Hearing: A Review of the New Basel Capital Accord
Wednesday, June 18, 2003
Statement required by 12 U.S.C. 250:
The views expressed herein are those of the Office of the Comptroller of the Currency and do not necessarily represent the views of the President.
Chairman Shelby, Senator Sarbanes, and members of the Committee, thank you for inviting the Office of the Comptroller of the Currency (OCC) to participate in this hearing on proposed revisions to the 1988 Capital Accord developed by the Basel Committee on Banking Supervision (Basel Committee). I welcome the efforts of the Committee to focus attention on these critical issues. The health of the U.S. commercial banking system is a critical element to a strong economy. Thus, it's essential that any regulatory changes that might affect the condition and competitiveness of our banking system be fully understood and carefully evaluated by the banking industry, the U.S. Congress and the American public.
The 1988 Accord, referred to as Basel I, established the framework for the risk-based capital adequacy standards applicable to internationally active commercial banks in all of the G-10 countries, and it has been adopted by most other banking authorities around the world. U.S. banking and thrift agencies have applied the 1988 framework to all U.S. insured depository institutions.
By the late 1990s, it became evident that Basel I had become outdated. The increased scope and complexity of the banking activities of our largest banking institutions over the last decade, and the unintended consequences of various provisions of the regulations, severely undercut the utility of the Capital Accord. Basel I simply does not provide a meaningful measure of the risks faced by large, internationally active banks or the capital they should hold against those risks.
Consequently, over the past several years, the Basel Committee has been developing a more detailed and risk sensitive capital adequacy framework to replace Basel I. The Committee's first draft document, Consultative Paper No. 1 (CP-1), was issued in June 1999. It laid the groundwork for the new capital adequacy framework (Basel II), but provided few details. The Committee provided additional detail on the specifics of Basel II in its January 2001 issuance of Consultative Paper No. 2 (CP-2). Although more detailed, CP-2 still left a number of key issues unaddressed and unresolved. The Committee's most recent paper, Consultative Paper No. 3 (CP-3), which I will discuss today, was issued on April 29 of this year.
As work on these consultative papers has progressed, the Basel Committee also has attempted to gauge the impact of its proposals on the required capital levels of banking institutions through a series of quantitative impact studies. In May, the Committee published the results of the most recent assessment, the third quantitative impact study (QIS-3). While the Committee concluded that the results were generally in line with the objectives of Basel II, the QIS-3 data still do not provide a sufficiently reliable estimate of the likely regulatory capital requirements for banks subject to Basel II. More work in this area is clearly warranted and I will discuss this later in my testimony.
The Basel Committee has outlined an aggressive timeline for the remaining actions leading to the adoption of Basel II. As a consequence, the U.S. banking agencies, the agencies responsible for the maintenance of capital adequacy standards for U.S. financial institutions, are faced with a daunting task. While we will work earnestly in this effort, the timeline should be seen as a means to an end, not an end in itself. As will be highlighted in my testimony, basic principles of safety and soundness demand that the banking agencies have a more complete understanding of the consequences of this proposal on the overall capital levels of affected institutions, the competitive effects on our financial system, and associated compliance costs and burdens before moving forward to finalize this proposal.
Our current primary focus in this effort is the development of U.S. implementing regulations and policies. As I will discuss later, the OCC and the other U.S. banking agencies will soon issue for comment proposed revisions to U.S. risk-based capital regulations to reflect the primary components of Basel II. Let me be absolutely clear about the integrity of this rulemaking process - the OCC, which has the sole statutory responsibility for promulgating capital regulations for national banks, will not begin implementing a final Basel II framework until we have conducted whatever cost-benefit and impact analyses that are required, and fully considered all comments received during our notice and comment process - as we would with any domestic rulemaking. If we determine through this process that changes to the proposal are necessary, we will not implement proposed revisions until appropriate changes are made. We made this point quite clearly to our Basel Committee colleagues before we agreed to go forward with CP-3. Indeed, many of them will also have to go through their own internal domestic processes before they can adopt the Basel II framework.
Current Basel Proposal
The Basel Committee deserves considerable credit for its articulation of Basel II in CP-3. The proposal is still exceedingly complex, but CP-3 is a clearer presentation of inherently difficult material than its predecessors. This is an important step, since regardless of the complexity of the proposal, it is important that the industry and other interested parties have a clear understanding of the proposed Accord.
The attachment to this written statement provides a summary of the substantive provisions contained in CP-3. As before, this iteration of the proposed new Accord has three mutually reinforcing "pillars" that comprise the framework for assessing bank capital adequacy. The first pillar of the new Accord is the minimum regulatory capital requirement. The Pillar 1 capital requirement includes a credit risk charge, measured by either a standardized approach or one of the new internal ratings-based (IRB) approaches (foundation or advanced), an operational risk charge, and a market risk charge. Again, the attached document provides a more detailed description of the various components of the Pillar 1 charge.
Pillar 2 addresses supervisory review. It is "intended to ensure not only that banks have adequate capital to support all the risks in their business, but also to encourage banks to develop and use better risk management techniques in monitoring and managing these risks." This pillar encourages supervisors to assess banks' internal approaches to capital allocation and internal assessments of capital adequacy, and, subject to national discretion, provides an opportunity for the supervisor to indicate where such approaches do not appear sufficient. Pillar 2 should also be seen as a way to focus supervisors on other means of addressing risks in a bank's portfolio, such as improving overall risk management techniques and internal controls.
The third pillar recognizes that market discipline has the potential to reinforce capital regulation and other supervisory efforts to ensure the safety and soundness of the banking system. Thus, the Committee is proposing a wide range of disclosure initiatives, which are designed to make the risk and capital positions of a bank more transparent. As a bank begins to use the more advanced methodologies, such as the Advanced IRB approach, the new Accord will require a significant increase in the level of disclosure. In essence, the tradeoff for greater reliance on a bank's own assessment of the building blocks of capital adequacy is greater transparency.
U.S. Implementation Actions
It is important to recognize that the Basel Accord is not self-executing in the U.S. Even when adopted by the Basel Committee, Basel II will not apply to U.S. institutions unless and until the U.S. banking agencies adopt regulations to implement it. In accordance with the Administrative Procedure Act, 5 U.S.C. 551, et seq., the U.S. banking agencies must publish notice and seek comment from all interested persons on any such proposal, and must fully consider those comments, before adopting a new capital regulation in final form. Obviously, the OCC and the other federal banking agencies intend to comply fully with these requirements. The importance of this rulemaking makes this comment process particularly critical to our success. Thus, we welcome this process as a means for positive contribution to this deliberative effort. We believe that the solicitation and assessment of comments is a critical step in determining the feasibility, effectiveness, and expected consequences of Basel II and related domestic capital regulations.
Next month, the U.S. banking agencies expect to jointly issue an Advance Notice of Proposed Rulemaking (ANPR) soliciting comment on proposed revisions to the existing domestic capital adequacy regulations that would implement Basel II. The ANPR will be largely based on CP-3, and will provide a description of proposed revisions to current capital regulations, while seeking comment on outstanding or contentious issues associated with the proposal. The ANPR will also request information on the cost of implementing the proposal, and will seek comment on the competitive implications in both domestic and international markets for banks of all sizes. In conjunction with the ANPR, the banking agencies will also issue for comment draft supervisory guidance articulating general supervisory expectations for banks seeking to implement Basel II-compliant methodologies for the Advanced Measurement Approach (AMA) to operational risk and Advanced IRB for corporate credits. Recognizing that CP-3 is a complex document, we understand the importance of providing U.S. banks an opportunity to review and comment on U.S. implementing documents as soon as practicable. By describing these concepts within the context of our existing regulatory and supervisory regime, the ANPR and draft guidance will provide a meaningful forum for a full discussion of Basel II.
After assessing comments generated during the ANPR process, the U.S. banking agencies will consider a complete cost analysis in accordance with applicable rulemaking requirements, including the standards of Executive Order 12866, discussed below, and will develop specific regulatory language for a joint Notice of Proposed Rulemaking (NPR). Again, the banking industry and other interested parties will have an opportunity to comment on this fully articulated proposal before any revisions to our capital regulations are finalized.
Let me now focus on two important, unique features of the U.S. regulatory capital regime that will be highlighted in the ANPR and NPR - the scope of application of Basel II and the content and structure of the proposed revisions to the capital adequacy regulations. First, the U.S. expects to set forth in the ANPR proposed criteria for identifying which banks in the U.S. will be subject to the new Accord. Despite language in the 1988 Capital Accord that permitted a more limited application, U.S. banking and thrift agencies applied the Basel framework to all U.S. insured depository institutions. As we will highlight in the forthcoming ANPR, the U.S. agencies have determined to apply Basel II concepts more narrowly. Specifically, consistent with the focus of the Basel Capital Accord on banks that compete in the global market place, we will propose applying Basel II concepts on a mandatory basis only to large, internationally active institutions that compete on a significant global basis with other financial service providers. Other institutions will have the opportunity to voluntarily opt into the Basel framework upon application to, and approval by, their primary federal supervisor.
Preliminary analysis by the U.S. agencies suggests that under the narrow approach we are proposing, there are currently fewer than a dozen U.S. banks that would be mandatorily subject to Basel II-based regulatory capital requirements. Of course, the approach of requiring only a small population of banks to comply with Basel II will be subject to notice and comment in the ANPR and will be definitively resolved only after the U.S. rulemaking process has been completed.
Second, in developing revisions to existing capital adequacy regulations, U.S. banking agencies recognize that the revised regulation, and interagency implementation policies, need not follow the literal structure and language of Basel II. While consistent with the objectives, general principles and core elements of the revised Basel Accord, the language, structure and degree of detail of U.S. implementing documents may be very different from Basel II. These implementation differences are reflective of the particular statutory, regulatory and accounting structures and practices in place in the U.S. It is important to note that U.S. implementation actions do not contemplate changes to many fundamental aspects of our regulatory/supervisory process, including a focus on regular on-site supervision, our prompt corrective action rules, and our minimum leverage ratio for capital adequacy. As described more fully in the attachment, the U.S. agencies will propose for notice and comment a Basel II-based regime incorporating only the Advanced IRB approach for credit risk, the AMA for operational risk, and the internal models approach for market risk.
We are also very cognizant that in connection with this, or any rulemaking, existing requirements may compel preparation of detailed analysis of the costs, benefits, and other effects of our regulations, depending on threshold determinations of whether the rulemaking in question triggers the substantive requirements of particular statutes or Executive Orders. Relevant requirements are set forth in the Regulatory Flexibility Act (RFA), the Unfunded Mandates Reform Act of 1995 (UMRA) and Executive Order 12866 (EO 12866). Issuance of the ANPR will help us identify and determine costs, benefits, and other effects of the proposed rulemaking, for purposes of complying with these requirements.
As I noted early on in my testimony, the Basel Committee timeline presents a daunting task to both the U.S. banking agencies and the banking industry. While it is clearly necessary to move forward in addressing the acknowledged deficiencies in the current Basel Capital Accord, the banking agencies must better understand the full range and scale of likely consequences before finalizing any proposal. The list provided below identifies the milestones the OCC must meet under the current Basel II timeline. Each step is critical in a prudential consideration of Basel II in the U.S.:
* Consideration of comments received by the Basel Committee on CP-3. The comment period on this document concludes on July 31.
* Finalization, issuance and consideration of comments on the U.S. ANPR. Based on current estimates, the notice and comment period will run from July to October.
* Finalization, issuance and consideration of comments on supervisory guidance on Corporate IRB and AMA methodologies. Based on current estimates, the notice and comment period will run from July to October.
* Development, issuance and consideration of comments on supervisory guidance on other substantive aspects of Basel II-based regulations, especially including retail IRB. Based on current estimates, the agencies hope to commence solicitation of comment on this guidance by year-end 2003.
* Participation in the Basel Committee's consideration of Basel II. Under the current timeline, the Committee is to consider approval of Basel II in December of this year.
* Development, issuance and analysis of results of additional agency efforts to evaluate the prospective effects of Basel II implementation. EO 12866 may compel the OCC and OTS to undertake such analysis prior to the issuance of an NPR. Even without regard to this requirement, however, it is essential that we have a reliable estimate of the impact of Basel II on the capital and competitive position of U.S. banks.
* Development, issuance and consideration of comments on the U.S. NPR. This document would only be issued after the Basel Committee finalizes its consideration of Basel II. If the existing timeline is maintained, solicitation of comment on the NPR would commence no earlier that the first quarter of 2004.
* Development and issuance of a U.S. final rule and supervisory guidance. Again, assuming the present timeline is maintained, our best estimate for the issue date of a final rule implementing Basel II is the third or fourth quarter of 2004.
* Completion of all necessary supervision-related steps to implement Basel II-based regulations in advance of the presently proposed December 2006 effective date. Most significantly, the agencies need to determine whether each bank subject to Basel II-based regulations has appropriate systems and procedures in place to qualify for using the A-IRB and AMA.
Status of Basel Proposal - Outstanding Issues
In commencing an objective assessment of the status of Basel II, it is important to reiterate and reaffirm the commendable work of the Basel Committee, and in particular, the strong and intelligent leadership of its former Chairman, William McDonough. The OCC firmly supports the objectives of Basel II. These objectives constitute a sound conceptual basis for the development of a new regulatory capital regime and should continue to serve as a useful benchmark to gauge our progress in this effort. Nonetheless, much of that conceptual basis has not been tested in practice in any manner approaching the magnitude of Basel II. We continue to be concerned about the potential for unintended or unanticipated consequences of the Basel II proposals.
At its foundation, the Basel II proposals permit qualifying institutions to calculate their minimum risk-based capital requirements by reference to their own internal systems and methodologies. While it is the hallmark of Basel II, a greater alignment of internal risk assessment with minimum regulatory capital derived through internal models represents a radical departure from our existing regulatory capital framework. As we will highlight in the ANPR and accompanying guidance, this reliance on internal risk assessment systems mandates changes in the way we structure our capital regulations and, in certain important respects, how we conduct our supervisory activities. The fundamental question for the banking agencies in assessing Basel II is the issue the OCC has previously identified - whether the regime will work in practice, as well as theory, as the basis for a regulatory capital regime.
For bank supervisors and other external stakeholders to be in a position to rely on a bank's internal process in the establishment of regulatory capital requirements, there must be a high degree of confidence that regulators can establish and enforce appropriate risk measurement and management standards consistently across the banks subject to a Basel II-based regime. The challenge for supervisors is to create a verifiably accurate system that appropriately balances the need for flexibility, to promote continued improvement in risk management practices, with the need for objective standards, to ensure consistency in application across institutions and supervisors, both foreign and domestic.
The capital rule we implement must respect the evolutionary nature of risk management. As regulators, we must acknowledge that we are still in the relatively early days of model-based credit and operational risk measurement and management, and we must recognize the inevitability of further innovation and improvements in this area. This respect for the evolutionary nature of this discipline must then be reconciled with the need for objective standards to ensure consistency in application. Much of the detail and complexity within Basel II derives from the need to establish more objective expectations for bank rating systems, control mechanisms, audit processes, data systems, and other internal determinations of risk by individual banks. In many cases, this has led to the establishment of supervisory standards in areas previously left to management discretion or supervisory judgment.
Not surprisingly, the regulatory community has struggled with the establishment of these standards. Failing to achieve the proper balance for these often conflicting objectives while moving forward with the radically different Basel II-based regime can have dramatic consequences. If our regulation and supervisory process is overly flexible, bank internal calculations of capital adequacy may prove insufficient, non-comparable, or both. If we err on the other extreme, we establish an excessively prescriptive supervisory regime that stifles innovation, imposes undue regulatory burden and inappropriately narrows the role of judgment.
This need to carefully balance dramatically opposed objectives, together with the significant uncertainties that still exist about the practical feasibility of these proposed changes to the Capital Accord, raise doubts about the achievability of the timeframe established by the Basel Committee.
A stated goal of the Basel Committee in developing Basel II was that "the Accord should continue to enhance competitive equality." Realistically, we are not yet in a position to assess definitively the full range of consequences from the implementation of Basel II, including its effect on competitive equality in the global financial marketplace. There are risks that Basel II may create or exacerbate relative advantages between domestic banks and foreign banks; between banks and non-banks; and between large domestic banks and mid-size/small domestic banks. It is imperative that the U.S. banking agencies remain sensitive to these concerns and assess, to the extent possible, any unintended consequences resulting from the implementation of Basel II.
One of the primary objectives of the Basel Committee itself is the reduction of gaps and differences in international supervisory coverage by national supervisory agencies, especially as it relates to large internationally active banks that compete on a significant global basis with other financial service providers. This principle of competitive equality and a level playing field for international banks is an admirable one, and an appropriate goal of the Committee's efforts. Yet, the very complexity of the rules themselves calls this objective into question. Bank supervision varies significantly from one country to another in approach, intrusiveness, and quality. Is it realistic to think that an enormously complex set of rules will be applied in an evenhanded way across such a broad spectrum of supervisory regimes? For example, the OCC has as many as 30 to 40 full-time resident examiners in our largest banks. They are intimately involved as supervisors in assessing the banks' operations and judging the banks' compliance with a myriad of laws, rules, and guidelines. Some other countries may send examiners in once a year to a comparably sized institution, or may examine such an institution thoroughly only every five years, or may put heavy reliance on the oversight of outside auditors.1
It's fair to ask, I think, in which type of supervisory regime detailed, prescriptive capital rules are more likely to be robustly and reliably enforced. The Basel Committee has not undertaken to set standards of supervision for member countries. Yet the attainment of competitive equity among internationally active banks is a bedrock principle of Basel II. Can we really achieve competitive equality without addressing disparities in supervision, particularly when we are operating on the assumption that the complex new rules we're writing will be applied in an evenhanded way throughout the world?
Another principle source of competition for many banks is not other insured depository institutions, but non-banks. This situation is especially pronounced in businesses such as asset management and payments processing. As you are aware, however, regulations implementing Basel II-based concepts in the U.S. will apply only to insured depository institutions and their holding companies. While differences in regulatory requirements for banks and non-banks exist today, many institutions have voiced concern that implementation of Basel II may unduly exacerbate the current differences. These concerns have been mainly focused on the effects on competition from the application of the operational risk proposal and the enhanced disclosures required under Pillar 3.
Finally, there is concern about the potential effect of Basel II on the competitive balance between large and small banks. As implemented in the U.S., Basel II would result in a bifurcated regulatory capital regime, with large banks subject to Basel II-based requirements and small and mid-sized banks subject to the current capital regime. This structure is premised on the belief that, to the extent possible, regulations should reflect the size, structure, complexity, and risk profile of banking institutions. The Basel II framework was developed to address the unique risks of large internationally active institutions. Mandatory application of such a framework to small banks, with its associated costs, was deemed inappropriate. In fact, the banking agencies sought comment from the banking industry, especially smaller institutions, on the development of a simplified capital framework specifically for non-complex institutions.2 Industry comments were overwhelming negative on the proposal - most institutions felt that the cost of adopting a new regulatory capital regime outweighed any potential benefits. Accordingly, the banking agencies tabled the proposal.
With that said, the banking agencies need to continue to assess the competitive effects of a bifurcated regulatory capital regime, and it is one of the areas on which we will seek guidance in our ANPR. There are several concerns in this regard. First, banks using a Basel II-based regime may have a lower minimum capital requirement, allowing those banks to grow and compete more aggressively with smaller banks for both assets and liabilities. To be sure, banks subject to the new Basel II requirements will incur very significant systems and compliance costs in preparing for the new regime. These concerns are discussed in more detail in the "Calibration" section below. Moreover, banks using a Basel II-based regime may have significantly higher or lower marginal regulatory capital charges than non-Basel banks for some types of loan products, resulting in potential pricing differentials. While Basel II might enable larger banks to compete more effectively for high quality credits, it could also result in larger concentrations of lower quality credits in smaller institutions. Finally, the potential implications on industry consolidation are simply not known. The banking agencies must continue to assess this situation and, if warranted, take steps to mitigate adverse effects on the competitive balance between large and small banks. We would be seriously concerned if, as an unintended consequence of the implementation of Basel II, we significantly alter the structure of banking in the U.S.
The first objective of the Basel Committee in embarking on the Basel II effort was to calibrate minimum capital requirements to bring about a level of capital in the industry that, on average, is approximately equal to the global requirements of the present Basel Accord. That calibration was to be designed to provide an incentive to banks to develop and maintain sophisticated and risk-sensitive internal ratings-based systems.
In order to gauge its success in meeting that objective, the Basel Committee attempted to measure the impact of its proposals on the required capital levels of banking institutions through several quantitative impact studies. On May 5, 2003, the Committee published an overview of the results of its most recent assessment, the third quantitative impact study (QIS-3). On the basis of QIS-3 results, the Committee concluded that the aggregate results were generally in line with the objectives established for Basel II.
Unfortunately, the QIS-3 data do not provide a reliable estimate of the likely regulatory capital requirements for banks subject to Basel II. Banks encountered several practical impediments to providing accurate estimates of the effect of the proposals on their measured ratios; thus, the estimated risk-based capital ratios were subject to a substantial margin of error. For example, in many cases, existing bank systems were not able to produce the data requirements necessary for inputs required by the new Accord. In some areas, the QIS-3 instructions were not sufficiently clear or were misinterpreted, and in other cases, the proposals were still in flux as banks were completing the survey. Most important, QIS-3 was completed without the rigorous supervisory validation and oversight that would occur when the proposal actually takes effect.
A key concern is that focusing on the overall results of the QIS-3 exercise masks the wide dispersion of results for individual institutions. In the U.S., measured against current risk-weighted assets, the use of advanced approaches yielded results that ranged from a decrease in regulatory capital requirements of 36% to an increase of 43%. Similarly broad dispersions are found in a great many of the underlying components that make up the total capital requirement. While some dispersion of results in a truly more risk-sensitive framework would be expected, we are not convinced that the wide ranges indicated by QIS-3 can be explained by relative differences in risk among institutions; it appears that comparability of QIS-3 results among different institutions may be severely lacking.
Finally, the quantitative studies that have been done to date have been based on unilateral inputs from the participating banks. We and other supervisors have had only very limited ability to review the veracity of the results. I want to be clear that we have no reason to believe that U.S. banks did not make every effort to provide results as accurate as possible given the constraints they were operating under. Nonetheless, it is certainly conceivable - I would say highly likely - that the results might change significantly, and not necessarily in any particular direction, when all the intricacies of real-world implementation come into play. It seems fair to assume that banks will have fewer incentives to take conservative stances and greater incentives to exploit any loopholes or gray areas in the final rules; the extent to which these effects might be offset (or exceeded by) greater supervisory oversight is unknown.
Notwithstanding the significant uncertainties noted above, it presently appears that the required capital levels of some U.S. institutions could drop significantly, even taking into account the temporary minimum floor capital requirements, discussed in the attachment. The OCC does not believe that some reduction in minimum regulatory capital requirements for certain institutions is, in and of itself, an adverse feature of Basel II. Such a result is only acceptable, however, if the reduction is based on a regulatory capital regime that appropriately reflects the degree of risk in that bank's positions and activities. Given the fact that relevant bank systems and procedures are still in development, the OCC is not yet in a position to make that determination as it relates to Basel II. As such, the OCC is not yet comfortable allowing national banks to materially lower their current capital levels simply on the basis of the output of the currently proposed Basel II framework.
The OCC expects that an additional quantitative study will be necessary after the Basel Committee's work on Basel II is completed. Ideally, this should take the form of another global study by the Basel Committee itself - i.e., a QIS-4. However, even if the Basel Committee does not undertake such a study, I believe that it is absolutely essential that the U.S. agencies do so prior to the adoption of final implementing regulations. I strongly believe that we cannot responsibly adopt final rules implementing Basel II until we have not only determined with a high degree of reliability what the impact will be on the capital of our banks, but we have made the judgment that the impact is acceptable and conducive to the maintenance of a safe and sound banking system in the U.S.
As I have indicated, the OCC firmly supports the objectives of Basel II - a more risk-sensitive and accurate capital regime. However, in light of the issues that been identified with the current iteration of Basel II, the U.S. banking agencies must now determine how best to proceed on this critically important issue. I believe the following are essential elements in the agencies' consideration of Basel II implementation within the U.S.
First, the agencies need to move forward with the solicitation of comments on a Basel II-related ANPR and associated guidance. That is the most effective mechanism to a have full and complete consideration of the proposal from all interested parties. The solicitation of comments on a proposed regulatory and supervisory structure for Basel II implementation will also permit supervisors to tangibly assess the feasibility of the proposal.
Second, the agencies need to undertake additional steps to evaluate the costs, benefits, and other effects of the proposal before moving forward with any final regulatory action. Frankly, we simply need additional information to reasonably address the numerous issues, concerns and uncertainties associated with Basel II implementation. We must better understand the likely consequences of this proposal on overall capital levels of affected institutions, the competitive effects on our financial system, and associated compliance costs and burdens. In determining the appropriate additional steps, the agencies should consider the obligations imposed under EO 12866, the other statutory requirements for consideration of costs and impact, lessons learned from QIS-3, and perhaps, a U.S. version of QIS-4.
Third, as I have consistently reiterated, if we determine through this process that changes to the Basel II proposal are necessary, the U.S. agencies must pursue those changes, both domestically and in the Basel Committee. In this regard, the U.S. agencies should not foreclose consideration of alternative proposals that address the acknowledged deficiencies of the 1988 Accord but that do not constitute such a radical departure from our existing regulatory capital framework.
Fourth, the overarching consideration for supervisors in moving forward on Basel II is the need to act in accordance with our primary mission - to ensure the continued maintenance of a robust and safe and sound banking system. We need to incent banks to continue to better measure and manage the full panoply of risks they face and to make use of new and evolving risk management practices. We must also ensure that prudential consideration of safety and soundness principles remain paramount.
As I said in the beginning of my statement, the OCC, the agency to which Congress has committed the authority to define capital requirements for national banks, will not sign off on implementation of a final Basel II framework until we have fully considered all comments received during our notice and comment process. Given the importance of this proposal, the significant issues that remain unresolved, and the prospect that whatever emerges from this process is likely to govern the financial landscape for years to come, we need to take whatever time is necessary to develop and implement a revised risk-based capital regime that achieves the stated objectives of the Basel Committee in both theory as well as practice.
I am pleased to have had this opportunity to provide our views on this important initiative, and I would be happy to answer any questions you may have.
Summary of Basel II: The Proposed New Accord
Office of the Comptroller of the Currency
The Basel Committee (the Committee) has been developing the new Accord over the past five years. During that time, three full-scale consultative papers (June 1999, January 2001 and April 2003) and numerous working papers supporting various elements of the new Accord have been released to the industry for comment. This summary is intended to convey a general idea of the structure and substance of the proposed new Accord, and does not attempt to provide a complete analysis. It is based on the most recent publications from the Basel Committee, notably the New Basel Capital Accord (Consultative Document) which is out for comment until July 31; the document can be found on the Committee's website at http://www.bis.org/bcbs/index.htm.
The new Accord will include menus of approaches for measuring the capital required for credit risk, market risk, and operational risk. For credit risk and operational risk, each of the proposed approaches is described briefly below; capital charges for market risk are unchanged in the new Accord and are not discussed here. Some of the approaches described are unlikely to be implemented in the U.S. and have been noted as such. Moreover, based on preliminary analysis by the U.S. agencies, currently there are less than a dozen U.S. banks that would be mandatorily subject to Basel-based regulatory capital requirements. While other banks would be permitted to opt in to the Basel rules (subject to meeting prudential qualification requirements), the U.S. capital rules will remain in place for the vast majority of U.S. banks that either are not required to or do not opt to apply the Basel II framework. Of course, any issues regarding U.S. implementation of the new Accord will be definitively resolved only after the U.S. rulemaking process has been completed.
The current structure of the Accord has been influenced by the results of several quantitative impact studies (QIS), the most recent of which was completed in December 2002. Approximately 20 US banks participated in the QIS exercise in December and the results have been factored into the most recent version of the Accord. Changes were made in several areas including the treatment of retail credits, specialized lending, securitization, and operational risk.
General Structure of the Proposed New Accord
The new Accord has three mutually reinforcing "pillars" that make up the framework for assessing capital adequacy in a bank. The first pillar of the new Accord is the minimum regulatory capital charge. In order to calculate the capital charge under Pillar 1, banks will have to determine the individual charges for credit, market, and operational risk. The new Accord offers a series of options for calculating credit and operational risk. Market risk will remain unchanged from a 1996 amendment to the Accord. The new options for credit and operational risk were designed to be available to a wide range of banks, from relatively simple to very complex. For credit risk, the Pillar 1 capital requirement includes both the standardized approach, updated since the 1988 Accord, and the new Internal Ratings-Based (IRB) approaches (foundation and advanced). Pillar 1 has been the focal point of much of the discussion and comment from the industry on the new Accord.
Pillar 2 covers supervisory review and banks' obligation to hold sufficient capital vis-à-vis their risk profile. The pillar is "intended to ensure not only that banks have adequate capital to support all the risks in their business, but also to encourage banks to develop and use better risk management techniques in monitoring and managing these risks." This pillar encourages supervisors to assess banks' internal approaches to capital allocation and internal assessments of capital adequacy. It provides an opportunity for the supervisor to indicate where such approaches do not appear sufficient. Pillar 2 is also a way to focus supervisors on other means of addressing risks in bank's portfolio, such as improving risk management techniques and internal controls.
The third pillar recognizes that market discipline has the potential to reinforce capital regulation and other supervisory efforts to ensure the safety and soundness of the banking system. Thus, the new Accord proposes a wide range of disclosure initiatives, which are designed to make the risk and capital positions of a bank more transparent. As a bank begins to use the more advanced methodologies for market and operational risk, the new Accord will require a significant increase in the level of disclosure. In essence, the tradeoff for greater reliance on a bank's own assessment of capital adequacy is greater transparency. This pillar has been subject to numerous changes as the Committee has worked to balance the need for robust disclosure with a recognition of the proprietary and confidential nature of some of the information.
Capital for Credit Risk
Under Basel II, banks must select one of three approaches to determine their capital for credit risk. The three approaches, from simplest to most complex are: the standardized approach, the foundation IRB and the advanced IRB.
The 1988 Accord introduced the standardized risk-bucketing approach for setting the minimum regulatory capital requirement, which is still used in the U.S. today. The approach has been subject to criticism that it lacks sufficient risk sensitivity. The revised standardized approach under Basel II enhances the 1988 Accord by providing greater, though still limited, risk sensitivity.
Key changes to create a more risk-sensitive framework include the refinement and addition of risk buckets, the introduction of external credit ratings, and a wider recognition of credit risk mitigation techniques. Risk weights are still determined by category of the borrower-sovereign, bank or corporate-but within each of these categories changes have been made to make the capital more reflective of the riskiness of the asset category. For example, the risk weight on mortgage loans has decreased from 50% to 35% and the risk weight on certain retail credits has moved from 100% to 75%. Risk weights for externally-rated corporate credits, currently 100%, will range from 20% to 150%. Sovereign risk weights are no longer dependent upon whether a country is a member of the Organization for Economic Cooperation and Development (OECD), but rather on the external rating identified for the country.
The standardized approach is not likely to be implemented in the U.S. U.S. supervisors believe that credit risk measured under the standardized approach of Basel II would generally not be appreciably different than that measured under current rules for most U.S. banks, and the marginal changes in capital requirements would not justify the cost of implementation.
Internal Ratings-Based Approach (Foundation and Advanced)
The IRB approach represents a fundamental shift in the Committee's thinking on regulatory capital. It builds on internal credit risk rating practices used by some institutions to estimate the amount of capital they believe necessary to support their economic risks. In recent years, as a result of technological and financial innovations and the growth of the securities markets, leading banking institutions throughout the world have improved their measurement and management of credit risks. These developments have encouraged the supervisory authorities to devote greater attention to introducing more risk-sensitive regulatory capital requirements, particularly for large, complex banking organizations.
Banks must meet an extensive set of eligibility standards or "qualifying criteria" in order to use the IRB approach. Because the requirements include both qualitative and quantitative measures, national supervisors will need to evaluate compliance with them to determine which banks may apply the new framework. The requirements vary by both the type of exposure and whether the bank intends to use the simpler foundation IRB framework or the more advanced IRB framework. The requirements are extensive and cover a number of different areas, including rating system design, risk rating system operations, corporate governance, and validation of internal estimates. A brief sample of actual criteria include:
* The board of directors and senior management have a responsibility to oversee all material aspects of the IRB framework, including rating and probability of default (PD) estimation processes, frequency and content of risk rating management reports, documentation of risk rating determinations, and evaluation of control functions.
* A one-year PD estimate for each grade must be provided as a minimum input.
* Banks must collect and store historical data on borrower defaults, rating decisions, rating histories, rating migration, information used to assign ratings, PD estimate histories, key borrower characteristics, and facility information.
As mentioned above, the requirements that a bank must meet are partially dependent upon which of the two IRB approaches a bank will use. The first methodology, called the foundation approach, requires fewer direct inputs by banks and provides several supervisory parameters that, in many cases, carry over from those proposed for the standardized approach. For a variety of reasons, the U.S. does not plan to introduce the foundation approach in its regulations. The second approach, the advanced IRB approach, allows banks much greater use of their internal assessments in calculating the regulatory capital requirements. This flexibility is subject to the constraints of prudential regulation, current banking practices and capabilities, and the need for sufficiently compatible standards among countries to maintain competitive equality among banks worldwide.
There are four key inputs that are needed under IRB, for both the foundation and advanced approaches. The first element is the PD of a borrower; the bank is required to provide the PD in both the foundation and the advanced approaches. The second input is the estimate of loss severity, known as the loss given default (LGD). The final two elements are the amount at risk in the event of default or exposure at default (EAD) and the facility's remaining maturity (M). LGD, EAD, and M are provided by supervisors in the foundation approach, but must be provided by banks operating under the advanced approach (subject to supervisory review and validation). For each exposure, the risk weight is a function of PD, LGD, and EAD.
The IRB approach envisions internal rating systems that are two-dimensional. One dimension focuses on the borrower's financial capacity and PD estimates that quantify the likelihood of default by the borrower, independent of the structure of the facility. The other dimension takes into account transaction-specific factors such as terms, structure, and collateral. These characteristics would determine the second dimension, i.e., the LGD. Implicit in this treatment is the assumption that when a borrower defaults on one obligation, it will generally default on all its obligations. (This assumption is relaxed with the IRB treatment of retail portfolios.)
Calculating the capital charge under the IRB approach involves several steps. The first of these steps is the breakdown of the bank's portfolio into five categories: corporate (including commercial real estate), retail, bank, sovereign, and equity. The IRB rules differ to varying degrees across these portfolios. As a result, the IRB capital charge is calculated by category, with the PD, LGD, and EAD inputs potentially differing across these categories. Supervisory approval is needed before banks can use the IRB approach for any of the five categories. The minimum requirements described above were written to apply across these five types of exposures.
Another important step is the determination by the bank of the PDs for its loan grading categories. The PD of an exposure is the one-year PD associated with the borrower grade, subject to a floor of 0.03% (excluding sovereigns). The determination of PDs for borrowers supported by guarantees or credit derivatives is more complex. Banks under the advanced approach would use their internal assessments of the degree of risk transfer within supervisory defined parameters, while those under the foundation approach would use the framework set forth in the new credit risk mitigation provisions. Overall, the PD must be "grounded in historical experience and empirical evidence," while being "forward looking" and "conservative." A reference definition of default has been developed for use in PD estimation and internal data collection of realized defaults.
Once the PD has been established, banks must then establish the dimensions of LGD based on collateral and M. Under the foundation approach, M is assumed to be 2.5 years. There are several options that may be selected for the advanced approach, but in general, M is defined as the greater of one year or the remaining effective maturity in years.
After the bank determines the PDs and LGDs for all applicable exposures, these combinations can be mapped into regulatory risk weights. The risk weights, which are calibrated to include coverage for both expected and unexpected losses, are expressed as a continuous function. The minimum capital charge is then determined by multiplying the risk weight by the amount expected to be outstanding at the time of default (EAD), and by 8%.
A final step in this process involves the ongoing review by the supervisors of the systems used to develop the IRB capital charge. Periodically, supervisors will need to validate these systems and review the internal controls that provide the foundation for the IRB approach. In addition, supervisors will also have to consider, under Pillar 2, whether the amount of capital generated by the IRB approach is commensurate with the bank's risk profile.
Implementation of the IRB Approach
In addition to the requirement that a bank meet the qualifying or eligibility criteria, the new Accord requires that banks using the IRB approach run parallel systems for one year before implementation. This means that a bank planning to implement the IRB approach in December 2006, will actually have to begin calculating results as of December 2005, while continuing to run its current systems.
Adjustments to the Capital Charge for Credit Risk
There are additional considerations that banks may have to factor in when determining the capital charge for credit risk. These additional considerations will further adjust required capital, outside of the requirements of the different approaches to credit risk. The two primary adjustments that might be made to the credit risk charge are for credit risk mitigation and asset securitization.
Credit Risk Mitigation
The new Accord provides a measure of capital relief for certain qualifying risk-mitigating techniques used by banks. However, it is important to note that most of the credit risk mitigation proposals in the new Accord are only directly relevant to the standardized or foundation IRB approaches, which are not likely to be used in the U.S. In the advanced IRB approach, credit risk mitigation must meet certain qualitative requirements, such as legal certainty. In addition, specific proposals related to maturity mismatches and backtesting requirements of certain model results are applicable to the advanced IRB approach. Otherwise, it is assumed that any credit risk mitigation efforts will be factored into the PDs and LGDs assigned by the bank.
With that caveat in mind, the section on credit risk mitigation in the new Accord attempts to provide rough approximations of the risk reduction attributable to various forms of collateralized credit exposures, guarantees, credit derivatives, and on-balance sheet netting arrangements. The Committee has proposed a conceptual approach to these risk mitigation techniques that, while recognizing their risk reduction benefits, attempts to capture the additional risks posed by such transactions.
The credit risk mitigation proposal provides both a simple and a comprehensive approach to dealing with collateral. The proposal expands the range of eligible collateral from that recognized in Basel I. It also discusses the appropriate treatment for maturity mismatches between the credit risk mitigant and the underlying credit exposure. The proposal introduces haircuts, which the bank may estimate, to cover the market price and foreign exchange volatility that may be inherent in collateral. The proposal allows banks to greatly reduce the capital requirements for exposures with large amounts of high quality collateral. There are strict quantitative and qualitative factors that must be met in order for a bank to be permitted to use its own haircut estimates. The proposal encourages the use of credit risk mitigation by expanding the type of collateral, guarantors, and transaction structures that are recognized for capital reduction. Different types of credit risk mitigation techniques pose different levels of additional risk; the proposal incorporates flexibility that recognizes these differences and adjusts the capital treatment accordingly.
Asset securitization is clearly an important issue in the U.S., as the securitization market is significantly greater than the securitization market of any other Basel-member country. The Committee believes that it is important to construct a more comprehensive framework to better reflect the risks inherent in the many forms of asset securitizations, including traditional and synthetic forms.
The securitization framework in the new Basel Accord applies generally when there is a transaction that involves the stratification or tranching of credit risk. The Committee has developed securitization approaches for both standardized and IRB banks. The level of complexity is significantly higher for IRB banks. The framework tries to focus on the economic substance of the transaction, rather than its legal form.
Under the proposal for the treatment of securitizations by standardized banks, the capital charge is generally determined by multiplying the amount of the securitization exposure by the risk weight mapped to the long- and short-term rating categories. Off-balance sheet exposures are subject to a conversion factor before the appropriate risk weight is applied. The proposal does allow for some recognition of credit risk mitigants provided on securitization exposures, but that recognition is permitted only when the bank meets a series of stringent criteria.
Banks that adopt the IRB approach for credit risk are generally required to use one of two methods for determining capital requirements for securitization exposures. One method is the Supervisory Formula Approach (SFA), under which capital is calculated through the use of five bank-supplied inputs: the IRB capital charge on the underlying securitized exposures (as if held directly on the bank's balance sheet); the tranche's credit enhancement level and thickness; the pool's effective number of loans; and the pool's exposure-weighted average loss given default (LGD). The second method is known as the Ratings Based Approach (RBA). Under this approach, capital is determined by multiplying the amount of the exposure by the appropriate asset-backed security risk weights, which depend on external rating grades, short- or long-term. Granularity of the pool and the level of seniority of the position are also considered.
The securitization proposal is one of the newest pieces of the Accord and its potential impact on the industry is still being assessed. In the December 2002 QIS exercise, banks were asked for the first time to provide data on the relative impact of the proposals. The QIS results did not provide entirely reliable results. However, the Committee has responded to some of the concerns raised during the QIS process by making changes to the securitization framework. One key change was the introduction of a simpler approach for liquidity facilities.
One of the most significant changes in the new Accord is the proposal for an operational risk charge. It is expected to represent, on average, 10-15% of the total minimum regulatory capital charge. The framework is based upon the following operational risk definition: the risk of loss resulting from inadequate or failed internal processes, people, and systems or from external events. This includes legal risk, but excludes strategic and reputational risks.
The Committee has proposed three approaches to calculate the operational risk charge, which represent a continuum of increasing sophistication and risk sensitivity. The Basic Indicator Approach (BIA) is the simplest of the three approaches; the capital charge is determined by taking an alpha factor decided by the Committee and multiplying it by an indicator, gross income. The next approach is known as the Standardized Approach and is similar to the BIA, but breaks out gross income into business lines. The Committee has introduced an Alternative Standardized Approach to address some of the concerns raised by the results of the December 2002 QIS exercise; this is not a separate approach, but rather a modification to the Standardized Approach. Because there is no compelling link between these measures and the level of operational risk, the U.S. does not plan to utilize the BIA or the Standardized Approach (including the Alternative Standardized Approach) to determine the capital charge for operational risk.
The Committee has made the most significant changes to the advanced approach since it was originally introduced in January 2001. At that time, the Committee envisaged a single, very prescriptive advanced approach for operational risk, similar to credit risk. However, after numerous comments from the industry, the Committee made substantive changes in the proposal to reflect the evolutionary nature of the operational risk framework. The Committee recognized that, unlike credit risk, there are very little data and no internal systems specifically designed to target operational risk; instead, banks and supervisors rely primarily on internal controls to deal with a myriad of banking risks that cannot be as readily quantified as credit and market risks.
The Committee considered the comments and analyzed the state of the art of operational risk and developed what is known as the Advanced Measurement Approaches (AMA). Rather than prescribing one methodology, the AMA will allow banks the option of designing the operational risk measurement framework that best suits their institution, subject to some broad criteria. The criteria will be the key to achieving a certain level of consistency and comparability among institutions, as well as providing a margin of comfort to supervisors who must assess these differing systems. The criteria currently identified in the new Accord include the need for internal and external data, scenario analysis, and consideration of business environment and internal control factors. Banks may also, under the AMA, consider the impact of risk mitigation (such as insurance), again subject to certain criteria set to ensure that the risk mitigants act as an effective capital-replacement tool.
Temporary Capital Floors
Two floors that have been established for the Basel II framework. In the first year of implementation, an institution's required minimum level of regulatory risk-based capital cannot be less than 90 percent of the minimum level of capital that would be required under the Agencies' general risk-based capital rules. In the following year, an institution's minimum level of regulatory risk-based capital cannot be less than 80 percent of the minimum amount required under the Agencies' general risk-based capital rules.
1See, Daniel E. Nolle, "Bank Supervision in the U.S. and the G-10: Implications for Basel II," RMA Journal, June 2003.
2 See Advance Notice of Proposed Rulemaking, Simplified Capital Framework for Non-Complex Institutions, 65 FR 66193 (November 3, 2000).