International Swaps and Derivatives Association
Jonathan G. Katz
RE: File numbers S7-22-03 and S7-21-03
February 4, 2004
Dear Mr. Katz,
The International Swaps and Derivatives Association appreciates the opportunity to comment on the U.S. Securities and Exchange Commission's risk-based capital proposals for Supervised Investment Bank Holding Companies (68 Fed. Reg. 62910, November 6, 2003) and for Consolidated Supervised Entities (68 Fed. Reg. 62872, November 6, 2003) (together, the "Proposals").
ISDA is an international organization, whose members include more than 600 of the world's largest commercial, universal, and investment banks as well as other companies and institutions with extensive activities in the area of swaps and other privately negotiated-commonly known as "over-the-counter" (OTC)-derivatives. ISDA pays close attention to regulatory developments in a variety of jurisdictions that might affect the ability of ISDA member firms to manage their risks effectively and at reasonable cost. And because ISDA member firms have a particularly important stake in the outcome of efforts to reform the risk-based capital regime, ISDA members and staff have participated actively in the Basel II process.
During 2003, for example, ISDA commented on the Basel Committee's Third Consultative Paper ("CP3")1, the U.S. Federal banking regulators' Advance Notice of Proposed Rulemaking on the New Basel Capital Accord (the "ANPR")2, and on the European Union's Risk-Based Capital Directive.3 ISDA is pleased to comment now on the SEC Proposals.
In this comment letter, we will address three issues, namely, counterparty credit risk, credit derivatives, and operational risk. We defer generally to the Securities Industry Association (SIA) on other issues raised in the Proposals.
Counterparty credit risk
ISDA member firms have a vital interest in the capital treatment of counterparty risk. And the U.S. investment banks have a particularly strong interest in the outcome because they rank among the largest OTC derivatives dealers worldwide. ISDA is therefore pleased to have the opportunity to present its views on counterparty risk to the SEC.
ISDA's efforts in pursuit of consistency between economic and regulatory capital measures of counterparty risk have been extensive. In its comment on the Basel Committee's Second Consultative Proposal in 20014, ISDA proposed a measure of counterparty credit future exposure, known as expected positive exposure (EPE), as an alternative to peak exposure. The Models Task Force of the Basel Committee found the EPE proposal sufficiently compelling as to request further elaboration and investigation. In response, ISDA (jointly with The Bond Market Association and the London Investment Bankers Association) conducted a survey of its members' counterparty risk management practices. We released the survey results and the results of additional work on the EPE concept in June 2003 (Appendices A and B).
The Basel Committee did not, however, make reform of the treatment of counterparty risk a priority in completing the New Basel Accord, so the New Accord treatment remains similar in most respects to that in the original Basel Accord. But Governor Caruana, current chairman of the Basel Committee, has committed the Basel Committee to revisit the counterparty credit risk issue as soon as the New Accord is signed this summer, and the Committee recently announced that it plans "a review of counterparty credit risk and trading book issues in coordination with the International Organization of Securities Commissioners".5 ISDA looks forward to working with the Basel Committee and other regulatory bodies on the issue, and hopes that a new capital treatment will be ready in time for implementation along with the New Accord in 2007.
It its Proposals, the SEC has moved beyond the current version of the New Accord by setting forth a new treatment of OTC derivatives counterparty risk. The methodology, which is found in Rule 17i-7 (68 Fed. Reg. 62938), calculates Credit Equivalent Amount as sum of Current Exposure and Maximum Potential Exposure (MPE). The SEC further defines MPE as:
Finally, the VAR model must "use a 99 percent, one-tailed confidence level with price changes equivalent to a one-year movement in rates and prices" (§240.15c3-1e(e)(2)(ii), 68 Fed. Reg. 62901). The SEC has requested comment "on whether the proposed method of calculating the credit equivalent amount is appropriate, or whether it should be changed" (68 Fed. Reg. 62921).
On the positive side, ISDA strongly supports the SEC in allowing firms to calculate current exposure and MPE at the counterparty instead of the transaction level, in recognizing the effect of netting arrangements, in taking account of collateral posted by the counterparty, and in recognizing the protection from credit derivatives. In addition, the Proposal allows a regulated firm to choose between the SEC treatment and that developed by the Basel Committee (68 Fed. Reg. 62920). This option allows for the flexibility for institutions to benefit by advances in counterparty risk treatment and avoids locking in one measure.
But on the negative side, ISDA disagrees with the choice of MPE as the measure of future exposure. We acknowledge that the SEC's choice of MPE instead of a notional add-on represents an advance in that it allows institutions to use their internal models to calculate required capital. But as we have pointed out to the Basel Committee, we believe that MPE models-whether they take the form of VAR or other worst-case models-are excessively conservative and that they overstate the actual risk involved in OTC derivatives. The problem is that MPE measures go beyond conservatism by implicitly and implausibly assuming that exposure to a counterparty would always be the worst case amount whenever default occurs. EPE-based measures, in contrast, anticipate default occurring over the entire distribution of exposures.
Further, the statement that "broker-dealers generally use maximum potential exposure to measure and manage the credit risk of their portfolios" (68 Fed. Reg. 62921) is not borne out by the evidence collected by the ISDA-TBMA-LIBA Survey of Counterparty Risk Management Practices (Appendix B, p. 5). There we found that financial institutions generally use EPE or adjusted EPE measures (such as loan equivalent exposure) for allocating economic capital. And we found that institutions generally use worst-case MPE measures for limit setting and not for allocating economic capital.
Along with asking for comment on whether the proposed measure of credit equivalent amount is appropriate, the SEC has asked for suggestions on how the measure should be changed. We propose that, instead of MPE, the SEC adopt the future exposure measure that ISDA has proposed to the Basel Committee and looks forward to discussing again this summer (Appendix A). That is, future exposure should be measured not as MPE but as EPE adjusted upward by a multiplier (_) that reflects non-zero pair-wise correlations between counterparty exposures and less than infinite granularity in an OTC derivatives portfolio. As in the SEC Proposal, the measure should also recognize the risk reduction benefits of netting, of collateral, and of credit derivatives protection. Along with providing a more plausible measure of counterparty exposure that more closely approximates economic capital, using a measure based on EPE is generally more consistent with practices at major dealers than is MPE.
Other than use of MPE, ISDA has one more concern regarding counterparty risk, namely, the loss given default (LGD) measures implied in risk weights in the Proposal (68 Fed. Reg. 62921). The weights are much higher than those proposed by Basel; the reason is that the Proposal assumes a 75 percent LGD. Such a loss rate is high relative to the historical experience; an LGD closer to 50 percent would be more appropriate.
Finally, ISDA has argued that OTC derivatives and securities financing transactions (SFT) such as repo often exhibit similar counterparty risk characteristics and should therefore receive uniform treatment under the New Accord. By adopting similar measures for OTC derivatives and SFT, the SEC has moved in the direction of consistent treatment. But we hope that the SEC-and the Basel Committee when it revisits the treatment of counterparty risk in OTC derivatives-maintains this emphasis on consistent treatment by ensuring that any improvements in the treatment of OTC derivative counterparty risks, such as use of EPE instead of MPE, apply also to the treatment of SFT.
The Proposal contains two approaches to capital treatment of credit derivatives. One is the substitution approach, which substitutes the risk weight of the protection seller in place of the risk weight of the reference credit (§17i-7 c(1)(ii)(E), 68 Fed. Reg. 62939). The other, which we will call the offset approach, reduces the amount of current exposure by the amount of protection provided by the credit derivative (§17i-7 c(1)(i)(D), 68 Fed. Reg. 62938). The offset approach apparently applies only to counterparty exposures, but it is unclear from our reading whether `counterparty positions' include only OTC derivatives or if they also include SFT and other trading book exposures giving rise to counterparty risk. ISDA would appreciate clarification on this point.
We appreciate the SEC's offering an offset approach as a potential alternative to substitution as noted above. ISDA has consistently opposed the substitution approach and has argued for alternatives that recognize double default effects arising from joint probability of default by both the reference credit and the protection seller.
ISDA found it encouraging that the Federal Reserve took double defaults seriously enough to issue a white paper on the subject during 20036; the U.S. bank regulatory agencies also requested comment on the double default issue in their ANPR. ISDA's comment on the white paper (with the International Association of Credit Portfolio Managers and the London Investment Bankers Association) is attached as Appendix C.
In its white paper, the Federal Reserve proposed an approach, known as the Asymptotic-Single-Risk Factor (ASRF) framework, that takes account of double default effects in setting capital charges for hedged exposures. ISDA supports the ASRF framework, and has proposed a calibration for the model (Appendix C, pp. 3-5). And recently, the Basel Committee announced that it has formed a working group to consider the issue:
ISDA looks forward to working with the Basel Committee on the issue, and is similarly eager to work with the SEC on the subject.
ISDA's views on the treatment of operational risk are reflected in an appendix to our CP3 response (Appendix D) and generally concur with the views of SIA.
We are particularly concerned that substantial uncertainties surround the practicability of the Advanced Measurement Approach (AMA). Continuous compliance with the 99.9 percent soundness standard, for example, is likely to be a persistent problem. Further, we are opposed to constraints on calculating capital requirements at group level and allocating the capital downwards as distinct from being required to calculate AMA numbers for each entity within a group (see Appendix D for more detail).
There are also serious questions about the functioning of the Standardized Approach. As currently structured, there is a perverse incentive built into the Standardized Approach. Firms built predominantly around the business lines that attract a `beta' factor of 18 percent-namely, Corporate Finance, Trading and Sales, and Payments and Settlements-end up with a higher capital charge than if they were to remain on the Basic Indicator Approach. They therefore have an incentive to stay with an approach with lower standards of risk management, which was certainly not the drafters' intention. ISDA believes it is essential to remove this structural defect and ensure that no firm systematically faces a higher Pillar One charge under the Standardized Approach than under the Basic Indicator Approach. ISDA stands ready to work with the SEC, using relevant data, to determine the effect of the Standardized Approach.
Again, ISDA is grateful for the opportunity to discuss the issues raised in the SEC Proposals. The process of developing a New Accord has been characterized by extensive openness and thoughtful consideration on the part of the regulatory bodies involved, and we are pleased that the SEC has approached the subject in the same way.
We look forward to your response and to further consultation. If you have questions, please feel free to contact David Mengle, ISDA, at firstname.lastname@example.org or 1-212-901-6017.