Remarks by John D. Hawke, Jr. Comptroller of the Currency before the International Monetary Conference June 8, 1999 Philadelphia, Pennsylvania Introduction Few would disagree that the financial world has changed substantially since the Basel Committee on Banking Supervision promulgated the Capital Accord in 1988. Internationally-active banks today are significantly more complex, more driven by technology, and more global in their scope. Clearly, for these more sophisticated banks, a “one-size-fits-all” approach to capital is no longer appropriate. Instead, we must find new ways to reflect credit risks in the capital framework. The Basel Committee has taken the first step in this direction by recently issuing a consultative paper looking toward a revision of the Accord. Today, I will focus my remarks on the issues raised by the 1988 Accord and the options we at the OCC see for the future of the regulatory capital framework, in light of the new Basel Committee proposal. I. Deconstructing the 1988 Accord In retrospect, there is no question that the 1988 Accord represented a significant step forward on several fronts. It has been credited with reducing international disparities in the regulation of capital adequacy, addressing the risks posed by the growth of off-balance sheet instruments, making banks' capital levels more transparent to market participants, and reversing what had been a prolonged decline in the capital levels of internationally active banks. However, as institutions have grown in complexity and have increasingly resorted to the use of sophisticated financial tools such as securitization and complex derivatives to reduce capital requirements, manage risk and allocate credit, the limitations of the Accord have become manifest. Many of the limitations of the current Accord are already well recognized. One fundamental problem is that the current system does not adequately or accurately assess risk. The current “risk bucketing” approach, under which assets are sorted into different buckets based on broad categories of risk, is a crude approach to allocating capital. It has resulted in a poor differentiation of credit risks, given rise to tremendous arbitrage opportunities, and led to distortions in the way that banks allocate credit and price products. Moreover, the current system does not take into account some of the common risk management techniques used by banks today, including diversification and hedging. Although these techniques have helped bankers to better manage risks, the regulatory capital framework has not been adjusted to reflect these improvements. Nor has it been adjusted to take account of the additional risks that stem from concentrations and other broader “portfolio effects.” Finally, the Accord has not kept pace with developments in risk management and capital allocation in the banking industry. It has become increasingly difficult to fit many new product offerings into the existing risk buckets--credit derivatives being a prime example. The inability of the Accord to stay abreast of market innovations has become one of the main reasons the Basel Committee is looking toward a substantial revision of the Accord. Given the problems posed by the current system, our goal and our challenge is to develop a system that is flexible and forward-looking, and that provides for a useful and rational assessment of risk upon which to base a capital charge. II. What a Revised Accord Offers The first step toward meeting that goal was taken last week with the release of the Basel Committee’s consultative paper describing elements for a new capital framework. This paper represents the first step in a two-step process: first, the development of the framework for the future Accord; second, the articulation of the details. In each case, industry comment will be sought, and this is an opportunity that the industry should grasp. Ultimately, the views of the industry will be of critical importance to the framing of any future Accord. The Committee has made it clear that it wants capital rules that reflect not only the risks, but also the realities of banking today. The new proposal will seek to update the 1988 Accord in a number of different areas. First, the proposal would expand the “standardized” or risk bucketing approach. Second, it looks to the use of new approaches to measuring credit risk, such as internal ratings. Third, it raises the issue of what other areas need to be covered by a revised Accord. Among the fundamental changes that the proposal would make is the introduction of the “three pillars” of capital. To date, the Accord has looked primarily at the quantitative aspects of capital. The new Accord will add two new “pillars” as integral parts of the regulatory capital framework: supervisory review and market discipline. The OCC has long believed that supervisory review and market discipline are important elements in the review of capital adequacy. However, this is not a view that is held around the world. Qualitative elements tend to be forgotten when viewed alongside the challenges posed by quantitative measures such as internal ratings and credit risk modeling. While most supervisors have ways to implement and enforce capital adequacy standards, there is now a need to enhance the role that supervision plays in assessing the qualitative aspects of capital and identifying the specific methods by which to do so. Market discipline must also play a role in any capital adequacy framework, as it rewards banks that manage risks effectively and penalizes those whose risk management is less prudent. Greater transparency will improve the market’s ability to make rational judgments about an institution’s risk management and overall soundness. Another change is the proposal for the addition of a capital charge for “other” risks, such as operational risk and interest rate risk. The main issue here is how to arrive at an appropriate capital charge. These “other” risks are not easily quantifiable. Even interest rate risk, while measurable, is not measured in a consistent way among banks or countries. The charge for “other” risks will be the source of much discussion within the Basel Committee, and, I expect, within the banking community, over the coming months. The third major change addresses the most serious shortcoming of the present Accord--the need to make the credit risk measurement criteria more sensitive to actual risk. III. The Future Methodologies Four approaches to credit risk measurement have been suggested: expansion of risk buckets; use of external ratings; use of internal ratings; and portfolio credit risk modeling. Expanded risk bucketing would not, in my view, be a major step forward or an option we should pursue avidly in the future. While it offers some opportunities for refinement, it would perpetuate some of the problems of the present Accord--lumping assets that inevitably have differing risk characteristics into fixed categories, with attendant opportunities for arbitrage. A second approach, the use of external ratings from widely recognized rating agencies, could be applied to both sovereign and corporate credits. However, there are still some issues that must be thought through before this approach could be used. For sovereign credits, there are currently only two risk buckets, one for sovereigns that are members of the Organization for Economic Cooperation and Development; another for those who are not. While there is wide agreement that the present framework for rating sovereigns should be dismantled, the track record of the rating agencies on sovereign credits has proved disappointing. We have seen during the Asian crisis of the past two years that ratings were often lagging indicators of emerging problems. The use of external ratings also presents some problems for corporate credits. The fact is that there are generally very few externally rated credits on the books of most U.S. banks, and outside of the United States the use of external ratings is much less common than in the US. While external ratings may prove useful as a part of a broader package, particularly as they become more common abroad, I do not believe that the external ratings approach alone will go very far in solving the problems of the current Accord. The other two approaches raised in the proposal--internal ratings and, over the longer term, full portfolio credit risk modeling-- offer the greatest promise for the most sophisticated internationally- active banks. But given the current state of the art of these methodologies, it is questionable whether they will be feasible options in the near term. The internal ratings approach is geared toward basing a capital charge on the ratings that banks themselves assign to the credits in their portfolios. Full portfolio credit risk modeling goes a step further, using the internal risk ratings as a starting point and then applying sophisticated modeling techniques to adjust the ratings for “portfolio effects” to arrive at a capital charge for the entire portfolio. Advances are being made in both of these methodologies by many institutions. However, there are still a number of difficulties to work out before either of these approaches can be reliably used. Two major challenges posed by the internal risk ratings approach are the lack of consistency among the internal ratings systems, and the need to “map” internal ratings to a uniform schedule of capital charges. Systems developed by individual banks can differ in a number of very important ways. For example, some institutions may define the credit risk attributable to “default” as the probability that a loan will go bad, while others may go further and derive a loss figure that would result if the credit becomes a problem, the so-called “loss given default.” These inconsistencies compound the problem of translating internal ratings into a generally applicable range of risk weightings. Of course, moral hazard must also be considered in connection with any methodology that attaches significant economic consequences to a bank’s own classification of its risks. Despite the issues raised by the internal ratings approach, it is still far closer to implementation than portfolio credit risk modeling, where challenges are much more difficult to overcome. A recent Basel Committee report highlighted two substantial difficulties with the current state of credit risk modeling--a lack of data and the inability to validate the models. As I noted, this approach has promise, but we are still a number of years away from being able to depend comfortably on credit risk models. The good news is that many of you in this audience are devoting significant resources to the development of systems that can overcome the difficulties that I have just described. I see this process as a continuing, collaborative effort between the public and private sector. IV. Conclusion The effort to amend the Accord in a way that both addresses problems already recognized and takes account of emerging technologies of risk assessment has substantial momentum behind it. The process will take some time to complete, but that is desirable for a number of reasons. As time passes, we will continue to see advances in credit risk measurement methodologies that will allow for a more precise calculation of the risks against which capital should be held. Despite the challenges that must be overcome, the internal ratings approach and, further down the road, portfolio modeling, offer the most promise for the future. At the same time that these new methodologies are being developed, the focus on qualitative approaches to capital, namely supervisory review and market discipline, will sharpen. Together, these elements will allow for a more supportable determination of capital adequacy, while retaining the flexibility to adapt to the changes--in risk management and in products and services offered by the banking community--that are inevitable in the future. Thank you.