U.S. Securities & Exchange Commission
SEC Seal
Home | Previous Page
U.S. Securities and Exchange Commission

Speech by SEC Commissioner:
Statement regarding the NRSRO Proposal before the Open Meeting


Cynthia A. Glassman

U.S. Securities and Exchange Commission

Washington, D.C.
March 3, 2005

Over the years, commenters on the various rule proposals and concept releases we've issued on credit rating agencies have made it clear that greater transparency is needed with respect to the process by which the staff assesses applications for NRSRO status and the criteria they use. The proposed rule, which would for the first time define the term NRSRO, makes incremental improvements in both regards, but it does not, in my view, address the anticompetitive aspects of the NRSRO process.

Since it was introduced in 1975 in the context of our net capital rule, the NRSRO concept has become ubiquitous. It has migrated into many other SEC rules as well as unrelated U.S. and foreign regulations. At the same time, our no-action process for recognizing NRSROs is opaque. The process can only be described as Kafkaesque. Rating agencies cannot be an NRSRO unless their ratings are nationally recognized, yet they cannot achieve national recognition without being recognized by the SEC staff as an NRSRO, a classic "chicken and egg" problem. Applicants typically wait years to get no-action letters. Applicants that do not meet the staff's criteria also wait years without being told, so they say, exactly why they don't meet the criteria. Our recognition process bestows the Commission's "seal of approval" on NRSROs with all the reputational and financial benefits that flow from it, yet the absence of statutory authority and the rating agencies' claims under the First Amendment prevent us from exercising substantive oversight.

Having a proposed definition of "NRSRO" in a Commission rule provides more transparency regarding the factors necessary to NRSRO recognition and more guidance as to how rating agencies can comply with the definition. At the same time, I do not believe that the proposed definition goes far enough in addressing the anticompetitive effects of NRSRO recognition. As the Department of Justice pointed out in a March 1998 comment letter, the requirement that a rating agency be nationally recognized acts as a "nearly insurmountable" barrier to entry for new credit rating agencies. Keeping the national recognition requirement entrenches the existing NRSROs and eliminates the possibility of NRSRO status for any innovative models.

I recognize that the proposed definition has been expanded to include rating agencies that are limited to particular sectors of the debt market or particular geographic sectors. Such agencies would still need to demonstrate, however, that their ratings are nationally recognized. The appeal of the "provisional" NRSRO category, which the staff is not recommending, is that new entrants that meet all other qualifications for NRSRO recognition except for national recognition would get the opportunity to develop national recognized ratings over time instead of being barred from entry.

In the same vein, I question the assumption that ratings based on qualitative factors, specifically, access to senior management, are necessarily more credible and reliable than ratings based on quantitative models. I am not aware of a substantive basis for this distinction, and I agree with commenters who view this requirement as another barrier to entry for smaller or newer entrants without access to higher levels of management.

Encouraging competition in the credit rating agency arena should be our major focus, and this means reducing SEC-created barriers to entry. Once the SEC created the NRSRO concept and required rating agencies to obtain the approval of the staff, market participants could not realistically rely on the ratings of any agency that didn't have our seal of approval. This is why the national recognition requirement is a barrier to entry and why we should be focusing on minimizing the barriers and increasing competition. More NRSROs and increased competition among them would likely improve the quality of ratings, reduce the cost of ratings and resolve many of the other problems that we are attempting to micromanage through regulation, including those related to conflicts.

Turning to the process, the proposal includes two improvements. First, we are proposing that no-action requests would be granted on a time-limited basis. This would mean that, after some yet to be determined period of years, the no-action relief would expire, and an NRSRO would have to make a new application for the relief. Second, to speed up our process, notice will be given to the Commission when an application for no-action relief is first received, and applicants will have some ability to bring their concerns directly to the Commission. In my view, these could be the best parts of an otherwise modest proposal.

So I support issuing the proposal for comment, although in reality it does little to change the status quo or remove impediments to competition. We do not know whether any NRSROs will participate in the voluntary framework or how effective it would be. At the end of the day, the Commission needs to decide whether we can live with essentially the status quo. If so, we need to make clear to the world that the term NRSRO means nothing more than that our staff thinks an agency meets our definition. If not, it will be up to Congress to decide whether to give us additional authority.

I have some questions.

  • What is the status of the voluntary framework?
  • Providing a definition of NRSRO would mean - at least theoretically -- that a rating agency that satisfies the elements of the definition would not need to obtain a no-action letter. Is it practical to expect, though, that agencies will bypass the no-action process?
  • The release describes "nationally recognized" as "wide acceptance in the marketplace" or acceptance by a "substantial percentage" of the relevant market. What is a "substantial percentage"? And how will new entrants demonstrate their "general acceptance in the financial markets"?
  • In the proposed definition, what is "current"? Is quarterly sufficiently current? If not, what is, and how do we know how frequently the agencies update their information and their assessments?
  • Why have we dropped the provisional NRSRO concept? Without it, we have not solved the chicken and egg problem.
  • Why have we not addressed the solicited vs. unsolicited ratings issue, especially given the inherent conflicts and evidence that unsolicited ratings tend to be lower?
  • If an agency is nationally recognized and accepted, however we define it, why do we care how it staffs itself and what specific techniques it uses? The point is that the market has already judged that.


Modified: 03/04/2005