View original document

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

Monetary Policy in a Low
Inflation Environment
J. Alfred Broaddus, Jr.

I

t’s a pleasure to be here today and to have this opportunity to comment on
conducting monetary policy in a low inflation environment. I’ve been at
the Fed for more than thirty-two years and have had the privilege of either
advising monetary policymakers or being a policymaker myself throughout my
career. It has been quite a ride. For much of this period the Fed was struggling
either to prevent inflation from rising further or to bring it down. As you know,
over the last several years we have succeeded in reducing the inflation rate to
1
about 1 2 percent as measured by the core personal consumption expenditures
(PCE) price index—today’s favored inflation index—and in stabilizing the
rate at that level. In the parlance of the day, we have finally attained “price
stability,” meaning both low actual inflation and the credible expectation in
the minds of financial market participants and the general public that it will
persist, which together constitute the monetary policy equivalent of finding
the Holy Grail.
In my brief remarks I want to do four things. First, I will compactly
review several key aspects of the evolution of monetary policy over the last
thirty years. To appreciate fully the nature of the challenge that lies ahead,
it is essential to understand how price stability was lost in the 1970s and regained in the ’80s and ’90s. Second, I will try to convey the essence of the
current strategy of Fed monetary policy. I’ll then close with a brief discussion
of the challenges monetary policymakers face in today’s low inflation environment, as I see them, and a pitch for explicit inflation targeting as a means
of preserving the substantial improvement in the effectiveness of monetary
policy during the Volcker-Greenspan years. As usual, these views are my own
This article is the text of a speech given by J. Alfred Broaddus, Jr., president of the Federal
Reserve Bank of Richmond, before the Public Policy Forum of the American Finance Association on January 3, 2003. The author wishes to thank his colleague Marvin Goodfriend for
his assistance in preparing the article. The views expressed herein are those of the author and
are not necessarily those of the Federal Reserve Bank of Richmond or the Federal Reserve
System.

Federal Reserve Bank of Richmond Economic Quarterly Volume 89/2 Spring 2003

1

2

Federal Reserve Bank of Richmond Economic Quarterly

and don’t necessarily reflect those of my Federal Open Market Committee
(FOMC) colleagues. Also as usual, my views have been strongly influenced
by discussions with, and the writings of, my longtime Richmond Fed colleague Marvin Goodfriend—in particular a preliminary version of a paper he
will deliver at a National Bureau of Economic Research conference on inflation targeting later this month. He is not necessarily responsible, however, for
anything I say here today.

1.

HOW PRICE STABILITY WAS LOST AND REGAINED

It is probably fair to say that, after periods of moderate inflation in the 1950s,
the economy returned to virtual price stability in the early 1960s. The core
CPI inflation rate fluctuated in a narrow 1.0 to 1.6 percent range between 1960
and 1965. Subsequently, as you all know, it increased steadily to double-digit
levels in the mid-1970s and again in the late ’70s.
This extraordinary and debilitating increase in inflation has been attributed,
in whole or in part, to many things: the two oil price shocks in the ’70s, excess
demand associated with the Vietnam War buildup and the Great Society social programs, the ineffectiveness of the Nixon Administration’s price control
program and the Ford Administration’s “Whip Inflation Now” program, and
even the failure of anchovy harvests off the coast of South America. Both
theory and historical evidence, however, indicate that inflationary monetary
policy was the central culprit.
The failure of monetary policy to contain inflation in this period can be
approached from several directions. Economists of a monetarist persuasion
argue that persistently above-target money supply growth, and the practice
of adjusting the money supply target’s base up each year to accommodate
the preceding year’s upside miss, was the principal operational deficiency.
Currently, the more mainstream view focuses on the failure of the Fed’s shortrun interest rate policy to counter the rise of inflationary pressures promptly
during business expansions. The 1970s and early 1980s are sometimes referred
to as the period of “go-stop” monetary policy. Concerned about the potential
impact of policy tightening on employment and production, the Fed would wait
until a broad public consensus emerged that inflation was a serious problem
before acting decisively to contain it. By then, however, it was generally too
late to bring inflation down via tighter monetary policy without at the same
time touching off a recession. The cycles surrounding the 1980 and 1981–82
recessions illustrate this pattern especially well.
More fundamentally, however, it is not an exaggeration to say that Fed
monetary policy lost all or most of its credibility as an effective force against
inflation in this period. As inflation began to rise, financial markets and the
public—even in the early stages of expansions—quickly revised their inflation expectations upward. This reinforced the upward pressure on current

J. Alfred Broaddus, Jr.: Monetary Policy in a Low Inflation Environment

3

inflation, pushed up long-term interest rates, and in general helped foster the
macroeconomic malaise described by the term stagflation.
Perhaps the most important lesson for monetary policy from this experience is how difficult and costly it is for the Fed to rebuild its credibility for low
inflation once it has been lost—especially when, for all practical purposes, it
has been totally lost, as in the late ’70s and early ’80s. Led by Chairman
Volcker, the Fed had to raise nominal short-term interest rates to unprecedented levels and essentially induce one of the longest and deepest recessions
in the entire post–World War II era just to begin the process of restoring its
credibility for low inflation. It is highly doubtful that the process could have
begun without this costly recession, in which real GDP declined 2.8 percent
and the unemployment rate rose to 10.8 percent. And it has taken the Fed
about twenty years—nearly a quarter century—to complete the process.
The essence of this process, in my view, has been the Fed’s demonstration,
particularly in two episodes, that it can preempt an increase in inflation without precipitating a recession, and its success in recent years in convincing the
markets and the public that it will routinely do so in the future. One of these
episodes came early in the process, in 1983 and 1984, as the economy recovered from the 1981–82 recession; the other was in 1994 when the recovery
from the 1990–91 recession finally began to gather momentum. In both cases,
incipient inflationary pressures were quickly picked up by financial markets,
which produced what Goodfriend calls “inflation scares,” characterized by
sharply rising nominal bond rates. In both instances the Fed acted swiftly and
decisively to preempt inflation. The 1994 episode was especially important
since it occurred at a time when Fed policy had become much more transparent than earlier, as evidenced by its decision that year to announce publicly its
federal funds rate target immediately following each FOMC meeting.

2. THE CURRENT STRATEGY OF MONETARY POLICY
The U.S. economy has now enjoyed virtual price stability since about 1996.
There seems to be a growing consensus currently among monetary policymakers, close observers of the policy process in financial markets, Congress
and the press, and individual Americans interested in policy that price stability
and the Fed’s credibility for low inflation should be sustained. This consensus
is based partly on a broader public appreciation of the high costs of reestablishing lost credibility, as described above. More fundamentally, however, it
appears to reflect a recognition that the Fed’s revived credibility is beneficial
to the economy—specifically, that the Fed can contribute meaningfully to an
improved longer-term U.S. economic performance in the form of an increase
in the sustainable growth of production and higher employment. Further,
the public seems less concerned than earlier about possible short-run costs
of low inflation, in terms of lower growth, perhaps because the transition to

4

Federal Reserve Bank of Richmond Economic Quarterly

low inflation has now been accomplished. And since low inflation is broadly
expected to persist, the public would be surprised and disappointed if it were
lost. Consequently, the consensus arguably sharpens the Fed’s accountability
for maintaining low inflation.
Against this background, I sense the emergence within the Fed of a more
cohesive strategy of monetary policy than at any other time in the last three
decades. To my mind it consists of two elements: (1) a strong commitment
to maintaining high credibility for low inflation permanently, and (2) active
management of real short-term interest rates to help stabilize the economy in
the short run. Regarding the first, Goodfriend argues in his forthcoming paper
that the Fed is now practicing “implicit” inflation targeting. As he points out,
with the core inflation rate in the 1 to 2 percent range since the mid-1990s, it is
hard to imagine the Fed now accepting a sustained inflation rate significantly
above 2 percent. Nor would it be likely to accept a sustained rate significantly
below 1 percent given the increased sensitivity to the risk of deflation and the
proximity of the zero bound on nominal interest rates.
I personally believe that “implicit longer-term inflation targeting” is an
accurate description of the first element of the Fed’s current strategy. It is
important to stress, however, that its ultimate objective is not price stability
and high Fed credibility for its own sake, but the optimal financial foundation
these conditions provide for strong real growth and high employment.
Moreover, these conditions enable the Fed to pursue the second element of the strategy: active countercyclical short-term interest rate policy.
When the Fed’s credibility was very low in the 1970s and early ’80s, it was
difficult—perhaps impossible—to conduct countercyclical interest rate policy
effectively. With the Fed’s long-run objective for inflation still unclear, the
public could not confidently deduce the longer-term ramifications of particular
short-term policy actions, and the Fed, in turn, could not confidently predict
the public’s reaction to its actions. With its credibility for low inflation now
well established, the Fed can act both more promptly and more aggressively to
counter the effects of unanticipated shocks and thereby stabilize the economy
in the short run. Beginning exactly two years ago today, the Fed began to ease
policy in response to the softening of the economy in the second half of 2000.
It accelerated the easing process in the wake of 9/11, and over the course of
1
the two-year period has reduced the federal funds rate 5 4 percentage points,
1
1
from 6 2 percent to its present level of 1 4 percent. This is arguably the most
aggressive series of policy easings taken to cushion a softening economy in
the Fed’s history and may well account for the apparent brevity of the recent
recession despite the extraordinary decline in the stock market, 9/11, and other
shocks.
The two elements of the Fed’s current strategy, then, are complementary
and mutually reinforcing. Implicit inflation targeting enhances the effectiveness of countercyclical interest rate policy. Conversely, active countercyclical

J. Alfred Broaddus, Jr.: Monetary Policy in a Low Inflation Environment

5

policy makes implicit inflation targeting acceptable, since the ability to act
aggressively to stabilize the economy in the short run provides a clear and
easily understood rationale for containing inflation.

3.

CHALLENGES IN A LOW INFLATION ENVIRONMENT
WITH IMPLICIT INFLATION TARGETING

After such a long struggle, one might expect that Fed monetary policymakers
would be relatively comfortable now that price stability has been achieved and
credibility for low inflation has been reestablished. And I think most, if not all,
policymakers are more confident that the Fed can contribute constructively to
the economy’s longer-term performance rather than retarding it, as occurred
when inflation was high and variable, and credibility was low.
But the Fed still faces significant policy challenges in the new low inflation environment. Historically, little practical attention has been given to
the possibility of excessively sharp disinflation and deflation. And with the
press here I need to emphasize at the outset that I do not believe deflation is
a serious present risk to the economy. But policymakers obviously need to
think more about how they would deal with a deflationary threat, should one
emerge unexpectedly, when inflation is in a 1 to 2 percent range than when it
is at 6, 7, or 8 percent. This is especially so with the nominal funds rate, our
principal short-term policy instrument, only 125 basis points above zero.
We have been thinking about it, and I am quite confident that we could
deal with a deflationary threat successfully. In October 1999 the Fed sponsored a conference in Woodstock, Vermont, on conducting monetary policy in
a low inflation environment attended by a large number of leading monetary
economists. The participants gave substantial attention to deflation and how
to deal with it should it arise in the future. More recently, Fed Governor Ben
Bernanke nicely summarized current thinking on this issue in a speech to the
National Economists Club. There is now broad agreement that the most effective way to deal with deflation is to prevent it from developing in the first
place. In the present situation, the Fed’s aggressive easing over the last two
years appears to have preempted any significant drift toward either excessive
disinflation or deflation. Moreover, even if disinflation unexpectedly intensified and the funds rate were reduced close to the zero bound, the Fed would
still have a number of channels available to reestablish a comfortably positive
inflation rate. For example, it could increase broad liquidity by purchasing
long-term bonds.
The other potential policy challenge I see in today’s low inflation environment is how to handle an incipient increase in inflation above its implicit
target range. This possibility is not on many radar screens currently, but it
is obviously a longer-term risk—arguably the most likely longer-term risk. I
believe that the policy experience of the 1970s, ’80s, and ’90s summarized

6

Federal Reserve Bank of Richmond Economic Quarterly

above argues strongly for prompt action to preempt any sustained increase in
inflation. If policymakers had precise, detailed foreknowledge of the relative
costs, in terms of lost production, of alternative paths back to price stability,
it might be feasible to tune the return more finely. There is little evidence,
however, that we have such knowledge. Hence, it seems reasonable to resist
any deviations from price stability promptly and strongly—and preferably
preempt them altogether.

4.

CONCLUSION—SUSTAINING THE PROGRESS

Hopefully these comments have convinced you that the conduct of monetary
policy in the U.S. has improved significantly during the last two decades, and
that this improvement of policy holds out the prospect of an improved longerterm economic performance going forward. The trick now is to sustain the
progress. Much of the progress, in my view, is due to the exceptionally strong
leadership since 1979 of, first, Paul Volcker and now Alan Greenspan. But,
ultimately, high-quality monetary policy—i.e., sustained credibility for low
inflation as a foundation for strong real growth—is too important to be dependent on exceptional leadership alone, which, after all, cannot be guaranteed
over the long pull. The progress in recent years needs to be institutionalized—
“locked in”—in some manner.
There are probably several ways this could be accomplished. Earlier I
referred to one element of the Fed’s current policy strategy as implicit inflation
targeting. My personal preference for “hardening” our credibility is to make
the implicit target both explicit and quantitative—specifically, 1 to 2 percent,
based on the core PCE index. Explicit, quantitative inflation targeting is
practiced by a number of other leading central banks around the world, and it
would be consistent with the continuing evolution of Fed policy toward greater
transparency and accountability. Most importantly, it would be a strong and
visible step toward ensuring that the Fed’s current high credibility for low
inflation will be maintained indefinitely so that we can make our strongest
possible contribution to maximum sustainable growth in the long run and
economic stability in the short run.

Banking and Commerce:
Tear Down This Wall?
John R. Walter

M

any U.S. firms include both commercial and nonbank financial
units. For example, General Motors Corporation encompasses
not only units that manufacture automobiles but also those, such
as General Motors Acceptance Corporation, that gather funding and make
loans to individuals and businesses. Firms that handle both commercial and
financial activities appear to reap significant benefits that create the appeal
of such combinations. One byproduct of a commercial firm’s activities may
be information about its customers’ financial situation. The financial affiliate
might then use this information to inexpensively target products to particular
customers, benefiting both the financial firm and its customers.
While finance/commerce combinations are widespread, combinations between banks and commercial firms are typically prohibited under various U.S.
laws. Banks are distinguished from other financial firms by their ability to
gather funding by issuing government-insured deposits such as checking and
savings deposits. Despite prohibitions of banking/commerce combinations,
firms have managed to find loopholes. Until recently the unitary thrift loophole was a popular means of circumventing the banking/commerce wall. The
loophole allowed commercial companies to start or buy one, and only one,
thrift (i.e., a savings bank or savings and loan association, both of which issue government-insured deposits), using the thrift as a conduit for providing
financial services.
The unitary thrift loophole was closed in 1999, but another loophole remains open. Federal banking law allows commercial firms to own industrial
loan corporations—essentially banks with somewhat restricted deposit-taking
powers. (For additional discussion of the unitary thrift and industrial loan corporation loopholes, see the Appendix.)
The author benefited greatly from discussions with Kartik Athreya, Tom Humphrey, Ray
Owens, and John Weinberg. The views expressed herein are not necessarily those of the
Federal Reserve Bank of Richmond or the Federal Reserve System.

Federal Reserve Bank of Richmond Economic Quarterly Volume 89/2 Spring 2003

7

8

Federal Reserve Bank of Richmond Economic Quarterly

Given the apparent benefits of combinations, why prohibit or restrict
them? What hazards result from banking/commerce combinations? Traditionally, discussions of the threats focused on conflicts of interest and diminished competition. More recently, observers have been concerned that
combinations might increase deposit insurance claims and expand the universe of economic activities protected by the government safety net. As will
be discussed presently, the traditional concerns seem less relevant given the
level of competition banks face in today’s banking market. Over the last
twenty-five years, competition has expanded as restrictions were eliminated
on banks’ ability to operate across state lines and to offer market rates on
deposits. Also, new nonbank firms have arisen offering financial products
competitive with most banking products. The concerns over increased deposit insurance claims and expansion of the safety net remain quite relevant.
Nevertheless, over the last decade a number of legislators have argued for
removal of the banking/commerce wall. I will analyze the threats and suggest
some restrictions that would be necessary if the wall were removed.

1. THE STATUTES FORMING THE WALL
The building blocks of the wall between banking and commerce are various
federal and state laws. The laws prevent banks from engaging in commercial activities. They also prevent banks from owning subsidiary commercial
companies and from being owned by companies conducting commercial activities. Specifically, the building blocks are the National Bank Act of 1864,
state banking laws, the Federal Deposit Insurance Corporation Improvement
Act of 1991, and the Bank Holding Company Act of 1956. (For a detailed
review of these statutes and their motivations see the Appendix.)
The National Bank Act limits the powers of national banks and their subsidiaries. National banks are those chartered and regulated by the U.S. Treasury Department’s Office of the Comptroller of the Currency. The Act states
that “a national banking association shall. . . have power to. . . exercise. . . all
such incidental powers as shall be necessary to carry on the business of banking” (12 U.S.C. 24). While over the years the courts and the Comptroller
have wrangled over the meaning of the phrase “business of banking,” national
banks have been allowed to engage in businesses similar to traditional banking
services but not other commercial activities. This restriction of powers extends not only to activities of banks, but to activities conducted by subsidiaries
owned by banks.
State banking statutes typically set limits on the nonbank activities of
state banks and their subsidiaries, similar to the limits on national banks. Yet,
over the years, a number of states have authorized activities well beyond those
allowed national banks, some of which typically would be considered commercial powers. Concerns that these new powers might endanger state-chartered

J. R. Walter: Banking and Commerce

9

banks, and thereby the taxpayer-backed deposit insurance fund, led Congress
to restrict state legislatures’ ability to grant powers to state-chartered banks.
Specifically, under the Federal Deposit Insurance Corporation Improvement
Act of 1991 (FDICIA), insured state-chartered banks are prohibited from engaging in any activities impermissible for national banks unless the FDIC rules
that such activities pose no threat to the deposit insurance fund.
Those who advocate removing the banking/commerce wall typically do
not argue for allowing banks to conduct commercial activities or own commercial firms. Instead, they focus on allowing companies that own banks—that
is, bank holding companies—to own commercial companies, too. In other
words, they do not condone direct bank involvement in commerce but find it
acceptable for bank holding companies to own commercial companies, allowing banks to affiliate with commercial companies.
The banking/commerce restrictions in the Bank Holding Company Act of
1956 were based on the view that “bank holding companies ought to confine
their activities to the management and control of banks.” The Act restricted
bank holding companies such that they “would no longer be authorized to
manage or control nonbanking assets unrelated to the banking business” (U.S.
Code: Congressional and Administrative News [1956, 2482, 2484]).
To enforce this restriction, the Act defines a bank holding company as any
company that owns a bank.1 It prohibits bank holding companies from engaging in nonbanking activities. The Act allows an exception to the nonbanking
prohibition in cases in which the Board of Governors of the Federal Reserve
System determines the nonbanking activity “to be so closely related to banking as to be a proper incident thereto” (12 U.S.C. 1843c). Typically the Board
defines activities closely related to banking as only those activities traditionally performed by banks. Certain additional activities also have been allowed,
however, in cases in which they are tied to banking. For example, the Board
has determined that a bank holding company may own a data-processing firm
if it is primarily engaged in processing financial, banking, or economic data
(Spong [2000, 156]).
The Gramm-Leach-Bliley Act, enacted in 1999, added securities underwriting and dealing, as well as insurance, to the list of activities in which
banks—through bank-owned subsidiaries—and bank holding companies could
engage. Those bank holding companies choosing to engage in securities or
insurance activities are called financial holding companies. The Act also expanded the activities of bank-owning companies to include most financial
activities, those determined by the Board and the Treasury Department to be
“financial in nature,” “incidental to a financial activity,” or “complementary to
1 The Bank Holding Company Act of 1956 applied only to companies owning two or more
banks. Amendments enacted in 1970 extended the Act’s provisions to single-bank holding companies.

10

Federal Reserve Bank of Richmond Economic Quarterly

a financial activity.” Nonetheless, while authorizing a wide range of financial
activities, the Act leaves in place the wall between banking and commerce.
Deciding whether an activity is commercial instead of banking or financial
and whether it can be conducted by a banking company is often not simple.
The decision was difficult under the old Bank Holding Company Act standard
of being closely related and remains so under the new standard of being
financial in nature, incidental to a financial activity, or complementary to a
financial activity. The placement of an activity on one side or the other of the
banking/commerce wall can be controversial and contentious. For example,
in December 2000 the Board of Governors of the Federal Reserve System
and the Secretary of the Treasury jointly released for comments a proposal to
permit bank subsidiaries and financial holding companies to engage in real
estate brokerage and management. Real estate industry trade groups quickly
objected to the proposal, arguing that it would amount to an illegal mixture of
banking and commerce. Banking trade groups argued in favor of the proposal.
Beyond the real estate industry’s objections, legislators introduced bills in both
the U.S. House of Representatives and the Senate to prohibit these real estate
activities in financial holding companies. In April 2002 the Secretary of the
Treasury announced plans to put off a decision on the proposal until 2003.

2.

RATIONALES FOR THE WALL

Three reasons are typically cited for maintaining a wall that prohibits banking/commerce combinations: conflicts of interest, monopoly power, and risk
to the taxpayer-backed deposit insurance fund.2 These three justifications will
be examined below. As it turns out, because banking markets appear fairly
competitive, the first two seem of relatively minor import. The third remains
quite significant.

Conflicts of Interest
Observers have at times raised concerns over conflicts of interest that might
arise if banks and commercial firms are owned by the same firm. They argue
that such concerns justify keeping banking and commerce separate. Three
conflicts have been described. First, a bank affiliated with a commercial firm
would tend to deny loans to the affiliate’s competitors. Second, a bank might
use access to insured deposits to provide below-market-rate funding to its
affiliates while charging higher interest rates to unaffiliated borrowers. Third,
in the legislative history of the Bank Holding Company Act, legislators noted
2 For discussions of these three reasons for maintaining the wall, see Krainer (2000, 21–23);
Halpert (1988, 490–517); and U.S. GAO (1987, 11–12).

J. R. Walter: Banking and Commerce

11

that a bank with a commercial affiliate might deny loans to individuals who
do not purchase goods from the affiliate.
It seems natural that removing the banking/commerce wall would allow
the first conflict to arise, since a bank with a commercial affiliate, say a restaurant, would not wish to provide funding to competing restaurants. Helping
the competitor would tend to lower the profits of the affiliated restaurant. Yet,
if competition is reasonably strong, denying loans to competitors only lowers overall profits of the consolidated banking/restaurant firm. If there are
alternative lenders over which the affiliated bank has no price advantage, the
competing restaurant would get a loan anyway and at the same interest rate
the affiliated bank would offer. So, by failing to make the loan, the bank loses
any profit it might have made on that loan, hurting the bank. Yet the affiliated
restaurant suffers a loss in profits regardless.3 Therefore, if competition is
strong, this potential conflict of interest is unlikely to present a problem and
cannot justify maintaining the banking/commerce wall.
But is banking competition strong? Since the 1970s, restrictions on bankversus-bank competition have been greatly reduced. Restrictions on banks’
ability to compete for deposits outside of their local markets, or at least outside
of their home states, were severe before the late 1970s. While these restrictions did not apply to bank lending, banks generate a good bit of their lending
in the same markets in which they gather loans. Consequently, these restrictions likely limited loan competition as well. Banks’ ability to open branches
statewide was greatly enhanced in the 1980s as many states removed branching restrictions. Restrictions on operating across state lines began to fall in
the mid-1980s and were almost completely removed by the Riegle-Neal Interstate Banking Act of 1994. As a result, banks that had been protected from
competition because of branching restrictions became subject to competitive
pressure from nonlocal banks by the mid-1980s.
While the elimination of branching restrictions opened local banking markets to greater competition, other market and technological developments have
expanded competition further. Consequently, if a bank denies a loan to a business firm because it competes with the bank’s affiliate, that firm can find
numerous alternative sources of funding in today’s more competitive loan
markets.
Competition among those who would lend to business borrowers has expanded along several dimensions. For large business borrowers, banks faced
growing competition from the debt markets as commercial paper and bond
issues increased significantly relative to bank lending over the last twentyfive years. While small businesses cannot issue commercial paper or bonds,
today’s small businesses have access to loans from a wide range of lenders.
3 Owens (1994) makes this argument for bank lending in real estate.

12

Federal Reserve Bank of Richmond Economic Quarterly

The largest banks aggressively court small business borrowers throughout the
country via their Web sites and toll-free phone lines. Further, small businesses
enjoy a range of choices of nonbank lenders, including finance companies and
leasing companies. Clearly, today’s borrowers, both large and small, have a
plethora of borrowing opportunities because of the competitive loan market.
If, in spite of these factors, some banking markets remain uncompetitive,
policymakers can address the problem directly by removing any remaining
barriers to entry. Alternatively, they might tackle monopoly power through
antitrust enforcement. Maintaining a wall that separates banking and commerce at best addresses a symptom of an uncompetitive market rather than the
lack of competition itself.
Still, using the restaurant example, one might argue that the bank with an
affiliated restaurant may for some reason have a cost advantage over its bank
competitors that lack such an affiliation. One reason for a cost advantage is
that the bank acquires information about the restaurant business through its
affiliation. While prohibiting affiliations might eliminate the advantage this
bank (and its affiliates) has over competitors, the restriction would diminish
economic efficiency because the least costly and most efficient means of producing banking services—through restaurant affiliation—would be denied.
Additionally, if bank/restaurant affiliations are allowed, banks that lack a
restaurant affiliate can simply overcome the cost disadvantage by affiliating.
This is just what the banking industry did when banks established branches
in grocery and discount stores, though they did it through leasing agreements
rather than affiliation. After perceiving the advantage gained by the innovative bank that first placed branches in such stores, other banks followed
suit to achieve the same advantage. Soon the advantage was dissipated by
competition.
Aside from the situation whereby a bank might deny loans to its affiliate’s
competitors, some observers note a second conflict of interest. They argue
that banks may have access to inexpensive funding because of underpriced
deposit insurance, and that this funding might be granted to banks’ affiliates
but not to other borrowers. Such funding would give affiliates an advantage
over firms not so affiliated.4 As long as the banking market is competitive,
however, every firm that borrows from a bank gets equivalent access to lowcost funding whether affiliated with a bank or not. Access is equivalent because
a bank only hurts itself (i.e., lowers its revenues) by not lending to its affiliate’s
competitors on equivalent terms to those offered its affiliate. If the bank with
an affiliate does not lend to its affiliate’s competitors, other banks would take
those customers and profit from doing so. Further, Section 23A of the Federal
Reserve Act, applicable to all banks, restricts the amount of such affiliate
4 For discussions of the argument that access to bank funding could give bank affiliates an
advantage, see Board of Governors (1987, 500) and Macey and Miller (1992, 377).

J. R. Walter: Banking and Commerce

13

funding to at most 10 percent of the bank’s capital. Section 23B of the Act
requires such funding be on market terms.
In the legislative history of the Bank Holding Company Act, members of
Congress describe a third possible conflict of interest. Specifically, they argue
that a bank with a commercial affiliate might deny loans to individuals who
do not purchase goods from the affiliate. The Senate Banking Committee’s
report that analyzes the features of the bill that later became the Bank Holding
Company Act describes the concern as follows:
The committee was informed of the danger to a bank within a bank holding
company controlling nonbanking assets, should the company unduly favor
its nonbanking operations by requiring the bank’s customers to make use
of such nonbanking enterprises as a condition to doing business with
the bank. The bill’s divestment provisions should prevent this fear from
becoming a reality. (U.S. Code: Congressional and Administrative News
[1956, 2486], as cited in Halpert [1988, 500])

Tying a loan (or other service) to the purchase of another product can
only benefit a bank if the bank has monopoly power in its loan market. If
it faces competition, denying loans to individuals who are not its affiliate’s
customers only hurts the bank, and so would not be undertaken. The bank is
hurt because it forfeits revenues and helps its bank competitors who would
make the loans (Owens [1994]). As noted earlier, when the Bank Holding
Company Act was passed in 1956, banking markets were heavily regulated.
Entry was restricted and prices were controlled. Monopoly power may have
been significant, but most such restrictions have been removed. Additionally,
even if banks maintain monopoly power in credit markets, the commercial
affiliate must also have market power in order for tying to make consumers
worse off. In the case in which the combined firm has market power in
the banking and commercial markets, only under limited circumstances are
consumers actually made worse off. In other cases consumers are unhurt by
tying (Weinberg [1996]). Regardless, current statutes make tying by banking
companies illegal.5

Proliferation of Monopoly
Some observers argue that, in addition to conflicts of interest, preventing the
exercise of monopoly power is another reason for the banking/commerce separation. The legislative history of the Bank Holding Company Act makes
it clear that Congress intended the Act to guard against the proliferation of
5 See Weinberg (1996) for a review of bank anti-tying statutes and the economics of tying.
See Krainer (2000, 22) for a discussion of banks denying credit to their commercial affiliates’
competitors. In 1997 the U.S. General Accounting Office analyzed banks for evidence of tying
bank loans to securities activities. It found little evidence of any such tying (U.S. GAO [1997]).

14

Federal Reserve Bank of Richmond Economic Quarterly

monopoly. For example, the Senate Banking Committee report on the conference bill notes that the Act was to provide “safeguards. . . against undue
concentration of control of banking activities. The dangers accompanying
monopoly in this field are particularly undesirable in view of the significant
part played by banking in our present national economy” (U.S. Code: Congressional and Administrative News [1956, 2482–83]). Because such language is vague, it is difficult to determine whether the undue concentration
discussed refers to horizontal or conglomerate concentration. As a result, it
is uncertain whether proliferation of monopoly was behind the Bank Holding Company Act’s banking/commerce restrictions. Horizontal concentration
means combining a number of banks under one bank holding company such
that this holding company controls a high percentage of banks. Conglomerate
concentration means combining both banks and nonbanks under one holding
company so that one conglomerate controls a significant percentage of business firms in banking and a nonbanking industry, or in several nonbanking
industries.
Observers since have argued that Congress was indeed concerned with
conglomerate concentration. A case in point was a 1974 Federal Reserve’s
denial, under the Bank Holding CompanyAct, of an application by BankAmerica Corporation to form an overseas joint venture with Allstate Insurance.
Here the Fed said that “close working relationships abroad between large
U.S. banking organizations and large U.S. insurance companies could in time
weave a matrix of relationships. . . that could lead to an undue concentration
of economic resources in the domestic and foreign commerce of the United
States. . . not. . . consistent with the purposes of the Bank Holding Company
Act.”6
Frequently, when advocating the separation of banks from nonbanks on
the basis of monopoly, proponents have argued that the combination would
allow the monopoly power that banks hold in their product markets to be used
by combined firms to raise prices in other areas.7 But as discussed earlier,
while in 1956 concerns about monopoly may have motivated Congress, during
the 1970s and 1980s competition expanded greatly, among banks and between
banks and nonbanks. Expanded competition significantly reduced any opportunity banks might have had to exercise monopoly power in banking services
and to expand it to other businesses with which they might combine. Congress
seems to have been cognizant of these changes. When Congress passed the
Gramm-Leach-Bliley Act in 1999, it allowed combinations of large banks
with large insurance and large securities firms. These were exactly the types
6 Board of Governors (1974, 519) (italics added for emphasis). For a similar argument on
another application, see Board of Governors (1981, 451), as cited in Halpert (1988).
7 Halpert (1988, 500–505) discusses the argument that banking monopoly might proliferate
into nonbank businesses.

J. R. Walter: Banking and Commerce

15

of combinations—that is, large banks with large nonbanks—denied earlier
by regulators based on undue concentration language in the Bank Holding
Company Act.

Safety Net Concerns
For the reasons discussed above, conflicts of interest and fears of expanding
monopoly power alone are probably insufficient reasons to maintain the current wall separating banking and commerce and deny firms the opportunity
to benefit from combinations. Nevertheless, there is another set of hazards
that could justify the continued presence of the wall separating banking and
commerce or at least require that significant precautions be taken if the wall
is removed. The hazards come in three forms, discussed in the following
paragraphs, and each involves an increased chance of bank failures and a
subsequent bailout financed by taxpayers. If bailouts occur, the government
safety net, meant to protect bank depositors, could be extended to creditors
of commercial companies. If extended, too many resources might flow to
bank-affiliated commercial companies, and economic efficiency would be diminished. The threat to the safety net could arise because (1) losses might
be shifted to banks to protect a combined firm’s reputation with investors, (2)
losses might be shifted to banks to take advantage of shareholder limited liability, and (3) the combined firm’s riskiest assets might be shifted to the bank.
Of the three hazards, the first two could justify continued banking/commerce
separation. The third cannot justify separation but is discussed below because
it is often mentioned as a hazard of bank/nonbank affiliations.
Loss Shifts that Protect Reputation

If banking/commerce combinations are allowed, a combined company can
be expected under certain circumstances to withdraw resources from its bank
to hide problems in its commercial subsidiary, damaging bank safety. The
holding company is likely to choose this course when it can hide commercial
subsidiary losses from investors and analysts by shifting commercial subsidiary losses to the bank. The holding company would benefit by hiding
the loss, which if revealed would likely be perceived as negative information
about the ability of the firm’s management and the riskiness of its operations;
that is, it would damage the firm’s reputation. Such negative information
would lead creditors to demand higher interest rates, lowering future profits.
Yet, as discussed presently, while shifts to hide losses can be detrimental to
banks, they can just as easily be beneficial: holding companies could choose
to shift bank losses to commercial subsidiaries. Consequently, a concern that
bank holding companies might engage in loss shifts is no reason to prohibit
banking/commerce combinations. Instead, if one is to argue that the danger

16

Federal Reserve Bank of Richmond Economic Quarterly

of shifts can justify the banking/commerce wall, one must believe that loss
shifts are more likely to flow toward bank than commercial subsidiaries.
Reputation-protecting shifts that can work to the detriment of bank health
are likely to occur when two conditions are met. First, the commercial subsidiary suffers a loss large enough to create its insolvency. Second, the loss, if
shifted to the bank subsidiary, would avert the bank’s insolvency. The second
condition would generally be met if the bank’s net worth is considerably larger
than that of the commercial subsidiary (before and after the shift). Under these
conditions, a shift of a commercial subsidiary’s loss to the bank would protect
the holding company’s reputation. An insolvency is certain to draw negative outsider attention, since it will likely involve either debt renegotiation or
bankruptcy. In contrast, a mere loss or perhaps just an increase in the bank’s
reported expenses can be expected to draw far less attention. Even so, a loss
shift necessarily weakens the bank.
Even if the commercial subsidiary experiences a loss that does not lead
to its insolvency but is still significant, such a loss could still shift to a larger
bank subsidiary. The bank holding company might choose to shift the loss
because it might be less noticeable on the books of a large firm than on those
of a smaller one. In addition, observers have traditionally argued that banks
may have more opaque balance sheets than do commercial firms, so that losses
can be better hidden in a bank subsidiary. Some recent research appears to
support this view of bank opacity.8 If banks are indeed more opaque, then
losses are more likely to go unnoticed if shifted to the bank.
Nevertheless, the existence of this incentive to shift losses in order to hide
them does not imply that commercial and banking firms should be kept separate. Under one set of circumstances already discussed—when the bank’s net
worth (meaning its capital) is larger than the commercial subsidiary’s—a bank
holding company can hide the loss by shifting it to the bank. However, under an equally likely set of circumstances—when the commercial subsidiary’s
capital exceeds the bank’s capital—there is no benefit from shifting commercial subsidiary losses to the bank. Instead, if the bank produces losses, the
bank holding company can benefit by shifting bank losses into the commercial
firm. Therefore, prohibiting banking/commercial affiliations will not necessarily improve bank safety or protect taxpayers and the FDIC from losses.
Further, creditors of banks as well as commercial firms affiliated with
banks are likely to be well aware of the incentive to shift losses in order to
8 Morgan (2000) finds that banks and insurance companies are inherently more opaque than
other firms. Morgan checks for opaqueness of banks and insurance firms versus nonfinancial firms
by measuring the frequency of disagreements between the two major ratings agencies in their
ratings of banks, insurance companies, and other firms. He finds that the two agencies disagree
more frequently over banks (and over insurance companies) than over commercial firms. Morgan
contends that the cause of the disagreement is the difficulty of evaluating opaque bank balance
sheets.

J. R. Walter: Banking and Commerce

17

hide them. By charging higher interest rates, both types of creditors will
penalize affiliations that might shift losses to the detriment of their debtor
(either commercial firm or bank). For example, a bank’s creditors would
be likely to view a combination with a risky commercial firm—that is, one
that might produce shiftable losses—as dangerous. They would demand the
bank pay an increased interest rate if such an affiliation were undertaken.
Moreover, if the affiliated commercial firm’s riskiness increased, the bank’s
creditors would impose an additional risk premium to account for the increased
risk of a loss shift. If the risk became large, the increased premium might be
sufficient to cause the holding company to divest either the commercial firm
or the bank. The affiliated commercial firm’s creditors would do the same.
Yet there is reason to think that shifts would tend more frequently to deplete bank resources rather than commercial firm resources. While creditors of
commercial subsidiaries of bank holding companies would demand higher risk
premia for affiliations likely to produce losses that can be shifted to the commercial affiliate, bank creditors have a reduced incentive to do so. Many of a
bank’s creditors—those holding insured deposits in the bank—would demand
no additional compensation when the bank affiliates with a risky commercial
firm. If losses are shifted to the bank, weakening it, its government-insured
deposits are no less likely to be repaid. Therefore, while the creditors of commercial affiliates penalize risky combinations, those of banks do not. Consequently, combinations that could lead to loss shifts toward banks are likely
to be more common than combinations that could produce loss shifts toward
commercial firms.
Note that if bank deposit insurance premia were closely tied to individual
bank riskiness and accounted for the risk of loss shifts, higher premia would
discourage affiliations that could be risky to banks. As discussed in more
detail below, observers typically argue that deposit insurance premia are not
closely tied to bank risk.
Loss Shifts that Take Advantage of Limited Liability

While the previous section discusses a set of incentives that could lead a
bank holding company to shift losses from a less capitalized subsidiary to
a more capitalized one, another set of incentives can produce the opposite
result. Under certain circumstances, by shifting losses from a more capitalized
subsidiary to a less capitalized one, the bank holding company can reduce
losses. The strategy, discussed below, is beneficial because of the protections
offered shareholders by the principle of limited liability. In some cases it
could work to the detriment of the FDIC, and ultimately, to the detriment
of taxpayers. As in the case of reputation-protecting shifts, shifts that take
advantage of limited liability seem at first to be just as likely to enhance
bank safety as diminish it, suggesting that this argument cannot be used as a
justification for maintaining the separation between banking and commerce.

18

Federal Reserve Bank of Richmond Economic Quarterly

On further analysis, however, it is clear that limited liability shifts would more
likely work to the detriment of banks and therefore to the detriment of the FDIC
and taxpayers. Therefore, maintaining the banking/commerce wall might be
justified as a means of preventing these shifts.
The following example shows that with limited liability, a bank holding
company can avoid losses if it shifts them. Suppose a holding company—
Alpha Conglomerate Inc.—owns two subsidiaries, Bravo Dry Cleaners and
Echo National Bank. Bravo has a net worth of $100 million, while Echo’s
net worth is $5 million. Alpha’s only assets are its investments in the stock
of Bravo and Echo, and it is the sole owner of both. Consequently, Alpha’s
net worth is $105 million, the sum of Bravo’s and Echo’s net worth. If Bravo
suffers a $10 million loss (say it has bankrupt commercial customers to which
it has made $10 million in loans), Bravo’s net worth falls to $90 million. Also
as a consequence of Bravo’s loss, Alpha’s stockholders suffer a $10 million
loss since Alpha’s net worth falls to $95 million (the sum of Bravo’s $90
million net worth and Echo’s $5 million).
Suppose instead that Alpha could arrange to have Echo take the loss.
Echo could take the loss by purchasing the loans made to Bravo’s bankrupt
commercial customers for $10 million, even though the loans are worthless.9
The shift of the $10 million loss to Echo, which had only $5 million in net
worth before the shift, drives it into insolvency. Bravo has a net worth of $100
million after the shift, and Echo has a net worth of negative $5 million. Based
on the principle of limited liability of shareholders, however, Alpha can suffer
a loss of no more than its investment in Echo, or $5 million. The shift has
saved Alpha’s shareholders $5 million. In this case the FDIC, which insures
Echo, suffers the remaining $5 million loss. In summary, a holding company
can benefit by shifting a loss when that loss is smaller than the loss-producing
subsidiary’s capital, but greater than the other subsidiary’s capital.
Alpha’s incentive to protect its reputation, as discussed in the previous
section, will tend to work to prevent it from employing a shift that will produce
an insolvency. Echo’s insolvency is certain to damage Alpha’s reputation
and raise its future borrowing costs. Further, such shifts could be illegal.
Nevertheless, when the benefit from shifting losses is large, the shift might be
undertaken regardless of reputation or legality.
While this incentive to shift losses could endanger bank health, and in
fact such a shift led to a bank failure in 1953, holding companies owning
bank and commercial affiliates initially seem no more likely to shift losses
9 As will be discussed presently, federal law restricts a bank’s ability to purchase loans made
by its affiliates to a small percentage of the bank’s net worth. Purchases of sufficient affiliate
loans that lead to the bank’s insolvency would be illegal under the restrictions. Nevertheless, in
some cases, such purchases have occurred.

J. R. Walter: Banking and Commerce

19

into banks than away from them.10 In other words, combinations of banking
and commerce are just as likely to enhance bank safety as reduce it. However,
there is a greater chance that shifts will work against banks; for while banks’
major creditors—insured depositors—are largely indifferent about the risks
that affiliations with nonbanks might impose, commercial affiliates’ creditors
are very interested. Creditors of commercial affiliates will penalize, with
demands of higher interest rates, affiliations that increase the likelihood of an
affiliate failure.
Since commercial firm losses tend to be shifted toward banks, undermining
bank health, banking/commerce affiliations increase the likelihood of FDIC
payouts and ultimately of taxpayer bailouts of the FDIC. Consequently, a limited liability motive for loss shifts could provide a reason to favor prohibiting
banking/commerce combinations.
Beyond the cost to the FDIC and perhaps to taxpayers, who provide the
backstop for FDIC insurance, there is an additional cost of loss shifts (motivated by limited liability as well as reputation protection). If creditors of
bank-affiliated commercial firms believe that these firms’ losses can be shifted
to banks and ultimately to the FDIC, then creditors will charge bank-affiliated
commercial firms lower interest rates than they would absent the perceived
ability to shift. As a result of this reduced cost of capital, affiliated firms would
regard projects as viable that without this taxpayer-provided subsidy would
be unprofitable. In sum, too much investment capital would flow to affiliated firms, and the economy’s resources would be wasted.11 This potential
for resource waste may provide further reason to prohibit banking/commerce
combinations, or at least to regulate combined firms to discourage shifts.
Risk Shifts

At first blush there appears to be one additional reason to maintain the separation between banking and commerce: combinations would allow risky assets
to be shifted from the commercial firm to the bank. Doing so increases the
bank’s riskiness, putting taxpayer funds at risk. This possibility has caused
some observers to raise concerns about affiliations between banks and nonbanks. As a justification for maintaining the banking/commerce separation,
however, the argument is unconvincing.
To lower its total funding costs, a bank holding company with a commercial subsidiary can shift the commercial firm’s riskiest assets to the bank. The
commercial firm must borrow using uninsured debt, while the bank can gather
10 See Walter (1996, 22) for a discussion of the case in 1953 in which a bank failure resulted

from shifts from a bank holding company’s nonbank subsidiary to its bank subsidiary, apparently
motivated by an attempt to take advantage of limited liability.
11 See Walter and Weinberg (2002, 373–75) for a more complete discussion of the economic
costs of government subsidization of private firms’ borrowing costs.

20

Federal Reserve Bank of Richmond Economic Quarterly

funds by issuing insured deposits. As a result, funding costs are lowered and
holding company profits are increased when the commercial firm’s riskiest
assets are shifted to the bank. (Note that risk shifts differ from loss shifts,
discussed earlier. Loss shifts occur when the bank purchases the assets from
the commercial firm at a price that produces a loss for the bank. Risk shifts
occur when the bank pays a price that produces no loss for the bank, since it
is insensitive to risk.) For banks’ costs to be less sensitive, deposit insurance
premia and other supervisor-imposed costs must be imperfectly sensitive to
bank riskiness. Observers argue that this could be the case for many banks.12
Risk shifts, however, do not justify the banking/commerce wall because
affiliation creates no more incentive to shift risks than would exist without
affiliation. If the penalty for holding risky assets is lower for banks than for
commercial firms (or, for that matter, for any uninsured firm), then risky assets
would flow into banks even if they have no affiliates. Banks would be willing
to pay more for risky assets than would other firms and would bid them away
from others.
For example, imagine that a commercial firm, Juliet Tool and Die, Inc., is
currently paying its creditors 15 percent in annual interest payments to raise
$100,000. It uses this $100,000 to make trade credit (i.e., a loan from a seller
to its customer used by the customer to purchase the seller’s goods) available
to its customer, Kilo Millwork. Juliet’s creditors charge this high rate because
they view Kilo as risky, such that Juliet’s loan to Kilo heightens the chance
that Juliet will itself fail; in other words, the trade credit is a risky asset.
Alternatively, Lima National Bank, which pays depositors only 10 percent,
can raise the $100,000 from depositors with which it can provide funds to
Kilo. Because of FDIC insurance, its depositors care little about the riskiness
of Lima’s assets. In such a case, Lima National can be expected to approach
Juliet and offer to buy its asset, the trade credit to Kilo. Lima would be willing
to pay more than the asset is worth to Juliet since Lima can fund the asset less
expensively than can Juliet.13
A holding company with a bank subsidiary and a commercial subsidiary
may well benefit from having its commercial firm sell its risky assets to the
bank subsidiary because of underpriced deposit insurance. But a commercial
firm with no affiliated bank would find that banks would want to buy their risky
assets just as well. So, if deposit insurance is underpriced, whereby it is less
expensive for banks than for commercial firms to hold risky assets, preventing
affiliation would not prevent risky assets from being shifted to banks.
12 See Walter (1998, 2–9) for a discussion of the means by which banks can receive riskinsensitive funding.
13 Lima would only be willing to pay more for the loan than it is worth to Juliet if Lima’s
deposit insurance premia do not completely account for the risk the loan adds. Still, as already
noted, for many banks, insurance premia may not accurately reflect their riskiness.

J. R. Walter: Banking and Commerce
3.

21

PROTECTIONS NEEDED IF THE WALL COMES DOWN

The previous section describes the hazard from corporate combinations between banks and commercial firms, and argues that the pertinent hazards arise
from (1) loss shifts to protect bank holding company reputation, that is, shifts
meant to hide the loss; and (2) loss shifts that take advantage of limited liability, allowing shareholders to avoid the loss by imposing them instead on
creditors or on the FDIC. In either case, economic efficiency can be diminished
and the loss can end up with taxpayers. One means of addressing the hazards
is to prohibit banking/commerce combinations; in other words, maintain the
legislative status quo. Alternatively, legislators may decide that the benefits
of combinations are worth bearing some danger of loss shifts. If legislators
took this latter view, what types of protections could they employ to reduce
the frequency of loss shifts into banks and thereby possibly to the FDIC or
taxpayers? Already in place are the firewalls established by Sections 23A and
23B of the Federal Reserve Act. These statutory provisions limit transactions
between banks and their affiliates. The firewalls are enforced by regular (once
every year or year and a half) supervisory examination and by the threat of
penalty if violations are discovered.
Beyond these current protections, which apply to any affiliations, including any commercial affiliations that might be allowed in the future, supervisors
might wish to mimic the types of limitations uninsured creditors would impose on risky affiliations. As noted earlier, one can expect uninsured creditors
to penalize the firm they fund (their debtor) and thereby potentially prevent
an affiliation that would tend to lend itself to loss shifts. If supervisors are
to mimic creditors’ actions, they will restrict the types of firms that banks
can affiliate with to those least likely to produce loss shifts in the first place.
In other words, supervisors would only allow banks to affiliate with healthy
commercial firms possessing strong capital at the time of affiliation.
While at the time of acquisition a new commercial affiliate may be strong,
its health could deteriorate or it could take on undue risks. If uninsured
creditors find that their debtor’s affiliates are suffering losses or assuming
risky endeavors, thereby increasing the chance of losses that might be shifted
to their debtor, they would demand higher interest payments to compensate
for their added risk. So creditors can be expected to monitor carefully the
health of their debtor’s affiliates. The Federal Reserve mimics private creditors
by performing such monitoring (called umbrella supervision by the Fed) of
holding companies owning a bank and a securities or insurance company,
under provisions specified in the Gramm-Leach-Bliley Act of 1999. Umbrella
oversight might well be desirable for combinations of banks with commercial
firms, but could be more difficult than umbrella oversight of financial firms, for
reasons discussed presently. If monitoring reveals that the bank’s commercial
affiliate has increased its risk, supervisors could impose a monetary cost on
the bank through the current mechanism by which insurance premia are set.

22

Federal Reserve Bank of Richmond Economic Quarterly

Because umbrella oversight of commercial firms may be more difficult
than oversight of financial firms, any legislation that might remove the wall
could add additional protections beyond those found in the Gramm-LeachBliley Act. For example, such legislation could also limit the size of commercial affiliates to those no larger than a fraction of the size of the bank. Such
a limit could be beneficial because as noted in an earlier section, shifts large
enough to sink the bank are most likely to derive from commercial affiliates
that are large relative to the size of the bank affiliate.

Firewalls
The 23A and 23B firewalls are intended to stop exactly those loss shifts that
present hazards for bank/commercial firm affiliations. Yet there have been several cases in which shifts have caused bank failures, regardless of firewalls.
Additionally, the firewalls have not been tested in a period of widespread affiliations involving nonbanks large enough to produce dangerous losses. Consequently, while in principle firewalls should prevent loss shifts, supervisors
will probably wish to take further protective steps if the banking/commerce
wall is removed.
Since 1933, banks have been protected against shifts of losses from affiliates by firewalls. Firewalls are found in Sections 23A and 23B of the Federal
Reserve Act and apply to all banks.14 They limit and place controls on transactions between banks and their affiliates.15 For example, the 23A firewalls
limit transactions, such as loans and asset purchases, between a bank and any
individual affiliate to 10 percent of the bank’s capital, and with all of its affiliates in total to 20 percent of the bank’s capital. The firewalls operate only in
one direction—they prevent transactions that might shift affiliate losses to the
bank, but do not prevent transactions that might shift bank losses to affiliates.
For instance, the firewalls prohibit loans to affiliates beyond 10 percent of
bank capital, but not the reverse—loans to the bank by affiliates. They also
require that purchases of affiliate assets by the bank be on terms at least as
favorable to the bank as market terms. In contrast, the nonbank affiliate can
purchase assets from the bank on terms unfavorable to the affiliate. Penalties
for firewall violations can be quite severe, extending to significant monetary
penalties imposed on banks and their managers and directors.
Bank failures caused by the shifts that firewalls were designed to prevent
have been infrequent, but there were at least two, one of which was quite large.
The 1955 Senate Report on the Bank Holding Company Act noted that “no
widespread abuse of this nature [loss shifts] has been brought to the attention
14 More specifically, Sections 23A and 23B apply to all insured banks and savings institutions.
They are found at 12 U.S.C. 371c and 12 U.S.C. 371c-1, respectively.
15 See Walter (1996) for a discussion of firewalls.

J. R. Walter: Banking and Commerce

23

of [Congress]” (U.S. Code: Congressional and Administrative News [1956,
2486]). The House Report on the Act did discuss one case, that of the 1953 failure of First State Bank of Elmwood Park, Illinois, which resulted from shifts
of bad loans from a nonbank loan company to its affiliate bank—apparently to
take advantage of limited liability protections (U.S. House [1955, 18–19]).16
Similarly, a 1983 study of the causes of bank failures for the previous ten years
found only one case out of 120 failures caused by transactions between a bank
and its nonbank affiliates. Still, this case, the failure of Chattanooga-based,
$461 million Hamilton National Bank in 1976, was the third largest in U.S.
history up to that time (Walter [1996, 23]).
Historically, then, firewalls have proven less than perfectly impervious.
Further, until recently, affiliations were quite limited, offering few opportunities to put the firewalls to the test. Bank holding companies were, for the most
part, restricted to owning nonbank financial firms that conducted activities that
were similar to banking. Until the Gramm-Leach-Bliley Act was enacted in
1999, banking companies were prohibited from broad securities and insurance
powers. Because of the limitations on the types of nonbank firms that bank
holding companies could own, nonbanks have typically been much smaller
than their bank affiliates. Since they were smaller, they were unlikely to be
capable of producing losses large enough to sink affiliated banks.

Due Diligence prior to Affiliation
Since the effectiveness of firewalls is uncertain, care must be taken to ensure that bank holding companies do not acquire especially risky nonbanks.
Currently, supervisors evaluate the nonbank’s financial health as part of their
review of applications from bank holding companies to acquire nonbanks.17
They conduct analyses similar to due diligence analyses performed by investment companies for unregulated acquirers. Supervisors look for many of the
same signals of problems that a creditor would, such as excessive debt and
weak earnings performance. Similar analyses would be necessary for bank
holding company acquisitions of commercial firms, should the wall come
down.
Still, it might seem that such analysis of commercial firms would be expensive for bank supervisors, requiring the development of a very different
16 See FDIC (1953, 7–8) for details of the First State Bank case beyond those provided in
the House Report.
17 The Gramm-Leach-Bliley Act of 1999 allows financial holding companies to dispense with
applying for supervisors’ approval of many nonbank acquisitions. Financial holding companies
simply notify supervisors of the acquisition, within 30 days of the acquisition (Spong [2000, 157]).
Therefore, acquisitions of nonbanks by financial holding companies often do not involve a preacquisition review of the nonbank’s financial health. For acquisitions by bank holding companies
that have not chosen, under rules specified by Gramm-Leach-Bliley, to become financial holding
companies, such reviews still occur.

24

Federal Reserve Bank of Richmond Economic Quarterly

skill set. Bank supervisors who specialize in application review are experienced in examining the health of financial firms, not of commercial firms.
Yet other supervisory employees—those who review banks’ loans for their repayment prospects—are practiced in analyzing commercial firms, since most
large bank loans go to such firms. Today’s bank examiners also engage in
industrywide analysis as part of their review of syndicated lending to large
commercial firms. Therefore, these skills might be brought to bear fairly
cheaply.
Under current procedures, if the supervisory review of the firm to be
acquired turns up potential risks, the supervisor can deny the application or
require that the risk be ameliorated. An application review of acquisitions of
commercial firms would likely include the same options.
Beyond these procedures, supervisors might add another requirement,
because analyses of commercial firms could be more difficult than analyses
of financial firms. Shifts of losses large enough to sink the bank and take
advantage of limited liability are most likely to occur when the commercial
firm is large relative to its bank affiliate. Consequently, supervisors might
also limit the size of commercial firms acquired to a fraction of the size of
affiliated banks. Doing so would reduce the chance of bank-sinking or bankendangering loss shifts. Such a requirement is not unprecedented, as relative
size limits were imposed on bank holding company merchant-banking acquisitions under provisions of Gramm-Leach-Bliley.

Umbrella Supervision
While careful analysis of potential affiliates might prevent bank holding companies from purchasing troubled or initially risky commercial firms, problems
at a commercial firm could arise well after its acquisition. Because of concern
for this possibility, supervisors may wish to maintain ongoing oversight of the
health of banks’ commercial affiliates. The aim of such oversight is to determine whether the commercial affiliate has suffered losses or is expanding
its riskiness. If supervisors find losses or heightened riskiness of commercial affiliates, they could indirectly impose a monetary penalty on the bank
by lowering its supervisory rating. All banks are graded by supervisors on
their financial health, riskiness, and management expertise. When a bank’s
grade (supervisory rating) declines, its insurance premiums can rise. Beyond
this monetary penalty, when commercial affiliates suffer losses or increases in
riskiness, supervisors might also watch more carefully for loss shifts (i.e., firewall violations). Further, they might even prohibit all transactions between
the bank and its troubled commercial affiliate. In doing so, the supervisor
mimics the monitoring that bank creditors would be expected to perform in
the absence of deposit insurance.

J. R. Walter: Banking and Commerce

25

Oversight of this sort currently occurs under provisions of the GrammLeach-Bliley Act, which make the Fed the umbrella supervisor of all financial
holding companies. In this role, the Fed is to ensure that problems in a
securities or insurance affiliate do not endanger the bank. For information
on the health of securities and insurance affiliates, the Fed relies on financial
reports from the Securities and Exchange Commission and state insurance
commissioners. In some cases the Fed will participate in examinations of
insurance companies performed by insurance commissioners. One main point
of its umbrella oversight is to ensure that bank resources are not being shifted
to nonbank affiliates. In the extreme, the Gramm-Leach-Bliley Act gives the
Fed the authority to require that nonbank affiliates are divested. Divestiture
provides the ultimate prohibition on loss-shifting transactions.
Umbrella oversight of commercial firms may be more difficult than oversight of securities and insurance firms. Securities and insurance firms already
face strict regulation by agencies with long-standing experience as supervisors. Moreover, insurance companies receive regular examinations for financial health. Most commercial firms are less regulated and are not subject
to examination by governmental supervisors. Developing such processes for
commercial firms affiliated with banks could be quite expensive for an umbrella supervisor of combined bank/commercial firms and could impose large
regulatory costs on the combined firms themselves.
Nevertheless, public firms—those whose securities trade in public
markets—must release a great deal of financial information. Such information could provide much of the data necessary to judge financial health.
While publicly available information may be less accurate and complete than
that typically available to bank regulators, who have the power to require the
release of any additional information they may deem useful, it is the set of
information private investors rely upon when deciding whether to invest.18
Therefore, for unregulated commercial firms, umbrella supervisors could rely
upon publicly available information to a significant extent. Additionally, in
passing the Sarbanes-Oxley Act, enacted in July 2002, legislators attempted
to enhance the reliability of disclosures made by publicly traded companies.

4.

CONCLUSION

Clearly, firms benefit from combinations of commercial and financial units,
since for years they have chosen to mix them. Yet many experts have maintained that banking and commerce should remain separate. Two predominant
reasons for maintaining the separation are concerns with conflicts of interest
and the proliferation of monopoly. The most credible reason—indeed, one
18 Of course, if private investors believe losses suffered by their bank-affiliated commercial
firms can be shifted, then they have less reason to demand accurate accounting information.

26

Federal Reserve Bank of Richmond Economic Quarterly

that poses a significant hazard from combining banking with commerce—is
that such affiliations could provide, at least under certain circumstances, incentives for loss shifts. While it turns out those circumstances are somewhat
limited, they are not inconsequential.
Loss shifts can impose costs on taxpayers and waste resources. If losses
are shifted from commercial firms to affiliated banks, taxpayer-funded bailouts
may result. If creditors become convinced that firms affiliated with banks
can shift losses to insured banks, then these firms will enjoy below-market
borrowing costs. Below-market funding means that too many resources will
flow to bank-affiliated firms. If so, productivity and financial market efficiency
are diminished; in other words, scarce resources are wasted.
Nonetheless, since the Gramm-Leach-Bliley Act of 1999 allowed securities and insurance firms to affiliate with banks, potentially producing the same
loss shifts that commercial affiliation might engender, why not also allow
commercial affiliations? One reason that legislators might prefer not to open
that door is that commercial firms are largely unregulated so the demands on
supervisory resources are likely greater when protecting against shifts from
largely unregulated commercial firms.
On the other hand, if legislators decide that the benefits of banking/
commerce combinations could outweigh the hazards, what means of protection might they employ to minimize them? Several come to mind, including
(1) careful analysis of the financial condition of commercial firms that bank
holding companies wish to acquire, prior to acquisition; (2) the maintenance
of firewalls to prevent loss shifts; and (3) umbrella supervision to provide the
means of reducing the hazard. In addition to these means, the requirement
that commercial firms be significantly smaller than any banks they affiliate
with offers further protection. Size limits are likely to be valuable since a
commercial firm is unlikely to produce a loss large enough to threaten a much
larger bank affiliate.

APPENDIX:

BACKGROUND ON THE STATUTES
FORMING THE WALL

National Bank Act of 1864
The National Bank Act restricts the opportunities for national banks to undertake commercial activities. National banks are those chartered and regulated by the U.S. Treasury Department’s Office of the Comptroller of the
Currency. The Act states that “a national banking association shall. . . have
power to. . . exercise. . . all such incidental powers as shall be necessary to carry

J. R. Walter: Banking and Commerce

27

on the business of banking; by discounting and negotiating promissory notes,
drafts, bills of exchange, and other evidences of debt; by receiving deposits;
by buying and selling exchange, coin, and bullion; by loaning money on personal security” (12 U.S.C. 24). While courts and the Comptroller have, over
the years, wrangled over the meaning of the business of banking clause, courts
have generally taken a fairly conservative view of activities that might qualify.
As decided in an influential court ruling, for example, banks are generally
limited to conducting businesses that are functionally interchangeable with
traditional banking services (M&M Leasing Corp. v. Seattle First National
Bank, 563 F.2d 1377, 1383 [9th Cir. 1977] as cited by Halpert [1988, 487]).
In sum, under the Act courts have allowed national banks to engage in businesses similar to banking but not other commercial activities. This restriction
of powers extends not only to activities of banks, but to activities conducted
by subsidiaries owned by banks (Halpert [1988, 486]).19

State Laws and FDICIA
For state-chartered banks, the banking/commerce wall is constructed of a mix
of elements from state laws, the Federal Deposit Insurance Corporation Improvement Act of 1991 (FDICIA), and the National Bank Act. State banking
statutes typically set limits on the nonbank activities of state banks and their
subsidiaries similar to the limits on national banks (Spong [2000, 37–41]).
Over the years a number of states have authorized activities beyond those
allowed national banks. Yet state banks’ opportunity to expand further than
the activities allowed under the National Bank Act was largely ruled out by
the FDICIA. Specifically, Section 24 of the Federal Deposit Insurance Act as
amended by the FDICIA prohibits insured state-chartered banks from engaging in any activities impermissible for national banks unless the FDIC rules
that such activities pose no threat to the deposit insurance fund (sec. 303 of
Public Law 102-242).

Bank Holding Company Act
Enacted in May 1956, the Bank Holding Company Act was based on the
view that “bank holding companies ought to confine their activities to the
management and control of banks.” Legislators appear to have been motivated
by two concerns. First, that conflicts of interest might arise if one company
19 Halpert (1988, 497) argues that it was of minor significance to Congress whether banks
engaged in nonbank activities when writing this language of the National Banking Act. He maintains that Congress “never affirmatively required banks to stay out of nonbanking business,” but
“[r]ather, subsequent interpretations of the statute by comptrollers of the currency and various courts
provided its restrictive cast.”

28

Federal Reserve Bank of Richmond Economic Quarterly

owned both a bank and a commercial firm. For example, such a conflict arises
when a bank receives a request for a loan from one of its commercial affiliate’s
competitors. Second, though the legislative history is less clear on this point,
legislators appear to have also been worried that combinations might lead to
the growth of monopoly power.
To address these concerns, the Act restricted bank holding companies such
that they “would no longer be authorized to manage and control nonbanking
assets unrelated to the banking business” (U.S. Code: Congressional and Administrative News [1956, 2484, 2492]). At the time the Act was passed, banking companies were growing rapidly through mergers. In a few cases these
companies included nonbanking businesses. The widest-ranging example was
found in Transamerica Corporation. It combined in one firm, banking, insurance, and a relatively small amount (as a percentage of Transamerica’s total
assets) of metals manufacturing and fish processing (Halpert [1988, 498]).
The Act required that companies wishing to purchase a bank first seek
approval from the Board of Governors of the Federal Reserve System. Further, the Act prohibited the Board from approving purchases by companies engaged in activities that were not closely related to banking, thereby prohibiting
commercial companies such as manufacturers from purchasing banks. Commercial firms like Transamerica that owned banks were given several years
in which to divest either the bank or alternatively their commercial activities. Through the next forty years the Board developed a list of activities
that would be considered closely related, excluding activities most observers
would consider commercial.
In 1999, the Gramm-Leach-Bliley Act was enacted. It added securities underwriting and dealing as well as insurance to the list of activities in
which banks—through bank-owned subsidiaries—and bank holding companies could engage. Until that time, banks and their subsidiaries and bank holding companies had been prohibited from the securities business by the 1933
Glass-Steagall Act.20 Insurance activities were likewise highly restricted before Gramm-Leach-Bliley by the Bank Holding Company Act and other laws.
The Glass-Steagall Act’s separation of securities activities from banking
was driven by legislators’ concerns over conflicts of interest, excessive stock
market speculation by bank-owned securities firms, and threats to the health of
banks from securities activities. Likewise the Bank Holding Company Act’s
separation of banking and insurance was part of that law’s general separation of
banking from nonbank activities, driven by concerns over conflicts of interest
and monopoly power. By the time the Gramm-Leach-Bliley Act was passed,
legislators and other observers had various reasons for removing the walls
that separated banks from securities and insurance activities. These reasons
20 Bank holding companies began to engage in limited securities activities starting in 1987
through a loophole in Glass-Steagall.

J. R. Walter: Banking and Commerce

29

fall into three categories. First, there is little evidence of conflicts of interest
or other problems when banks were combined with nonbank firms. Second,
market developments, such as growing competition in banking markets, had
rendered these problems less important by the 1990s. Third, the concerns
could be dealt with effectively by regulating the combined firms.
Ultimately, the Gramm-Leach-BlileyAct specified that only healthy, fairly
low-risk bank holding companies were to be allowed to undertake this broader
array of financial activities. Those that do so are called financial holding companies. Further, the Act allows these new financial holding companies to engage in merchant banking, whereby under certain conditions financial holding
companies may purchase the equity of (in other words, become owners of) any
type of corporation, commercial or otherwise. Financial holding companies’
merchant banking subsidiaries are restricted to holding the equity of firms
for a limited period of time and are prohibited from active management of
the firms. Beyond securities and insurance, Gramm-Leach-Bliley allows the
Board of Governors, in conjunction with the Treasury Department, to also authorize financial holding companies to undertake additional activities that are
“financial in nature” or “incidental to financial activities.” It also authorizes
the Board to approve activities that are “complementary to a financial activity.” So the Gramm-Leach-Bliley Act expands the activities of bank-owning
companies beyond those previously allowed by the Bank Holding Company
Act to include most financial activities, but leaves in place the wall between
banking and commerce.
Loopholes in the Bank Holding Company Act Section
of the Wall

Loopholes have been employed to allow banking/commerce combinations, at
least to a limited extent. The unitary thrift loophole, closed by the GrammLeach-Bliley Act in 1999, was one such opening. Through it, companies owning only one thrift (thus the phrase unitary thrift) could also own commercial
firms. The loophole existed because thrift institutions (meaning primarily
savings and loans, and savings banks) are not covered by the Bank Holding Company Act, which prevents banking/commerce ties. Instead, thrifts
are regulated under the Savings and Loan Holding Company Amendments
of 1967, which allow commercial activities in unitary thrift holding companies (Seidman [1998, 7]). Gramm-Leach-Bliley closed the loophole though
it grandfathered existing unitary thrift holding companies, allowing them to
continue to engage in commercial activities.
An additional loophole was partially closed in 1987, but remains open to a
limited degree. Before 1987, the Bank Holding Company Act defined a bank
as a firm that both offered demand deposits (a type of checking account) and
made commercial loans. This definition prevented commercial firms from
owning a typical bank, which offers both demand deposits and commercial

30

Federal Reserve Bank of Richmond Economic Quarterly

loans. Nevertheless, commercial firms could form a bank that did not offer
one or the other. By doing so, commercial firms could own banks that did
not fall within the Bank Holding Company Act definition of a bank and could
circumvent the Act’s prohibition of mixing banking and commerce. These
banks, known as nonbank banks, did not fit the Act’s definition of a bank
but did offer most banking services. A number of firms established nonbank
banks, both as a means of combining banking and commerce and as a means
of banking across state lines, which was difficult until the 1990s. In 1987,
Congress closed the loophole by tightening the definition, but allowed states
with existing laws authorizing the chartering of industrial loan corporations
(a type of nonbank bank that funds itself with insured deposits but does not
offer demand deposits) to continue to charter these ILCs. Several states had
such laws as of 1987. This option remains in force as a means of combining
banking and commerce in these states.

REFERENCES
Bank Holding Company Act. 1956. U.S. Code. Title 12, sec. 1843c.
Board of Governors of the Federal Reserve System. 1974. “Order Denying
Investment in Allstate International S.A., Zurich, Switzerland.” Federal
Reserve Bulletin 60 (July): 517–19.
. 1981. “Order Approving Proposed Bookkeeping and Data
Processing Activities and Denying Proposed Finance, Loan Servicing,
Leasing and Insurance Activities.” Federal Reserve Bulletin 67 (May):
449–51.
.1987. “Order Approving Applications to Engage in Limited
Underwriting and Dealing in Certain Securities.” Federal Reserve
Bulletin 73 (June): 473–508.
Federal Deposit Insurance Corporation (FDIC). 1953. Annual Report.
Washington, D.C.: FDIC.
Federal Deposit Insurance Corporation Improvement Act (FDICIA). 1991.
U.S. Public Law 102-242, sec. 303. 102nd Cong., 1st sess.
Federal Reserve Act. 1913. U.S. Code. Title 12, secs. 371c, 371c-1.
Halpert, Stephen K. 1988. “The Separation of Banking and Commerce
Reconsidered.” Journal of Corporation Law 13 (Winter): 481–533.
Krainer, John. 2000. “The Separation of Banking and Commerce.” Federal
Reserve Bank of San Francisco Economic Review: 15–25.

J. R. Walter: Banking and Commerce

31

Macey, Jonathan R., and Geoffrey P. Miller. 1992. Banking Law and
Regulation. Boston: Little, Brown and Company.
Morgan, Donald P. 2000. “Rating Banks: Risk and Uncertainty in an Opaque
Industry.” Federal Reserve Bank of New York Staff Report 105 (April).
National Bank Act. 1864. U.S. Code. Title 12, sec. 24.
Owens, Raymond E. 1994. “Commercial Real Estate Overbuilding in the
1980s: Beyond the Hog Cycle.” Federal Reserve Bank of Richmond
Working Paper 94-6 (May).
Seidman, Ellen. 1998. Statement on Financial Modernization before the
Committee on Banking, Housing, and Urban Affairs. U.S. Senate. 105th
Cong., 2nd sess. 25 June.
Spong, Kenneth. 2000. Banking Regulation: Its Purposes, Implementation,
and Effects, 5th ed. Kansas City, Mo.: Federal Reserve Bank.
U.S. Code: Congressional and Administrative News. 1956. St. Paul, Minn.:
West Publishing Company.
U.S. General Accounting Office. 1987. “Bank Powers: Insulating Banks
from the Potential Risks of Expanded Activities.” Washington, D.C.:
U.S. GAO. April.
. 1997. “Bank Oversight: Few Cases of Tying Have Been
Detected.” Washington, D.C.: U.S. GAO. May.
U.S. House. 1955. Report to Accompany H.R. 6227. 84th Cong., 1st sess.
H.R. 609.
Walter, John R. “Firewalls.” 1996. Federal Reserve Bank of Richmond
Economic Quarterly 82 (Fall): 15–39.
. 1998. “Can a Safety Net Subsidy Be Contained?” Federal
Reserve Bank of Richmond Economic Quarterly 84 (Winter): 1–20.
Walter, John R., and John A. Weinberg. 2002. “How Large Is the Federal
Financial Safety Net?” The Cato Journal 21 (Winter): 369–93.
Weinberg, John A. 1996. “Tie-in Sales and Banks.” Federal Reserve Bank of
Richmond Economic Quarterly 82 (Spring): 1–19.

Unemployment Insurance
and Personal Bankruptcy
Kartik Athreya

P

ersonal bankruptcy allows households to stop or delay the repayment of
debts. In so doing, bankruptcy provides a form of insurance to households. In particular, bankruptcy allows households some flexibility in
timing repayments in a way that allows for sudden unforeseen contingencies.
As an implicit form of insurance, bankruptcy may augment, substitute for, or
even limit other forms of insurance. Conversely, the presence of other forms
of insurance against life’s vicissitudes may enhance or limit the usefulness of
bankruptcy. In this article, I investigate the interaction between one of the
largest social insurance schemes, the U.S. unemployment insurance system
(UI), and the personal bankruptcy system.
An overwhelmingly large proportion of those filing for bankruptcy (over
two-thirds) have recently experienced a job disruption (Sullivan et al. [2000]
and Domowitz and Sartain [1999]). Further, Cochrane (1991) finds that prolonged spells of unemployment are poorly insured and therefore result in large
drops in consumption levels. How does the level of unemployment insurance
available to workers affect the benefits of bankruptcy protection? Conversely,
how do the benefits of bankruptcy alter the benefits generated by UI? Lastly,
how does the presence of bankruptcy alter the consequences of scaling back
unemployment insurance?
My findings are as follows: First, in the benchmark economy, introducing
bankruptcy under even low UI replacement ratios lowers welfare. Second,
reducing the UI replacement ratio increases bankruptcy rates.1 Additionally,
The author would like to thank Tom Humphrey, Ned Prescott, John Weinberg, Roy Webb, and
seminar participants at Colgate University and Hamilton College. The views expressed herein
are those of the author and do not necessarily represent the views of the Federal Reserve
Bank of Richmond or the Federal Reserve System.
1 One reason why this should not be seen as obvious is that penalties for bankruptcy involve
ejection from credit markets. Therefore, lowering the replacement ratio hurts bankruptcy filers more
than before, which may imply fewer, not greater, annual filings.

Federal Reserve Bank of Richmond Economic Quarterly Volume 89/2 Spring 2003

33

34

Federal Reserve Bank of Richmond Economic Quarterly

reducing the UI replacement ratio worsens consumption smoothing less when
bankruptcy is allowed than when it is not. However, while welfare falls slightly
with the replacement ratio, the fall is nearly independent of bankruptcy law.
Third, bankruptcy lowers asset trade, which in turn implies a more equal longrun distribution of wealth (as fewer households hold either very low or very
high asset levels to deal with income shocks). However, asset trading behavior is not affected greatly by changes in the UI replacement ratio. Fourth,
UI is more important than bankruptcy: if society must choose either UI or
bankruptcy, it should choose UI. Last, bankruptcy’s role in providing insurance is clearly dependent on the existing social safety net. In summary, unemployment insurance appears to materially affect the desirability of bankruptcy
protection, but allowing bankruptcy does not, in the benchmark economy, alter
the consequences of scaling back UI.
The environment here is an extension of the environment studied inAthreya
(2002b), augmented to include unemployment. Athreya (2002b) examines the
welfare implications of recent “means-testing” proposals. The present work
is perhaps closest to the work of Livshits et al. (2002) and Fisher (2002). The
work of Fisher (2002) is the first empirical study of the effects of public insurance on the personal bankruptcy decision. With respect to bankruptcy, this
work is also related to recent research of Chatterjee et al. (2001) and Li and
Sarte (2002). With respect to positive analyses of unemployment insurance
and its consequences, the work is related to, but simpler than, the models of
Hansen and Imrohoroglu (1992) and Alvarez and Veracierto (2001). The two
preceding articles study unemployment insurance in general and the effect of
severance payments on job security, respectively.
The sudden fall in earnings associated with a layoff or firing or an inability
to continue working due to illness has long been cited by bankruptcy scholars
as an important correlate of bankruptcy.2 Thus, it stands to reason that the
treatment received by those who become separated from their employers will
influence their decision whether or not to file for bankruptcy. I turn now to
a simple dynamic general equilibrium model of consumption and savings in
the presence of some uninsurable income risks, including the risk of losing
one’s job. To simplify matters, I abstract from production decisions as well
as the impact of moral hazard in increasing the costs of administering an
unemployment insurance system. In ongoing research (Athreya [2002a]), I
pursue a more complete analysis to incorporate moral hazard and production.

1. THE MODEL
Bankruptcy allows a borrower to essentially design a state-contingent repayment plan, whereby repayment is made only when outcomes for the borrower
2 Of course, this is no more causal than is having too little income or too much debt, given
one’s income stream.

K. Athreya: Unemployment Insurance and Personal Bankruptcy

35

are relatively good. In this sense, the amount of a household’s income dedicated to loan repayment can be varied, allowing it to apply limited income in a
difficult period towards consumption rather than debt service. Unemployment
insurance and antipoverty programs, conversely, act directly on the income of
the household and help it remain above a threshold. Both of these programs
can help households insure themselves within a period against uncertain job
or health prospects. However, both programs must be paid for.
Allowing bankruptcy implies paying more for loans, as households are
also purchasing the right to suspend or completely avoid repayment, subject
to penalties. The high rate on loans also means that as households attempt
to avoid borrowing, each saves so much that the return to savings may fall
relative to an economy without bankruptcy. In turn, this fall mutes the effectiveness of savings to carry consumption across periods. Unemployment
insurance, for its part, must be paid for via (possibly distortionary) taxes. Furthermore, as is well known, UI may introduce inefficiency, as both the effort
expended by currently employed households and the job search efforts of currently unemployed households may fall. Moreover, a major penalty for filing
for bankruptcy is exclusion from credit markets. In contrast, while UI may
directly lower the need for borrowing and subsequent bankruptcy, generous
UI makes exclusion from credit markets less painful. Thus, while bankruptcy
and UI act in different ways, the presence of each is likely to affect the other.

Preferences and Endowments
Individuals maximize the present value of expected lifetime utility, given by
∞

βt

E0
t=0

ct1−α − 1
,
1−α

(1)

where E0 is the expectations operator, conditional on time 0 information,
β ∈ (0, 1) is the discount factor, c is consumption, and α is the measure of
both risk aversion and the desire for intertemporal consumption smoothing.
A full description of the household’s optimization problem will be given after
more notation is introduced.
Consumers in this market, intended to represent U.S. households, are
assumed to be risk-averse price takers. They face uncertain labor incomes
and other uninsurable idiosyncratic risks. The economy is composed of many
long-lived households. At the beginning of each period, all households receive
a random level of labor income that depends on their employment status.
Households in the economy retain employment in each period with probability
ρ and are subject to the risk of losing employment in a given period with
probability (1 − ρ). Once employment is lost, regaining employment occurs
with probability ξ . An unemployed worker receives unemployment insurance

36

Federal Reserve Bank of Richmond Economic Quarterly

in only the first period of unemployment, that is, when the worker is newly
unemployed. In subsequent periods, there is a subsistence level of income
given to households. The endowment structure for unemployed households
is meant to reflect the current practice of the use of a flat “replacement ratio”
and the limited length of UI benefits in the United States. Newly unemployed
households receive θ Y , where Y is mean labor income and θ ∈ [θ , 1] the
replacement ratio. After the first period of unemployment, households, if
unemployed, will receive the subsistence transfer of Ymin > 0.3
Given the exogenously imposed flow of households out of unemployment,
the replacement ratio for UI benefits θ, and the long-run average employment
rate µe , it is easily shown that the per-period lump-sum tax ηu necessary to
finance the UI system is given by (1 − ρ)µe θY .4
The endowments of employed households are random and cross-sectionally
independent but are serially dependent. Agents are identical ex ante in terms
of expected income, assets, and consumption. When employed, the after-tax
endowment of a household in period t can take two values, Y = yl and Y = yh ,
where the subscripts h and l denote high and low labor income, respectively,
such that yl < yh .5 Defining unemployment as a separate state for the endowment process is what allows for an analysis of how UI benefits interact with
bankruptcy law.
There is a transition function over the income of employed households
whereby P (y = yl |y = yl ) = pll and P (y = yh |y = yh ) = phh . That is, pll
is the probability that the labor income shock remains low in the next period,
given that it is low in the current period. Similarly, phh is the probability that
the labor income shock remains high in the next period, given that it is high in
the current period. The assumption of serially dependent income introduces
anticipation effects in asset holdings and default behavior and determines the
effectiveness of using assets to smooth consumption. The parameters of the
income process will be chosen to be broadly consistent with post-transfer
income variability in U.S. data.
3 To focus on the interaction of bankruptcy and explicitly financed unemployment insurance,
I avoid tracking the collection of taxes with which subsistence income payments are made. When
analyzing changes in bankruptcy law, Ymin will remain fixed, so there is no harm in treating it
as an endowment.
To keep matters simple, I did not specify UI replacement to depend on previous income.
Doing so would entail tracking households flowing into employment separately, which increases
the cost of computing solutions. Moreover, as job loss is exogenous with respect to income, the
average household flowing into unemployment will have Y as the previous period’s labor income.
4 This is a simple example of a “bathtub” model of unemployment, whereby the exogenous
flows into and out of employment are set such that there is a constant level of employed (and
unemployed) households in the economy. See Ljungqvist and Sargent (2000). The flow into unemployment is given by (1 − ρ)µe , and the cost of insurance payments to each household is θ Y .
Therefore, per capita taxes, ηu , must satisfy ηu = (1 − ρ)µe θY .
5 I define endowments as “after-tax” income for simplicity of notation and exposition.

K. Athreya: Unemployment Insurance and Personal Bankruptcy

37

Assets
Agents may save using risk-free private bonds or risk-free government debt
and may borrow on an unsecured credit market. Government debt is incorporated both for descriptive accuracy as well as to avoid artificially constraining
households to the use of private borrowing and lending alone.6 Household
borrowing is subject to a liquidity constraint, and households may default
on previously acquired debt. The stock of private risk-free debt is issued by
diversified competitive financial intermediaries in order to finance loans to
households. The market for privately issued unsecured credit in the United
States is characterized by a large, competitive marketplace where price-taking
lenders issue credit through the purchase of securities backed by repayments
from borrowers. These transactions are intermediated principally by credit
card issuers. As the typical credit card contract is described by a fixed interest rate and credit line, the interest rates charged by credit card issuers may
be viewed as being set to cover the aggregate default rate rather than being
individually tailored for each account. Further, interest rates do not appear
to vary systematically with individual debt levels, even though the marginal
likelihood of default may change.7
There will be two prices quoted for assets: a loan rate, r l , for those
who borrow and a deposit rate, r d , for those who save. These two rates are
different, because with the bankruptcy option, a certain fraction of households
will default in equilibrium, and in order to break even, financial intermediaries
will have to charge higher interest rates on loans than they pay for deposits.
The stock of government debt is denoted D and is financed by a lump-sum tax
ηD = r d D on all households, where r d is the interest rate on risk-free savings
deposits.
The assumption that all debt is unsecured is less restrictive than it may
seem. The model best represents the section of U.S. households with little or
no collateral and higher than average labor income risk that rely on unsecured
debt to smooth consumption. Therefore, the welfare implications developed
here apply directly to a population most affected by bankruptcy reform. Additionally, Gropp, Scholz, and White (1997) argue that in many cases those
considering filing use unsecured credit to pay off secured debts and then discharge this debt in bankruptcy, thereby making the distinction between these
types of debt less clear in practice.
6 The stock of government debt per capita will, however, be held fixed throughout all the
policy experiments I conduct.
7 On the existence of competitive equilibrium in a model where interest rates on loans cover
average repayment rates, see Dubey et al. (2000).

38

Federal Reserve Bank of Richmond Economic Quarterly

Bankruptcy
Bankruptcy in the model will most closely resemble Chapter 7 “total liquidation” bankruptcy. If a household files for bankruptcy, its income and assets
become known to the credit market, and if it qualifies, its unsecured debt
is discharged but is then constrained for an uncertain period of time from
borrowing. Households may, however, save during this time. The principal
motivation for a random period of restricted credit access is that it reduces significantly the computational burden of solving the household’s optimization
problem. Specifically, the assumption allows one not to distinguish between
households on the basis of the length of their credit market exile.8 In each
period following a bankruptcy, a borrowing-constrained household remains
constrained with probability (1 − ψ). Therefore, the average time that a
household is constrained from borrowing and prohibited from filing again is
given by 1/(1 − ψ).
The Cost of Bankruptcy and Deadweight Loss

Bankruptcy involves three types of costs. First, as was just discussed, it results
in at least some exclusion from credit markets. Second, there are explicit time
costs arising from court dates and other legal proceedings. Finally, societal
disapproval or “stigma” may play a role (see Dubey et al. [2000]; Fay et al.
[1996]; and Gross and Souleles [2000]).
An important drawback of using bankruptcy to provide insurance is that
the penalties listed above typically do not involve a transfer of wealth from
debtors to anyone, let alone creditors. I denote by λ all costs of bankruptcy
beyond credit market exclusion. That is, λ represents the “deadweight” costs
of bankruptcy. I will set λ to match observed bankruptcy filing rates among
homeowners, given the current average length of credit market exclusion.

The Household’s Problem
At any point in time, households belong to one of two mutually exclusive
classes of credit market status and three mutually exclusive classes with respect to employment status. For credit status, households are either solvent or
constrained from borrowing. Solvent households are those that have full access to credit markets and have the option of filing for bankruptcy. Borrowingconstrained households are those that have filed for bankruptcy in the past but
have not yet been readmitted to credit markets.9 With respect to employment
8 For more on stochastic punishment spells in bankruptcy, see Athreya (2002b) and Chatterjee
et al. (2001).
9 Therefore, while the move from solvent to borrowing-constrained status is a choice for
households, the release from borrowing-constrained status is exogenous.

K. Athreya: Unemployment Insurance and Personal Bankruptcy

39

status, households are either employed, newly unemployed, or unemployed
for more than one period.
In each period, given their current income and beginning-of-period assets,
households must choose consumption, c, and asset holdings to carry forward
into the next period, denoted a . From the individual’s point of view, all saving
is risk-free and earns the same rate of return. Therefore, the household makes
no distinction between government debt and private bonds when choosing how
to allocate its savings. Depending on whether it chooses to be a net borrower
or lender, it faces either the net rate of interest on loans, r l , or deposits, r d ,
where r l > r d .
I restrict borrowing according to a household’s credit status as follows.
For solvent households, assets a must be greater than a S , a negative number
indicating that solvent households may borrow. Households that have filed
for bankruptcy face a more severe restriction than solvent households in their
ability to borrow. Their borrowing limit, denoted a B , therefore is given by
a ≥ a B , where a B > a S . Similarly, households that are constrained from
borrowing are also restricted in their borrowing, with a limit denoted a BC ,
whereby a ≥ a BC , where a BC > a S .
When a household is solvent, it must first choose whether or not to file.
It then chooses assets subject to the constraints for solvent or borrowingconstrained households, depending on its employment status and default decision. The current period state vector, conditional on credit status, is denoted
(e, a, y), indicating employment status, asset holdings, and current income,
respectively.
Current labor income is denoted y(e), where e denotes beginning-ofperiod labor market status. A worker’s employment status belongs to one of
three categories, that is, e ∈ {e0 , e1 , e2 }, where e0 denotes an employed worker,
e1 a newly unemployed worker, and e2 a worker who has been unemployed
for more than one period. The law of motion for labor income is simple. In
any period, an employed worker may lose his job with probability (1 − ρ). He
is then classified as “newly unemployed” and is eligible for UI benefits. In the
following period, he finds employment with probability ξ , in which case he
receives a (random) endowment of y(e0 ).10 If he fails to find employment in
this period, he is classified as “unemployed” and is therefore no longer eligible
for UI benefits and receives labor income Ymin > 0.
I denote the value of being solvent by V S , the value of not filing for
bankruptcy as W S , the value of filing for bankruptcy as W B , and the value
of being borrowing-constrained as V BC . The value of solvency is given as
follows:
10 This income is drawn from the conditional probability distribution of income, as if the
household had received income shocks while unemployed. This simplifies the analysis by avoiding
the use of a separate income process once released from unemployment.

40

Federal Reserve Bank of Richmond Economic Quarterly

V S (e, a, y) = max[W S (e, a, y), W B (e, a, y)],

(2)

W S (e, a, y) = max{u(c) + βEV S (e , a , y )}

(3)

where

s.t.
c+

a
≤ y(e) + a
1 + r d,l

(4)

s.t.
a ≥ aS .

(5)

When the household chooses to file for bankruptcy, it has its debt removed,
pays the nonpecuniary cost, λ, and then is automatically sent to the borrowingconstrained state, where it obtains value V BC . Therefore, the value of filing
for bankruptcy, W B , satisfies
W B (e, a, y) = max{u(c) − λ + βEV BC (e , a , y )}

(6)

s.t.
a
≤ y(e)
1 + rd

(7)

a ≥ aB .

c+

(8)

s.t.

To define V BC above, note that households in the borrowing-constrained
state face a lottery, whereby with probability ψ, they are returned to solvency
(i.e., they are free to borrow and default in the following period), and with
probability (1 − ψ), they are still restricted from borrowing or defaulting.
Thus, we have
V BC (e, a, y) = max{u(c) + ψβEV S (e , a , y ) + (1 − ψ)βEV BC (e , a , y )}
(9)
s.t.
a
c+
≤ y(e) + a
(10)
1 + rd
s.t.
a ≥ a BC .
I turn now to the definition of equilibrium in the model.

(11)

K. Athreya: Unemployment Insurance and Personal Bankruptcy

41

Equilibrium
The consumer choice problem above captures the decisions of a very large
number of households. However, given the absence of perfect income insurance, households that have received many bad income shocks are likely to find
themselves in debt, while those that have been lucky may have large levels of
savings. Their choices are governed by a decision rule, which, for a household
of type i, specifies asset holdings as a function of interest rates, employment
status, income, current assets, and borrowing constraints.
An equilibrium consists of a decision rule for each type of agent and
interest rates r l and r d such that four requirements are met. First, given
these interest rates, decision rules solve the optimization problem described
above for each type of household. Second, total economy-wide borrowing
by households equals total economy-wide saving. Third, the spread between
loan and deposit rates is such that financial intermediaries exactly cover their
costs, given the observed bankruptcy rate. Fourth, the payments to newly
unemployed households each period must be covered by tax revenues (i.e.,
the government runs a balanced budget while maintaining the stock of debt D).
In addition, I restrict attention to steady state equilibria where the bankruptcy
rate and the proportion of agents in the population with a given level of assets
are stationary, that is, the same at every date.

Welfare Measurement
The welfare criterion used here measures the percentage change in consumption, in all states and at all dates, that would make a household indifferent between living in an economy in which a given policy experiment prevailed and
one in which the benchmark setting prevailed. Let this increment/decrement
to consumption be denoted by φ. A negative value for φ implies that households are worse off, and a positive value implies the reverse. Multiplying φ
by mean household income then converts φ into a dollar measure of annual
welfare gains or losses per household.11
11 With the utility function used here, the welfare measure is given as follows. The desirability of outcomes will be evaluated according to the following expression:

=

V (x)dµ,

(12)

X

where V (x) is the maximal attainable utility from being in a given state x and µ is the long-run
stationary distribution (C.D.F.) of households across states. Therefore,
is the expected value
function of households over assets, income, and credit status. This is a utilitarian social welfare
function that weights all households equally. It measures ex ante welfare. I use this measure to
estimate the increment/decrement to consumption under a given bankruptcy policy, at all dates and
states, that makes households indifferent between the economy defined by the proposed bankruptcy
policy and the benchmark economy. I denote this increment/decrement φ. Let bench denote
benchmark welfare, and policy denote welfare under a proposed policy. Given the preferences
used here, φ will satisfy the following:

42

Federal Reserve Bank of Richmond Economic Quarterly

Beyond this measure of welfare, I will also examine the behavior of some
other statistics in assessing the interaction between bankruptcy and UI. Given
that changes in the replacement ratio alter the mean level of after-tax income
for households in the model, it is useful to have a measure of consumption
volatility that does not depend on average income, such as the coefficient of
variation (denoted c.v.). The c.v. will also be useful when exploring the role
of bankruptcy in altering the level of asset accumulation and decumulation.
To measure inequality, I use a traditional tool, the Gini Coefficient. Roughly
speaking, this coefficient measures the departure of a given distribution of
wealth, consumption, or income from a perfectly equal distribution. A Gini
of one indicates, for example, that the very richest household holds the entirety of wealth, while a coefficient of zero indicates that all households hold
exactly equal levels of wealth. A more disaggregated measure of inequality
is the distribution of income, consumption, and wealth by various percentiles,
which I also report below.

Parameterization
The model parameters are set to match observed bankruptcy rates under plausible levels of income shock persistence and volatility and are summarized in
Table 1. For brevity, rather than including a full discussion here, I refer the
interested reader to the details in Athreya (2002a, b).
With respect to unemployment insurance, I follow Hopenhayn and Nicolini (1997), who use the estimates of Meyer (1990). In particular, Meyer (1990)
finds that the average length of insured unemployment is thirteen weeks, with
a replacement rate of 66 percent and a 10 percent chance of reemployment
at the end of the spell. I therefore set the model period at thirteen weeks, set
θ = 0.66, and set ξ = 0.10. The credit limit is set by noting that median unsecured debt among bankrupt households in recent years has fluctuated between
one-fourth and one-half of annual median income (see Sullivan et al. [2000,
65–66, 122]). Credit card debt, to which the debt in the model corresponds
most closely, was approximately $9,500 in 1997, equal to U.S. median quarterly income (Sullivan et al. [2000]). Given the period of thirteen weeks, or
one quarter, I therefore set a s = Y . For simplicity, I set a B and a BC to zero.
An important parameter in the model with respect to bankruptcy is the
one governing credit market exclusion, ψ. While ψ is not easily observable,
lenders in the unsecured credit market still allow agents access to loan markets

φ=

1
policy +
(1−α)(1−β)
1
bench +
(1−α)(1−β)

1
1−α

− 1.

(13)

Under this criterion, φ > 0 implies that households are better off under a proposed policy than in
the benchmark case, and φ < 0 implies the reverse.

K. Athreya: Unemployment Insurance and Personal Bankruptcy

43

Table 1 Parameters
Parameter
β (annual)
α
i
phh
i
pll
ρ
µe
ξ
θ (benchmark)
yh
yl
Ymin
λ
as
a B , a BC

Value
0.914
2.00
0.74
0.74
0.006
0.943
0.10
0.66
1.25
0.75
0.40, 0.10
2.80
−Y
0

Source
Calibrated
Aiyagari (1994)
Heaton and Lucas (1997)
Heaton and Lucas (1997)
Calibrated
Alvarez and Veracierto (2001)
Meyer (1990)
Meyer (1990)
Heaton and Lucas (1997)
Heaton and Lucas (1997)
Calibrated
Huggett (1993); Sullivan et al. (2000)

following default or bankruptcy within a year or two. I set ψ = 0.25 such
that the average period of exile from credit markets is four model periods, or
one year.12 The level of income received by unemployed households after
unemployment benefits are exhausted, Ymin , is set in the benchmark case to
0.40, to provide 40 percent of median household income, as a proxy for the
various income support and transfer programs available to U.S. households.
This level amounts to $1,332 per household per month.13 Subsequently, Ymin
will be set to a much lower 0.10, or $333 per household per month, to examine
the role played by social insurance beyond unemployment compensation.
The parameter λ, which is the cost of bankruptcy in excess of credit market
restrictions, will be inferred by the level that it must take in order to match observed bankruptcy filing rates. In terms of bankruptcy rates, total nonbusiness
bankruptcy filings have been stable at roughly 1.3 million annually. Of these,
roughly 70 percent are Chapter 7, “total liquidation” bankruptcies, implying
an annual incidence of 0.9 percent.

2.

RESULTS

To begin, I define the benchmark case, against which policy experiments will
be compared.
12 In this model, exclusion from borrowing hurts the households without helping anyone else.
It is therefore a deadweight penalty and could have been left unmodeled by combining it with
the general nonpecuniary penalty, λ.
13 The transfers received by households beyond UI come from the major public assistance
programs in the United States: Supplemental Security Income (SSI), General Assistance, Medicaid,
and Temporary Assistance to Needy Families (TANF).

44

Federal Reserve Bank of Richmond Economic Quarterly

Table 2 Welfare Effects of Introducing Bankruptcy
θ
0.66
0.66
0.50
0.50
0.40
0.40

rl
4.39%
13.00%
4.21%
13.00%
4.06%
12.55%

rd
4.39%
2.57%
4.21%
2.35%
4.06%
2.31%

Bankruptcy
Rate
—
0.90%
—
0.99%
—
1.03%

Welfare
Change ($)
—
−$66.88
—
−$70.46
—
−$62.70

Utility
−37.43
−37.57
−37.49
−37.63
−37.52
−37.65

Definition 1 Throughout the analysis, the “benchmark” economy is defined
specifically to be the case where bankruptcy is allowed, the replacement ratio,
θ, is set at 0.66, and Ymin = 0.40.
I first study the consequences, when Ymin = 0.40, of introducing bankruptcy into a setting where unemployment is already partially insured. The clear
conclusion in this case is that bankruptcy protection is harmful, as seen in
Table 2. Introducing bankruptcy is damaging even when the unemployment
insurance system is very strict. The quarterly cost to the household of introducing bankruptcy ranges from $66.88, when θ = 0.66, to $70.46, when
θ = 0.50, to $62.70, when θ = 0.40. With respect to prices, I find that when
bankruptcy is introduced, the interest rate on savings falls, while the rate on
borrowing rises. For example, when θ = 0.66, r d falls from 4.39 percent to
2.57 percent, while r l rises sharply from 4.39 percent to 13.00 percent. Such
changes in interest rates are associated with worsened consumption smoothing, as the return to savings is low, while borrowing becomes very expensive.
On the other hand, the option of bankruptcy allows households new consumption smoothing possibilities. On net, however, welfare appears to suffer. The
welfare measure reported in Table 2 captures the change in welfare generated
by the introduction of bankruptcy, holding the replacement ratio fixed. This
result is summarized in Result 1.
Result 1 In the benchmark economy, introducing bankruptcy under even low
UI replacement ratios lowers welfare, increases interest rates on loans, and
reduces interest rates on savings.
Perhaps unsurprisingly, lower replacement ratios produce systematically
lower utility levels. For example, when bankruptcy is not allowed, the expected utility of households falls from −37.43 to −37.49 to −37.52 as θ
drops from 0.66 to 0.50 to 0.40 (see Table 2). The intuition here is simple. As
the replacement ratio falls, the income risk faced by households rises, leaving
more room for bankruptcy to be a useful form of implicit insurance.

K. Athreya: Unemployment Insurance and Personal Bankruptcy

45

Table 3 Effects of Lower UI Replacement Ratios
Panel A: Welfare Effects of Lower UI Replacement,
without Bankruptcy
θ

rl

rd

0.66
0.50
0.40

4.39%
4.21%
4.06%

4.39%
4.21%
4.06%

Bankruptcy
Rate
—
—
—

Welfare
Change ($)
—
−$23.13
−$40.68

Welfare Change ($)
Rel. to Benchmark
+$66.88
+$43.71
+$26.14

Panel B: Welfare Effects of Lower UI Replacement,
with Bankruptcy
θ

rl

rd

0.66
0.50
0.40

13.00%
13.00%
12.55%

2.57%
2.35%
2.31%

Bankruptcy
Rate
0.90%
0.99%
1.03%

Welfare
Change ($)
—
−$26.82
−$36.60

Because bankruptcy causes less harm when the replacement ratio is low
than when it is high, it appears that bankruptcy does play an insurance role.
To see this, consider Panel A of Table 3 for the results when bankruptcy is not
allowed. Welfare (relative to the case where θ = 0.66, and bankruptcy is not
allowed) falls slightly with the replacement ratio, by the equivalent of $23.13
when θ falls from 0.66 to 0.50, and by $40.68 when θ falls from 0.66 to 0.40.
As seen in Panel B of Table 3, when bankruptcy is allowed, the bankruptcy
rate rises systematically when θ falls from 0.66 to 0.40, from 0.90 percent in
the benchmark to 1.03 percent, an increase of 100,000 filings annually. This
effect is supported in recent empirical work of Fisher (2002), who finds that
higher UI benefits are associated with lower bankruptcy rates.
When welfare is measured relative to the benchmark economy, as shown
in Table 4, the welfare effect of eliminating bankruptcy, while always positive,
becomes smaller as θ rises. The gain from eliminating bankruptcy, relative
to the benchmark, is $66.88 when θ = 0.66 but drops to $26.14 when θ falls
to 0.40. As noted earlier, all else equal, the effect of an increased interest
rate on borrowing and a lowered rate on savings deposits would be to worsen
consumption smoothing. Yet such interest rate changes are actually associated
with small improvements in consumption smoothing, as seen in the column
“c.v.-Cons.” Panels A and B of Table 4 show that when θ = 0.66, the c.v. of
consumption falls slightly, from 0.1347 without bankruptcy to 0.1336 when
bankruptcy is allowed. This suggests that bankruptcy must be providing some
offsetting consumption benefits. Nonetheless, the costs of implementing a
bankruptcy system, from both the socially wasteful penalty of credit market

46

Federal Reserve Bank of Richmond Economic Quarterly

Table 4 Distributional Effects of Lower UI Replacement Ratios

θ

Panel A: Distributional Effects of UI, without Bankruptcy
Gini-Cons.
c.v.-Cons.
c.v.-Assets
Gini-Assets

0.66
0.50
0.40

0.0663
0.0667
0.0671

θ

Panel B: Distributional Effects of UI, with Bankruptcy
Gini-Cons.
c.v.-Cons.
c.v.-Assets
Gini-Assets

0.66
0.50
0.40

0.0694
0.0699
0.0698

0.1347
0.1363
0.1374

0.1336
0.1349
0.1352

1.6878
1.6809
1.6829

1.4211
1.4211
1.4189

0.9497
0.9465
0.9477

0.8027
0.8029
0.8014

Avg. Borr.
(% of Y )
−17.21%
−17.17%
−17.12%

Avg. Borr.
(% of Y )
−11.64%
−11.63%
−11.67%

exclusion against filers, as well as the nonpecuniary costs, cause overall welfare to fall.
Interest rate spreads are relatively stable, but the deposit rate does fall
from 2.57 percent in the benchmark to 2.35 percent and 2.31 percent as θ falls
from 0.66 to 0.50 to 0.40, respectively (see Panel B of Table 3). The fall in
deposit rates is the consequence of households needing to save more in the face
of greater income loss from unemployment than before. As all households
attempt to save more, the interest rate on savings falls. Conversely, as the cost
of funds for banks falls, the increased bankruptcy rate does not result in an
increase in the level of the interest rate on loans, relative to the benchmark. In
terms of consumption smoothing, however, the presence of bankruptcy helps
in the face of reduced replacement ratios. In Panel A of Table 4, the c.v.
of consumption rises from 0.1347 to 0.1363 to 0.1374, with reductions in θ ,
when bankruptcy is not allowed. When bankruptcy is allowed (see Panel B
of Table 4), the c.v. of consumption rises by less, from 0.1336 to 0.1349 to
0.1352. We therefore have the following:
Result 2 Reducing the UI replacement ratio lowers welfare slightly and increases bankruptcy rates. However, the fall in welfare is nearly independent of
whether or not bankruptcy is allowed. Additionally, reducing the UI replacement ratio worsens consumption smoothing less when bankruptcy is allowed.
By making repayment optional, bankruptcy has the potential to reduce
the need to actively accumulate and decumulate savings in the face of income
shocks. Furthermore, in an economy where bankruptcy is allowed, the interest rate on loans might be prohibitively high, while that on savings very low,
thereby retarding the ability of households to smooth consumption by borrowing and saving frequently. Indeed, for both reasons, bankruptcy appears to

K. Athreya: Unemployment Insurance and Personal Bankruptcy

47

significantly lower asset trade. In particular, whenever bankruptcy is allowed,
the volume and volatility of asset trade fall sharply, as seen in Panels A and B
of Table 4. For example, compare the case when bankruptcy is allowed under the benchmark replacement ratio (Panel B) to the case where bankruptcy
is eliminated under benchmark replacement ratios (Panel A). The coefficient
of variation of assets jumps from 1.42 to 1.68 and the Gini Coefficient for
assets similarly rises from 0.80 to 0.95. The average volume of borrowing,
denoted “Avg. Borr.,” also jumps from a roughly 11.6 percent debt-income
ratio (approximately $4,000 per household annually), which is close to the
8.5 percent level found in the data (CBO [2000]), to roughly 17 percent of
median annual income (or $7,000 per household).14 Note, however, that in
all cases, the response of asset trading to reductions in the replacement ratio
is very modest. Therefore, we have the following:
Result 3 Bankruptcy lowers asset trade and makes the distribution of wealth
more equal. However, changes in the UI replacement ratio do not greatly alter
asset trade.
As seen above, when bankruptcy is prohibited, the premium on borrowing
falls. This fall is in turn associated with a great deal more borrowing. For
equilibrium to obtain in the credit market, however, it must also be the case
that households save more in good times. In turn, one might expect the interest
rate (recall that there is only one interest rate in the absence of bankruptcy)
to rise. Indeed, when bankruptcy is eliminated, the rate of interest on bank
deposits rises sharply from 2.57 percent under benchmark UI replacement
ratios to 4.39 percent when bankruptcy is eliminated.
Given that both unemployment insurance and bankruptcy protection provide some insurance, it is useful to ask the following: If households had to
choose either one, but not both, which would households prefer? Table 5 shows
the results for four polar cases. Not surprisingly, it is unemployment insurance
that is quantitatively much more important than bankruptcy. Welfare is lowest
when bankruptcy is allowed and UI is driven down to Ymin by setting θ = 0.40.
The latter generates a utility level of −37.65 units. When bankruptcy is allowed but θ = 0.66 (the benchmark case), utility rises to −37.57 units. When
bankruptcy is not allowed and θ = 0.40, welfare climbs further to −37.52
units. Last, allowing UI alone, with θ = 0.66, produces the highest welfare, −37.44 units. In dollar terms, the quarterly welfare consequences range
from −$36.60 when bankruptcy is allowed and θ = 0.40, to $66.88 when
bankruptcy is not allowed and θ = 0.66, to +$26.14 when bankruptcy is not
allowed and θ = 0.40. For exposition, let W elf (Bk = {Y es, No}, θ ) denote the welfare under a regime where bankruptcy is either allowed (whereby
14 Specifically, this measures, conditional on borrowing, the mean level of unsecured debt
held by households.

48

Federal Reserve Bank of Richmond Economic Quarterly

Bk = Y es), or not (Bk = N o), and a UI replacement ratio, θ . We can express
the following rank ordering for welfare:
Result 4 W elf (N o, θ = 0.66) > W elf (N o, θ = 0.40) > W elf (Y es, θ =
0.66)[Benchmark] > W elf (Y es, θ = 0.40). Therefore, if society must
choose either UI or bankruptcy, it should choose UI. Furthermore, even if
it could choose both UI and bankruptcy, a society should choose UI alone.
Also, as mentioned above, not allowing bankruptcy even when UI is very
strict (θ = 0.40) improves welfare relative to allowing bankruptcy when UI is
generous (θ = 0.66). This is the sense in which bankruptcy is quite damaging.
The intuition for this is that better UI coverage mutes the consequences of
exclusion from the credit market and makes bankruptcy more attractive. This
raises a more general issue.
Remark 1 Any program that smooths a household’s income lowers the need
for access to credit markets. Therefore, bankruptcy becomes most attractive
precisely when it is least necessary.
Thus far, I have held the subsistence level of income, Ymin , fixed while
altering the replacement ratio and bankruptcy law. The subsistence level of
income is meant to represent the combined effects of all social insurance
programs beyond unemployment insurance. One abstraction is that the period
is thirteen weeks long, when eligibility for unemployment benefits is typically
at least twenty-six weeks. In the benchmark setting, Ymin could be thought
of as representing these extra benefits in the remaining thirteen weeks (if one
qualifies), after which other income support programs might take over. I now
briefly note the effects of cutting UI off after one period, followed by only
minimal public assistance. To this end, I set public assistance to cover just
10 percent of median household income, whereby Ymin = 0.10. In this case,
the household that is no longer qualified for UI receives the equivalent of only
$333 monthly in public assistance.15 I will not discuss these results in detail,
but will note the following findings: First, both savings and borrowing interest
rates fall, as precautionary savings rise. Second, welfare rises as bankruptcy
is allowed. Third, welfare rises by increasing amounts as the replacement
ratio falls, consistent with an increased insurance role. Last, the reductions
in welfare emerging from reductions in the UI replacement ratio are smaller
when bankruptcy is allowed than when it is not. Therefore, because the results
for Ymin = 0.10 reverse those where Ymin = 0.40, we are led to the following
conclusion:
Result 5 Bankruptcy’s role in providing insurance is clearly dependent on the
existing social safety net. For example, when Ymin is lowered to 0.10, allowing
bankruptcy improves welfare relative under all UI replacement ratios.
15 Assuming a $40,000 median annual income.

K. Athreya: Unemployment Insurance and Personal Bankruptcy

49

Table 5 Bankruptcy and UI: Four Polar Cases
Bkrptcy.,
θ
Bench
[Yes, θ = 0.66]
Yes, θ = 0.40
No, θ = 0.66
No, θ = 0.40

rl

rd

Bkrptcy.
Rate

c.v.Cons.

GiniCons.

13.00%
12.55%
4.39%
4.06%

2.57%
2.31%
4.39%
4.06%

0.90%
0.99%
—
—

0.1336
0.1352
0.1347
0.1374

0.0694
—
0.0698 −$36.60
0.0663 +$66.88
0.0671 +$26.14

Welfare
Change
($)

Utility

−37.57
−37.65
−37.44
−37.52

Corollary 1 The debate over bankruptcy protection (in the presence of existing
insurance programs) should be centered on the quantitative aspects of income
uncertainty.
This result is also consistent with the recent work of Livshits et al. (2002),
who find, in a life-cycle setting, that the presence of large uninsured medical
shocks allows bankruptcy to play a role in improving welfare.

3.

FINAL REMARKS

I have developed a stylized model of employment, unemployment, and bankruptcy in order to better understand how the consumption “insurance” provided by bankruptcy interacts with that provided by explicit unemployment
insurance programs.
Five results are worth noting. First, in the benchmark economy, introducing bankruptcy under even low UI replacement ratios lowers welfare. Second, reducing the UI replacement ratio lowers welfare slightly and increases
bankruptcy rates. Although the fall in welfare is nearly independent of whether
bankruptcy is allowed or not, reducing the UI replacement ratio worsens consumption smoothing less when bankruptcy is allowed. Third, bankruptcy
lowers asset trade and makes the distribution of wealth more equal. However,
asset trading behavior is not affected greatly by changes in the UI replacement
ratio. Fourth, UI is more important than bankruptcy: If society must choose
either UI or bankruptcy, it should choose UI.
Last, bankruptcy’s role in providing insurance is clearly dependent on the
existing social safety net. In summary, unemployment insurance appears to
materially affect the desirability of bankruptcy protection. Were other social
assistance to be scaled sharply back, the results suggest that bankruptcy could
serve a useful insurance role in the United States. However, as currently
practiced, income risk, broadly defined, does not appear high enough to justify bankruptcy in the presence of unemployment insurance. Indeed, when

50

Federal Reserve Bank of Richmond Economic Quarterly

unemployment insurance is set to current levels, bankruptcy actually appears
to harm the efficacy of the UI system.
A potentially important abstraction in the model is the absence of moral
hazard that could limit the extent of socially desirable insurance protection.
Specifically, unemployed households in the model do not alter their job search
efforts in the face of insurance payments but rather face an exogenous probability (ξ ) of return to employment. Hansen and Imrohoroglu (1992) find
that when households are allowed to reject job offers while unemployed, but
are subject to random (or imperfect) auditing by the government, the welfare
maximizing level of insurance is much lower than otherwise. Furthermore,
effort expended by workers while employed may fall with the promise of generous unemployment insurance. Also, the availability of bankruptcy will help
reduce the incentive effects of strict unemployment insurance and may further
increase moral hazard. The experience rating of employers lowers the willingness of firms to fire lazy workers, leading again to the possibility of reduced
effort. Moral hazard in an economy where output is explicitly produced using
labor leads in turn to lower output, quite unlike the pure endowment setting
employed here. The model also places fixed limits on credit availability that
do vary with bankruptcy law. It is possible that a strict bankruptcy code would
improve access to credit.
Nevertheless, there are reasons to suspect that the simple environment
developed here does provide a useful first pass at the interactions between the
UI system and the personal bankruptcy system. In particular, the unemployment insurance in the model is strictly capped at one period, and the re-entry
probability of 0.10 by no means provides comfortable income prospects for
those who fail to find work. With respect to the robustness of using fixed
credit limits, note that the elimination of bankruptcy is treated here synonymously with the prohibition of default. To the extent that informal default
would become more prevalent were bankruptcy outlawed, the expansion of
credit availability might be limited. With that said, in ongoing work (Athreya
[2002a]), I augment the model developed here to include moral hazard in job
search effort, as well as capital accumulation and the production of output
where labor effort matters. This article is therefore a first step in the analysis
of how the interactions between bankruptcy and an existing social insurance
program determine the desirability of changes to each one in isolation.

APPENDIX:

DEFINING EQUILIBRIUM

A stochastic stationary equilibrium is defined as follows: Let X = A× Y ×CS
denote the state space for households, where CS = {S, BC}. Let χ B be the
Borel σ -algebra on X. The household’s asset decision rule is denoted a(x).

K. Athreya: Unemployment Insurance and Personal Bankruptcy

51

The decision rule and the uncertainty of income together imply a stochastic
process for consumption and asset holdings, with an associated transition
function Q(x, Z), ∀Z ∈ χ B on the measurable space (X, χ B ). This transition
function implies a stationary probability measure µ(Z) for all Z ∈ χ B . This
is a measure on subsets of X that describes the joint distribution of households
on asset holdings, current income, and credit market status. For a measure to
be stationary, it must satisfy the following fixed-point condition:
µ(Z) =

Q(x, Z)dµ.
X

This implies fixed interest rates on loans and deposits and a constant
fraction of bankrupt households. Not every stationary probability measure,
however, qualifies as part of an equilibrium. Since the private bond market
must clear, aggregate holdings of private bonds must be zero. Additionally, all
public debt must be held in equilibrium. Therefore, market clearing requires
that the aggregate supply of bonds equals the stock of public debt, D.
Next, as the banking sector is competitive, profits also must be zero.
The zero-profit constraint is motivated as follows: First, let Xneg = {x ∈
X|a < 0} denote the subset of the state space X such that households hold
negative asset balances. In the stationary state, there is a time-invariant mass of
households, whose total borrowing is given by Xneg a(x)dµ. The total revenue
for the intermediary will therefore be (1 + r l )(| Xneg a(x)dµ|). Analogously,
the total cost of funds for the intermediary is determined by total borrowing
times the gross deposit interest rate, |(1 + r d ) Xneg a(x)dµ|. The losses from
default are on both interest and principal from those who borrow. Define
π(x) to be the probability that a household in state x will default. Total
principal losses are therefore | X a(x)π (x)dµ|. The zero profit condition on
intermediaries is then: (1 + r l ) ((| Xneg a(x)dµ|) − (| Xneg a(x)π (x)dµ|)) −
|(1 + r d ) Xneg a(x)dµ| = 0. (Note that the aggregate default rate is then
given by ≡ X π (x)dµ.) Lastly, the unemployment insurance system must
collect revenues equal to outlays, i.e., ηu = (1 − ρ)µe θ Y . The following five
equations will therefore define equilibrium.
Definition 2 A stationary equilibrium of the model is a four-tuple, (a(x),
π(x), µ(Z), (r l , r d )), that satisfies four conditions.
1. The decision rule, a(x), is optimal, given r d and r l .
2. µ(Z) is stationary: µ(Z) = X Q(x, Z)dµ for all Z ∈ χ B .
3. Asset market clearing: X a(x)dµ = D.
4. Zero profits: (1 + r l ) ((| Xneg a(x)dµ|) − (| Xneg a(x)π (x)dµ|)) −
|(1 + r d ) Xneg a(x)dµ| = 0.
5. Unemployment insurance fund breaks even: ηu = (1 − ρ)µe θY .

52

Federal Reserve Bank of Richmond Economic Quarterly

I use simple discrete state approximations to the value functions, conditional on income and credit market status, and then use Monte Carlo integration
with antithetic variates to compute all integrals. I then bisect on both r l and
r d until I simultaneously clear markets and satisfy the zero-profit condition.

REFERENCES
Aiyagari, S. R. 1994. “Uninsured Idiosyncratic Risk and Aggregate Saving.”
Quarterly Journal of Economics 109: 659–84.
Alvarez, F., and M. Veracierto. 2001. “Severance Payments in an Economy
with Frictions.” Journal of Monetary Economics 47: 477–98.
Athreya, K. B. 2002a. “Personal Bankruptcy and Unemployment Insurance.”
Federal Reserve Bank of Richmond. Mimeo.
. 2002b. “Welfare Implications of the Bankruptcy Reform
Act of 1999.” Journal of Monetary Economics 49: 567–95.
Chatterjee, S., D. Corbae, M. Nakajima, and J. V. Rios-Rull. 2001. “A
Quantitative Theory of Unsecured Consumer Credit with Risk of
Default.” University of Pennsylvania. Mimeo. Available at
http://www.ssc.upenn.edu/˜vr0j.
Cochrane, John. 1991. “A Simple Test of Consumption Insurance.” Journal
of Political Economy 99: 957–76.
Congressional Budget Office. 2000. “Personal Bankruptcy: A Literature
Review.” CBO paper. Washington, D.C.
Domowitz, I., and R. Sartain. 1999. “Determinants of the Consumer
Bankruptcy Decision.” Journal of Finance 54(1): 403–20.
Dubey, P., J. Geanakoplos, and M. Shubik. 2000. “Default in a General
Equilibrium Model with Incomplete Markets.” Cowles Foundation
Discussion Paper 1247. Available at http://cowles.econ.yale.edu/P/cd/.
Fisher, J. D. 2002. “The Effect of Transfer Programs on Personal
Bankruptcy.” Bureau of Labor Statistics Working Paper 346. Available at
http://www.bls.gov/ore/orecatlg.htm.
Hansen, G., and A. Imrohoroglu. 1992. “Optimal Unemployment Insurance
in an Economy with Liquidity Constraints and Moral Hazard.” Journal
of Political Economy 100: 118–42.
Heaton, J., and D. Lucas. 1997. “Market Frictions, Savings Behavior, and
Portfolio Choice.” Macroeconomic Dynamics 1: 76–101.

K. Athreya: Unemployment Insurance and Personal Bankruptcy

53

Hopenhayn, H., and J. P. Nicolini. 1997. “Optimal Unemployment
Insurance.” Journal of Political Economy 105: 412–38.
Huggett, M. 1993. “The Risk-Free Rate in Heterogeneous-Agent
Incomplete-Insurance Economies.” Journal of Economic Dynamics and
Control 17: 953–69.
Li, W., and P.-D. G. Sarte. 2002. “The Macroeconomics of U.S. Consumer
Bankruptcy Choice: Chapter 7 or Chapter 13?” Federal Reserve Bank of
Richmond Working Paper 02-1 (September).
Livshits, Igor, James MacGee, and Michele Tertilt. 2001. “Consumer
Bankruptcy: A Fresh Start.” Federal Reserve Bank of Minneapolis
Working Paper 617 (December). Available at woodrow.mpls.frb.
fed.us/research/wp.
Ljungqvist, L., and T. Sargent. 2000. Recursive Macroeconomic Theory.
Cambridge: MIT Press.
Meyer, B. 1990. “Unemployment Insurance and Unemployment Spells.”
Econometrica 58: 752–82.
Sullivan, T., E. Warren, and J. Westbrook. 2000. The Fragile Middle Class:
Americans in Debt. New Haven: Yale University Press.

Economic Fundamentals
and Bank Runs
Huberto M. Ennis

R

ecently there has been a renewed discussion in the literature about the
determinants of bank runs. Two alternative theoretical explanations
are usually provided. According to the first theory, bank runs are
exclusively driven by changes in economic fundamentals, such as a deterioration in the return on investment. The second theory views bank runs as a
consequence of the existence of multiple equilibria. In the latter case, which
equilibrium obtains depends on the realization of an extrinsic random variable,
often called “sunspots.” Extrinsic uncertainty is uncertainty in economic outcomes that does not originate directly in changes of economic fundamentals
(see Shell and Smith [1992]). The word “sunspots” is intended to convey the
idea that these random variables do not directly influence the economic fundamentals of the economy.1 However, sunspots can still influence economic
outcomes to the extent that people believe they do. In this sense, sunspots can
be viewed as coordination devices for agents’ expectations in decentralized
market economies. This is the view adopted in the bank-run literature and in
this paper.
Some scholars have recently argued that the multiple-equilibria-plussunspots explanation of bank runs is inconsistent with available evidence
Research Department, Federal Reserve Bank of Richmond, huberto.ennis@rich.frb.org. Some
of the ideas discussed in this article are the product of my joint work with Todd Keister
from Instituto Tecnol´ gico Aut´ nomo de M´ xico. I would like to thank Emilio Espino, Tom
o
o
e
Humphrey, Ned Prescott, John Weinberg, Alex Wolman, and especially Todd Keister for comments on an earlier draft. All errors are of course my own. The views expressed here do
not necessarily reflect those of the Federal Reserve Bank of Richmond or the Federal Reserve
System.
1 Shell and Smith (1992, 602) write: “The ‘sunspot’ terminology is a bit of a spoof on the
work of Jevons (1884) and his followers, who related the business cycle to the cycle of actual
sunspots. To the extent that actual sunspots do affect economic fundamentals this is intrinsic
uncertainty, but the overall effects of actual sunspots on economic fundamentals are probably not
major. Then, if actual sunspot activity does have substantial impacts on the economy, it must be
that it serves a role beyond its effects on fundamentals. Cass-Shell (1983) sunspots are highly
stylized; by definition, they represent purely extrinsic uncertainty.”

Federal Reserve Bank of Richmond Economic Quarterly Volume 89/2 Spring 2003

55

56

Federal Reserve Bank of Richmond Economic Quarterly

showing that bank runs have historically been strongly correlated with deteriorating economic fundamentals (see Gorton [1988]; Allen and Gale [1998];
and Schumacher [2000]). In this paper I will argue that such a conclusion is
not well justified. More specifically, I will show that the multiple-equilibria
model of bank runs, combined with a reasonable (and well-accepted) equilibrium selection concept, can provide theoretical justification for the correlation
observed in the data. In other words, the presence of an empirical correlation
between bank runs and poor economic fundamentals cannot be used to discriminate between the two competing theories. Furthermore, the equilibrium
selection story presented here strongly accords with the long-standing belief
that some bank runs can be characterized as events resulting from exogenous
waves of pessimism and that those mood shifts are more likely when economic
conditions are bad or deteriorating.
The empirical evidence that links bank runs to economic conditions has
been well documented. Gorton (1988) discusses what he calls the “recession
hypothesis,” according to which bank panics are closely associated with the
business cycle. In a related paper, Miron (1986) presents evidence in favor
of the “seasonal hypothesis,” which is that bank runs tend to be correlated
with seasonal fluctuations in the liquidity needs of depositors. Saunders and
Wilson (1996) and Schumacher (2000) discuss evidence on the selectivity of
depositors: not all banks are equally likely to experience a run during a panic,
and in particular a questionable solvency position prior to the run tends to
increase the probability of depositors running on a particular bank.2
Gorton (1988) studies bank panics during the National Banking Era (1865–
1914). Using data for national banks, Gorton investigates whether the model
and variables that explain the behavior of depositors during no-panic situations also explain their behavior during panics. In this sense, panics would
not be purely random events; rather, they would be directly correlated with
the arrival of new information that determines depositors’ desire to withdraw
funds from the bank. Gorton finds no evidence for something special happening during panics that cannot be explained by the model that describes the
behavior of depositors in no-panic situations. Instead, the evidence seems to
suggest that panic events are just the consequence of extreme realizations of
the circumstances that explain behavior during normal times. It is important
to note, however, that Gorton finds examples in which shocks of equal magnitude to those usually associated with runs did not cause a panic (for example,
the November 1887 spike in the liability of failed businesses did not induce a
2 Calomiris and Mason (1997) find evidence of depositors’ confusion during the June 1932
bank panic, but they also find that solvent banks were able to support each other to avoid failure.

H. M. Ennis: Economic Fundamentals and Bank Runs

57

Table 1 Financial Panics, 1890–1908 (Miron, 1986)
Major Panics

September 1890
May 1893
December 1899
May 1901
March 1903
October 1907

Minor Panics

February 1893
September 1895
June 1896
December 1896
March 1898
September 1899
July 1901
September 1901
September 1902
December 1904
April 1905
April 1906
December 1906
March 1907
September 1908

panic, while the smaller increase in June 1884 did). Finally, in Tables 1 and
2 we can see that there is some disagreement as to what constitutes a panic.
For example, Gorton does not consider the episodes of May 1901 and March
1903 as panics. Furthermore, and more germane to this paper, Tables 1 and
2 suggest that there were several bank panics in periods with no economic
recession. Of course, seasonality may be part of the answer in those cases (as
discussed by Miron [1986]).3
These are interesting findings, but they are not enough to rule out the possibility that, in some cases, banking panics are associated with the existence of
multiple equilibrium outcomes (that is, situations where both the panic and the
no-panic outcomes are possible). These stylized facts refute only the simplest
way of modeling multiple equilibria and even then only under fairly specific
conditions. Showing that reasonable theories of multiple-equilibria bank runs
are not refuted by the available evidence is important since policy prescriptions depend on the assessment of the economic conditions that generate those
bank runs. It would be helpful for policymakers to be able to conclude that
multiple-equilibria bank runs are not the norm. However, as I will show here,
the evidence discussed above does not allow us to reach that conclusion.
3 Gorton (1988) finds no evidence of seasonal effects as causes for panics using his definition.

58

Federal Reserve Bank of Richmond Economic Quarterly

Table 2 Business Cycle and Bank Panics (Gorton, 1988)
NBER Cycle (Peak-Trough)

Panic Date

October 1873 - March 1879
March 1882 - May 1885
March 1887 - April 1888
July 1890 - May 1891
January 1893 - June 1894
December 1895 - June 1897
June 1899 - December 1900
September 1902 - August 1904
May 1907 - June 1908
January 1910 - January 1912
January 1913 - December 1914

September 1873
June 1884
No panic
November 1890
May 1893
October 1896
No panic
No panic
October 1907
No panic
August 1914

The paper is organized as follows. In the next section I discuss a simple
model of bank runs that is now standard in the economic literature. I then study
the conditions under which multiple equilibria arise, and I review different
theories of how an equilibrium is selected in those cases. I show that some of
the more appealing equilibrium selection mechanisms are indeed compatible
with the available evidence. Finally, in the conclusion I discuss some policy
implications.

1.

MODELING BANK RUNS

The Environment
The environment is similar to that in Diamond and Dybvig (1983), except that
the return on investment is stochastic. There are two time periods, t = 1, 2,
and a large number of ex ante identical agents (a continuum of agents with
unit mass). Each agent is endowed with a consumption good at the beginning
of date 1 and none after that. Agents are uncertain about their preferences:
some will be impatient and will need to consume at the end of period 1; the
rest will be patient and can wait to consume in period 2. At the beginning
of period 1 agents do not know whether they will be patient or impatient, but
they know that the probability of being impatient at the end of the period is u.
Preferences are represented by the following utility function:
v (c1 , c2 ) =

1
γ

1
γ

(c1 )γ
(c1 + c2 )γ

with probability
with probability

u
,
1−u

where c1 is consumption at the end of period 1, c2 is consumption at period
2, and γ < 1. The realization of preference types is independent across
agents, implying that u will also be the fraction of the population that becomes
impatient. Agents’ types are not observable and hence patient agents can

H. M. Ennis: Economic Fundamentals and Bank Runs

59

Figure 1 Timing

always pretend to be impatient if they wish to do so (impatient agents could
pretend to be patient, but this is never the case for the contracts studied below).
There are two saving technologies available: storage and investment. One
unit of consumption placed in storage yields one unit of consumption at any
future time. For the investment technology, one unit of consumption placed in
investment at the beginning of period 1 yields R units in period 2. The return
on investment R is a random variable taking values greater than unity and with
a probability density function given by f (R). Note that the expected value of
R is necessarily greater than one and hence investment is a better technology
than storage to save consumption for the second period (that is, for funds that
are needed with certainty in the second period). If investment is liquidated
early (at the end of period 1), then it yields x < 1 units of consumption per
unit invested. Hence, investment is an illiquid asset that yields a higher return
than storage if held to maturity, but a lower return if liquidated early.

Timing
Since agents do not know their preferences until after the opportunity to invest
has passed, they pool their endowments in banking coalitions. These banks
then allocate some resources into the illiquid investment and provide insurance
to their members in case they happen to become impatient at the end of period 1.
Competition in the banking industry drives the banks to offer the best
possible available contract to consumers. I restrict the type of contracts that
banks can offer to simple deposit contracts that are subject to a sequential
service constraint (Wallace [1998]). Under this type of deposit contract, an
agent gets the right to either a fixed payment at the end of period 1 (as long as
the bank has funds) or a contingent payment in period 2. The sequential service
constraint prevents the bank from adjusting the payment to early withdrawers
according to the number of agents that decide to withdraw early. The bank
must pay a fixed amount until it runs out of funds. This kind of contract is in

60

Federal Reserve Bank of Richmond Economic Quarterly

the tradition of Diamond and Dybvig (1983) and Cooper and Ross (1998). I
use it here mainly because of its simplicity and potential descriptive content.4
The timing of events is as follows. At the beginning of period 1, the bank,
without knowing the value of R, chooses a deposit contract and a portfolio
of assets (investment is possible only at this point). This choice can be summarized by the pair (a, η), where a is the payment that the bank will give
to depositors if they decide to withdraw early and η is the proportion of total
deposits that the bank decides to keep in storage (with (1−η) being the proportion that the bank puts in the illiquid investment technology). Also at this time,
agents decide whether or not to deposit their funds in the bank. At the end of
period 1, the uncertainty about preferences and technology is resolved: agents
find out whether or not they are impatient and the value of R is revealed.5 At
this time, then, agents decide whether or not to go to the bank to withdraw
their deposits. Impatient agents have no choice but to withdraw early. Patient
agents, however, could choose to wait until period 2, which they will do if
they are not better off imitating the impatient agents. Whether a patient agent
would be better off withdrawing his or her deposits early depends, in general,
on what all the other patient agents are doing. Hence, patient agents play a
strategic game at the end of period 1. Following Peck and Shell (2003), I
shall call it the “post-deposit game.” In period 2, the return on the illiquid
technology is realized and those agents that did not withdraw their deposits
early (at the end of period 1) go to the bank and share the total remaining
resources equally.

2. THE POST-DEPOSIT GAME
The source of multiplicity of equilibria in the model lies in the post-deposit
game played by patient agents. The expected outcome of this game will
determine the bank’s investment decisions and the willingness of agents to
make deposits in the bank. The details of those problems are presented in
Section 4. What is important here is to understand that solving those problems
requires knowing what could happen in the post-deposit game. For this reason,
I turn next to the study of this game.
4 See also Ennis and Keister (2003b). In this environment, there are potential gains from
making the early payments contingent on the realization of the return on investment R. The
contracts studied here do not allow for this possibility. Gale and Vives (2002) and Allen and
Gale (1998) do not assume sequential service, but the optimal contract has a structure similar to
the deposit contract in the sense that for high values of R the payoff to early withdrawers is
not contingent. This is because investment cannot be liquidated (it has zero liquidation value),
and for high enough values of R (so that late consumers get more than early consumers), early
consumers just divide the available liquid funds among them, resulting in a fixed quantity for each,
independent of the value of R. The costly state verification literature provides another justification
for the debt contracts (see, for example, Williamson [1986]).
5 This value of R is common to all investment in the economy. No diversification is possible.

H. M. Ennis: Economic Fundamentals and Bank Runs

61

Table 3 Notation
u
γ
R
x
f (R)
a
η
u
R∗
R
pr
π

Probability of being impatient
Coefficient of relative risk aversion
Return on the risky investment
Return from early liquidation
Probability distribution of R
Bank payment for early withdrawal
Proportion of total deposits held in storage
Probability of getting paid in case of run
Multiple-equilibria threshold for R
Risk-dominance threshold for R
Risk factor of the bank-run equilibrium
Probability of a bank run

I concentrate only on symmetric pure strategy equilibria.6 At the end of
period 1, the patient agents are faced with the decision of whether to withdraw
their deposits early or leave them in the bank until period 2. Let r denote the
decision to go to the bank to withdraw (i.e., to run) and n the decision to wait
until the next period (i.e., not to run). Let us define as Pij (R; a, η) the payoff to
a patient agent following action i (i = r, n) given that all other patient agents
are following action j (j = r, n). We need only to consider those payoffs
because we are looking at symmetric equilibria, where all patient agents act
in the same manner. The normal form of the post-deposit game played by
patient agents is given by the following matrix:
Other Patient Agents
Patient
Agent

Run
No Run

Run
Prr (R; a, η)
Pnr (R; a, η)

No Run
Prn (R; a, η)
Pnn (R; a, η)

Note that the payoff Pij (R; a, η) depends on the return on investment
R and on the deposit contract chosen by the bank (a, η). (Note also that
deviations by a single player do not change the payoff to the rest of the players
because we are assuming that there is a large number of players.)
It is easy to state conditions under which this game has multiple equilibria.
In particular, if Prr (R; a, η) > Pnr (R; a, η) and Pnn (R; a, η) > Prn (R; a, η),
then running to the bank at the end of period 1 and waiting until period 2
to withdraw are both equilibria of the game. To see this, note that when
Prr (R; a, η) > Pnr (R; a, η) holds, if the patient agent thinks that all other
patient agents will run to the bank, then it is in her best interest to run as
6 Symmetry implies that in equilibrium all impatient agents play the same strategy and all
patient agents play the same strategy (but perhaps different from the one played by the impatient
agents). Pure strategies are those strategies that do not involve randomization over different possible
actions (each agent plays a single action with probability one).

62

Federal Reserve Bank of Richmond Economic Quarterly

well. Therefore, if all patient agents believe that a run will occur, the run
does occur and running is a Nash equilibrium of the game. Likewise, when
Pnn (R; a, η) > Prn (R; a, η) holds, if the patient agent thinks that no other
patient agent will run to the bank, then it is in her best interest not to run.
Therefore, if all patient agents believe that there will be no run, there is indeed
no run and not running is a Nash equilibrium of the game. In equilibrium,
then, all players play the same strategy, and I will denote each equilibrium by
the strategy being played in it. Thus, I call the run equilibrium (if it exists)
“equilibrium r,” and the no-run equilibrium “equilibrium n.”
Another important characteristic of this post-deposit game is that the multiple equilibria are usually Pareto-ranked.7 One equilibrium is better than
another equilibrium in the Pareto sense if all players in the former receive a
payoff at least as high as in the latter and one or more players receive a strictly
higher payoff. In the game studied here, if Pnn (R; a, η) > Prr (R; a, η), then
the no-run equilibrium n is Pareto-preferred to the run equilibrium r.
Given the possibility of multiple equilibria, the natural next step is to ask,
how does one of the equilibria get selected? I will discuss the answer to this
question in the next section.
Before going into the equilibrium selection issue, it is worth noting that we
can further characterize the payoff matrix of the post-deposit game. Studying
these payoffs will give us a better idea of the conditions that determine the
existence of multiple equilibria in the game.
Since the bank chooses the contract before observing the return R, the
values of η and a depend only on the probability distribution of R and not on
the particular realizations of R. The bank will never choose a contract such
that ua > η holds. In such a case, the bank will be certain to need to earlyliquidate some of the investment in order to pay depositors (even if no patient
agent runs). Since early liquidation is costly, this contract is never optimal.
I will study the problem of the bank later, but for now let us assume that the
distribution of R is such that the bank chooses a contract (a, η) that satisfies
η + x(1 − η) < a. This inequality implies that if every agent goes to the bank
early, then the bank would run out of resources before being able to pay the
promised amount a to each withdrawer. Furthermore, if the inequality does
not hold, then there would be no runs in equilibrium. These two inequalities
allow us to determine the value of waiting when there is a run, Pnr (R; a, η),
and the value of running when there is no run, Prn (R; a, η). First, we have that
Pnr (R; a, η) = 0 because if (almost) every agent goes to the bank to withdraw
early, then the bank will run out of funds and no payments will be made in the
second period. Second, we have that Prn (R; a, η) = Prn (a) = a γ /γ because
when only impatient agents withdraw early, total withdrawals are equal to ua
7 Games with multiple Pareto-ranked equilibria are called “coordination games” in the literature (for a general review, see Cooper [1999]).

H. M. Ennis: Economic Fundamentals and Bank Runs

63

and the bank has access to enough liquid funds, η + x(1 − η), to cover that
amount.
Let us now define u ≡ [η + x(1 − η)] /a < 1 as the probability of being
paid when every agent goes to the bank early. This formula is a direct consequence of assuming that agents take random positions in the line formed at the
bank’s window and that there is a sequential service constraint. Thus, we have
that Prr (R; a, η) = Prr (a, η) = ua γ /γ . It is important to note that Pnr , Prn ,
and Prr are not functions of the particular realization of R. The only payoff
that is a direct function of the realization of R is that for late withdrawals when
there is no run, that is
1 R(1 − η) + (η − ua) γ
Pnn (R; a, η) =
.
γ
1−u
Note that Pnn (R; a, η) is a continuous, increasing, and unbounded function
of R. Hence, there exists a threshold value R ∗ such that if R > R ∗ , we have
that Pnn (R; a, η) > Prn (a) = a γ /γ and the post-deposit game is a multipleequilibria coordination game. If R < R ∗ , the post-deposit game has a unique
equilibrium in which all agents withdraw their deposits at the end of period 1.
In summary, the payoff matrix for the post-deposit game is:
Other Patient Agents
Run
Patient

No Run

0

No Run
1 γ
γa

1
γ
γ ua

Agent

3.

Run

1
γ

R(1−η)+(η−ua) γ
1−u

EQUILIBRIUM SELECTION IN THE POST-DEPOSIT GAME

There is an extensive literature on equilibrium selection in games. This literature has concentrated some attention on 2 × 2 games with multiple equilibria.
The post-deposit game of the previous section can be thought of as just an
example of a 2 × 2 symmetric game with the potential for multiple equilibria
(i.e., a 2 × 2 symmetric coordination game).8 In this section, I will review
some of the basic ideas from this literature and discuss how they apply to the
bank-run problem at hand.
It is useful at this point to introduce the concept of equilibrium selection
mechanism (ESM). An ESM is a probability distribution that assigns, to each
equilibrium of the game, a probability indicating how likely it is to be the
result of play. For the post-deposit game under consideration, an ESM is
a function that for each possible triplet (R, a, η) assigns a probability π to
8 Usually we refer to a 2 × 2 game as a game that is played by two individuals who each
have two possible pure strategies that they can choose to play. In the post-deposit game, agents
play a “game” against the population that is often called a “macroeconomic game.” See Cooper
(1999) for an extensive discussion on the subject.

64

Federal Reserve Bank of Richmond Economic Quarterly

the run equilibrium (r) and a probability (1 − π ) to the no-run equilibrium
(n). These probabilities must be feasible in the sense that, for given values
(R, a, η), if the run equilibrium does not exist, then π = 0, and if a run
is the only equilibrium, then π = 1. It is important to note that there is
a degree of coordination being assumed from the outset: agents know that
the only possible outcomes are those where all the rest of the agents play in
the same manner (and this coordination is common knowledge). The ESM
provides some structure to the coordination problem but does not explain why
and how coordination arises. In this sense, the concept of an ESM can be
thought of as a generalization of the traditional sunspot approach: there is
still in place an exogenous coordination device on which all agents base their
actions. The innovation is that the general ESM allows for the probability
of each equilibrium to depend on exogenous and endogenous variables in the
model.
The next natural question is, where does the function π (R, a, η) come
from? In the traditional sunspot approach, the function π is a constant between zero and one when feasible (i.e., when both equilibria exist). Another
commonly used criterion for equilibrium selection is to assume that the best
equilibrium (in the Pareto sense) will be selected. In this case, the ESM is
such that the probability π is equal to zero as long as the no-run equilibrium
exists and switches to unity when only the run equilibrium exists. Yet there are
other possible forms that the function π may take and that can be reasonably
justified. I review some of these forms next.
Let us start by defining the risk factor of equilibrium j , for j ∈ {r, n} as
the smallest probability p such that if a player believes that with probability
strictly greater than p all the other players are going to play action j , then
action j is the unique optimal action to take (see, for example, Young [1998]).
Hence, the risk factor of the run equilibrium (r) is given by the solution to the
following equation:9
pr Prr + (1 − pr )Prn = pr Pnr + (1 − pr )Pnn .
Therefore,
pr =

Pnn − Prn
(Prr − Pnr ) + (Pnn − Prn )

is the risk factor of the run equilibrium. When both equilibria exist (run and
no-run), the only payoff that depends on R is Pnn , and this payoff is increasing
in R. Hence, pr is an increasing function of R. This result is rather intuitive. It
says that the higher the return on investment R, the higher the belief probability
of a run p must be in order to induce a patient agent to run on the bank.
9 The payoffs are still a function of the triple (R, a, η), but I choose not to explicitly write
this dependence in order to simplify notation.

H. M. Ennis: Economic Fundamentals and Bank Runs

65

An equilibrium j is p-dominant if the equilibrium action j is the unique
best response to any belief of the player that puts probability at least p ∈ [0, 1]
on the other players playing action j (see Morris, Rob, and Shin [1995]).
Hence, the run equilibrium is pr -dominant.
If the risk factor of the run equilibrium pr is less than or equal to onehalf, then the run equilibrium is risk dominant (Harsanyi and Selten [1988]).
Risk dominance has been used as a criterion for equilibrium selection: the
risk-dominant equilibrium will be the one selected and played. This criterion
has an appealing interpretation. If each player is uncertain about the action of
the other players, it is plausible that he or she would assign equal probability
to each of the possible outcomes (a flat or diffuse prior). If the risk factor of
equilibrium j is less than one-half, that is, if equilibrium j is risk dominant,
and if players have flat priors about the actions of the other players, then
equilibrium j will be the one played. In the post-deposit game, when each
player assigns equal odds to all of his or her opponents playing either action
r or n, then the players will choose to play the action of the risk-dominant
equilibrium. In terms of the definition of ESM, the risk dominance criterion
assigns probability one to the risk-dominant equilibrium.
Another way of motivating an equilibrium selection rule in games with
multiple equilibria is to study learning dynamics under repeated iterations of
the static (stage) game. See, for example, Kandori, Mailath, and Rob (1993);
Young (1998); and Matsui and Matsuyama (1995). These papers concentrate
on games with two players and assume that there are frictions limiting the
ability of agents to adjust their strategies. Kandori, Mailath, and Rob also
assume bounded rationality on the part of the agents playing the dynamic game
(in the form of myopic behavior and some propensity to make mistakes). It is
interesting to note that the learning dynamics under these assumptions tend to
select (as the frictions or the probability of mistakes vanish) the risk-dominant
equilibrium as the one most likely to be played. Temzelides (1997) extends
this work and applies it to the bank-run model.
Ennis and Keister (2003a) study a learning model that induces a probability
distribution over the possible equilibria of a 2×2 macroeconomic coordination
game. We show that the probability of equilibrium j induced by this learning
process is strictly decreasing in the risk factor of equilibrium j and can take
values strictly lower than one even when equilibrium j is risk dominant. In
terms of the previous ESM terminology, we have that the function π is a
decreasing function of pr and may take values strictly between zero and one.
Since pr is an increasing function of R (the fundamentals), we have that the
probability of a run π is a decreasing function of R. That is, the better the
fundamentals (R), the less likely is a bank-run event. In Ennis and Keister
(2003b) we apply these ideas to study the effect of bank runs on economic
growth.

66

Federal Reserve Bank of Richmond Economic Quarterly

Let us now go back to the case of equilibrium selection based on the traditional sunspot approach. Assume that the return on investment R takes values
only in the interval (R ∗ , ∞), where R ∗ is the threshold such that for values of
R greater than R ∗ there are multiple equilibria of the post-deposit game. In
other words, assume that the contract is such that the no-run equilibrium exists
for every possible value of R. Assume also that a binomial sunspot random
variable determines which equilibrium is selected. Because both equilibria
exist for every value of R, the probability of a bank run is always given by the
constant probability associated with the sunspot realization that coordinates
agents to “run” to the bank. This is the sense in which the previous literature
on bank runs has dismissed the sunspot explanation for not conforming with
the observed correlation of bank runs with economic fundamentals.
However, note that if R can be below R ∗ with positive probability, then
for those realizations, regardless of the sunspot variable, the probability of a
run will be equal to unity. In such a case, even though sunspots still play an
important role in coordinating the agents when there are multiple equilibria,
the probability of bank runs will be higher for lower values of R, and indeed
the probability of observing a bank run will be the highest (equal to one) when
the fundamentals deteriorate sufficiently (that is, when R < R ∗ ). In this sense,
even the traditional sunspot approach can account for some of the correlation
of bank runs with economic fundamentals. Economic fundamentals determine
whether multiple equilibria exist, and then probabilities have to adjust to reflect
this fact.10
Furthermore, the traditional sunspot approach seems too simplistic for this
environment, and the risk-dominance-based selection mechanism appears to
be a reasonable extension. We can think that the risk dominance ESM is the
case where the particular sunspot variable that coordinates patient agents to
run to the bank is correlated, in a specific way, with the stochastic variable R
determining fundamentals. Risk dominance provides discipline and intuition
to this correlation.
In particular, the risk dominance criterion divides the support of the distribution of R into two sets: the set where R < R, in which the run equilibrium
is risk dominant, and the set where R > R, in which the no-run equilibrium
is risk dominant. We can think that there is an associated sunspot random
10 Ironically, the model in the second part of the paper by Allen and Gale (1998) can be

used to provide a good example of this situation. For some parameter values their model has
multiple equilibria. Their equilibrium analysis delineates three relevant regions for the possible
realization of the return on the risky asset R. When R is very low, the equilibrium has a bank
run; when R is very high, there are no bank runs in equilibrium; and for intermediate values of
R, there are multiple equilibria: both having a bank run and not having a bank run are possible
equilibrium outcomes. Therefore, just using a simple sunspot variable to determine which of the
two equilibria will be observed in the intermediate region of R would deliver the historical correlation: as fundamentals deteriorate (as R goes from high to low levels), the probability of bank
runs first goes from zero to positive (the value associated with the sunspot) and then to unity
when fundamentals are so poor that a bank run is unavoidable.

H. M. Ennis: Economic Fundamentals and Bank Runs

67

variable s, perfectly correlated with R, such that whenever R takes values in
the interval [1, R], the variable s takes the value r, and whenever R takes values in the interval (R, ∞), the sunspot variable s equals n. If agents associate
values of s = r with a run situation and values of s = n with a no-run situation,
the equilibrium selection process is still driven by sunspots (the variable s),
but it generates a correlation of bank runs with the behavior of fundamentals.
It is worth noting that for most values of R, both equilibria still exist, even
though one of them is risk dominant. What determines which equilibrium will
be played is a matter of how agents get coordinated. Coordination is driven
by the sunspot variable s. Risk dominance can be thought of as the justification for why the particular sunspot random variable s has been selected as
a coordination device over all possible variables that may be available. Note
that there is a higher level of coordination among agents in the choice of the
relevant sunspot variable. This interpretation of sunspots is in fact associated
with another argument that has been used to explain the appearance of such
coordination devices: sunspots can be viewed as the limiting case of situations in which agents are overreacting to some small movement in economic
fundamentals. Manuelli and Peck (1992) formalize this argument.
Finally, it should be clear at this point that the more general ESM approach
(Ennis and Keister [2003a]), in which the probability of a bank run π is
a decreasing function of R, is also consistent with both the multiplicity of
equilibria and the correlation of bank runs with economic fundamentals. In
fact, with this approach the probability π can be strictly between zero and unity
and at the same time be dependent on R. This feature seems very appealing,
since the historical correlation was never perfect: sometimes bank runs did
not occur even though economic fundamentals were as bad as or worse than
in periods where a bank run did occur (see Gorton [1988]).

4. THE BANK’S PROBLEM
In Section 2 we assumed that agents would be willing to deposit their funds
in the bank and that the bank would choose a contract with some specific
properties. This section provides the justification for those assumptions.
Given that the banking system may be subject to runs, agents might choose
not to participate in the banking system.11 In that case, their payoff would be
given by the following “autarky” problem
VA ≡ max
η

11 For
the bank or
just part of
where bank

u

(η + R(1 − η))γ
(η + x(1 − η))γ
+ (1 − u)
γ
γ

f (R) dR,

the sake of simplicity, I am restricting agents to deposit either all their resources in
nothing at all. Ennis and Keister (2003b) consider the case where agents can deposit
their initial resources in the bank. This is an important extension in environments
runs can happen with positive probability, as is the case in this paper.

68

Federal Reserve Bank of Richmond Economic Quarterly

subject to 0 ≤ η ≤ 1. At the beginning of period 1, the agent decides how to
split the endowment between storage (η) and investment (1 − η). At the end
of period 1, the agent finds out whether she is patient or impatient. If she is
impatient, then she liquidates the investment and consumes (funds are useless
for her in the second period). If she is patient, then she stores the liquid funds
and consumes in the second period both the liquid funds and the return on
investment (recall that we are assuming that R > 1 > x).
A bank could always choose a contract that eliminates the possibility of
experiencing a run. I will call the best contract with such property the “runproof contract.” A contract is run-proof if there is enough liquidity in the bank
to pay all agents the amount a at the end of period 1. But because the contract
is run-proof, patient agents actually wait until the second period to withdraw.
The problem of a bank choosing the run-proof contract is the following:
1
aγ
u + (1 − u)
γ
γ

VRP ≡ max
a,η

R(1 − η) + (η − ua)
1−u

γ

f (R) dR,

subject to
a ≤ η + x(1 − η), a ≥ 0, and 0 ≤ η ≤ 1.
The first constraint is the run-proof constraint. It says that even if all agents go
to the bank in the first period (i.e., early), the bank will not run out of funds.
Finally, after having studied equilibrium selection in the post-deposit
game, we are now in a position to write down the problem faced by the bank
at the beginning of period 1. It is important to note that the probability of
a run may depend on the contract chosen by the bank and hence the bank
will take this effect into account when determining the best possible contract.
Formally, the bank’s problem is given by
V

≡ max
a,η

[π (R, a, n)Prr (a, η) +

(1 − π (R, a, η)) u

aγ
+ (1 − u)Pnn (R, a, η) ]f (R) dR,
γ

subject to a ≥ 0 and 0 ≤ η ≤ 1. Note that Prn does not enter the problem
directly. It may, however, enter the problem indirectly through the determination of π (R, a, n), as in the case of the ESM based in risk dominance or
adaptive learning.
When we have VA < max{V , VRP }, the agents will choose to deposit
their funds at the bank. When we have V > VRP , the bank will choose the
contract that allows for the possibility of bank runs according to the ESM that is
operating in the economy (that is, according to the given function π (R, a, η)).
It is important to note that if there exist values of R such that R < R ∗ and
f (R) > 0, then for those values of R we must have that π (R, a, η) = 1
because the post-deposit game has a unique (run) equilibrium for those values
of R.

H. M. Ennis: Economic Fundamentals and Bank Runs

69

Diamond and Dybvig (1983) show that when the return on investment R
is not stochastic (and greater than unity) and the probability π is arbitrarily
set at zero, the bank chooses a contract (a, η) for which a bank run is a
possible equilibrium of the post-deposit game played by the patient agents.
Hence, using arguments of continuity, it can be shown that there exist functions
f (R) and π (R, a, η) > 0 such that a bank solving the problem V described
above will also choose a contract that admits runs (that is, a contract such that
η + x(1 − η) < a holds).

5.

CONCLUSION

I have shown that even when bank runs are driven by self-fulfilling expectations
in environments with multiple equilibria, the historical correlation of bank runs
with poor economic fundamentals can still be accounted for. More evidence
would be necessary to reject the case of bank runs originating in situations
with multiple equilibria. For now, when we observe a bank run, we cannot
in principle confidently discard the possibility that another equilibrium with
no bank run was also possible. This conclusion is important from a policy
standpoint. In some cases, multiple-equilibria bank runs can be avoided by the
design of off-equilibrium policies that are hence never observed. For example,
the suspension of convertibility could make the run situation I have presented
no longer an equilibrium of the post-deposit game (as proposed by Diamond
and Dybvig in their original paper). But because suspension would occur
only when there is a run and runs are not equilibrium outcomes anymore,
the suspension of payments will not be observed. An important qualification
is that, like many other off-equilibrium threats, this policy entails a certain
ability of the bank to commit to actually implementing the policy if it becomes
necessary.
There is another important policy implication of the ideas presented here.
In the multiple-equilibria case, bank runs are usually not optimal and in general
the policymaker would like to avoid them (or at least lower their probability).
Contrary to this position, Allen and Gale (1998) present the case of bank runs
that are not the consequence of a coordination failure and that are in fact part
of the optimal arrangement for risk sharing in the economy. The policymaker
would not want to avoid the Allen-Gale type of bank runs. Determining which
of the two cases is driving a particular episode is an important issue that the
policymaker would need to carefully evaluate.

70

Federal Reserve Bank of Richmond Economic Quarterly

REFERENCES
Allen, Franklin, and Douglas Gale. 1998. “Optimal Financial Crises.”
Journal of Finance 53 (August): 1245–84.
Calomiris, Charles, and Joseph Mason. 1997. “Contagion and Bank Failures
During the Great Depression: The June 1932 Chicago Banking Panic.”
American Economic Review 87 (December): 863–83.
Cass, David, and Karl Shell. 1983. “Do Sunspots Matter?” Journal of
Political Economy 91 (April): 193–227.
Cooper, Russell. 1999. Coordination Games. Cambridge: Cambridge
University Press.
, and T. W. Ross. 1998. “Bank Runs: Liquidity Costs and
Investment Decisions.” Journal of Monetary Economics 41 (February):
27–38.
Diamond, Douglas, and Philip Dybvig. 1983. “Bank Runs, Deposit
Insurance, and Liquidity.” Journal of Political Economy 91 (June):
401–19.
Ennis, Huberto, and Todd Keister. 2003a. “Government Policy and the
Probability of Coordination Failures.” Centro de Investigaci´ n
o
Econ´ mica Discussion Paper 03-01, ITAM (February).
o
. 2003b. “Economic Growth, Liquidity, and Bank Runs.”
Forthcoming in the Journal of Economic Theory.
Gale, Douglas, and Xavier Vives. 2002. “Dollarization, Bailouts, and the
Stability of the Banking System.” Quarterly Journal of Economics
(May): 467–502.
Gorton, Gary. 1988. “Banking Panics and Business Cycles.” Oxford
Economic Papers 40 (December): 751–81.
Harsanyi, John, and Reinhard Selten. 1988. A General Theory of
Equilibrium Selection in Games. Cambridge: MIT Press.
Jevons, W. Stanley. 1884. Investigation in Currency and Finance. London:
Macmillan.
Kandori, Michihiro, George Mailath, and Rafael Rob. 1993. “Learning,
Mutation, and Long Run Equilibria in Games.” Econometrica 61
(January): 29–56.
Manuelli, Rodolfo, and James Peck. 1992. “Sunspot-Like Effects of Random
Endowments.” Journal of Economic Dynamics and Control 16 (April):
193–206.

H. M. Ennis: Economic Fundamentals and Bank Runs

71

Matsui, Akihiko, and Kiminori Matsuyama. 1995. “An Approach to
Equilibrium Selection.” Journal of Economic Theory 65 (April): 415–34.
Miron, Jeffrey. 1986. “Financial Panics, the Seasonality of the Nominal
Interest Rate, and the Founding of the Fed.” American Economic Review
76 (March): 125–40.
Morris, Stephen, Rafael Rob, and Hyun Song Shin. 1995. “p-Dominance
and Belief Potential.” Econometrica 63 (January): 145–57.
Peck, James, and Karl Shell. 2003. “Equilibrium Bank Runs.” Journal of
Political Economy 111 (February): 103–23.
Saunders, Anthony, and Berry Wilson. 1996. “Contagious Bank Runs:
Evidence from the 1929–1933 Period.” Journal of Financial
Intermediation 5 (October): 409–23.
Schumacher, Liliana. 2000. “Bank Runs and Currency Run in a System
Without a Safety Net: Argentina and the ‘Tequila’ Shock.” Journal of
Monetary Economics 46 (August): 257–77.
Shell, Karl, and Bruce Smith. 1992. “Sunspot Equilibrium.” In The New
Palgrave Dictionary of Money and Finance. Vol. 3. New York:
Macmillan, 601–7.
Temzelides, Ted. 1997. “Evolution, Coordination, and Banking Panics.”
Journal of Monetary Economics 40 (September): 163–83.
Wallace, Neil. 1988. “Another Attempt to Explain an Illiquid Banking
System: The Diamond and Dybvig Model with Sequential Service Taken
Seriously.” Federal Reserve Bank of Minneapolis Quarterly Review 12
(Fall): 3–16.
Williamson, Stephen. 1986. “Costly Monitoring, Financial Intermediation,
and Equilibrium Credit Rationing.” Journal of Monetary Economics 18
(September): 159–79.
Young, Peyton. 1998. Individual Strategy and Social Structure. Princeton,
N.J.: Princeton University Press.