Editor's note: This article is based on remarks made at a symposium
sponsored by the Federal Deposit Insurance Corp. on "The Future
of Financial Regulation: Structural Reform or Status Quo?" March
13, 2003, Washington, D.C.
Observers have long been troubled by the "excessive number"
of government agencies that supervise and regulate banking organizations.
Analysts have also expressed concern about an "artificial division"
between the supervision and regulation of banking organizations and
other financial intermediaries such as insurance and securities firms.
Plans to reduce the number of bank supervisors (or all the agencies
that govern financial intermediation) through consolidation go under
the rubric of regulatory restructuring. Discussions of restructuring
were reenergized in 2002 with some senior officials at the Treasury
Department and the Federal Deposit Insurance Corp. offering their
observations on and suggestions for regulatory restructuring.1
As these officials recognize, the numerous restructuring proposals
put forward over the years have failed to pick up momentum despite
extensive study and discussion. Given this lack of momentum, I think
it productive to consider methods of achieving the objectives of regulatory
restructuring without going through the process itself. Specifically,
I will suggest several potentially significant improvements to regulation
and supervision that do not require regulatory restructuring, but
stay true to the purposes behind proposed reforms. (Of course, these
are my views and not the views of the Federal Reserve System.)
In these comments, I first suggest why regulatory restructuring has
failed in the past, then briefly discuss what I see as the underlying
objectives of regulatory reform and finally describe some options
for achieving these objectives without venturing into the morass of
The failure of regulatory restructuring
Justifications to restructure the supervision and regulation of
financial institutions almost always include a flow chart depicting
the multiple government bureaucracies that interact with firms engaged,
at the broadest level, in the same types of activities. Such pictures
purport to demonstrate duplication of effort on the one hand and inefficient
division of responsibilities on the other. Whatever the optimal organization
of supervisors and regulators should be, one would be hard-pressed
to argue that the current systemwith its apparent cost inefficienciesis
one that we would build from scratch.
Given this negative assessment of the current arrangement, why do
efforts at restructuringincluding those reforms aimed directly
at cost reductionfail to achieve appreciable momentum? I see
three possible explanations:
First, some observers see nonpecuniary benefits in the current system.
They argue that creating multiple agencies to supervise banks creates
competition among those agencies and encourages them to innovate and
adapt. Proponents of this view believe the benefits of this innovation
exceed the costs of potentially inefficient operations.
Second, any proposal to alter bureaucratic responsibilities naturally
raises concerns about "turf." The many agencies potentially
affected by restructuring create a natural opposition to reform. In
contrast, the public is indifferent to the current structure and changes
to it. The political calculus is therefore not favorably disposed
Third, the benefits of regulatory reform could be significantly smaller
in practice than in theory. Moreover, these benefits could also seem
small relative to other reforms that could be pursued without encountering
the hassle of restructuring. The premise that restructuring would
clearly pass a benefit/cost test is perhaps weaker than "Rube
Goldberg" diagrams of banking supervision suggest.
Intuition, along with a process of elimination, leads me to believe
that this last explanation does the best job of explaining why serious
regulatory reform stalls. Supervisors have tried to innovate, but
it is far from clear that such attempts have kept up with the industry.
I also seriously doubt that the existence of multiple supervisors
explains the innovation that has occurred. From what I can see, supervisory
agenciesindividually and collectivelyinnovate when their
practices fall sufficiently behind the industry. And while turf certainly
matters, there are many counterexamplessuch as military base
closings and the dismantling of much economic regulationwhere
such resistance has been overcome.
In contrast, the behavior of the supervisors themselves suggests that
the benefits of restructuring might be overstated. When needed, the
agencies turn inward to achieve cost savings through traditional means
such as job reductions and regional office closings. When the agencies
do work collectively, they tend to do so on major policy issues (for
example, Basel II) rather than on reductions in back-office costs.
Such cost reductions presumably could be achieved simply by contracting
out with each other.
If the cost reduction benefits of restructuring or consolidation per
se are not high, one might still find objectives associated with the
effort that are worthwhile. In fact, such worthy objectives exist
and we now turn to them.
Objectives of regulatory reform
As I see it, there are two underlying objectives of regulatory reform
beyond the cost-saving rationale just discussed. The first is to improve
the performance of banking supervision and regulation. More specifically,
supporters of restructuring hope to make supervisors more sensitive
and responsive to bank risk taking. They hope restructuring allows
supervisors to identify risk more accurately and respond to it more
The second underlying objective of restructuring is to improve economic
efficiency. Restructuring could improve the allocation of society's
resources such that net benefits are maximized. Of course, careful
attention to duplication (and other) issues that may raise costs unnecessarily
is part of this objective, but the goal is much broader in scope than
simple cost efficiency.
If this is a reasonable characterization of the underlying objectives
of regulatory restructuring, are there meaningful steps to achieve
them short of consolidation? I would suggest there are, and they include:
Focus on systemic risk;
Enhance use of market data;
Move to accrual budgeting; and
Review the benefits and costs of regulation.
Achieving the objectives of reform
The reforms discussed below explicitly do not require a consolidation
of bank supervisory agencies. Nonetheless, I think their focus and potential
results capture the reasons why restructuring remains a topic of enduring
Leverage expertise. Some insured depositories have concentrated
their activities in what a decade or two ago would have been seen as "nontraditional"
businesses. These activities include securities underwriting and trading,
processing of payments and securities transactions, lending to consumers
and especially to borrowers with blemished credit histories. Not only
might the concentration of activity increase the riskiness of these banks,
but also these activities could be particularly risky in their own right.
Some might view these institutions as posing an outsized risk to the deposit
insurance fund. The number of banks with such high-risk specialties is
small, but the potential for imposing loss is high, and it may also be
that such specialties make effective supervision of these entities more
challenging than that of "plain vanilla" organizations.
There are many regulatory approaches to this issue, including higher capital
standards and limits on growth and, clearly, the supervisory approach
to these institutions will also be important. A lynchpin of supervisory
strategy, in my view, should be to charge examiners with appropriate specialized
skills to review these potentially high-risk banks. Having each supervisory
agency develop this human capital separately could be ineffective. A cross-agency
team of examiners focused on these specialized institutions should lead
to staff with the desired backgrounds, significant experience and adequate
To be sure, there are recent examples of effective sharing of resources
in examination of specialized entities and in examination of and supervisory
action against large banking organizations. But we are proposing going
further, and having supervisory agencies rely on pooled expertise rather
than building their own human capital whenever the situation demands.
This approach also has the potential to help the supervisory agencies
react more swiftly to new risks at specialized institutions and provide
a more consistent response to them.
Focus on systemic risk. The discussion of specialized institutions
makes the obvious point that risk is not evenly distributed across banks.
In the same vein, the risk that the failure of one bank will lead to the
failure of other institutions or otherwise lead to reductions in real
economic output is not relevant for most institutions. The largest banks,
or banks focused on payments system processing, in general pose systemic
risks while other banks do not. Given that the prevention of systemic
risk is one of the primary justifications for bank regulation and supervision,
policymakers should ensure that resources are allocated disproportionately
to those banks that raise this threat.
Despite some shifting of resources to more systemically important banks,
it is my impression that supervisory resources are still largely allocated
according to the distribution of banks rather than banking assets. Because
there are many more small than large banks, supervisors end up using the
bulk of their resources to supervise banks that no one could claim pose
a systemic threat. The failure of a small bankwhile potentially
costly to employees, some creditors and certain borrowersis unlikely
to have ramifications for large economic units like states or countries.
This consideration suggests that supervisory resources could be reallocated
to better match the threat to the real economy.
The Federal Reserve System provides an example of potential misallocation
of resources. We devote about 75 percent of our total domestic institution
supervision hours to institutions with under $10 billion in assets. Yet
these banks hold only 30 percent of aggregate banking assets, and it seems
unlikely that such relatively small institutions pose any great risk to
the system as a whole.
While supervisors might not know precisely which banks are systemically
important, they have a pretty good idea and can allocate resources accordingly.
But identifying the systemically important banks is only one step. Presumably,
supervisors should want to focus resources on those important banks which
are in trouble, or likely to become so. This is not an easy task, but
it leads me to some additional steps that supervisors can take to better
assess bank risk taking.
Enhance use of market data. We believe that market data can be
a valuable and low-cost complement or adjunct to standard bank examination
data in the assessment of bank risk taking. We have detailed this position
elsewhere, so I will only touch on the major issues here.2
Market data can be valuable in assessing and responding to bank risk taking
because they have many traits that the financial data traditionally used
by bank examiners do not. Specifically, market data are available on a
timely basis, are forward looking, can serve to confirm (or alternatively
to raise cautions) about supervisory judgments and can be used effectively
to communicate examiner findings to bank management and to boards of directors.
To be sure, incorporation of market data into supervisory assessments
is no panacea. For several reasons, including explicit and implicit government
support for bank creditors, market signals are not all they could be.
Moreover, there will certainly be a learning curve in determining the
best ways to use market data in the supervisory process. Nevertheless,
the scholarship suggests that market data add value even for examiners
who have access to private, internal information about banks.
Obviously, increased utilization of market data does not require regulatory
reform. But if one bank regulatory agency pushes ahead with it, I would
expect that competition would bring others along as well. There has already
been a bit of interagency work on this topic with staff at the Federal
Deposit Insurance Corp. and the Federal Reserve producing some of the
scholarship to which I've just referred.
Accrual budgeting. Another constructive proposal that does not
require regulatory reform but could improve supervisory effectiveness
pertains to the federal budget and the signals and incentives it provides.
Currently, the budget records the cash flows associated with deposit insurance
in order to measure the exposure of taxpayers to bank failures. When current
cash flows into the deposit insurance fund (premiums and investment income)
exceed the cash outflows (funds used to resolve bank failures), policymakers
receive a financial signal from the budget that everything is okay.
However, these cash flows might well have little connection to the current
or future financial condition of the banking industry. Instead, they largely
reflect previous bank actions. Thus, at precisely the time that premiums
exceed resolution costs, banks could be taking on significant risk that
will result in future losses to taxpayers. Under current practices, when
policymakers and supervisors should in fact be taking steps to restrict
bank risk taking, the financial signal they'll be seeingbased on
cash flowswill be pointing in the wrong direction.
An accrual budgeting system, in contrast, would rely on estimates of the
long-run exposure of the government (taxpayer) to bank failures. These
estimates would form the basis for current budgetary charges. When it
appears that future costs are increasing, policymakers would have to add
charges to their budgets today. Supervisory agencies and lawmakers would
thereby gain a heightened sense of the need to respond to risk taking
in the current period by enacting changes in law, policy or supervisory
strategy. To the degree that such changes reduce expected future costs,
agencies and officials would avoid additional current charges.
While this proposal might appear arcane, it is worth noting that Federal
Reserve Board Chairman Greenspan has recently made identical points with
regard to other federal programs.3
Officials at the Office of Management and Budget engaged in federal budgeting
on a full-time basis have called for similar budgeting practices in the
past.4 While such a system
poses technical and accounting issues, they seem surmountable with existing
technology, posing no more of a challenge than that which exists in the
current system and offering a more accurate assessment of current and
future conditions. Again, one should note that these benefits occur across
supervisory agencies and do not require changes to structure.
The rules, not the rulemakers. Another commonly voiced concern
about the current regulatory structure, with its multiplicity of regulators,
is its potential to limit competition in banking and in financial services
more broadly. Each regulator might issue rules that increase the safety
and soundness of the firms it governs but which, for example, make entry
of new service providers more costly. If this actually occurs, it means
the financial system is less efficient than it could be, and hence customers
are not as well served as they should be.
An alternative, although not mutually exclusive, view is that such inefficiencies
are minor relative to the costs of rules that have little if any effect
on the ability of firms to effectively compete. Moreover, these costly
rules could very well find support whatever the regulatory structure.
I find this alternative view compelling, and I would guess that the source
of inefficiency is likely in the number and details of the rules already
in place and not in the number of rulemakers. And while I admit my argument
relies on assertion rather than on empirical evidence, relatively rigorous
evaluation suggests the costs of regulation are high. Analysis by Board
of Governors staff puts the cost of activities undertaken by banks solely
because of regulation at about 6.5 percent of aggregate noninterest expense
(which amounted to $7.5 billion for the year  studied and would
amount to around $13 billion today).5
We already have a forum to take on excessively costly regulation, namely,
the Federal Financial Institutions Examination Council (FFIEC). Among
other duties, the FFIEC coordinates the release of new regulations. Instead
of looking to the FFIEC to issue new regulations, perhaps it should be
charged with sponsoring research to assess the benefits and costs of existing
regulations. The results of such research could then be provided to appropriate
This commentary discussed some suggestions, and their associated rationales,
to accomplish the objectives of regulatory reform while avoiding at least
some of the aggravation associated with wholesale restructuring. Essentially,
there are two objectives: Improve the effectiveness of supervision and
reduce the economic inefficiency resulting from the current bank regulatory
system. In this vein, I have made five proposals:
Focus on systemic issues;
Utilize market data;
Introduce accrual budgeting; and
Assess costs and benefits of regulations.
The first three recommendations could help address both objectives, namely,
effectiveness and efficiency, at least to some degree. The fourth recommendation,
the introduction of accrual budgeting, should also improve supervisory
effectiveness to the extent that it heightens decision-maker sensitivity
to the potential costs of bank failures. My final recommendation, on benefit/cost
analysis, clearly goes to the issue of economic efficiency, and it may
also increase effectiveness if it leads to improved resource deployment
and better supervision.
Of course, this list of recommendations is selective, and partial. Others
might choose a different set of tools to achieve the two objectives I
laid out. Thus, I would encourage interested parties who do not find the
reforms suggested here attractive to focus on the underlying objectives
of restructuring. In this light, the contribution of this list is not
the specific items I supportalthough, of course, I hope others rally
to thembut rather the impetus it might give to other policymakers
to consider and assemble their own recommendations. We then would have
an alternative path if restructuring falters once again.
See Donald E. Powell, Chairman of the Federal Deposit Insurance Corp.,
"Why Regulatory Restructuring? Why Now?" Comments before the
Exchequer Club, Washington, D.C., Oct. 16, 2002; and Peter R. Fisher,
Undersecretary for Domestic Finance, U.S. Department of the Treasury,
"The Need to Reduce Regulatory Arbitrage." Comments before the
Brooklyn Law School, Center for the Study of International Business Law,
Sept. 20, 2002.
2 For example, see the September 2001
Region which includes presentations at a symposium on "Using
Market Data in the Supervisory Process."
3 See Testimony of Chairman Alan Greenspan,
Federal Reserve Board's semiannual monetary
policy report to the Congress before the Committee on Banking, Housing,
and Urban Affairs, U.S. Senate, Feb. 11, 2003.
4 Office of Management and
Budget. 1992. The Budget of the United States Government, Fiscal Year
Elliehausen, "The Cost of Banking Regulation: A Review of the Evidence,"
Federal Reserve Board of Governors Staff Study, Number 171 (April 1998).