based on the above procedures regarding issuer payment of fees, ancillary services and selective disclosure, Fitch believes the rating agencies adequately address any potential conflict of interest. In fact, we believe that the suggestions proposed in the concept release to protect against conflicts of interest have already been in large part adopted by the rating agencies. However, Fitch would not oppose narrowly tailored conditions to recognition that ensure that these standards continue.
F. Anticompetitive, Abusive and Unfair Practices
Fitch believes that our emergence as a global, full service rating agency capable of competing against Moody's and S&P across all products and market segments has created meaningful competition in the ratings market for the first time in years. Fitch's challenge to the Moody's/S&P monopoly has enhanced innovation, forced transparency in the rating process, improved service to investors and created much needed price competition.
Academic research confirms our belief that innovations in the ratings industry have often "been initiated by the smaller rating firms [Fitch and its legacy firms], with the larger two [Moody's and S&P] then following."4 At Fitch, we are particularly proud of the work that we have done in the development of innovative methodologies to analyze new structured finance securities. These innovations in the securities markets have had substantial economic benefits. For instance, academic research has found that securitization has had a positive impact on both the availability and cost of credit to households and businesses.5
Fitch firmly believes in the power of competition. We also believe that there is always a demand for insightful, independent credit research.
As noted above, while the NRSRO system is often cited as a barrier to entry for new rating organizations, we believe that the debate over the NRSRO system ignores the single most important barrier to entry in the ratings market: the Moody's/S&P monopolies.
Moody's and S&P are a dual monopoly, each possessing separate monopoly power in a market that has grown to demand two ratings. Each engages in practices designed to perpetuate its market dominance and extend it to otherwise competitive markets such as structured finance. As we have publicly stated for two years, through their discriminatory practice known as "notching", Moody's and S&P successfully altered competition in the commercial and residential mortgage-backed securities market by leveraging their monopoly position in other markets.
If the SEC wishes to address barriers to entry in the ratings market and ensure competition in the ratings market, the Commissioners should enact rules prohibiting anticompetitive conduct by NRSROs and precluding NRSROs from discriminating against the ratings by other NRSROs for the purpose of preserving market share. Fitch believes that this is an area which would benefit from SEC regulation to protect NRSRO competition. Fitch believes that any NRSRO found to be using anti-competitive practices or unfair business practices should have their NRSRO designation revoked.
The Commission also has expressed concerns that certain practices regarding unsolicited ratings may be anti-competitive or constitute unfair practices. Fitch does not believe that the issuance of unsolicited ratings, in and of themselves, is anti-competitive or unfair.
As members of the financial media, rating agencies must be free to publish opinions as to any issuer or issue that they determine to be of interest to the investing public. Fitch will only publish a so-called unsolicited rating if we conclude that there is sufficient information available to us to allow us to express our opinion, and in all cases, such ratings are uncompensated and Fitch does not assess or seek fees for the analysis done in connection with these ratings. In 2001, Fitch introduced its Fitch Initiated Ratings program. Fitch Initiated Ratings target high profile market participants or issuers about which there is a discrepancy in market opinions not traditionally rated by Fitch. Ratings initiated under this program are identified as such in the original publication concerning the rating.
There is no difference in the analytical process or criteria used for Fitch Initiated Ratings, although the level of management involvement varies. Procedures relative to the publication of the ratings are also the same and we contact the issuer prior to publishing a new rating or subsequent rating action in accordance with our regular practices.
An NRSRO needs to retain the ability to withdraw a rating for whatever reason it deems necessary, including lack of payment. A contrary rule would increase the potential for conflicts of interest because there might be a greater reason to seek to please the issuer. In addition, issuers who stop paying or refuse to pay for a rating often do not provide adequate information for the agency to maintain the rating.
G. Information Flow
Fitch believes strongly in transparency in the ratings process. Accordingly, at Fitch, there are hundreds of criteria reports published highlighting the methodology we use to rate various types of entities and securities, together with detailed sector analysis on a broad array of sectors, companies, and issues, all available free on our web site (www.fitchratings.com). Fitch has also been a leader in publishing so-called presale reports in the areas of structured finance, global power, project finance and public finance where our published analysis of various transactions of interest to the market is made available free of charge on our web site prior to the pricing of the transaction. In addition, Fitch makes available free of charge on our web site all of our outstanding ratings. Announcements of ratings actions are also distributed through a variety of wire services as mentioned above. Criteria and methodology reports, as well as industry and general topical research, also are freely available on our websites.
Certain of our publications and data are only available to our paid subscribers. We commit extensive time and resources to produce our publications and data and we believe they are valuable to anyone interested in objective credit analysis. In this practice we are no different than other members of the financial media, such as Bloomberg, Dow Jones, Thomson Financial and others, that charge subscribers for access to their publications and data services.
Currently, Fitch issues a press release when it withdraws a rating or ceases covering an issuer, unless an issue is called in for redemption or matures. Fitch believes that the market should be aware when we cease rating an issuer as a result of our own decision making process. However, when an issue is called in for redemption or matures, there is no added benefit to the market in issuing a press release explaining the obvious fact that we are withdrawing our rating on the debt. This should not be required.
We also believe that recognition should be conditioned upon the widespread public dissemination of ratings and rating actions and annual publication of transition and default studies setting forth the performance of ratings over time by ratings categories.
However, we do not believe that it is necessary or appropriate to require the disclosure of whether a rating is unsolicited, whether the issuer has cooperated in the rating process or key bases of, and assumptions underlying, the rating. We believe that any such requirements inappropriately interfere in the editorial process of the rating agencies and demean the value of ratings done using publicly available information and alternative sources of views of the creditworthiness of an issuer or issue (i.e. other market participants, customers, suppliers, competitors, regulators). Ultimately, the best oversight of the effectiveness of unsolicited and uncooperative ratings is monitoring the overall performance over time of the ratings of the organizations that issue them.
While we believe that for the most part credit rating agencies have adequate access to the information they need to form an independent and objective opinion about the creditworthiness of an issuer, improved disclosure by issuers would be welcomed by Fitch. As we found in our recently published study of the use of credit derivatives in the global market, financial reporting and disclosure with respect to areas such as credit derivatives, off-balance sheet financing and other forms of contingencies vary greatly by sector and comparability is further obscured by differences in international reporting and accounting standards. The Commission should continue to advocate for clear and complete disclosure of the true financial position of issuers without regard to the vagaries of accounting standards.
Please call me at (212) 908-0626 with any questions that you might have on our comments or to discuss this matter further at your convenience.
1 All of the trading data with respect to credit default swaps provided herein is based on data provided as of July 11, 2003 by GFI Group Inc.
2 For a further discussion of the use of benchmarks in evaluating ratings, see The New Basel Accord (April 2003), Basel Committee on Banking Supervision, Bank for International Settlements.
3 "It is important to recognize, however, that there are fundamental differences between the work of equity analysts, whose recommendations are often directed at specific groups of investors to influence their judgment on effecting a securities transaction, and credit rating analysts, who do not make recommendations to specific investors, but rather publish their opinion on the creditworthiness of a particular company, security, or obligation, as f a specific date (emphasis added)." Memorandum to Chairman Donaldson from Annette L. Nazareth, Director, Division Market Regulation, June 4, 2003 at Page 7.
4 Lawrence J. White, The Credit Rating Industry: An Industrial Organization Analysis, June 2001 (paper presented at the conference on "Rating Agencies in the Global Financial System", presented at the Stern School of Business, New York University, June 1, 2001.
5 Mark M. Zandi, The Securitization of America, Regional Financial Review, February 1998; Ali Anari, Donald R. Fraser and James W. Kolari, The Effects of Securitization on Mortgage Market Yields: A Cointegration Analysis, Real Estate Economics, 1998.