View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Remarks by Governor Laurence H. Meyer

At the Risk Management Association’s Conference on Capital Management,
Washington, D.C.
May 17, 2001

The New Basel Accord: Challenges for Banks and Their Supervisors
It is a particular pleasure for me to join you this morning. The Risk Management
Association has been a significant source of expertise and insight for the banking agencies.
The RMA’s comments on regulatory and supervisory proposals have uniformly been well
thought out and responsible. Your meetings with staff and your cooperation on special
surveys to help flesh out proposals have been exemplary, if not essential, especially on
recent efforts to modify the Basel Capital Accord. I should, of course, stop heaping praise
on the RMA lest its bank membership begin to believe that the organization shares the
blame with us for any regulation or proposal that is not to their liking. Nonetheless, thank
you for your help, and I know we can continue to rely on you for both candor and empirical
insights, insights along the lines of your conference theme: the enterprise perspective on risk
I suspect that the recently proposed changes to the Basel accord will affect your working
lives rather dramatically. The proposals touch upon many aspects of banking and many
different kinds of risk, but much of the focus is on credit risk, and RMA’s roots are firmly in
credit risk. Many of you will be key players in your organization’s response to, and
comments on, the new accord--Basel II. As background to your efforts, I would like to
review how we got where we are today and offer some views of the way forward.
Basel II is a response to changes flowing from the increasing sophistication of risk
measurement and management, but the shaping of the response has motivations that have
become more multifaceted over time. It is fair to say that the supervisory community
initially was motivated to move toward Basel II by the erosion of the Basel I rules through
capital arbitrage. Simply put, “capital arbitrage” refers to strategies that reduce a bank’s
regulatory capital requirements without a commensurate reduction in the bank’s risk
exposure. One example of such arbitrage is the sale or other shift off of the balance sheet of
assets with economic capital allocations below regulatory capital requirements and the
retention of those for which regulatory requirements are less than the economic capital
burden. Aggregate regulatory capital thus ends up being lower than the economic risks
require; and although regulatory capital ratios rise, they are, in effect, merely meaningless
statistical artifacts. The fear of capital arbitrage got us moving, and the new risk
measurement technologies offered a way to make capital regulation more risk-sensitive.
Thus, a significant aspect of Basel II is mending what has been broken. But, as we began to
shape the new accord, we became ever more convinced that encouraging the adoption of
advanced risk management is enormously desirable in its own right. Internally, good risk
measurement is crucial to the control of risk postures and to choices about capitalization.

Until the new technologies came along, and still today at many banks, choices about
portfolio risk were the outcome of internal debates and struggles between those wishing to
push the boundaries of prudence in search of short-term volume and those who saw the risks
clearly. Traditionally, the right balance was not always achieved because it was hard to
distinguish the boundary between reasonable and excessive risk. This is unfortunate because
a major mistake in just one business line, such as commercial real estate, can be fatal. The
new risk-management technologies offer the promise of a common system of risk
measurement that will support more rational, efficient, reliable internal control systems and
choices about risk.
Moreover, under Basel II, the practices that a strong and rigorous bank management would
use on its own will be used to guide regulatory minimums. Basel II seeks to eliminate the
practice of keeping, as it were, two sets of books, by requiring that the parameters used to
determine regulatory capital be the same as those that management uses to run the bank. To
be sure, the supervisors will have to be comfortable with the internal ratings systems and
other techniques used by banks to measure risk exposures and potential losses. But once
verified--and periodically tested--bank managers and regulators will be using the same
concepts and the same techniques. And as the measuring techniques improve, many of them
may be incorporated in bank management practices--through improved estimates of loss
characteristics--without changing the regulatory structure. Change and innovation will
improve the process, rather than undermine it.
Improved risk measurement also promises better-informed investors and greater discipline.
If good risk measurement can improve internal discipline, then disclosing internal risk
measures will help the market to understand banks’ risk postures and to react rationally. It is
important to understand that improved market discipline does not mean uniformly harsher
treatment of banks by the market. Bank stocks tend to trade at lower multiples of earnings
than the equities of many other industries, and one conventional explanation for this is that
banks are quite opaque. It has become increasingly clear that investors cannot easily know
the level of risk they are acquiring with any particular bank equity position. When it comes
to assessing share values, uncertainty can often breed doubt: It seems reasonable to expect,
as a long-run proposition, that more-accurate market valuations of bank equities will follow
better disclosure.
The three pillars discussed in the Basel consultative documents flow directly from this
multifaceted view, and the rank-ordering of their importance in the long run is not indicated
by the number of pages devoted to each in the consultative documents. In the long run, the
supervision pillar remains critical, but the supervision must be even more risk-focused and
increasingly concerned with validating systems. Line supervisors will evaluate the quality of
risk management and examine the adequacy of the risk measures. Without good supervision,
the other pillars cannot stand. For example, accurate internal risk-based (IRB) inputs are
surely crucial to obtaining reasonably accurate regulatory measures of capital adequacy.
And the market will not believe or use risk disclosures unless it believes that the underlying
risk measures, like ratings and the probabilities of default, have been validated. Thus,
supervisors must validate the risk measures to support both capital regulation and market
Market discipline will become increasingly important. Given informative and comparable
disclosures of internal risk measures, the market will react more quickly and appropriately
than any regulator to variations in risk postures, and such responses will help banks strike

the right balance between risk and reward.
Formal capital regulation is the most familiar of the three pillars, and the one to which the
most pages are devoted--no doubt because it, by definition, contains more “rules” that must
be explained. Capital ratios will serve as a trigger for corrective actions for banks that get
into trouble. Projecting ahead many years, better risk management may mean that troubled
banks will be far fewer than they have been historically.
Ideally, the three pillars will work in an integrated way to strengthen banking systems.
Nonetheless, during the transition to a working Basel II, we may have to lean on capital
regulation more than we hope to in the long run. In navigating that transition, our
overarching goal should be to advance the practice of risk management and encourage its
wider adoption. And we all should accept that a lot of work remains both for banks and for
supervisors. Let me say that again for emphasis: Our goal should be to advance the practice
of risk management and encourage its wider adoption, and we have a lot of work to do.
Partly in pursuit of that goal, the qualification standards have been set very high for the
Advanced IRB approach and, in truth although perhaps not in public perception, for the
Foundation IRB approach as well. The standards are an amalgamation of the best practices
of many banks. No one bank follows all of them and, hence, at this moment, no bank meets
all the IRB standards, especially for the advanced approach. Let me also say that again: At
this moment and with current systems, no bank in the United States likely would qualify to
use the Advanced IRB approach.
Two considerations led the regulatory community to set standards at the cutting edge. First,
the IRB approach gives primacy to internal risk measures in setting regulatory capital
requirements. The regulators need to be comfortable that the risk measures are reasonably
accurate and consistent, and the draft standards represent our perception of the practices that
will make us comfortable. The standards are tough because the demands on the risk
measures and the incentive pressures on the underlying systems will be large. Moreover, to
the extent we are comfortable that a bank meets the standards, other aspects of supervision
can be less intrusive. A comfortable supervisor is a less intrusive supervisor!
Second, we believe the standards represent precepts for good risk management, worthy in
their own right even in the absence of an IRB approach to regulatory capital. By making
them a requirement for the IRB approaches, we hope to encourage their wider adoption.
It is our hope and expectation that the large, complex banking organizations (LCBOs) will
continue to enhance their risk management practices so that they might be prepared to adopt
the advanced IRB approach. Other banks that can meet the standards are also welcome to
use the advanced approach. But we do not plan to relax the standards to ensure that all the
LCBOs qualify. LCBOs that believe they will get a free pass into Advanced IRB simply by
virtue of being large and by promising to someday improve will be, I think, disappointed.
Of course, there is a tension between setting high standards and also expecting wide
adoption of the advanced approach by LCBOs. Is there a danger that the U.S. banking
industry will simply stick with the standardized approach and turn a cold shoulder to IRB?
Perhaps, but I believe that the market will help here. If any LCBO adopts the advanced
approach--and I believe some intend to--it seems likely the market will pressure all the large
banks to adopt it or be judged as having something to hide or having standards not quite up

to snuff.
There will also be direct economic incentives to adopt IRB approaches. The effective
average risk weight for a bank as a whole should decline with the more sophisticated
approaches depending on the extent of capital arbitrage already accomplished. If true, such
banks would achieve lower total regulatory capital charges and, consequently, a higher
reported risk-weighted capital ratio. At the same time, given the different risk profiles at
individual banks, capital requirements almost certainly would vary more widely under the
new risk-based capital ratios than under today's measure. However, banks would then
presumably respond to changes in their risk-based capital ratios.
For example, a bank with a relatively low risk portfolio would find that its risk-weighted
capital ratio increased because its risk-weighted exposures had declined. It would, as a
result, presumably reduce its capital, or increase its leverage, or even increase its risk
exposure. A bank with a higher level of risk-weighted exposures resulting in diminished
risk-based capital ratios would presumably do the opposite: raise more capital, or reduce its
leverage, or reduce its risk exposures. At the end of this adjustment process, we might
again--for either competitive reasons or because of the incentives from the promptcorrective-action structure--have a relatively tight configuration of risk-weighted capital
ratios, especially at LCBOs, but riskier banks would be holding more absolute capital to
support their risks. Indeed, that is the whole purpose of the exercise.
In contemplating your own bank’s effort to prepare for adoption of the IRB approach, I
would suggest that a pivotal step is warehousing your own institution’s credit-loss
experience. To be candid, the scarcity of data on losses has been a major stumbling block in
the development of the IRB approach. Let me urge all of you to follow the example of some
of you and move quickly in this direction. Data storage is extraordinarily cheap these days,
but we all know that extracting and organizing the right data from existing systems is very
expensive and often painful. Nonetheless, the contribution of better data to sound risk
management will undoubtedly prove to be extremely valuable.
As you develop your own plans, let me emphasize that supervisors will be looking for a
committed and integrated response by IRB banks. Focusing IRB efforts in a compliance unit
that seeks to meet the letter but not the spirit of the IRB standards will not be welcome. The
details of what we will expect are, of course, still being developed. Based on our discussions
with banks thus far, areas of special attention may include the breadth and depth of internal
audit and loan review and the consistency and timeliness of internal ratings. The overall
supervisory review process is likely to be familiar: verification of internal processes and
procedures, coupled with some degree of transaction testing. The overarching goal will be to
assess the effectiveness of a bank’s processes and systems in achieving sound risk
management and adherence to IRB standards rather than to dictate the particular form that
internal checks and balances must take.
Supervisors will also carefully review the manner in which loss characteristics are estimated
and the estimates themselves--such as the probabilities of default, the losses given default,
and the exposures at default--that are provided by IRB banks. This is an area where
supervisors have not tread very often and with many open issues. The issues include the
suitability of internal and external data sources, the appropriate historical period to use in
producing estimates, and the manner in which historical data and loss estimates should be
linked to the bank’s current exposures. These and other issues will be challenging and will

rightly occupy much of our attention over the coming months and years. In this context, the
industry and supervisors must work together to develop and use reasonable standards for
IRB quantification. Neither we nor banks can afford a “race to the bottom” in this critical
area. The expertise of the RMA will be particularly sought and appreciated.
Supervisors obviously have a lot of work to do. We are at the very beginning of efforts to
train our staff to understand all of the issues raised by Basel II and to make sound judgments
about banks’ adherence to the standards. I hope you will be patient with our examiners as
they come up to speed, just as we intend to be patient with you. And you should not
interpret my earlier remarks as a signal that we intend to be very inflexible. We recognize
that many systems, methods, and organizational forms can satisfy the spirit of the IRB
standard, and that what is appropriate for one bank may not be appropriate for another. We
expect and welcome variety, in part because that is necessary to support further innovation.
And innovation is absolutely necessary. Indeed, many technical issues are very unsettled for
all of us and must be worked out during the next few years so that standards can be applied
with reasonable consistency.
I know that many of you are especially concerned about the disclosure requirements because
of the danger that the market may be misled by inconsistencies across banks in the basis of
the numbers that are disclosed. Partly, it is the supervisor’s job to promote consistency by
ensuring that the IRB inputs are comparable across banks. To ensure this, we all must
strengthen our current understanding. I hope banks and regulators can work together on
these problems, as we have so often in the past.
The Office of the Comptroller of the Currency and the Federal Reserve have begun a series
of pilot reviews of a few banks to get a more up-to-date understanding of how practices and
standards are changing and how they relate to the proposals in train. Based on those reviews
and banker responses, we then must develop examination practices, not to mention retrain
and develop staff.
And of course the regulators must finish the proposal. We must carefully review your
comments, evaluations, and suggestions for improvements to the Bank for International
Settlements’ consultative papers, and then make the necessary modifications. The proposals
in several areas--such as credit cards and commercial mortgage finance--still require
fleshing out.
Basel II as a framework and a concept is nearly finished, but its application will be in a
constant state of flux--from the inception onward--as both banks and supervisors improve
their procedures and the applications of those procedures. Basel II will spread those changes
more rapidly, to your benefit, to our benefit, and to the benefit of our economy. But, for
present purposes, I would be remiss if I did not underline that some of the details are still
being developed.
2001 Speeches
Home | News and events
To comment on this site, please fill out our feedback form.
Last update: May 17, 2001, 10:00 AM