Recently the Staff of the Division of Trading and Markets (the “Staff”) of the U.S. Securities and Exchange Commission (the “Commission”) published a report to Congress outlining the findings from the study the Commission was required to conduct by Section 939F of the Dodd-Frank Act (the “Report”).1 The Commission was directed to deliver a report not later than 24 months after the date of enactment of Dodd-Frank containing the findings of a study on, among other things, matters related to (1) assigning credit ratings for structured finance products and (2) any recommendations for regulatory or statutory changes that the Commission determines should be made to implement the findings of the study.

In creating the Report, the Staff gathered information, conducted meetings with proponents of various models and other stakeholders, and reviewed 32 comment letters – six from nationally recognized statistical rating organizations (“NRSROs”) and 26 from other interested parties, including investors, trade organizations, non-profit organizations, brokerage and financial services firms, academics and individuals.

The Commission’s Task

In particular, the Commission was tasked to examine:

  • the credit rating process for structured finance products and any conflicts of interest associated with the issuer-pay and the subscriber-pay models;
  • the feasibility of establishing a system in which a public or private utility or a self-regulatory organization (“SRO”) assigns NRSROs to determine the credit ratings for structured finance products, including:
    • an assessment of potential mechanisms for determining fees for NRSROs for rating structured finance products,
    • appropriate methods for paying fees to NRSROs to rate structured finance products,
    • the extent to which the creation of such a system would be viewed as the creation of a moral hazard by the federal government, and
    • any constitutional or other issues concerning the establishment of such a system;
  • the range of metrics that could be used to determine the accuracy of credit ratings for structured finance products; and
  • alternative means for compensating NRSROs that would create incentives for accurate credit ratings for structured finance products.

Subsequent to submission of the Report, Section 939F provides that the Commission shall conduct rulemaking to implement an initial credit rating system as it determines is necessary or appropriate in the public interest or for the protection of investors, in a manner that prevents the issuer, sponsor or underwriter of the structured finance product from selecting the NRSRO that will determine the initial credit rating and monitor such credit ratings.

The Report summarized several different mechanisms for initial credit rating assignments, as well as a number of alternative compensation models. For each, the Staff organized the pros and cons of each program as pointed out by commenters, then weighed them, and came up with an ultimate evaluation of three “programs” – the 15E(w) system, the 17g-5 system, and “alternative models”.

The 15E(w) System

The 15E(w) system, based on the Franken Amendment,2  which was proposed as part of the Dodd-Frank Act, is the default system contemplated by Congress through the Dodd-Frank Act. Under a 15E(w) system, a “CRA Board” would be created to assign a Qualified NRSRO (defined below) to each issuer of a structured finance product who seeks an initial credit rating. NRSROs would be required to become pre-qualified with the CRA Board to be eligible to rate all or specific types of structured finance products based on criteria set forth by the CRA Board (a “Qualified NRSRO”). Issuers would also be able to secure additional credit ratings outside of this process by hiring non-assigned NRSROs. The selection of a Qualified NRSRO would be based on some combination of random chance, demonstrated accuracy in past ratings, and potentially the specific expertise needed for that type of structured finance product. The CRA Board would also be charged with qualifying NRSROs, monitoring or setting fees, as necessary, and evaluating the accuracy of each NRSRO’s ratings performance.

The Staff found that the primary benefit of this system would be to mitigate any issuer-pay conflict of interest. Issuers would be removed from the initial selection process, which the Staff indicated would result in less pressure on NRSROs to cater to issuers to ensure repeat business. The second benefit of this system, according to the Staff, would be the CRA Board’s ability to reward good performance and punish poor performance, as the CRA Board could consider past performance in assigning future ratings assignments. Other benefits identified by the Staff included providing smaller NRSROs the opportunity to develop their reputations through obtaining random assignments and the representation of the investor community on the CRA Board.

The Staff was not without concerns about this system. Principally, the Staff noted that such a system may not fully address ratings shopping, as issuers could hire additional NRSROs to provide supplemental initial credit ratings, so both the NRSROs they hired, as well as assigned NRSROs, might feel pressure to provide favorable ratings so that they might be hired in the future. Second, some NRSROs could feel that they have enough market presence already and choose not to become Qualified NRSROs, on the assumption that they will be hired to rate transactions outside of the 15E(w) system. The Staff suggested that this nonparticipation could result in the pool of Qualified NRSROs lacking the capacity or expertise to rate certain types of structured finance products brought to market, which could delay or stall the issuance of those types of securitizations.  Additional concerns mentioned by the Staff included the continued existence of investor requirements that investments be rated by certain NRSROs, increased costs to issuers and investors associated with obtaining ratings, operational complexity, the high costs of the CRA Board needed to implement and administer such a system, and First and Fifth Amendment questions. Finally, the Staff noted such a system could run counter to the Dodd-Frank mandate of reducing reliance on credit ratings by creating a category of “government-sanctioned” NRSROs and even an implicit government guaranty of credit ratings, creating the potential for moral hazard.

The Staff also set forth another set of concerns that were touched off by the requirement that assignments be distributed based on merit. First, the Staff noted that this merit-based incentive could lead NRSROs to be overly conservative in their initial credit rating, to avoid the need for a downgrade later on. Additionally, the Staff suggested that NRSROs could be reluctant to downgrade a credit rating, since that might damage their standing with the CRA Board. The Staff was concerned that these possibilities could combine to lead to decreased accuracy of credit ratings. Additionally, the Staff recognized concerns that this system could discourage innovation and improvements to methodologies by NRSROs and lead to homogeneity of credit ratings to satisfy the particular metrics chosen by the CRA Board. Finally, the Staff noted that merit assignments would put the CRA Board in the position to evaluate performance of NRSROs, which could require significant expertise and diligence on its part, require the creation of models that considered market factors, require a substantial dedication of resources and result in new conflicts of interest within the CRA Board.

The 17g-5 System

Rule 17g-53 of the Exchange Act, which is currently in effect, creates a mechanism wherein an NRSRO not hired by an issuer to rate a given transaction can nonetheless obtain the same information the hired NRSRO has access to, and then can simultaneously and independently develop an additional initial credit rating and conduct surveillance on that credit rating along with the hired NRSRO. The program was designed to prevent the arranger of a structured finance product from selecting a single NRSRO to exclusively determine an initial credit rating.4 Under this system, NRSROs are required to rate 10% of the transactions they view that they have not already been hired to rate.5 The Commission considered possible enhancements as described below to the 17g-5 system as it currently functions.6

As with the 15E(w) system, the Staff found that the primary benefit of the 17g-5 system is to mitigate any issuer-pay conflicts of interest. The Staff indicated that, if it is used to produce unsolicited ratings by non-hired NRSROS, it would encourage hired NRSROs to provide more accurate, unbiased credit ratings, given that any bias or inaccuracy might be exposed by a non-hired NRSRO. Additionally, according to the Staff, such a system would give investors alternatives to the hired NRSRO, as investors would be able to hire their NRSRO of choice to rate any transaction, as the investor’s NRSRO also has complete access to information. Other benefits identified by the Staff include decreased barriers to entry in the NRSRO market, increased accountability and transparency in the credit rating process, and the small administrative and financial burdens to implementing and perpetuating such a system, as it closely tracks the system already in place.

The Staff noted that the 17g-5 system is not without its own set of shortcomings. Their foremost concern is that most NRSROs have not been using it to produce unsolicited credit ratings, which has been attributed to the requirement that NRSROs rate at least 10% of the transactions they view. Reasons given by commenters for this nonparticipation include that ratings are too costly to produce for free, from both the initial production and maintenance standpoints, and that NRSROs are not compensated for or protected from future potential liability due to inaccurate ratings. Other concerns identified by the Staff include a concern that issuers will stop hiring NRSROs if they can expect several free ratings, the alleged difference in confidentiality requirements imposed on hired versus non-hired NRSROs and the inability of non-hired NRSROs to request additional information they might need to use in their particular methodology.

The Staff acknowledged the most needed improvement to the 17g-5 program may be to modify the 10% requirement. The Staff suggested that the 10% requirement could be deleted entirely, or reduced to a smaller number, like 5% or 2%. Alternatively, the Staff indicated that the requirement could be changed from a requirement to issue a credit rating, to a requirement to provide unsolicited commentary to the extent an NRSRO disagrees with a rating issued by a hired NRSRO. According to the Staff, this would be less of a burden on non-hired NRSROs, as they would not incur the initial costs of producing a rating, nor the costs of maintaining the rating. This would maintain the incentive for hired NRSROs to provide more accurate ratings, given the chance of being contradicted, and would lower the barrier to entry even further, allowing non-hired NRSROs to build a track record through their commentary and analysis, according to the Staff. Other suggestions for improvement mentioned by the Staff included making sure that hired and non-hired NRSROs are treated the same by an arranger when it comes to confidentiality agreements and dealing with concerns about liability for non-hired NRSROs, potentially through the establishment of statutory or regulatory safe harbors.

Other Systems

The Staff briefly described a number of other systems, including an investor-owned credit rating agency model, a stand-alone model, a designation model, a user-pay model, and a number of hybrid models. The Staff noted that some of these models might help to mitigate any issuer-pay conflict of interest to varying degrees, but often at the expense of adding one or more different potential conflicts of interest. Other potential benefits identified by the Staff from some of these models included increased quality of credit ratings, increased competition among NRSROs, and increased market choice for users of credit ratings. On the other hand, the Staff found that many of these alternatives would not be feasible to administer, that the Commission lacked Congressional authority to enact many of them, and that there were not clear workable compensation models for NRSROs, which could lead to decreased NRSRO participation.


Now that the Staff has published the Report, the Commission is required to go through the rulemaking process to implement the 15E(w) system as it determines is necessary or appropriate in the public interest or for the protection of investors, unless the Commission determines that an alternative system would better serve the public interest and the protection of investors. 

Ultimately, the Report recognized that none of the proposed systems are workable as currently conceived, and recommended that prior to any rulemaking, the Commission conduct an additional study of relevant information, including information, as applicable, related to the costs and benefits of the course of action and the consideration, as applicable, of public comment.

The Staff recommended that the next step be to convene a roundtable at which proponents and critics of the three courses of action are invited to discuss the study and its findings. The timing of this next step has not yet been announced. As a result, any rulemaking, or final rule is likely far off on the horizon. It is worth noting that an SEC Commissioner recently commented that as part of this continuing discussion about ratings of structured finance products, it might be worthwhile to analyze whether investors are relying too much on ratings, and if so, why they are.