Concerns and Considerations for the Practical Implementation of the New Basel Accord

I am pleased to participate in this conference on such an important subject as the proposed new Basel Accord. Your topical agenda and the expertise of your speakers are impressive, and I am sure that the discussions will provide new insights.

I have been asked to talk today about some of the issues associated with the practical implementation of Basel II. Before I start, let me reiterate that I remain optimistic that a new accord will emerge from the Basel Committee in a timely fashion, so that national review procedures can go forward. Although I will discuss concerns and considerations, these will not, in my judgment, threaten the basic momentum of the process.

Basel I

There is no better place to begin a discussion of some of the practical concerns and considerations of the new accord than at the beginning--with Basel I. Many have forgotten that the first accord had its origins in complaints that the globalization of banking had distorted competitive balance. Banks domiciled in jurisdictions whose supervisors required a more prudent level of capital perceived that they were disadvantaged, certainly in their home markets, by banks whose home supervisors were less aggressive in their minimum capital standards. Basel I was intended to level the playing field for banks that operated across national boundaries by establishing consistent standards on how minimum regulatory capital was to be determined in individual countries and what was to constitute capital.

We should not lose sight of the continuing imperative, both economic and political, to ensure that a revised accord is perceived by all to maintain a level playing field for banks operating not only across national boundaries but also domestically. I will return to this issue later.

Basel I--on which the world's regulatory capital regime has been based for more than ten years--was a genuine step forward for most countries' capital rules and a watershed for international cooperation among the world's supervisors. But it is certainly clear to those of you at this meeting that globalization, technology, and innovation have accelerated so dramatically that Basel I cannot provide the industrial world's largest and most-complex banking organizations with a regulatory capital requirement that reflects their underlying risk exposures. Basel I, with its modest risk sensitivity, treats most loans as if they come from one quality category. It ignores techniques that the largest banks have adopted to mitigate risk. It erroneously treats some transactions that only appear to reduce risk as if they in fact do. Its overly simple risk weights induce large banks to game the rules by shifting to the market those exposures that the market judges require less capital than the regulations do and by retaining exposures with a regulatory capital charge that is lower than the market perceives is necessary. Such capital arbitrage has greatly reduced the usefulness of regulatory capital ratios at the largest banks and provides little useful information to the public or the supervisor. For these institutions, the regulatory rules must be changed. This is the practical reality that bank supervisors face.

Change and Complexity, Flexibility and Comparability

This audience is well aware of the reality that I have just described. But a second practical problem is that no one likes change, at least change imposed on them. Change can be expensive and, worse, requires new constructs that have the potential for unintended implications. For this reason, more than any other, the development of Basel II has taken a long time. Not only are the issues complicated, but also the banks and their supervisors have engaged in a continuing conversation about this complexity and its implications. Both parties have shared a common goal of trying to develop a cost-efficient and workable system. On occasion they have differed on the tradeoffs between cost and complexity and between comparability and flexibility.

Another practical problem is that the world's supervisors are trying to do more than develop a better risk-based capital standard. They are also trying to harness modern risk-measurement and risk-management techniques to the regulatory system, and they are trying to construct a framework that can evolve as the science and the art of risk measurement and management evolve. I have previously called this evolutionary potential of Basel II its "evergreen" element, and I believe it is one of the many attractive features of the proposal.

Modern asset-pricing theory may have begun in academia, but its growth and application have quickly taken root in money, capital, and banking markets in new ways of thinking about risk, ways that essentially meant measuring risk in quantifiable ways. The largest banks, operating in multiple product and geographic markets, simply could not operate without applying the new evolving principles in measuring, shifting, and managing their risks.

Several implications follow. Harnessing large banks' internal practices to the supervisory process implies that supervisors could use such cutting-edge risk measures to evaluate bank risk profiles and relate them to a truly risk-sensitive capital standard. It also suggests that a bank's adoption of best-practice risk measurement and management could be used as a supervisor-imposed prerequisite for the application of the new capital standard; setting such a prerequisite would accelerate the development and adoption of these new techniques, which counterparties are increasingly demanding in any event. Stating the point more directly: Basel II is designed to harness the best new techniques but also to ensure their application by those banks that have been less aggressive in adopting them. That is, Basel II, at least in its more advanced form, is as much a proposal for strengthening risk management as it is a proposal for improving capital standards; these considerations are, as they should be, inseparable. That inseparability, in turn, as I will discuss later, is an important factor leading to the bifurcated application of Basel II proposed in the United States.

The use of the more advanced techniques does not go far enough for those sophisticated organizations that want to use their own models to estimate their own risk functions, particularly their estimates of portfolio correlation effects. Some are also concerned that the hard-wired risk functions and correlation parameters will prohibit better practices when they are available.

In my view, a full models regime could not, over the planning period for Basel II, produce capital charges that would be sufficiently comparable for a regulatory capital standard. Specifically, determinations of portfolio correlation effects tend to be as much art as science; and their subjectivity makes them difficult, if not impossible, to validate credibly. To ensure broad comparability of regulatory capital ratios across banks, therefore, the risk functions embedded in the internal risk-based (IRB) framework employ correlation parameters that are specified by the Basel Committee. In effect, we have sacrificed some flexibility and potential risk sensitivity for greater comparability. We have faced similar tradeoffs repeatedly during the development of the Basel II proposals, and in many ways one of the Basel Committee's greatest challenges has been reaching consensus on the right balance between flexibility and comparability.

In addition, there is evidence that credit risk models at many large, complex banking organizations have not attained the sophistication and robustness that would be consistent with their use for regulatory capital. Although the state of the art is progressing rapidly, many banks still must resolve some fundamental questions and finish compiling sufficient data before they can adopt their own full credit models that produce reliable and accurate results.

Ultimately, Basel II proposes a capital standard that implies the need for reasonable comparability across banks and across boundaries to ensure the competitive balance to which I referred earlier. Nonetheless, looking beyond the current proposals, I believe that over time the regulatory capital functions we have hard-wired into Basel II, along with their embedded correlation assumptions, will give way to individual bank-developed models that are verifiable by supervisors. That development will probably be a central part of the evergreen process. But the time for individual bank models is not now, and I see no reason to hold up improving the regulatory capital regime until the state of the art permits that innovation. Until then, I think we have developed a workable balance between flexibility and comparability. As the old saying goes, the perfect should not be the enemy of the good.

Level Playing Fields: Treating Comparable Banks Comparably

Basel II, as I noted, tries to develop a standard so that banks' risk exposures can be treated comparably. Such a standard is central to the required level playing field. But comparability requires not that all banks be treated the same but rather that comparable banks be treated comparably, raising practical issues of definition, of unintended consequences in competitive markets, and of unavoidable differences in national regimes.

General Issues. As you know, Basel II proposes three options from which banks in national jurisdictions may choose--the standardized, the foundation, and the advanced internal risk-based (A-IRB) approaches. The increasing risk sensitivity of the three options has the potential for the application of different capital charges across banks for the same exposures and implies that banks with a more or less average mix of corporate and retail exposures may have reductions in their overall minimum regulatory capital requirements as they move across the three options. Any differences in regulatory capital charges have the potential to distort both national competitive positions, an issue for domestic authorities, and international competition, an issue for both national authorities and the Basel Committee.

In considering all of these issues, one must, of course, make a distinction among minimum regulatory capital based on regulatory rules; economic capital based on the bank's own assessment of risk and capital needs; and actual capital held, which includes buffers above both the other capital measures for reasons varying from reduced cost of funding to counterparty demands and to desired contingency flexibility. Minimum regulatory capital is the lowest of the three capital "requirements," and thus the degree to which it may or may not affect competition across banks is an important conceptual and practical issue involving, at bottom, the way the price of credit gets determined.

When all is said and done, the nations participating in the Basel Committee are now developing a proposal to establish a revised capital standard for banking organizations that compete across national boundaries. Basel I was negotiated on the same basis. After Basel I, all the participating nations nevertheless chose to apply the new standard to all the banking organizations under their jurisdiction. The Europeans and the Japanese apparently have made the same decision for the proposed Basel II. In contrast to Basel I, however, all banking organizations will be subject to Basel II, but each organization will choose among the three variants.

Application in the United States. The U.S. authorities have made a somewhat different decision. Consistent with the letter and the spirit of the Basel II proposal, the latest U.S. proposal states that all U.S. banking organizations with meaningful cross-border exposures--at least $10 billion--will be required to adopt Basel II. In addition, any banking organization with consolidated assets of at least $250 billion will similarly be required to adopt Basel II. If these criteria were applied today, about ten or so U.S. entities would meet one or both of these criteria and hence would be among the "core" group of U.S. banking organizations required to adopt Basel II. To be sure, the actual number of mandatory U.S. banks may change before actual implementation--and among them could be U.S. subsidiaries of foreign banking organizations that meet the core bank standards. In addition, we initially assumed that about ten other large entities might choose to opt in to Basel II; we now believe that number may well be an underestimate, but we are still in the process of surveying our larger banks to determine their plans.

Even the twenty banks counted currently as mandatory and opt-in would account, we estimate, for more than 99 percent of the foreign exposure of U.S. chartered banks and more than two-thirds of the domestic assets of these entities.

Any non-mandatory bank in the United States that can estimate the internal risk parameters for its credit exposures--that is, measure and manage its risk exposure to the satisfaction of the supervisor--may opt in to Basel II in the United States. But if a bank chooses not to meet this test or not to adopt Basel II, it will remain under the current, unchanged, capital regime.

Banks that are required or that choose to adopt the Basel II rules in the United States will have only one option: the A-IRB approach for credit risk and the advanced measurement approach (AMA) for operational risk. Neither the standardized and foundation approaches for credit risk nor the basic indicator and standardized approaches for operational risk will be available in the United States.

The authorities in the United States proposed the bifurcated application of Basel II (with one group under Basel II and most banks remaining under the current capital requirements) and rejected the trifurcated approach (with banks choosing for themselves among the three Basel II variants for credit risk, as well as three variants for operational risk), which looks to be preferred in other countries, for three basic reasons. First, Basel II, as I noted, requires that those adopting it apply it to their internationally active banks. The data to which I referred are evidence that the U.S. framework for applying Basel II would meet that test. Second, as I also noted earlier, Basel II capital requirements are intended not only to be more sensitive to risk but also to link that risk-sensitivity to a significant increase in the standards for risk measurement and management at larger banks. Only the A-IRB and the AMA approaches fully impose that prerequisite on the large entities. The U.S. authorities believe that the largest U.S. banking organizations should adopt best-practice risk measurement and management for reasons of safety and soundness. Third, Basel II is not without cost. Most of the thousands of U.S. banks that are neither in the core set nor in the likely opt-in set have operations that, in the U.S. authorities' view, would not require the dramatic changes in credit risk measurement and management associated with either the A-IRB or the foundation approach. Additionally, the increased risk sensitivity of the standardized version seemed modest to us relative to the additional costs of system changes. Regarding operational risk, the arguments are even stronger that the AMA would impose undue burden on smaller banks. In short, Basel II does not seem to have a favorable cost-benefit ratio for most American banks.

The decision that most banks would remain under the current regime in the United States was also conditioned on some institutional facts that were perhaps not well known elsewhere. Minimum Basel I requirements are not the only capital regulations in the United States. They are supplemented by benefits established by statute for banks with higher tier 1 and tier 2 capital ratios and by legislatively imposed minimum tier 1 leverage ratios. In addition, statutory prompt corrective action has induced banks to carry buffer capital to avoid losing regulatory benefits that come with holding capital above regulatory minimums. Moreover, the market demands that our thousands of smaller banks hold substantial equity capital, an amount significantly above the minimum standards. Nearly 95 percent of U.S. small and medium-sized banks have capital ratios in excess of 10 percent and most likely would not be required to hold more under Basel II. U.S. supervision of these banks already includes substantial Pillar 2 elements, and of course, these banks operate almost totally within the United States.

The arguments that banks other than core banks, and especially the small and medium-sized banks, ought to be free to choose between Basel II and their current regulatory requirements seemed overwhelmingly convincing to us. That is, banks should be free to choose to bear the costs of implementation for the benefits of greater capital risk-sensitivity. Neither for international agreement nor for domestic supervisory reasons did imposing Basel II on all U.S. banks seem reasonable. But for international agreements and for domestic supervisory reasons it seems only reasonable for us to require nothing less than A-IRB and AMA for those banks that adopt Basel II.

Unintended Consequences and Competitive Distortions. Having made our bifurcated proposal in the United States, public comments and congressional oversight have made clear the purely domestic concern that banks remaining under the current regime, even though they avoid the costs of adoption, may be disadvantaged relative to Basel II banks. Specifically, the argument goes, Basel II will give the largest banks, if not a lower overall capital requirement, then lower capital charges on certain credits with which banks not adopting Basel II will have to compete. Focus has been placed on residential mortgages, small business loans, and credit cards. Concern has also been voiced that Basel II banks will use any newly created excess regulatory capital to acquire smaller banks, whose capital can be used more efficiently by the larger Basel II banks. In short, creating international competitive balance under Basel II carries the potential that domestic competitive balance will become distorted.

At bottom, these concerns raise empirical questions about how credit is priced, about the locus of competition, about the determinants of actual capital held, and about other matters. Questions have been raised in the comment and oversight process and these questions will have to be addressed. We are still reviewing the evidence presented to us, and Federal Reserve staff members are conducting empirical research on the issues that will be made public in the months ahead. We will need to review the available options for addressing any of these concerns that are supported by the evidence. These options include changes in U.S. Basel II rules, where national discretion is allowed; modifications to the proposed bifurcated application; and changes in the current capital regime in the United States. We will look seriously at each and all of these steps and will not be precluded from proposing any measure that we believe is necessary to ensure a level playing field in our domestic banking market. But first, as I noted, we must examine the evidence, just as other nations are undoubtedly reviewing the unintended consequences of a trifurcated application in their own markets.

Even though the objective is a level playing field internationally, the practical problem that differing supervisory regimes and procedures may distort international competitive positions remains. Frankly, the U.S. authorities hear arguments that some foreign supervisors have neither the resources nor the experience to apply the rules as rigorously as they anticipate will be the case in the United States. And foreign authorities have heard complaints from their banks about the competitive implications that will be created by U.S. insistence on A-IRB and AMA treatment for their large U.S. subsidiaries that meet the cross-border or scale criteria to be core banks.

Real and perceived differential treatment existed under Basel I, and it will remain under Basel II. However, the Accord Implementation Group (AIG) in Basel, made up of line supervisors from member countries, has already had some success in trying to ensure similar--or at least consistent--treatments across national boundaries under the new proposal. Tensions in applications by different national authorities are a fact of life that we have to address when we can. The test is whether we will be better off and whether competition will be fairer under Basel II even though it is not perfect.

Level Playing Fields: Home and Host Issues

As I just said, treating comparable banks comparably within a national jurisdiction may have unintended consequences for foreign banks. U.S. subsidiary banks owned by foreign parent organizations will be subject to the same criteria as those for mandatory A-IRB banks in the United States. Others may feel it desirable to opt in to such versions for their U.S. operations. Such national treatment in the host country--treating comparable banks comparably--is reasonable, but it may create genuine operational complexities if these banks plan to use one of the other Basel II options in their home or in third countries.

For their part, the U.S. authorities have clearly stated that they will be flexible during a transition period so that the relevant foreign banks can more easily meet the required standards, but national treatment policies will apply. Discussions within the AIG on this issue--as well as bilateral discussions with affected banks and their supervisors--have been productive, and I think the associated problems will be addressed.

However, one issue appears to be particularly difficult in more than one jurisdiction: the allocation of capital for operational risk among the legal entities within and across jurisdictions. Problems do not come more practical than this. Operational risk is generally measured on a consolidated basis, often by business line. That diversification benefits exist on a consolidated basis suggests that operational risk estimated from the bottom up--by, say, legal entity--would not only be more difficult but might well add up to more than the total from the top down. The problem is only made more complicated by the understandable focus of supervisors in each jurisdiction on ensuring that the entities under their supervision are sufficiently capitalized to absorb risk. We have come a long way in global banking, with deference to home consolidated supervisors, but the legal entity supervisor still needs the assurance that capital is protecting risk in the individual unit.

Basel II is not going to succeed or fail on this practical problem or on similar problems--and this will certainly not be the last. Through the AIG a compromise will be developed on the issue of operational risk capital to ensure that capital is sufficient and is allocated in a reasonable way. Other issues, including home-host tensions, will be resolved in similar ways as the member countries hear comments, conduct analysis, and then seek out a viable compromise solution.

The Process

The practical give and take of discussions is exactly how the first and now the second capital accord have been and will be developed. We are at a critical stage of that process with changes in significant provisions being studied and modified. Examples are the shift in standard to unexpected loss only and the review of securitization, credit card, and credit risk mitigation, all of which are now being actively studied.

It may seem strange that this far into the process--approaching year six--that such changes are being made. These changes, however, show how seriously the Basel Committee and its participating national authorities regard the public comment process and benefit from the analysis of bankers and other interested parties. When analysis and evidence were put forward, the committee responded.

The United States is considerably through its review of public comments on its own advanced notice of proposed rulemaking (ANPR), and other countries are going through their own process of translating a consensus proposal into rules. The processes in the United States and in other nations may well bring up further proposals for change, provided that supportable arguments are forthcoming.

In the United States, after this next round of committee discussions, additional procedural steps still are required, with opportunity for comment at each stage, before a final rule is in place. To be clear, the need for more procedural steps in the United States should not be taken as an indication of a lack of commitment to the Basel process. Rather, it is a sign of our attempt to develop these proposals on as transparent a basis as possible and to hear a wide range of comments. The Congress of the United States has also held hearings on the development of the new accord. The interest and oversight by our Congress in these discussions is appropriate and welcome, particularly for an undertaking as extensive as Basel II. The U.S. regulators appreciate the fact that we are able to operate as independent agencies but also realize that we have an obligation to keep the elected representatives informed of our progress in this major effort. Of course, it is expected that other countries and the European Union will be following their own procedures as well.

Overall, I have been impressed by the efficacy of the comment process and believe it offers an open forum for all parties to voice their opinions. Debate and discussion on such an important undertaking are essential, and they provide an opportunity for enhancements to the final product. The current proposal is better because the comments have elicited good ideas that have been seriously considered.

At the same time, we do need to ensure that momentum is maintained. In October, the committee members committed to work promptly to resolve outstanding issues by mid-2004. In January, the Basel Committee will meet to address further analysis of outstanding issues and review the timetable for completing and implementing the committee's work on the accord. Using the agreement reached in the committee as a template, U.S. supervisors then plan to conduct another Quantitative Impact Study (QIS) to gauge more clearly the effect of the Basel proposals. We believe that we must carefully test the new proposal to determine its effects on individual institutions and to ascertain the need to fine-tune the proposal further, a process that could include recalibrating some of the risk-weight functions. Our sense is that other countries will be doing the same. Because timing is a function of the degree of changes required by the QIS, the U.S. agencies will then conduct a full notice of proposed rulemaking once again to seek and then evaluate public comments before adopting a final rule.

Conclusions

Basel I is no longer a viable supervisory tool for the large, complex banking organizations of the industrial world and needs to be replaced as soon as feasible by a new capital accord. The proposal being developed at Basel builds on best-practice risk- measurement and risk-management techniques and, at least in its advanced versions, is as much about ensuring that banks use such techniques as it is about a more risk-sensitive capital approach. Indeed, these considerations are two sides of the same coin.

Any complex change will induce opposition and concern, in part because of fear of change and comfort with the known rules, in part because of preference for other alternatives by those affected, and in part because of anxiety that it may upset current competitive positions. But Basel II builds upon modern techniques and is entirely consistent with the directions that both large banks and their counterparties are now moving. And it can use future advances to evolve into an even more flexible and sophisticated supervisory tool.

Basel I and the proposed Basel II are designed to provide a level competitive playing field for banking organizations meeting in international competition. And though proposed application procedures--both the bifurcated approach in the United States and the trifurcated approach in other countries--by and large maintain that objective, they may have the unintended consequence of distorting domestic competitive equity. The empirical analysis on whether they will, especially when banks maintain capital well in excess of regulatory minimums, is yet to be completed, but the authorities are required to investigate these concerns and adjust the proposal if the analysis requires it.

Tensions in cross-border applications exist and cannot be avoided in a world of national regulatory authorities, especially where separate legal entities exist within national jurisdictions that apply national treatment. The AIG has been successful so far in smoothing differences that may develop in cross-border application, but these will inevitably involve compromises on difficult issues.

Return to top

2003 Speeches