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

Before the Annual Washington Conference of the Institute of International Bankers
Washington, D.C.
March 2, 1998

Financial Globalization and Efficient Banking Regulation
The highly publicized recent events in the world's financial system have served to make
certain things abundantly clear. In particular, we now have further evidence that there is a
large and growing disparity between the risk management practices of what might be called
the "best practice" financial institutions and those of their competitors around the globe.
This disparity, moreover, runs much more deeply than the weaknesses exposed during the
Asian financial crisis. In that event, bankers were seen to make the kinds of basic mistakes
that have been oft repeated at other times and in other places. For example, loans were made
to government-supported enterprises, either at the behest of the government itself or under
the assumption that official support would be provided if the loans turned bad. At times,
these loans were made to highly leveraged companies whose underlying financial ratios did
not justify the origination of the loans on the terms under which they were made. To add
insult to injury, the offending banks sometimes borrowed in dollars and lent in their home
currencies, without hedging, believing that their government could and would continue to
stabilize exchange rates.
While these practices are troublesome, I am much more concerned with what I believe is
both an exciting and disturbing aspect of the evolution of financial markets. Spurred by
improvements in computer technology and advances in financial theory -- most notably in
option-theoretic models -- new financial products, as well as the markets supporting
traditional banking products, are becoming ever more sophisticated and ever more global in
nature. While financial innovation and globalization can only be applauded for their salutary
impact on market efficiency, they present some difficult problems for market practitioners
and, where the practitioners are regulated entities, their supervisors.
Today, I should like to concentrate on three themes, or principles, related to the evolution of
financial markets: First, there exists a significant and dynamic connection running between
market innovation and market regulation. Financial innovation often occurs in response to
regulation, especially when such regulation does not make economic sense. Conversely, the
evolution of regulation often is spurred by advances in the market. Second, the globalization
of financial markets means that mistakes in risk management made by one or more
significant players in world markets can result in real losses not only to the entity making
the mistake, but also to other participants and to other countries' banking systems. Third, the
economic efficiencies that are potentially associated with financial innovation can be
negated by inefficient banking regulation. Efficient banking regulation, by contrast, not only
provides the background against which financial advances can occur, but also permits
governments to achieve social objectives where otherwise they might not, or might achieve
them only at higher cost.

To demonstrate these three principles, we need discuss only one aspect of banking
regulation, albeit the most important -- namely prudential regulation as currently embodied
within the international capital standard for banks. The Basle Accord of 1988, while it was
critical to reversing the decades-long decline in bank capital ratios, has come under frequent
and strong attack in recent years, both by regulators and those that are regulated. In
particular, there is considerable concern that technological advances and rapid evolution in
financial products are reducing the meaningfulness and effectiveness of the capital
standards, at least for the largest, most sophisticated institutions.
The deficiencies of the Accord are well known, but bear repeating here: First, while
intended to be "risk-based," the formal capital ratio requirements nevertheless lump most
bank risk positions into a single "bucket" corresponding to a rather arbitrary, minimum total
capital requirement of 8 percent against the book value of the position. Second, the capital
rules do not explicitly account for certain risks that may be important, such as operating risk.
Finally, portfolio composition, hedging, and general portfolio management techniques are
explicitly considered only within the market risk requirements for trading account activities,
not for the credit or other risks that dominate within the banking book.
This arbitrary, one-size-fits-all minimum capital ratio has spurred what can only be termed
an avalanche of financial innovations aimed at either evading or taking advantage of the
capital standard. Such regulatory capital arbitrage, as we call it, currently is carried out
primarily via the securitization markets. While securitization may serve useful economic
purposes having nothing to do with regulatory arbitrage, a properly structured securitization
conduit can assist the sponsoring bank in lowering its effective regulatory capital
requirement against a group of assets or other risk positions. In many cases, the
securitization results in the bank retaining essentially all of the risk of the underlying assets,
through the provision of credit enhancements to the conduit, but at lower capital
requirements than if the assets remained on the bank's books. This is accomplished, for
example, by having the conduit "remotely originate" credits, thus allowing the bank to
circumvent recourse capital requirements that apply only to assets sold to the conduit.
Alternatively, the bank can provide indirect credit enhancement to the conduit by, for
example, supplying backup lines of credit to the obligors that use the conduit to raise funds.
To a significant degree, the growth in securitization and other forms of regulatory arbitrage
has been spurred by the inadequacies of the international capital standard. This has occurred
largely because, over the last decade, many of the larger banks have developed fairly
sophisticated internal models for formally quantifying risk, including credit risk within the
banking book. These models are used to calculate internal economic capital allocations for
various sub-portfolios of the bank, and it is because these internal capital allocations often
differ substantially from the 8 percent regulatory standard that the problems arise.
In the typical case, the bank attempts to formally measure each major type of risk associated
with a product or business line -- credit risk, market risk, and operating risk. In the credit
risk arena, for example, risk is measured as the estimated shape of a loss probability
distribution over a particular horizon, generally one year. Economic, as opposed to
regulatory, capital is then allocated against this loss distribution in an amount necessary to
meet some corporate goal for insolvency probability. For example, several large banks
allocate enough capital internally for credit risk so as to reduce to 0.03 percent the
probability that credit losses will exceed allocated capital. Why is this 3 basis point standard
chosen? Because that is the historical average default probability, over a one-year horizon,

for double-A rated corporate instruments. In other words, the banking firm wants to hold
enough capital so that the chances of it becoming insolvent are low enough to win a doubleA rating on the bank's own liabilities.
The problem is that, when these economic capital calculations are made, they result in a
very wide range of internal capital allocations for individual positions or subportfolios -- as
low as several basis points up to more than 30 percent of the carrying value of the risk
position. When a group of loans is assigned an internal capital requirement that is very low
compared with the 8 percent regulatory standard, the bank has a strong incentive to
restructure the positions to allow them to be reclassified into a lower regulatory risk
category, by using securitization or other devices. If the bank doesn't do this, it cannot make
a market rate of return on the regulatory capital of 8 percent on the loans.
Regulatory arbitrage, from the perspective of proper resource allocation, can be a good
thing. If there were no way for the bank to avoid the uneconomically high regulatory
requirement, it would need eventually to exit its low risk businesses because of insufficient
returns to equity. In the long run, this would serve no purpose other than causing the
regulated entity to shrink in size relative to its unregulated competitor. At the extreme, the
one-size-fits-all capital standard, if there were no arbitrage safety valve, would cause the
bank to engage in only those activities for which the economic capital requirement is
greater than the 8 percent regulatory standard. That is, the regulatory standard would induce
risk-taking -- perhaps excessive risk-taking
While regulatory arbitrage can be useful in negating improperly high regulatory capital
requirements, it can also be used to mask the true riskiness of the bank. In the United States,
for example, the top 50 bank holding companies have a mean total risk-based capital ratio of
12.1 percent. The standard deviation of this ratio across the 50 institutions is only 0.8
percent. In other words, everyone seems to be holding about the same amount of capital.
Indeed, since a bank is declared to be "well-capitalized" when its total risk-based capital
ratio is over 10 percent, it is not surprising that we see no top-50 banking company with its
ratio less than 10 percent. But do all these banks have equally low insolvency probabilities?
One simply can not tell much of anything by looking at capital ratios. It is perfectly possible
that a bank may hold 12 percent capital when a more carefully constructed internal risk
model would call for holding 15 percent, or even 18 percent, capital to meet the bank's
internal insolvency standard. Or, the bank could have a great model, but simply have a
preference for risk that is unacceptable to regulators. Such a bank may be holding risky
positions for which even its own model would call for more capital, if the bank were to
adhere to a lower insolvency probability standard. For such a bank, the regulatory "wellcapitalized" designation may provide little comfort to supervisors or to the taxpayers we are
supposed to protect. That is why, in this country, we have placed a great emphasis on the
bank-by-bank supervisory process, as opposed to the formal capital regulations that apply to
all banks.
Just as the most sophisticated large banks have gone through a rapid evolution of their risk
measurement, management, and pricing systems, so must supervisors follow suit. At the
Federal Reserve we have ongoing projects aimed at providing supervisors with better tools
to assess banks' internal risk systems and, ultimately, to make determinations regarding the
real adequacy of bank capital on a case-by-case basis. Among these efforts is a review of the
credit risk aspects of asset securitization at our major banking companies. Also, we are
studying the possible uses within the supervisory process for the internal rating systems used

by almost all large banks. In the past, supervisors made risk distinctions only among and
between classified assets, not pass assets. Now, we are studying the possibility that future
deterioration in asset quality can be foreseen to some extent by changes in the average
rating, or the distribution of ratings, in a bank's pass assets.
We are also spending considerable effort in tracking and understanding the developments in
risk modeling, including the modeling of credit risk. At last week's conference on bank
capital, hosted by the Federal Reserve, the Bank of England, and the Bank of Japan, our
economists discussed the prospects of moving to a full-fledged, models-based approach to
bank capital standards for the largest banks. In my personal view, moving from a ratio-based
capital standard to an internal models based standard for our most complex institutions,
should be high on our agenda. For the first time, we would be setting a maximum
insolvency probability standard rather than simply a minimum capital ratio. This may be the
only avenue before us if we wish to achieve an efficient regulatory system. In the absence of
a thorough revamping of the international capital standard, we will continue to be plagued
by regulatory capital ratios that, on the one hand, say little about insolvency probability,
while on the other hand induce banks to engage in sometimes inefficient regulatory arbitrage
simply to avoid an inherently uneconomic capital rule.
Please do not misinterpret my remarks. I believe, like almost all risk practitioners, that there
is no substitute for good human judgment and experience when making credit decisions.
Total reliance on models is neither feasible nor desirable. But failure to use the best possible
tools at hand is to fall further and further behind the best-practice techniques of the industry,
with a resulting decline in risk-adjusted profitability and, inevitably, an increase in
insolvency probability. We must remember that any improvement in the accuracy with
which risk is measured is tantamount to a reduction in risk. Continual improvements in risk
measurement techniques, therefore, should be the norm for all banks that intend to play in
the global financial marketplace. Institutions, and entire banking systems, that do not adhere
to this principle are doomed to repeat the blunders of the past.
In this new world of financial complexity and globalization, it is extremely important that
the large institutions among the developed nations all strive to keep up with the bestpractice frontier. These institutions are the ones that are the price-leaders, the drivers of
markets locally and internationally. If a group of important institutions in only one or two
countries fails to keep pace with risk measurement practices, all banking systems are placed
at risk. This risk, moreover, is not simply that a large bank failure in one country can cause
counterparty failures in other countries. Systematic under-pricing of credit and other risks
can be damaging to all players, not only to the bank making the pricing errors.
Fortunately, the free-market mechanism for the dissemination of best-practices appears to be
functioning reasonably smoothly, at least in the global sense. No single developed nation
appears to have a monopoly on best-practice risk measurement techniques, if innovations in
complex financial products are any indication. For example, European banks were market
leaders in introducing CLOs, or collateralized loan obligations. In the field of asset-backed
commercial paper facilities, U.S. banks were the initiators, but now European and Japanese
institutions are significant players. And the ubiquitous consulting firms around the world
can be relied upon to spread the word of worthwhile advances in risk techniques.
Still, the individual bank in each country must face the proper incentives to keep up with the
most cost-effective risk techniques. Lax supervisory practices -- or, worse, government

support of banks with poor risk practices -- do not provide these proper incentives. Thus,
each supervisory authority in each developed nation must be ever vigilant that the disparity
between the world's best-practice institutions and those large banks that are "inside" the
best-practice frontier does not grow wider. Indeed, an important function of supervisors is to
act as something of a clearinghouse for best practices. If the supervisor perceives a
deficiency in practice, it is his responsibility to engage the bank manager in a discussion as
to whether the shortcoming really exists and, if so, how to fix it.
I will conclude by reiterating the three points I made at the beginning. First, there is a strong
and dynamic thread running between regulation and market innovation. While the Basle
Accord of 1988 was entirely appropriate to its time and circumstances, it is now clearly in
danger of becoming outmoded by the pace of financial innovation. Conversely, the
regulation has contributed to some market innovations that appear to be driven, if not solely,
at least primarily by the need to engage in regulatory capital arbitrage.
Second, the reality of globalization must be accounted for in designing and implementing
our regulatory and supervisory systems. Especially among developed nations, we cannot
afford a growing disparity in the quality of risk practices at our important institutions, nor a
disparity in the quality of supervision of those institutions. As bankers like to say, the worst
competitor is an uninformed competitor -- and that goes doubly for the competitor's
supervisor.
Third, given that financial markets are constantly evolving, this means that our regulatory
framework must also continually evolve. The international capital standard has not changed
in basic form for almost a decade -- it is still a ratio-based rule. While it may still be
adequate for the vast bulk of banking institutions, it clearly is inadequate for the world's
most complex banks. For these institutions, high capital ratios do not necessarily equate with
low insolvency probabilities. Thus, the ratio-based standard is inefficient in achieving the
supervisors' objective of limiting bank failure to acceptable levels. Worse, it may be
fostering other inefficiencies in the banking system, to the extent the capital standard
encourages regulatory arbitrage that entails significant transaction costs.
In the absence of any viable alternatives, it is my view that we should begin now to plan for
a models-based successor to the Accord. Inevitably, this will take a tremendous effort, given
the complexity of the subject and the differences across institutions and between countries.
Moreover, a models-based system of capital regulation would require a degree of
cooperation among supervisors, quite apart from having similar written rules, that is
unprecedented. But I believe the effort will be worth it.
Thank you for the opportunity to air these concerns and I am looking forward to continuing
the dialogue on this subject.
Return to top
1998 Speeches
Home | News and events
Accessibility | Contact Us
Last update: March 2, 1998, 3:00 PM