Reevaluating Approaches to Regulation
President and Chief Executive Officer
Federal Reserve Bank of Atlanta
Financial Markets Conference
Coral Gables, Florida
February 22, 1996
Good evening! I'm really pleased to welcome each of you to this conference on financial markets. As the new president of the Federal Reserve Bank of Atlanta, I couldn't be happier than to be presiding over this important convocation that brings together practitioners, regulators, and academics to discuss financial markets and related public policy issues. During the past five years, we have attracted more and more of the best minds in finance to this conference--thanks in large part to the efforts of former Atlanta Fed officers Sheila Tschinkel and Curt Hunter, both of whom are with us tonight. I'd like to thank them for their past efforts and also to thank those current officers and staff members of the Atlanta Fed who are continuing the hard work to keep this conference on the cutting edge.
Tonight, I'd like to discuss the regulatory aspect of financial markets. While doing so, I'll evaluate both some old approaches to regulation and some new ones. Although I recognize that the old paradigm of regulation worked well in a stable world, it's no longer completely suitable in today's dynamic world. As financial markets have become marked by rapid change and uncertainty, we can't just update parts of our regulatory approach that are essentially fixed or static--and even capital-based procedures fall into this category. The important point I would like to make is that we--and the "we" here refers not only to bank regulators or to financial regulators more broadly, but to firms in the industry as well--we must open our minds to new ways of achieving our ongoing goal. Together, we must find the best public policy approaches for today's financial markets--approaches that will promote efficiency, flexibility, and innovation while still keeping the overall financial system safe and sound.
It might help to think of regulatory approaches as an evolutionary process on two tracks: one track going from static to dynamic and the other from piecemeal to holistic. At one end of that continuum are the bank supervision practices of several decades ago. That was a time when examiners focused on verifying assets and measuring credit losses at the time of an exam. This snapshot approach worked satisfactorily in a static environment that was marked by a fair degree of certainty. The past 10 years represent the middle range of the continuum. This has been a time when a new set of circumstances has demanded a new way of thinking. We have learned that static measures of risk and capital can't capture the whole panoramic picture in the new era of dynamic financial markets. Therefore, we've moved toward analyzing the risk management process at banks. As we evolve, we're moving fastest toward a more dynamic and goals-oriented approach for those parts of banks that deal in tradable securities. Looking forward and trying to guess where our evolution might take us, I would think that we might learn how to apply this goals-oriented approach to all other bank operations as well. Naturally, I recognize that we don't yet have the capability to apply such an approach to a bank as a whole. But I don't believe that this current limitation should stop us from being open to new ideas in regulation.
BANKS AS A MODEL FOR REGULATION OF RISK
In my remarks tonight, I'll cover three areas: How we used to regulate, how we do it now, and how we might do it in the future. As a prelude to discussing possibilities for the future, let me set the context by recalling the time before the 1970s--a time when banks played a larger part in credit markets than they do today.
In those days, legislation restricted permissible activities to a narrow field and set clear limits on the fees and rates that banks could charge and pay. Legislators, however, didn't go so far toward regulating the regulators. As a result, bank examiners could exercise a certain amount of discretion when reviewing banks' activities. Change in positions and risks was relatively slow. Credit risk predominated. Interest rate risk and market risk were considered to be of secondary importance. Examiners spent their time testing mainly for credit quality and the capital adequacy of the lender. They didn't expect institutions to change quickly. Deposit insurance and Fed policy limited catastrophic losses, at least for banks and their depositors. In short, regulators monitored and exercised their judgment. And if a bank didn't meet regulatory standards, regulators might examine the bank more frequently or make good use of other forms of "supervisory attention" and its associated costs to discipline the bank.
Then came more than a decade of profound and interrelated changes. These included high inflation followed by significant disinflation, some deregulation of depository institutions, increased competition between bank and nonbank financial intermediaries, rapid intellectual and technological innovation, and the increasing internationalization of financial markets. The financial services industry was battered by these changes. The savings and loan crisis, as well as the sharp increase in bank failures and portfolio problems, reflected this battering. All these developments combined to move Congress to put in place more stringent regulations for the financial services industry. Legislation, like parts of the Federal Deposit Insurance Corporation Improvement Act (FDICIA), encouraged a substantially more intense level of supervision and regulation.
HOW REGULATION IS EVOLVING -- THE 1990s AND BEYOND
Banks have evolved in response to market pressures nonetheless. For instance, as this conference highlights, they've dramatically expanded their portfolio of tradable claims. Interest rate risk and credit risk that previously couldn't be unbundled are now routinely unbundled. Even traditionally nontradable assets, such as credit card receivables, are being securitized.
So here we are in the middle of the 1990s with a good deal of overhang from what I regard as overly detailed legislation and regulation. At the same time, we've come to realize that regulators are always going to have a difficult time keeping up with technology and market innovations. One major reason for the inevitability of this lag is that there are so many more resources--and much greater upside gains--available to practitioners as compared with regulators. Fortunately, regulation is already evolving, and it will continue to evolve in response to changes in financial markets. For example, at the Fed, we're increasing the focus on banks' internal risk management.
More generally, the risk-based capital proposal for assessing market risk in trading accounts exemplifies the regulatory evolution. The current internal models approach was adopted late last year for implementation over the next several years. Banks with active trading accounts already have risk management models that calculate value at risk, or VaR. This internal models approach seeks to take advantage of banks' models to determine appropriate capital set-asides. It does so by using a multiple of VaR to determine the level of capital that has been set aside in order to protect against trading losses.
One of the appeals of this method is that it addresses criticisms of regulatory micromanagement. Clearly, relying on banks' own internal models is a giant step--a giant step away from inflexibility, rigidity, and a one-size-fits-all premise. The internal models method is a solid indication of the evolution in regulatory thinking. Where is this evolution headed? Toward more flexible oversight, toward significant self-determination for banks. Yet, at the same time, there will continue to be an important role for supervisors. Examiners will continue to review the effectiveness of banks' risk management on an ongoing basis.
Needless to say, the internal models approach has been a big shift for the Fed (and other regulatory authorities abroad). Nonetheless, I hope we can progress beyond that point. There are a few ideas that have already been talked about that might prove to be useful in the future. One such idea that's been proposed is the precommitment approach. This concept, in some senses, parallels the idea behind the procedures used by the Fed in daylight overdrafts. The Fed's system for handling daylight overdraft risks gives institutions a major say in determining their own debit limits on Fedwire. It also penalizes them for exceeding these limits. The daylight overdraft system has been phased in since the late 1980s, and it seems to be working reasonably well. The precommitment approach attempts to bring this same kind of thinking to market risk in banks' trading accounts.
The precommitment approach is a proposal that would allow banks to specify the maximum expected loss over a given period and to set aside an amount of capital based on this level. Trading losses exceeding this amount would subject the bank to a penalty regime. Regulators wouldn't dictate the modeling methodologies that underlie the precommitment levels. Supervisors would, however, review the models.
This method has a number of potential advantages. The main one is that, properly designed, it would align a bank's incentives with regulatory goals. It would also reduce the importance of regulators' reviewing the accuracy of internal models and VaR calculations. Examiners could shift their focus instead to observing a breach of the precommitment levels. Information about such a breach might be made known to the public in the interests of generating additional market pressure. Of course, the integrity of the bank's underlying risk management model could be periodically scrutinized as part of the examination process. In addition, this precommitment approach would avoid the incentive for banks to engage in so-called gaming of the model, for example, by using models they know tend to underestimate their exposure to risk.
Naturally, there are potential problems as well. It would be necessary to create a penalty that would be strong enough. It would have to be a credible deterrent to keep a bank from under-reporting its risk exposures through the precommitment level. Setting the right penalty isn't a trivial challenge. Penalties would need to be structured to avoid compounding problems for individual banks that are experiencing unexpected losses. In a closely related vein, penalties would also have to be structured to avoid leading, in extreme circumstances, to systemic problems. The feasibility of doing so deserves more research. This topic is currently being studied and discussed in the Federal Reserve System and elsewhere.
THE CHALLENGE AHEAD
So where do we go from here? First, I think we'll need to keep a range of regulatory approaches. Traditional approaches may be suited to situations when risks need to be limited in advance. This is true even though these traditional approaches may impose costs on many financial activities that are important to banks in the 1990s. The Barings case immediately comes to mind as an example of a situation when effective regulatory encouragement of stronger internal controls might have made a difference.
Second, I'm the first to acknowledge that incentives-based approaches have their limits as well. Markets differ in liquidity. Consequently, they also differ in their capabilities to provide independent checks on prices and to allow for trading out of risk exposures. Where reasonably accurate valuation is possible, it seems that regulations could focus on performance measures. Sometimes, though, performance or valuation information is likely to be tenuous. In those cases, procedural regulation might continue to be more appropriate.
On the whole, however, I believe that regulation based on incentives applied in appropriate situations warrants further study and debate. I hold to this belief, in part, because such an approach may help us to transcend the vicious cycle in which we find ourselves. We all know what this cycle is: legislators and regulators write elaborate rules to circumscribe banks' activities. This body of statutes and regulations, in turn, prompts banks to find convoluted and inefficient ways to conduct their business. These legal maneuverings are deemed necessary in order to compete and survive. If regulators can find effective ways to focus on the results of banks' decisions, then banks should have substantially less incentive to find ways around regulations.
Overall, though, as I mentioned at the beginning of my remarks, these incentives-based approaches can now be applied only in a piecemeal fashion to certain parts of financial institutions. But a bank's total risk is not merely the sum of its parts. The next logical evolutionary step of regulation would be toward looking at the financial institution as a whole rather than piece by piece. Along these lines, I hope for a time when perhaps we can move away from an accounting measure of capital ratios to a test of viability based on the riskiness of future earnings. A change like this in our way of thinking would move us from a static point of view to one that is dynamic.
In conclusion, I'm challenged to evolve in my thinking as a regulator rather than to remain static. At the same time, though, I challenge each of you to help to formulate approaches to regulation that will keep our financial markets vibrant yet safe, innovative yet sound. Only by sharing the creativity we all have and by advancing the debate can we develop a new paradigm of risk management that will serve our shared purposes. Thank you for being with us for this conference. I look forward to learning a lot, to being stretched into new ways of thinking, and to coming away with some solid ideas about where we need to go next. I hope you, too, find the information you take away from here to be both useful and provocative.