Federal Deposit
Insurance Corporation

Re: Risk-Based Capital Guidelines; Implementation of the New Basel Capital Accord

Ladies and Gentlemen:

The member banks of The New York Clearing House Association L.L.C. (“The Clearing House”)¹ appreciate the opportunity to comment on the Advance Notice of Proposed Rulemaking (“ANPR”) regarding the implementation of the New Basel Capital Accord (the “New Accord”) in the United States and the related draft supervisory guidance (the “Draft Supervisory Guidance” or “DSG”) published by the Office of the Comptroller of the Currency, the Board of Governors of the Federal Reserve System, the Federal Deposit Insurance Corporation and the Office of Thrift Supervision (together, the “Agencies”). We commend the Agencies for their quick response to the Basel Committee on Banking Supervision’s (the “Basel Committee”) third consultative paper. We value highly the consultative process that has characterized the New Accord and we support the Agencies’ critical role in that process. We believe, however, that much work remains in order to accomplish a new regulatory capital framework that appropriately aligns regulatory capital with actual risk and thereby avoids economic inefficiencies.

Our comments on the ANPR are set forth below. In addition, several of our member banks (or their parent companies) have submitted individual comment letters to the Agencies.² Our comments below outline common concerns of The Clearing House member banks.

A. Bifurcated Regulatory Framework

We believe that the bifurcated regulatory framework proposed in the ANPR is a realistic and sensible approach, as it recognizes the substantial differences between large, internationally active banks and the thousands of smaller banks in the United States. We question whether this approach creates significant competitive concerns, but, if it does, the response should be to adjust specific elements of the capital rules, rather than abandon the basic framework. We believe that the suggestion that this bifurcated framework will in itself cause consolidation in the banking industry is particularly far-fetched.

The more serious competitive issues arise in the context of global consistency and potentially differing standards for banks and non-banks. They relate to (1) the risk that facially similar capital requirements on a global basis will be undermined by less robust accounting rules for loan write-offs and in other areas and (2) undue capital requirements for specific lines of business.

B. Home-Host Implementation

We are concerned that the ANPR and the New Accord may not adequately address the need for systematic cooperation among supervisors for internationally active banks that are regulated by supervisors in multiple jurisdictions. International regulatory coordination is essential in creating a level playing field and preventing unnecessary regulatory burden and duplicative regulation. We are encouraged by the publication of High-level principles for the cross-border implementation of the New Accord by the Basel Committee in August and the establishment of the Accord Implementation Group (“AIG”) to address these concerns. We applaud the Agencies’ leading role in the AIG and trust that the Agencies will continue their work in this important area. We hope that the AIG’s efforts will lead to standards for consistent regulatory supervision of internationally active banks. Please refer to our member banks’ individual comment letters for specific suggestions regarding international implementation of the New Accord.³

C. Treatment of Expected Losses

We believe strongly that assigning capital for expected loss as well as unexpected loss is not only completely inconsistent with standard industry practice, but fails to recognize the economic underpinning of how expected loss is a function of pricing rather than capital. We are encouraged by the Basel Committee’s recent announcement that it is re-evaluating this aspect of the New Accord and will adopt an approach based only on unexpected losses, and we appreciate the ANPR’s contribution by acknowledging the differences of opinion on this issue. At the same time, however, we believe it essential that a formal rejection of expected loss coverage not be accompanied by a “back door” adjustment of the capital requirements to accomplish the same purpose.

Specifically, the Basel Committee’s current proposal is insufficient because, although it eliminates the expected loss component of the capital requirement, it also limits the recognition of reserves in excess of expected loss on both performing and non-performing assets and imposes a ceiling for qualifying reserves at 20% of Tier 2 capital. This approach ignores the fact that pricing covers expected loss on performing assets. We respectfully suggest that banks should be allowed to recognize the amount of loan loss reserves in excess of expected loss on defaulted assets as Tier 1 capital, because these reserves are the “first line of defense” against loss. There should be no limit on the amount of reserves that qualifies as capital.

If expected losses are to be covered, then all relevant financial resources available to cover losses should be recognized.

D. Prescriptiveness of the ANPR and the DSG

We are pleased that the Agencies have moved towards a more principles-based approach in the ANPR than was the case under Basel I or the previously proposed versions of the New Accord. We believe that a principles-based approach will allow the regulatory capital regime to adapt to innovations in the financial marketplace and will encourage continued development of best risk management practices. Unfortunately, the requirements set forth in the ANPR and the Draft Supervisory Guidance remain, in many cases, far too prescriptive and undermine the Agencies’ objective of allowing banks to operate in accordance with their internal models (subject to certain minimum standards). As examples, and as discussed further below, the definition of default is inconsistent with current industry use and regulatory guidance; the complex, highly technical rules set forth in the proposed securitization framework would, if adopted, limit a bank’s ability to innovate; and the extensive disclosure requirements proposed in Pillar 3 are counterproductive.

We also believe that the New Accord is too prescriptive in dictating the risk management processes banks must follow. Individual banks should have the ability to make their own determinations about the form, structure and priority of risk management processes and system enhancements. We agree with the Agencies’ statement in the DSG that institutions ultimately “must have credit risk management practices that are consistent with the substance and spirit of the standards in this guidance."⁴ Unfortunately, the detailed and prescriptive text after each supervisory standard in the DSG would, if implemented, serve to constrain significantly a bank’s ability to use the risk management framework it deems best.

In our view, it is critical to the success and longevity of the New Accord that its requirements are sufficiently flexible to adapt to evolution in banks’ financial products and risk management practices. Consequently, we suggest that the Agencies publish rules and guidance that truly serve as a guide to appropriate practices, rather than as a set of determining requirements. We hope that, in the near future, the Agencies will allow banks to use fully internal models to determine the regulatory capital for credit risk.

E. Definition of Default

The definition of default should be simplified to correspond more closely to the definition more commonly used by risk managers. Default for the corporate model should be defined as entry into non-accrual or charge-off status. The definition of default for the retail model should conform to the Uniform Retail Credit Classification standards published by the Federal Financial Institutions Examination Council.⁵

One example of our objection to the definition of default in the DSG is the inclusion of loss on the sale of a credit obligation: “The bank sells the credit obligation at a material credit-related economic loss.”⁶ Although we appreciate the Agencies’ efforts to alleviate the impact of non-credit related changes in market value by adding the term “credit-related,” we remain concerned that it will often not be feasible, or at least subjective, to determine whether the loss is credit-related. There is a wide range of events that could result in economic loss despite the absence of default or near-default. Loan sales are portfolio management operations motivated by concentration management, balance sheet usage, market liquidity and many other factors. Discounts result from a variety of factors including interest rates, market liquidity and supply and demand issues. We believe it would be difficult and ultimately arbitrary to try to disentangle these effects from credit-related changes. We recommend that this element of the definition of default be deleted.

We are also concerned about the emphasis the Agencies have placed on capturing so-called “silent” defaults. Capturing data on credits that are well secured and in the process of collection adds little value in determining loss. The exception for these credits is universally utilized and applied precisely because there is a strong expectation of zero loss. Accordingly, including silent defaults would artificially increase probability of default (“PD”) and decrease loss given default (“LGD”) with a negligible net result.

More generally, we believe that imposing the definition of default proposed in the Draft Supervisory Guidance will be extremely expensive without a justifiable benefit. The accounting and supervisory regime already addresses, with both quantitative and qualitative factors, which loans should be placed on non-accrual and charged-off, and banks take appropriate steps to identify these loans. Implementing the proposed definition of default would result in banks having to modify their loan accounting systems to accommodate the new definitions, modify data feeds and warehouses to gather the relevant information, train operations personnel to understand the difference between current and new definitions of default, and other expensive measures that ultimately will not reduce risks. Consequently, we recommend that the Agencies rely on current accounting and supervisory definitions to identify default.

F. Cumulative Effect of Conservative Assumptions

We support a prudent level of conservatism in applying capital standards to ensure that risks to individual institutions and the industry as a whole are adequately considered and addressed. However, we do not believe that the ANPR sets the true minimum capital standard that the New Accord is intended to accomplish. The ANPR incorporates many conservative individual decisions (including choices with regard to parameter values, formula alternatives and constraints), and the cumulative effect of these choices is an unnecessarily high standard that will result in regulatory capital requirements that well exceed a true minimum requirement, both for individual institutions and the industry as a whole. Some examples of the conservatism in the ANPR are as follows:

• The 99.9% confidence level chosen as a reference point for measuring credit risk and operational risk is the standard used by banks for internal economic capital purposes to be considered well capitalized. This confidence level is unduly conservative in light of the goal of setting a minimum solvency standard consistent with an investment grade rating. In our view, a more appropriate confidence level for a minimum standard is 99.5%, which is approximately the border between investment and non-investment grade.

• The asset value correlation ranges prescribed for retail exposures are greater than industry standards, by more than 50% in many cases.

• Generally, we do not agree that floors on PD or LGD are necessary or desirable. In particular, the 10% LGD floor for mortgages does not take into account mortgage insurance and the possibility of low loan-to-value ratios. Floors on PD and LGD undermine banks’ ability to use data-driven calculations that are grounded in the actual PD and LGD of each risk segment and discourage legitimate risk mitigation strategies because the cost does not produce the true economic benefit.

• The ANPR does not recognize the benefits of diversification among asset classes, business lines, geographic regions and risk types. This is inconsistent with industry practice in risk management. Diversification mitigates both the possibility and magnitude of loss, and we believe that banks should be allowed a capital credit or risk-weighted asset reduction for such diversification.

• As discussed further below, the ANPR does not give sufficient recognition to the full economic benefits of credit risk hedging.

We urge the Agencies to consider the cumulative impact of the many conservative decisions in the ANPR, and revise those decisions to develop a regulatory capital regime that represents a true minimum standard.

G. Limited Recognition of Credit Risk Hedging

Over the last decade, credit risk mitigation techniques have evolved significantly in effectiveness, type and volume. We appreciate that the Agencies have attempted to capture the benefits of credit risk hedging and guarantees where the guarantor is a superior credit to the borrower by providing that the default probability of the guarantor can be substituted for that of the borrower in determining risk weightings. However, the Clearing House member banks believe that the recognition of credit risk mitigation remains significantly incomplete and thereby discourages better risk management practices. We believe the Agencies should revise their approach in three key respects.

First, the ANPR fails to capture the cumulative benefit of credit hedges. In order for a bank to experience loss on a hedged exposure, both the borrower and the guarantor must default on their obligations. In addition, banks may be able to recover from both counterparties. We continue to believe strongly that the Agencies should allow banks to recognize the lower probability of double default and the lower LGD of multiple sources of recovery in the treatment of credit risk mitigation techniques. Please see our member banks’ individual comment letters for specific proposals for the appropriate treatment of the double probability of default.⁷

Second, the treatment of maturity mismatches set forth in the ANPR is overly conservative and unnecessarily complex. The proportional adjustment mechanism is much more conservative than the maturity treatment for corporate exposures. We see no justifiable basis for this difference.

Third, hedges with a remaining maturity of less than one year provide significant, albeit not total, protection against loss on the underlying obligation. Accordingly, the prohibition on capital relief for these hedges should be eliminated. Although the value of the hedges decline as they reach maturity, they do not reach zero until maturity. We recommend that the continued benefits of the hedges be taken into account using a modified formula for determining the capital relief for such hedges during the final year. Please see our members’ individual comment letters for specific proposals for these calculations.⁸

H. Maturity Adjustments

We appreciate that the Agencies have incorporated a maturity adjustment in the risk-weighting formulas to differentiate the risks of instruments with different maturities. Unfortunately, the Agencies have limited the risk-correlated impact of this element of the ANPR by providing an adjustment only for maturities from one to five years, with limited exceptions. We strongly believe that the regulatory capital requirement should reflect the effective remaining maturity of all transactions, including above five years and below one year. The maturity limitation is particularly important for short-term transactions of the type that many borrowers require. Generally, we believe that the maturity adjustments for short-term transactions should be based on an adjustment to PD to reflect lower default risk. Please see our members’ individual comment letters for suggestions on how this could be achieved.⁹

We understand the Agencies’ concern about potential capital arbitrage if banks were to continuously roll over short-term transactions in order to take advantage of lower capital requirements rather than originating a long-term transaction. However, we believe that this is a legitimate risk management strategy if, on a frequent basis, banks truly reassess the decision to extend credit based on the customer’s evolving credit quality, and that banks should receive the capital benefit under these circumstances. In our view, controlling credit risk exposure to borrowers and counterparties by limiting the maturity of transactions is an effective risk management technique.

I. Retail Calibration Issues

We support the Agencies’ attempt to reflect important differences in the appropriate correlation of asset values through the introduction of separate risk-weighting curves for mortgages, revolving credits and non-mortgage non-revolving credits. We remain concerned, however, regarding the calibration of capital requirements for retail assets. Compared to the results of internal models by The Clearing House member banks and an industry study conducted by the Risk Management Association, the capital requirements for consumer assets under the proposed approach are generally materially higher than justified by the level of risk.

In our view, the primary flaw in the calibration is the inclusion of expected loss in the capital formula, discussed above. Including expected loss distorts the absolute level of capital and the relative levels of capital for assets of different credit quality. In addition, as noted, we believe the 10% floor on the LGD for mortgage portfolios should be eliminated to take into account such factors as low LTVs and private mortgage insurance.

J. Complexity of the Securitization Framework

We believe the proposed securitization framework is complex and conservative, and will be burdensome to implement. We are pleased that the Basel Committee has announced that it will revise the treatment of securitization set forth in its third consultative paper and replace the Supervisory Formula Approach with a less complex approach. We support any attempts to simplify and clarify the securitization framework. Generally, we believe that banks should be allowed to use their internal ratings and systems in determining the appropriate levels of capital for securitization activities. Banks’ internal systems have been developed to evaluate the risks of securitized asset pools, and those systems are subject to third party validation and to periodic regulatory review. Please refer to our member banks’ individual comment letters for specific comments and suggestions regarding various aspects of the securitization framework.¹⁰

K. Counterparty Credit Risk

We understand that the Basel Committee is willing to reconsider the method for calculating the capital charge for counterparty credit risk. We recognize that supervisors are in the early stages of reassessing the current approach, which requires add-on factors for potential future exposure. This add-on approach is inconsistent with the best practices of leading banks, and we strongly urge the Agencies to address this issue as they proceed with their review. The International Swaps and Derivatives Association, Inc. (“ISDA”) recently submitted a paper to the Committee, which we believe provides a good starting point for discussions of changes to the treatment of counterparty credit exposures. We encourage supervisors to consider the ISDA paper during the review process.

L. Operational Risk

The Clearing House member banks believe that appropriately addressing operational risk remains an important issue. We applaud the Agencies on the continuing dialogue with industry participants, and we encourage the Agencies to maintain that dialogue. The Clearing House member banks do not all agree on the appropriate treatment of operational risk, but all agree that at least some changes are necessary to the proposed framework.

Many of our members banks are concerned that there will not in all cases be sufficiently detailed relevant data, either internally or externally, to allow an appropriate Advanced Measurement Approach (“AMA”) calculation for capital to be performed at the subsidiary level. In addition, many of the scenario analyses and control assessments the AMA would require are only relevant at a consolidated firm or business line level. Lastly, regulatory capital requirements for the firm in total would be overstated by the sum of the subsidiary capital needs because the benefit of a capital reduction caused by diversification would be ignored. Given this, requiring full AMA calculations for all subsidiaries is impractical. Banks should be permitted to perform a group, product or business line level calculation for capital, and the capital requirement for subsidiaries should be determined through an apportionment of the group, product or business line level requirement. We appreciate the international regulators’ concerns regarding the appropriate capitalization of individual legal entities and we encourage the Agencies to work actively with the industry to develop satisfactory risk-sensitive solutions for this issue. Such solutions are expected to recognize appropriately the significant diversification benefits that should be reflected in the calculation of operational risk capital requirements.

Please refer to our members’ individual letters for more detailed comments on operational risk.¹¹

M. Disclosure Requirements

We believe that disclosure has a very important role to play in the effective implementation of the New Accord, and that it should be clear, transparent and understandable. We appreciate the steps the Agencies have taken to reduce the amount of required detailed disclosure from that previously considered. Unfortunately, the remaining disclosure requirements proposed in the ANPR would result in a significant increase in banking organizations’ reporting burden, even for organizations that currently publish much of this data. We strongly believe that the risk of misinterpretation of this information, in large part because of the danger of information overload, and the burden its distribution will place upon banks far outweigh any benefit it may have.

The market’s ability to accurately evaluate a bank’s risk exposure would be better facilitated by the clear presentation of important information than by the publication of voluminous, highly technical data. Providing data without analysis could lead to inaccurate assessments, and even undermine the safety and soundness of individual banks and the industry as a whole, by presenting data that the market cannot interpret correctly. Many market participants lack the depth and breadth of understanding of the institution and, in some cases, the industry that is required to evaluate the proposed disclosure. Even with extensive explanatory notes, they will struggle to assess the relative importance of the various required disclosures and are likely, we believe, to draw inappropriate conclusions from the information. Rather than encouraging market discipline, we believe that the proposed volume of disclosure will slow the market’s absorption of information and increase the likelihood of inappropriate or contradictory conclusions by investors. Neither of these results is consistent with the functioning of an efficient system of market discipline, which should be the goal of the disclosure requirements.

Furthermore, given the amount of work involved in compiling the necessary information, it will be nearly impossible to meet the 30-day deadline following quarter-end for Call Report and filings with the Securities and Exchange Commission that will be effective by the time the New Accord is implemented. Banks generally announce their financial results long before the 30-day deadline. Under the current regime, banks present risk-based capital ratios and supporting detail when earnings are announced. The level of disclosure proposed in the ANPR is not possible within that same timeframe.

We are also disappointed by the frequency with which the disclosures are proposed to be required. We believe that annual disclosure is appropriate for most information unless there is a material change that makes year-end data misleading. We suggest that full disclosure be required annually, with quarterly updates for any subsets of information that have materially changed.

We believe public companies have the central role to play in summarizing and analyzing data for their shareholders. We strongly recommend that the Agencies, in association with the industry, the investor community and other relevant bodies, identify a smaller subset of key disclosures that will convey a bank’s risk profile without inundating the market with irrelevant information or risking misinterpretation of overly technical information. Remaining disclosures should be left to the judgment of each institution based on the relevance of the information to the current financial condition of the bank and the demands of its investors.

*                *                  *

The Clearing House appreciates the opportunity to comment on the ANPR. If the Agencies would like additional information regarding these comments, please contact Norman R. Nelson, General Counsel of The Clearing House, at (212) 612-9205.


Sincerely,
Jeffrey P. Neubert
President and CEO
The New York Clearing House Association
New York, NY