Subject: File No. S7-12-03 July 28, 2003


Feedback 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 corporate credit rating plays a central role for an international company like Siemens, as an efficient credit rating process enables rapid, flexible and cost-effective access to the financing resources of the international capital markets. It is essential, given the great significance of ratings, that credit ratings are fair, current, clear and comparable.

To ensure these features of ratings, there must be a sufficient level of competition between agencies, agencies must employ a clearly defined and transparent rating process, a confidential dialog between the respective issuer and agencies must be in place and agencies must be immune to inappropriate attempts to exert influence over their actions.

A. Alternatives to the NRSRO Designation (in response to Question 1)

The use of credit ratings in connection with the federal securities laws and, as a necessary condition therefore, the registration of rating agencies with the SEC should be retained. Credit ratings issued by rating agencies employing high level of diligence according to standards of best practice are a valuable source of information for investors. Use of reliable credit ratings also provide an economically expedient approach of distinguishing between securities of different grades of creditworthiness for regulatory purposes (for example the Net Capital Rule). We have yet to see any practical alternatives to the use of credit ratings for such purposes.

In our view, the concept of "NRSRO" provides an appropriate instrument to impose the essential regulations and compliance to ensure that credit ratings of registered rating agencies achieve the characteristics essential to the operation of an efficient credit rating system. Nevertheless, the existing concept of the NRSRO should be enhanced by incorporating the amendments listed under items B to F.

We would like to assume that if and when other international institutions come to an equally relevant regulatory power that in turn their judgment should be equally recognized by the SEC.

B. Recognition Criteria (in response to Questions 11-16, 18-21, 23, 26, 27)

Active competition between rating agencies is in our view the single most important driver to ensure the quality of ratings and rating analyses. A transparent recognition process can provide unregistered agencies with a clear understanding of the requirements for registration and may therefore lower barriers of entry. Yet, the required capabilities and generally accepted standards of diligence must be clearly delineated. Furthermore, application for SEC recognition should be publicized by the Commission; SEC approval or rejection should be announced 90 days after receipt of all required information. In case of rejection, the Commission should provide a clear feedback to applicants in which areas improvement is necessary to gain recognition.

In order to maintain the desired rating qualities, the following criteria are of particular importance:

  • Clearly defined, readily comprehensible ("plain English"), published and auditable rating methodology:

    - Presentation of the specific underlying rating methodology used; the methodology should consist of a quantitative as well as a qualitative part - for example numerous large international diversified companies, such as ours, are too diverse and too complex for ratings to be derived predominantly on the basis of computerized statistical models;

    - The general methodology of the quantitative element has to be well-defined by each agency, including definitions, meaning and relevance of ratios, the respective formulas to compute these, applied adjustments, benchmarks and ratio guidelines for the appropriate rating category within a industry, as well as a short summary. The qualitative element should carefully lay out, among other things, relevant assumptions for future prospects of the company like competitive environment, growth path of the industries and the respective issuer, margin development, etc. Agencies should also be required to publicize any changes in the underlying rating methodology in advance (at least three months before the changes involved come into effect);

    - Involvement of senior management of rated companies should be an inherent part of the rating process in order to permit a qualified assessment of credit risks. One implication of this is that rating agencies are insiders and must comply with the corresponding regulations (subject to the provisions of Regulation FD);

    - Duty to publish for each single rating the way the methodology was applied for assessing the credit quality of the respective issuer, the individual results and relative weighting (including the rationale) of the quantitative and qualitative aspects of the analysis, analysts and members of rating committee involved in the process to allow better understanding of the rating by all market participants, reconciliation of ratios computed by the agency to the disclosed financial figures (in particular publication of any adjustments and projections used in the quantitative analysis);

  • Internal documentation of every single rating process (primarily in order to document consistent adherence to the rating methodology).

  • The size, experience and training of the rating agency's staff, in particular to ensure ratings consistency for example along all relevant generally accepted accounting principles or industries.

  • Sufficient financial resources to help guarantee independence of rating agencies from any particular issuers.

  • Organizational structure and technological infrastructure of the rating agency

  • Sufficient information flow to investors and issuers (see F).

A general rating methodology should not be prescribed by any legislative body. It is important that the agencies retain the authority to issue an independent opinion on creditworthiness and, where necessary, the ability to react quickly to changing market conditions by amending the methodology used. However, any such amendments made must be made public and explained.

Rating agencies specializing in particular market segments and/or regions should also be able to register with the SEC. These specialist ratings are likely to further encourage competition between agencies and enhance the quality of information available to investors.

C. Examination and Oversight of NRSROs (in response to Questions 28-32, 34)

To ensure that the quality of rating analyses remains constant, compliance with the recognition criteria should be reviewed and published annually by the Commission. For detailed controls, random checks of the agencies compliance with their published methodology and process, which might also include viewing internal notes or interviewing the analysts and members of the rating committee.

From our perspective, at least once a year each recognized agency should publish indicators allowing an assessment of the quality of its ratings (for example default statistics, transition matrices, overview of rating trend of issuers in default, number of downgrades of two notches or more).

The current rating process, especially in light of the limited number of rating agencies, is without sufficient checks and balances to avoid arbitrary or grossly flawed conclusions in a rating. There are few - if any - instances where the weight of a judgement is so significant while at the same time its background can be quite opaque. Issuers cannot appeal to an independent authority when they encounter conduct felt to contravene the regulations or when they have good reason to question the fairness and objectivity of a rating process. To this end, an appeal process should be established in form of a neutral board or "Ombudsman".

Such a board of appeal must, of course, have the necessary powers to investigate the complaints received (for example prompt access to the documentation relating to a specific rating process, interviews with analysts and members of the rating committee) and have appropriate resources to fulfill its mandate. Penalties up to and including the withdrawal of SEC registration should be available as sanctions in the event that any systematic or gross negligent breach of the recognition criteria for approval is identified.

D. Conflicts of Interest (in response to Questions 39-41, 43)

Rating agencies are already required to avoid all conflicts of interest, as this is a long recognized prerequisite to ensuring the availability of fair, current, clear and comparable credit ratings.

The generation of credit ratings on behalf of and at the expense of issuers does not present inappropriate conflicts of interest so long as (i) there is sufficient competition between the agencies and (ii) the agencies have independent financial strength and (iii) the agencies are subject to market discipline for failures to provide neutral and objective credit ratings. The following conditions should be enforced:

  • Unsolicited ratings should not be generated by recognized agencies; at minimum such ratings should not be relevant for regulatory purposes such as in connection with the federal securities laws and must be clearly identifiable from the rating itself (for example by the prefix "PI-": based on public information).

  • Commercial, organizational and personal isolation of credit rating analysis from consulting and advisory services has to be guaranteed, this also applies to rating evaluation services for which a charge is made.

  • Communication between analysts and investors helps to improve understanding in, and the quality of, rating decisions and should therefore not be prohibited. The dialogue between analysts and investors should be effected during pre-announced events which could be attended by the interested public, e.g. conference calls, analyst meetings etc. Such events should be carried out an regular basis.

  • Fees paid by each respective issuer should have no decisive economic relevance for an agency registered with the SEC. The SEC's proposal to limit the fees to a certain percentage of total revenue should be pursued.

  • Analysts' compensation should be independent of the level of the fees paid by the issuers covered by the individual analyst or such analysts development of new business from issuers. Analysts must be isolated organizationally from other parts of the agency, e.g. marketing department.

The implementation of measures to avoid conflicts of interest are also to be incorporated into the record-keeping duty of the agencies. Disregard of these conditions should be published by the Commission, in severe cases the Commission should have the authority to withdraw the recognition.

E. Alleged Anticompetitive, Abusive, and Unfair Practices (in response to Questions 47, 48)

The SEC recognition should be conditioned on agencies not engaging in unsolicited ratings, consulting or advisory services and any actions which jeopardize, or otherwise adversely affect, competition between agencies. This should help prevent anticompetitive, abusive and unfair practices.

Furthermore, issuers should have the opportunity to invoke a neutral board of appeal if there is good reason to question the fairness and objectivity of its rating (see section C above).

F. Information Flow (in response to Questions 49-53)

SEC-registered rating agencies should be compelled to disseminate to the public its ratings as well as the basis of, and assumptions underlying, each single rating decision, the employed rating methodology and every amendment to this methodology, the specific realizations of the recognition criteria and measures to prevent conflict of interest and anticompetitive, abusive and unfair practices (compare sections B, C, D, E). Furthermore indicators allowing measurement of rating quality (section C) should also be published so that investors and issuers can form their own opinions as to the quality of the rating process and credit ratings of individual rating agencies. The board of appeal should also publish the complaints received, the results of follow-up investigations and further measures on an appropriate website.

Where issuers commission SEC-registered rating agencies to generate credit ratings, all investors should have access, free of charge, to these corporate ratings and the reports published by the agencies on the website of the agency concerned.

The agencies should also be required to promptly disclose proposed changes in the quantitative and qualitative analysis so that issuers have the chance to engage in a proactive dialog with the agencies and are not taken by surprise by the timing of rating changes. This also applies in connection with regular rating reviews by rating committees. Threshold values for rating metrics that trigger a change in the ratings must be communicated clearly. These should perhaps also be accompanied by the qualitative expectations of the agencies so that issuers and market participants can more fully understand and predict changes to credit ratings.