DONALD E. POWELL
FEDERAL DEPOSIT INSURANCE CORPORATION
NEW BASEL ACCORD
SUBCOMMITTEE ON DOMESTIC AND INTERNATIONAL MONETARY
POLICY, TRADE, AND TECHNOLOGY
COMMITTEE ON FINANCIAL SERVICES
U.S. HOUSE OF REPRESENTATIVES
February 27, 2003
Room 2129, Rayburn House Office Building
Thank you, Mr. Chairman and members of the Subcommittee. I welcome the opportunity to testify on behalf of the Federal Deposit Insurance Corporation on the development of revised international capital standards. In addition, I would like to express my appreciation to the members of the Basel Committee on Banking Supervision and their staff for their dedication and hard work over the past five years in formulating the Basel II Accord. I came into this process eighteen months ago and recognize that much of the ground work had been done before I became involved.
This agreement could have far-reaching effects on the management and supervision of the largest, most complex banking organizations in the world. It is imperative, therefore, the end result of this Accord is better regulation. It is essential that our process be thorough and inclusive; that our deliberations and documentation be transparent; and that the impact of our actions to the greatest extent possible be widely understood by Congress, the regulators, and America's financial institutions.
Bank capital is subject to federal regulation because of its critical importance to the health and well-being of the U.S. financial system. Debt financing creates liabilities banks must satisfy regardless of the severity of external economic events, but capitalessentially the funds contributed by shareholderscan absorb losses without causing a bank to fail. An adequate capital cushion enhances banks' financial flexibility and their ability to weather periods of adversity. The FDIC, as insurer, has a vital stake in the adequacy of bank capital.
The conceptual changes being considered in Basel II are far-reaching. First, the new Accord contemplates a two-tiered regulatory capital standard for America's financial institutions: one set of rules for the large, complex and internationally active institutions, and another set for the rest of the banks in the country. Second, the Accord represents a significant shift in supervisory philosophy. Rather than emphasizing pre-set minimum numerical capital standards established by the regulators, the new Accord envisions banks using their own internal risk estimates as inputs to regulator-supplied formulas with the supervisors providing oversight and evaluation of the banks' ability to measure risk. Third, minimum capital requirements for credit risk would generally be reduced, with additional capital held based on more flexible elements of the Accord, such as an operational risk charge or the imposition of discretionary supervisory capital buffers.
These are all important issues worthy of the attention of this Committee, the regulators and the financial services industry at large. Each of these fundamental developments raises questions, however, that must be addressed before this new capital structure can be considered a success. There are good arguments for moving forward and the FDIC will continue to support the Basel II process. However, it is important that certain fundamental issues be resolved satisfactorily in the coming months in order for the FDIC to give its full support to the new Accord.
I will focus my testimony on the weaknesses of the capital framework that the largest banks are operating under today and the logic behind the new capital Accord. I will then address the threshold issues that must be resolved prior to a decision by the U.S. to adopt this new Accord, and conclude with a brief discussion of several issues involved that may arise during the implementation phase.
Weaknesses of the Current System
Under the 1988 Capital Accord1 as implemented in the U.S., assets and off- balance sheet contracts are risk-weighted based on their relative credit risk using four broad categories or buckets. Overall, institutions are required to maintain a minimum risk-based capital ratio of at least eight percent. Most unsecured corporate loans are placed in the 100 percent risk weight bucket, which requires an eight percent risk-based capital charge. Lower risk assets are given lower risk weights. For example, qualifying single family mortgage loans are generally risk-weighted at 50 percent and require only a four percent risk-based capital charge. In addition to the risk-based capital requirements, all U.S. institutions must comply with minimum leverage ratio requirements of Tier 1 capital-to-average total consolidated on-balance sheet assets2 and all U.S. institutions are subject to the statutorily mandated Prompt Corrective Action (PCA) regulatory capital ratios.3
Since 1988, this system has generally worked well for most small, non-complex banks. However, the activities of the largest banks have reached a degree of complexity not easily addressed under the existing Capital Accord. The business of banking, risk management practices, supervisory approaches and financial markets have undergone significant transformation. The regulators were forced to respond by piecemeal regulatory amendment. Banks were able to take advantage of the rigid "bucket" approach of the 1988 Accord and structure their balance sheets so as to minimize regulatory capital charges. The bucket approach lacks proper sensitivity to risk and is disconnected with internal bank practices. This formula has hobbled the 1988 Accord's ability to match the industry's innovations and has reduced the regulatory capital incentives for better risk-management.
An important argument in favor of a new regulatory capital framework for large banks is that the current system simply ignores most of the best available information about the credit risks faced by these banks, namely, the risk-related information generated by the banks themselves. Large banks generate a wealth of useful information pertinent to evaluating their own credit risks, and finding a way to use this information is an important component of the new capital Accord.
The FDIC believes there are three issues that need to be addressed before a commitment is reached to implement Basel II in the U.S. First, the Accord must ensure that appropriate minimum capital requirements are maintained. Second, the new Accord must ensure that the internal risk estimates used as inputs to the new capital formulas are estimated in a sound and conservative fashion and are evaluated consistently going forward using a uniform interagency process. In addition, the competitive impact of the new Accord must be fully explored and assessed.
As a result of an extensive data collection exercise just recently completed, the regulators have a sense of how Basel II might affect minimum required capital at the largest banks if applied today. The agencies are considering whether and how, in light of these results, the Basel II Accord should be adjusted prior to formal issuance for public comment.
As the process continues, the FDIC's focus will be to ensure that minimum capital requirements under Basel II are not unduly diminished. Substantial reductions in minimum capital requirements for the largest U.S. banks would be a grave concern to the FDIC. Lower capital minimums in conjunction with a flexible operational risk charge and supervisory discretion to impose additional capital may work well in theory, but in practice it may be difficult to enforce adequate discretionary capital buffers in cases where a bank itself does not agree that such a buffer is necessary. For the supervisory process Pillar 2 of the Accord to be fully effective, it must rest on a foundation of agreed-upon regulatory capital minimums. Congress recognized this when the PCA requirements were established in the Federal Deposit Insurance Corporation Improvement Act. I want to be clear that this is a critical issue for the FDIC.
A noteworthy aspect of the PCA regulation is the minimum leverage capital requirement. To be considered well-capitalized, a bank must have a ratio of Tier 1 capital-to-total assets (the leverage ratio) of at least five percent. Banks with leverage ratios under four percent are considered undercapitalized. There is an exception: "strong" banks (CAMELS "1" that are not experiencing significant growth) are not considered undercapitalized until their leverage ratio falls below three percent. To my knowledge, this exception has never been used since no bank with a leverage ratio less than four percent has ever met the standards for a "strong" bank.
U.S. banks subject to Basel II will certainly be interested in how the new Accord affects their regulatory compliance with PCA, particularly if their risk-based capital requirement decreases to such an extent that the PCA leverage test becomes the binding regulatory capital constraint. Just as when the PCA regulations were first written, there will probably be arguments that the leverage ratio should not be used as a PCA test because it is not sufficiently risk-focused.
While no one to our knowledge has suggested weakening the PCA leverage regulations, we believe the issue will have to be confronted if Basel II moves forward in its current form. We believe this is one of the discussions that should take place before we commit to adopt Basel II in the U.S.
There are a number of compelling reasons to maintain the leverage ratio as a key capital indicator. The risk weighted assets number that capital is measured against in Basel II is based on bank risk models, which vary according to their assumptions and can on occasion be wrong. During economic booms, model inputs are likely to become more optimistic. The estimated base of risk-weighted assets under these conditions could shrink, and the satisfaction of a capital standard of "eight percent of risk weighted assets" could become less and less meaningful.
The measurement of a leverage ratio, in contrast, is much less subject to model error and the creeping effects of economic euphoria. The base against which leverage capital is measured, total assets, is determined outside the bank regulatory process by the application of Generally Accepted Accounting Principles, subjecting the bank regulatory agencies to a valuable discipline. Moreover, recent legislative and regulatory changes raising the bar on corporate governance standards, enhancing internal controls and disclosure practices, and compelling changes to accounting standards will bring greater scrutiny to the determination of what assets and liabilities are on balance sheet and increase the value of the capital discipline provided by the leverage ratio.
While a leverage ratio provides the institution and the deposit insurance funds with valuable protection, it is certainly not sufficient in itself. Equally important under a Basel II regime is identifying processes that ensure banks' internal risk estimates are estimated soundly and conservatively, and that they are evaluated consistently.
The capital required by the Basel II risk-weight curves is quite sensitive to assumptions about the risk parameters of individual credits. How will examiners evaluate the validity of those assumptions? In this respect, it is important not to place exclusive reliance on quantitative methods and models. Internal risk estimates are likely to be as robust as the credit culture in which they are produced. A rigorous corporate governance structure, effective internal controls and a culture of transparency and disclosure can all play an important role in ensuring the integrity of internal risk estimates. These qualitative elements must be accompanied by agreed-upon processes that examiners can use in assessing the soundness and conservatism of banks' internal risk estimates. These processes are being developed by the agencies but the work here is not final. I raise the issue of validation of risk-estimates today to emphasize that it is important enough to make my short list of threshold issues which must resolved.
There is a second critical dimension to the issue of evaluating bank models and model inputs, and that is the issue of uniform supervisory standards. We must avoid a situation where there are differences in regulatory capital among banks utilizing Basel II that have nothing to do with differences in underlying risk profiles. An example of such an undesirable scenario would be where Bank A faces higher regulatory capital than Bank B simply because it uses a more conservative approach to measuring the same risks or because its supervisor differs in its approach to implementation. As more banks qualify for Basel II over time, the potential for inconsistent regulatory capital requirements among banks will be magnified.
In Basel II, the quest for supervisory consistency is currently met by the development of lengthy, detailed and comprehensive standards and technical guidance. Basel II relies upon highly prescriptive standards to ensure consistent interpretation and uniformity in application. While these standards have added immeasurably to the Accord's complexity, the fact remains that, even with detailed rules and standards, independent supervisory judgment will be required on a case-by-case basis. The capital requirements generated in a Basel II framework will be driven by the day-to-day rating of credits by lending officers and independent risk management review processes. These processes, although subject to detailed regulatory guidance and related interagency documents, must be assessed on an ongoing basis. Supervisory review and validation of an individual bank's internal rating and grading systems will be necessary, and key aspects of the internal system not fully addressed or foreseen in the written standards will require the exercise of informed examiner judgment.
Given the level of complexity and detail, it is likely that significant differences in application and supervision at the institution level will be unavoidable under Basel II unless the federal banking agencies enhance existing interagency processes and find effective methodologies to ensure a level playing field in the supervisory oversight of Basel II capital allocation systems.
Another example of the need for enhanced interagency coordination is the monitoring and controlling of the procyclicality concerns already identified in the Basel II framework. Procyclicality refers to the tendency of the capital framework to require less regulatory capital in "good times" and more regulatory capital in "bad times" possibly exaggerating phases of the economic cycle. Basel II's reliance upon banks' internal ratings could result in progressively less capital being assessed during the upswing phase of the economic cycle and conversely, progressively more capital being assessed during an economic downturn. This could result in more expansionary lending during an upswing, thus exaggerating the economic boom. On the other end of the cycle, the capital requirements could constrain the supply of credit and further an economic decline. As a result, minimum capital requirements under Basel II may tend to reach their low point at the height of the economic cycle, when a peak has been achieved and a strong economy is on the verge of a downturn.
Under Basel II, supervisory control and oversight is relied upon to moderate any negative side-effects of this procyclical capital framework. It is essential that the federal bank regulators closely coordinate their consideration of procyclicality under Basel II and develop uniform and transparent supervisory responses and guidance. From the FDIC's standpoint as deposit insurer, participation and input into these capital adequacy deliberations will be a high priority.
Thus far, my discussion of uniform supervisory treatment has been confined to the Basel II banks. Resolving these issues will be critical for ensuring safety and soundness, and for maintaining the credibility of our large bank supervision programs. From the standpoint of public policy towards the U.S. financial system, however, another issue could loom even larger the issue of competitive equity between Basel II banks and other institutions.
Basel II will most likely be mandatory only for a group of large, complex and internationally active U.S. banking organizations. This mandatory group of institutions does not include numerous large regional banking institutions, as well as thousands of smaller, community-based banks and thrifts. Basel II banks will compete with other institutions for business ranging from large corporate customers to small business loans, credit cards and mortgages. If Basel II provides the largest U.S. institutions some material economic advantage as a result of lower capital requirements, the "non-Basel" institutions may find themselves at a competitive disadvantage in certain markets. Lower capital requirements could give "Basel banks" an advantage in the pricing of loans, the ability to leverage, or the cost of capital. Some banks also have expressed a concern about the impact of being considered a "second tier" institution by the market, rating agencies, or sophisticated customers such as government or municipal depositors and borrowers. If significant, such disparities could accelerate the trend towards industry consolidation.
For these reasons, Basel II is potentially relevant to a larger universe of banks than those that may wish to qualify. With respect to how relevant the competitive effects will be, bankers know who their competitors are and will need to decide for themselves the potential impact on their businesses. Thus far, the documentation of Basel II has been largely technical and conceptual in nature. The work of translating the technical material into dollars-and-cents information about the capital that a bank or its competitor may be required to hold will, consequently, be very important for purposes of facilitating informed comment. We have worked to 'demystify' the Accord with a symposium on the Basel II process last summer in New York and with a series of informational papers, beginning last month, on various aspects of the Accord. We will continue this effort in the months ahead.
In summary, the threshold issues that must be addressed before the U.S. implements the proposed Basel II Accord are: (1) assuring appropriate minimum capital standards for banks regardless of the results of the models; (2) establishing a consistent supervisory process for ensuring that banks' internal risk estimates are sound and conservative; and, (3) vetting any potential competitive effects with all interested persons.
Presuming these threshold issues are satisfactorily resolved, numerous Accord implementation issues still need to be decided. This testimony concludes by touching briefly on a few of these issues: the operational risk capital charge, the complexity and burden of the new Accord, and the scope of its application.
Operational risk is defined as the risk from breakdowns in technology, systems or employee performance (including fraud), and spans a wide range of significant risk exposures to banks. Many recent bank failures were directly tied to fraud, and most included some failure of internal controls. Since the conclusion of the savings and loan crisis, the single, largest loss to the Bank Insurance Fund resulted from fraud (First National Bank of Keystone, September 1999). The failure of Barings Bank an insolvency of international consequence also resulted from fraud and poor internal controls. Fraud contributed to eight of the eleven U.S. bank failures in 2002 and was the direct cause of failure in several of these cases. In short, major operational losses caused by external or internal fraud or breakdowns in internal controls are, regrettably, a common cause of recent bank failures.
We believe a capital standard is not the sole or complete solution to confronting operational risks. Active federal supervision, independent auditors, effective internal controls, and strong bank management are obvious key components of a sound risk management program. It is clear, however, that adequate capital must be allocated for operational risk and, as long as banks hold adequate overall capital relative to the risks they assume, the FDIC's interests will be served.
With respect to the much discussed distinction between Pillar 1 and Pillar 2 treatment of operational risk, it should be noted that the currently contemplated Advanced Measurement Approach provides much of the same flexibility as would a Pillar 2 treatment. It is worth emphasizing, however, that the supervisory imposition of a new and untested science for Pillar 1 measurement of operational risk should not drive significant structural change to the internal risk management processes and control systems in the U.S. banking industry.
The treatment of operational risk is only one aspect of the overall cost-benefit tradeoff that banks will need to assess when deciding whether they wish to join the group of institutions that will use the proposed Basel II approach. In this regard, some bankers have pointedly asked how much capital reduction will be permitted for banks meeting the Basel II standards. I have already discussed the critical significance of where we draw the line in terms of banks' overall capital. It is, nevertheless, defensible that there could be at least some additional capital flexibility granted to banks that have substantially improved their risk-management programs. The question is, how do we identify the necessary improvements that qualify banks for a capital regime that allows for this additional flexibility.
Under Basel II, in order to implement the IRB (internal ratings based) framework, banks will need to obtain and aggregate default and loss data for each type of loan class in their portfolio. The data will need to span a period of several years in order to effectively gauge credit risk through an economic cycle. In addition to the systems that must be developed to fully adopt an IRB framework, banks must also invest in staff expertise, internal controls and make other structural changes driven by the high qualitative standards that append to the Basel II standards.
As a result, compliance with Basel II's IRB framework will require a significant investment in time and resources, systems and people. The entire banking organization could be affected by a conversion to an IRB framework.
The Basel II framework, especially the IRB standards, impose lengthy, detailed and complex requirements. The qualification standards, under development by the Accord Implementation Group for banks required to implement the IRB approach, will add a further layer of complexity and detail. For each level of complexity, an additional increment of burden is added to the regulatory framework. There is, indeed, a demand for complexity as banks seek to have capital tailored to their individual risk profiles. In short, in order to implement Basel II, a greater degree of complexity and associated burden is unavoidable. These burden considerations, and the desirability of testing the waters with the new Accord, suggest that the universe of Basel II banks initially should be relatively small.
The ideal Basel implementation would be an Accord that ensures adequate capital in the system while correcting the deficiencies of the 1988 Accord with respect to the regulation and supervision of large, complex institutions. The new Accord should ensure that the complexity needed to achieve the necessary risk focus is not so great as to stand in the way of effective implementation or supervision. In addition, it should provide incentives for better risk management and avoid such significant regulatory and capital discontinuities between Basel and non-Basel banks as to tilt the financial services playing field in major unintended ways.
The FDIC will work to ensure these goals are being met as the process moves forward.
Thank you for the opportunity to present the views of the FDIC.
1 "International Convergence of Capital Measurement," issued in July 1988, describes the framework. The Agencies' risk-based capital standards implementing the 1988 Accord are set forth in 12 CFR part 3 (OCC), 12 CFR parts 208 and 225 (Board), 12 CFR part 325 (FDIC), and 12 CFR part 567 (OTS).
2 In general terms, Tier 1 capital includes common stockholder's equity, qualifying noncumulative perpetual stock (for bank holding companies it also includes limited amounts of cumulative perpetual preferred stock), and minority interests in the equity accounts of consolidated subsidiaries.
3 Under the PCA regulations mandated by Congress, institutions are classified into categories based on their regulatory capital ratios. The minimum leverage ratio for strong institutions is 3 percent, and is 4 percent for other banks. As directed by the Federal Deposit Insurance Corporation Improvement Act of 1991, institutions with the highest capital ratios (i.e., at least 10 percent total risk based, at least 6 percent Tier 1 risk based, and at least 5 percent leverage) are categorized as "well-capitalized," while institutions with lower capital ratios are assigned lower capital categories. Institutions that are less than well-capitalized have restrictions or conditions on certain activities and may also be subject to mandatory or discretionary supervisory actions. These PCA requirements are unique to U.S. banks and reflect Congressional intent to reduce the cost of bank failures and reduce opportunities for bank supervisors to practice forbearance towards thinly capitalized institutions.