The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Contagion and Bank Failures During the Great Depression:
The June 1932 Chicago Banking Panic
By CHARLES W. CALOMIRIS AND JOSEPH R. MASON*
We examine the social costs of asymmetric-information-induced bank panics in
an environment without government deposit insurance. Our case study is the
Chicago bank panic of June 1932. We compare the ex ante characteristics of
panic failures and panic survivors. Despite temporary confusion about bank asset
quality on the part of depositors during the panic, which was associated with
widespread depositor runs and bank stock price declines, the panic did not produce significant social costs in terms of failures among solvent banks. (JEL G28,
G21, N22, E58, E32)
Recent work in banking theory and history
has helped to define the potential causes and
costs of bank panics, which various authors
have argued can be traced to speculative attacks on the numeraire (Barry A. Wigmore,
1987; R. Glenn Donaldson, 1992), illiquidity
shocks (Douglas W. Diamond and Philip H.
Dybvig, 1983; Donaldson, 1993), or shocks
to bank asset values when there is information
asymmetry between bankers and depositors
about the incidence of those shocks (Calomiris
and Gary B. Gorton, 1991; Calomiris and
Charles M. Kahn, 1991; Calomiris and Larry
Schweikart, 1991; Sudipto Bhattacharya and
Anjan V. Thakor, 1993; George G. Kaufman,
1994). In the latter case, when depositors cannot observe whether individual banks are sol-
vent, but can observe a shock that affects
banks' portfolios, they may initiate runs on all
banks, both solvent and insolvent.
Bank panics are short-lived phenomena,
historically measured in days or weeks. However, a panic still can have important longlived costs if it results in the disappearance of
solvent banking institutions. That concern is
often invoked to justify the significant expansion of the government safety net for U.S.
banks during the 1930's (Anthony J. Saunders
and Berry Wilson, 1994). In this paper, we
take a close look at one of the clearest examples of an asymmetric-induced bank panic during the Great Depression, and ask whether
solvent banks failed during that panic.
The answer to this question has important
public policy implications. Studies of bank
panics argue that panics induced by asset
shocks and asymmetric information can be
hard to resolve with monetary policy alone. In
contrast, speculative attacks on the numeraire
induced by uncertainty about its future value
can be stopped by policies that resolve uncertainty about monetary policy (for example, by
a devaluation, as in the United States in March
1933). Similarly, bank panics that result from
shocks to liquidity preference and a limited
supply of aggregate liquid bank assets can be
resolved by traditional monetary policy in the
form of open-market purchases of securities
by the central bank (Bruce Champ et al.,
1991). But bank panics caused by asymmetric
information about the condition of banks
* Calomiris: Graduate School of Business, Columbia
University, New York, NY 10027 and National Bureau of
Economic Research; Mason: Bank Research Division, Office of the Comptroller of the Currency, Washington, DC
20219. The authors thank Richard Grossman, Ed Kane,
George Pennacchi, Berry Wilson, two anonymous referees, seminar participants at Columbia University, Georgetown University, and Universidad Torcuato di Telia, and
conference attendees at the NBER Development of the
American Economy meetings, the American Economic
Association/Cliometric Society meetings, and the Financial Management Association meetings for helpful
comments. We are also grateful to Ven Vitale for research
assistance, and to the National Science Foundation (SBR9409768) for funding. Opinions expressed are the authors', not those of the Office of the Comptroller of the
Currency.
863
864
THE AMERICAN ECONOMIC REVIEW
cannot be resolved by monetary policy. Values
of individual bank loan portfolios (about
which uninformed depositors are concerned)
are not controllable by monetary policy
(Frederic S. Mishkin, 1991; Calomiris, 1994).
To resolve the problem of asymmetricinformation-induced runs one must limit the
risk depositors face as the result of asymmetric
information, and thus remove the incentive for
depositors to demand immediate withdrawal.
This can either be done privately or publicly.
Privately, banks can either act individually to
reassure depositors or agree temporarily to
stand behind each other's liabilities. So long
as depositors are confident that the coalition of
mutually insuring banks is solvent collectively, that collective action can bring the
panic to an end without resort to suspension of
convertibility. Alternatively, the government
can provide insurance of deposits, either in the
form of a commitment to pay depositors, or by
lending cash to banks against their illiquid assets at a subsidized rate.
This paper addresses the empirical question
of whether private bank actions to stem
asymmetric-information runs are adequate, or
whether government deposit insurance is
needed. Specifically, we ask whether private
institutions can prevent the failure of solvent
banks during a bank panic. We examine that
question in the context of the banking crises
of the Great Depression. The example we focus on is the Chicago panic of June 1932. We
choose this example for three reasons. First,
we argue it is a quintessential example of an
asymmetric-information-induced panic. Second, this was one of the most publicized
examples of a run on banks during the banking
crises of the early 1930's, which coincided
with the federal government's decisions to establish the public safety net for banks. Third,
by focusing on a particular location and episode, we are able to clearly identify the
origins of the panic and to control for the effects of location, time, and macroeconomic
environment—factors that might otherwise
complicate our analysis.
Our strategy in the paper is as follows. We
use a variety of measures to judge whether
banks that failed during the Chicago panic
were likely to have been solvent. We investigate whether they were predictably weaker ex
DECEMBER 1997
ante (and thus more vulnerable to asset price
decline) relative to banks that survived the
panic. We employ data from individual bank
failure experience, balance sheets, income and
expense statements, and stock prices for. failing and surviving Chicago banks before and
during the panic. We analyze characteristics of
failing and surviving banks to determine
whether the banks that failed during the panic
were similar ex ante to those that survived the
panic. Wefindthat panic failures were weaker
than panic survivors, and argue that panic failures can be attributed to asset value decline of
failed banks rather than to depositor confusion
about the value of bank assets.
While depositors did confuse panic survivors with panic failures, the failure of solvent
banks did not result from that confusion. One
reason such failures were avoided may be that
solvent banks knew each other's condition better than depositors, and had the incentive and
the ability to help each other avoid failure during the crisis. Private cooperation by the Chicago clearing house banks appears to have
been instrumental in preventing the failure of
at least one solvent bank during the panic.
Section I provides historical background for
the Chicago panic to support our use of the
June crisis as an example of an asymmetricinformation-induced bank panic, and our identification of the panic event window. Section
II presents our empirical analysis of the characteristics of panic failures, panic survivors,
and banks that failed outside the panic window. Section III concludes with a summary of
our findings and a discussion of their importance and limitations.
I. The June 1932 Banking Crisis in Chicago
Our discussion of the Chicago banking crisis establishes five "facts" that support its use
for testing the value and limitations of private
cooperation during asymmetric-informationinduced panics. Together, these five facts establish that the Chicago panic resulted from
location-specific asset shocks that were relevant for bank portfolios; that it was a true
asymmetric-information-induced panic in that
all banks (ex post solvent and ex post insolvent) experienced heavy withdrawals and
stock price declines during the panic; that (at
VOL. 87 NO. 5
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
865
Number of
Chicago Bank
Failures, Daily
FIGURE 1. NUMBER OF FAILED OR SUSPENDED BANKS (DAILY AND MONTHLY), JUNE 1931-DECEMBER 1932
Notes: Bank suspensions are from the Federal Reserve Bulletin, various issues. Bank failures consist of receiverships
and voluntary liquidations, and come from the Annual Report of the Comptroller of the Currency, and the Statement
Showing Total Resources and Liabilities of Illinois State Banks at the Close of Business, Superintendent of Banking of
the State of Illinois, various issues. In order to accommodate the logarithmic scale on theright-handside, 0.1 is substituted
for observed values of zero in the monthly data.
least in one case) banks were willing to support each other against the uninformed runs of
depositors; and that Chicago bank failures that
occurred outside the panic window did not coincide with similar (panic) events.
A. Was the June Crisis in Chicago
a Unique Event Nationally?
As Figure 1 shows, mid-to-late June of 1932
witnessed concentration of bank failures in
Chicago, whether measured by the number or
total assets of failed banks.1 The number of
bank failures in June 1932 was not particularly
1
Asset plots are available from the authors upon
request.
high at the state, Federal Reserve District, or
national level in comparison to previous
months. In contrast, Chicago experienced a severe concentration of failures during the week
of the panic. Of the 49 bank failures in the state
of Illinois during that month, 40 took place in
Chicago, and 26 of these failed in the week of
June 20-27 {Commercial and Financial
Chronicle, July 2, 1932 p. 71 ). 2
Deposit outflows indicate a similar pattern.
As shown in Figure 2, Chicago banks saw an
unusually large decline in their deposits during
2
The reported * 'failure dates'' in the Commercial and
Financial Chronicle of June 20—June 27 correspond to
failure dates of June 21-June 28 reported by state and
national bank regulators.
866
THE AMERICAN ECONOMIC REVIEW
DECEMBER 1997
Chi Total
Deposits
(3MA)
NY Total
Deposits
(3MA)
Note: There were 40 failures in
Chicago during June, 24 of which
occurred in the panic week
preceding June 29
FIGURE 2. PERCENT CHANGE IN THE THREE-PERIOD MOVING AVERAGE OF TOTAL DEPOSITS, WEEKLY-REPORTING
BANKS IN CHICAGO AND NEW YORK CITY, JUNE 1931-DECEMBER 1932
late June, and banks in other areas of the country (notably New York City, thefinancialcenter) did not share in that precipitous decline.
B. Solvent and Insolvent Banks Both
Suffered During the Panic
Contemporary chroniclers and economic
historians have pointed to the June banking
crisis in Chicago as an important example of
how a systemwide attack by depositors on
banks can produce pressure on solvent and insolvent banks alike. In 1932, the crisis received national, as well as local, attention in
the press. Contemporary reports clearly indicate that depositors ran ex post solvent as well
as ex post insolvent banks en masse.
The Commercial and Financial Chronicle
(July 2, 1932 pp. 70-71) provided a detailed
account of the runs on Chicago banks, and specifically noted that even healthy banks (including, for example, First Chicago) were affected.
These reports emphasized that long lines of individual depositors formed at banks. Some
banks that were reported to have experienced
large withdrawals (including First Chicago and
Continental) were able to withstand their runs
and remain open, while other banks (including
one Loop bank—the Chicago Bank of Commerce), were forced to close.
Initially (before June 22), bank distress was
limited to a few banks, but this soon spread
and was associated with a dramatic decline in
aggregate deposits in Chicago banks. The dramatic withdrawals from downtown banks began on June 22 and reached their peak on
Friday, June 24. F. Cyril James (1938 p. 1034)
distinguishes the panic in late June from previous periods of banking distress in Chicago:
[previous] runs ... were directed against
particular banks that were known to be
enfeebled; this one was directed against
the whole Chicago money market and
VOL. 87 NO. 5
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
the First National group, in the center of
the battle, still had more than a hundred
and twenty-five million dollars of cash
resources available, even though it had
paid out fifty millions since Tuesday
night. In the case of earlier runs, the
crowds had been drawn from a particular
locality or a special group: this time people from all parts of the city seemed to
converge on the Loop in hysterical fear
and anxiety.
Bank stock prices fell during the panic, although not all Loop banks experienced large
declines. Central Republic, First Chicago, and
Continental were among the Loop bank survivors whose stock prices declined the most. First
Chicago stock fell from 150 (bid) at the close
on June 18 to a low of 131 (bid) at the close
on June 24. Continental's stock fell from 70
(bid) on June 18 to a low of 60 (bid) on June
27. Central Republic saw its stock price fall
from 52 (bid, June 18 close) to 47 (bid) on
June 25, and then its stock plummeted to a price
of 4 (bid) on Monday, June 27. Afterwards, its
stock price, and those of the other surviving
Loop banks, rebounded rapidly. The stock of
the only Loop bank to fail during the panic, the
Bank of Commerce, was trading at 9 (bid) on
June 18-June 24 for a $20 par value. On June
25 its price ceased to be reported.
C. Local Adverse Economic News
Precipitated the Panic
Contemporary discussion of the crisis emphasized the adverse local economic news that
had precipitated it. The panic did not reflect exogenous liquidity-demand shocks. Rather than
withdrawing deposits to spend them, depositors
often redeposited those withdrawals in thenpostal savings accounts. James (1938) argues
that the panic was triggered by several factors,
all of which combined in rapid succession to
undermine depositors' faith in the value of Chicago banks' assets. The bad news included falling prices for local utility stocks and other
corporate assets, a well-publicized local case of
bank fraud and mismanagement, and a municipal revenue crisis for the city of Chicago (Chicago Tribune, June 26, 1932 p. A17).
The city government's revenue problem
weakened the banks in three ways. First, it
867
meant that the banks were forced to bear increased risk on their bond portfolios as the
flow of coupons was interrupted. Second, Chicago banks were called upon to purchase illiquid tax warrants to help keep the municipal
government afloat. Third, city workers were
forced to draw down their bank deposits to pay
normal living expenses, thus reducing bank reserves and increasing the proportion of (risky)
loans and securities in bank portfolios. Not
surprisingly, Chicago bankers saw the viability of the banking system and the financial
problems of the city as closely related. A delegation of Chicago city officials and citizens
visiting Congress to request federal government assistance for the city in June 1932
included many prominent bankers. The delegation's request for $80 million in aid was rebuffed by Congress on June 22 (Chicago
Tribune, June 23, 1932 p. 1).
The municipal revenue crisis was symptomatic of the deep contraction in local asset
prices and economic activity. Among the
many victims was the Insull utility empire in
Chicago, whose stock and debt were widely
held by institutions and individual investors.
Chicago utility companies grew dramatically
during the 1920's in anticipation of growing
demand and were caught short by the sudden
decline in the local economy. The prices of
Commonwealth Edison (ComEd) and Insull
stocks and bonds show precipitous declines in
the first six months of 1932. From January
through March, Insull stock declined by 76
percent, Insull bonds fell by 62 percent, and
ComEd stock lost 25 percent of its value. From
March through early June the declines accelerated. Insull debt lost 98 percent of its value,
Insull common stock lost 89 percent, and Insull preferred lost 47 percent.
Chicago's economic problems were reflected in several local financial disasters and
cases of bank fraud (an activity traditionally
more pronounced in bad times than in good)
that made front-page news in Chicago day after day just prior to the crisis.3 John Bain, a
defendant in the most important bank fraud
3
Milton Esbitt (1986) emphasizes the importance of
management practices for explaining bank failures in Chicago in 1931.
868
THE AMERICAN ECONOMIC REVIEW
case, was a local real estate developer who
owned a chain of banks. On June 9, the 12
banks in the chain failed to open for business
(James, 1938 p. 1033; Elmus R. Wicker, 1993
p. 15). Not until June 23, however, did it become clear just how large the losses from fraud
had been in the Bain chain. On that date the
court released its estimate that the value of the
banks' assets was roughly $3.5 million, compared to total deposits of $ 13 million (Chicago
Tribune, June 24, 1932 p. 9).
An important additional piece of evidence
that the panic was induced by fears of bank
insolvency, rather than by exogenous liquidity
demand by depositors, is that much of the
funds withdrawn from banks were redeposited
immediately in the form of riskless postal savings. The Commercial and Financial Chronicle (July 2, 1932 p. 71) reports that, during
the week of June 20-June 27, "The Postal
Savings Department, which normally has 25
or 30 windows in the Chicago post office ...
increased the number to around 100 ... to accommodate deposits." That report noted that
"... about $1,000,000 had been received [in
postal savings on June 27 ] , compared to
$2,000,000 and $3,000,000 a day at [the] peak
[of the crisis], and [compared to] a normal
daily average of $200,000."
In summary, by June 23, Chicago bank depositors had witnessed, in a matter of only two
weeks, the collapse of some of the largest businesses in their city, several enormously costly
cases of bank fraud, and a deepening of the
municipal financial crisis as the result of the
denial of relief to their city government by federal authorities. All of these stories were frontpage news day after day in the two weeks
leading up to the banking crisis, and the news
grew progressively worse. In this light, it is not
surprising that depositors became increasingly
concerned about the ability of banks to pay out
their deposits, and transferred bank deposits to
riskless postal savings accounts.
D. Interbank Cooperation Helped to
Preserve Solvent Banks Under Pressure
As already noted, the Loop banks experienced severe stock price decline and deposit
withdrawals during the crisis. Although two
banks, Central Republic and the Bank of Com-
DECEMBER 1997
merce, experienced unusually severe declines
in their stock prices, the Bank of Commerce
failed while Central Republic survived. Its survival, however, depended on cooperation by
large Loop banks to resolve its distress.
During the crisis, Central Republic was
nearly taken down by its depositors. As doubts
about Central Republic's solvency grew and
deposit withdrawal pressure mounted, the
bank's management prepared to close the bank
voluntarily to avoid the risk of its being closed
by bank depositors. Solvent banks that had lost
depositors' confidence had an incentive to
close preemptively to preserve stock value by
avoiding the costs of liquidating bank assets
during a run, and the transaction costs associated with having the bank taken over by a
regulator. This incentive was particularly
strong in 1932, when bank stockholders faced
the threat of double liability on deposits, which
meant that liquidation costs borne by stockholders could conceivably exceed the complete loss of the bank's capital.
Other Chicago banks saw the prospect of Central Republic's voluntary liquidation as a potential disaster for depositor confidence in their
banks. Bankers clearly believed that depositors
were reacting to fears of bank insolvency and
were unable to distinguish between solvent and
insolvent banks. Prominent bankers from Chicago and New York met as a group to devise a
plan to defend the Central Republic Bank and
Trust Co. from the continuing withdrawals.
Fearing the spillover effects of a decision to liquidate Central Republic, these bankers managed
to persuade General Dawes (its Chair) to continue operating by offering an arrangement to
infuse Central Republic with new liquidity.
The initial plan for the loan to Central Republic provided for $10 million in back-up liquidity from New York and Chicago banks
and $80 million from the Reconstruction Finance Corporation (RFC), but the final deal
involved assistance only from Chicago banks
and the RFC. The deal that emerged combined
liquidity assistance from the RFC with, in essence, a back-up credit enhancement by the
Chicago banks.4 RFC liquidity support for the
4
It is important to keep in mind that the RFC was the
only entity charged with helping avoid the insolvency of
VOL. 87 NO. 5
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
Chicago banks—like all RFC lending during
this period—was fully collateralized by very
high-quality, liquid assets; credit risk on the
RFC loan to Central Republic was borne in
largest part by the Chicago banks that formed
the lending syndicate.5 Importantly, the RFC
agreed to allow municipal tax warrants—$30
million of which had been sold to Loop banks
(Chicago Tribune, June 25, 1932 p. 6)—to
qualify as collateral for its loan.6 Once the crisis passed, Central Republic saw its deposit
outflows cease and its stock price increase.
Central Republic was a solvent bank saved
from failure by the collective intervention of
other Loop banks.
E. The June 20-27 Crisis
Was a Unique Event
Having argued that the June banking crisis
is an example of a panic induced by asset value
decline and asymmetric information, we turn
to the question of whether there were other
such episodes in Chicago during the first six
months of 1932. That question is relevant for
our discussion in Section II, where we will use
the absence of panics during that period (outside the window of the June crisis) as an identifying restriction to investigate whether
failures during the panic were similar to failures outside of the panic.
Bank failures during 1932, and more generally during the Great Depression, for the
most part have not been identified by historians as resulting from panics or asymmetric in-
individual banks. At this time, Federal Reserve Banks did
not view the prevention of bank insolvency as their mandate. This is in sharp contrast to more recent experience
during the 1980' s. For a review of the history of Fed lending policy, see Anna J. Schwartz (1992).
5
Joseph R. Mason (1995) argues that prior to its use
of preferred stock purchases to assist banks, the RFC was
not effective in stemming bank failures. James (1938 p.
1044) cites the view, common at the time, that because of
the strict collateral requirements on RFC lending, RFC
assistance often increased the credit risk faced by bank
depositors.
6
Abbreviated bank balance sheets were routinely reported in newspapers following call dates. Thus the June
30 Reports of Condition published in the Chicago Tribune
on July 2 provided further evidence of the soundness of
Chicago banks (Chicago Tribune, pp. 18-24).
869
formation. With the exception of the June
panic, no contemporary chronicler or scholar
of which we are aware has identified any other
time interval during 1932 as a panic or banking
crisis in Chicago. Neither has anyone identified any nationwide banking panics as having
occurred during 1932 (Milton Friedman and
Schwartz, 1963; Wicker, 1980, 1993; Ben S.
Bernanke, 1983; Calomiris, 1993; Clifford
Thies and Daniel Gerlowski, 1993; Eugene N.
White, 1984).
We have reviewed a variety of facts that
identify the Chicago bank panic of June 1932
as a quintessential example of an asymmetricinformation-induced panic, resulting from local economic problems that affected bank
asset values. Just as important for our purposes is the uniqueness of the panic window.
The crisis, as reflected in sudden deposit
withdrawals and stock price declines, and
widespread coverage in the local and national
press, was brief and was surrounded by times
in which bank failures did not coincide with
a panic.
II. Failures and Survivors During the Panic
We now return to our central question of
whether the absence of government deposit insurance promotes the failure of solvent banks
during asymmetric-information-induced panics. Specifically, we investigate whether solvent banks were able to survive withdrawal
pressures (partly via private coordination)
during the June 1932 Chicago bank panic. To
answer that question we make use of the fact
that the panic was a unique event during 1932.
We assume that outside the panic window (in
early 1932) banks that failed were actually insolvent. We then use the characteristics of
those nonpanic failures to evaluate the likely
solvency or insolvency of banks that failed
during the panic.
If banks that failed during the panic were
just as strong (according to some set of criteria) as those that survived during the panic,
that would provide evidence in favor of the
null hypothesis that confusion on the part of
depositors about bank quality produced random bank failure. If, on the other hand, banks
that failed during the panic were weaker than
bank survivors, then panic failures cannot be
THE AMERICAN ECONOMIC REVIEW
870
viewed as purely random. That implies a rejection of the null hypothesis.
It is not difficult to reject this "strong" version of the null hypothesis, but that does not
prove that only insolvent banks failed during
the panic. Depositor confusion might have produced the failure of some solvent banks, with
potentially large social costs, even if on average
banks that failed during the panic were weaker
than those that survived it. Thus in our empirical work we also address a "weak" form of
the null hypothesis—that panic-induced failures were not purely random, but were importantly random. This version of the null
hypothesis is inherently difficult to reject formally on the basis of ex ante statistical tests of
means, medians, and regression coefficients.
Combining these with additional evidence,
however, we argue that the social costs of the
unwarranted closure of solvent institutions (if
any) must have been very small.
Our empirical discussion divides into two
parts. First, in subsection A we present evidence that leads to the rejection of the strong
form of the null hypothesis. Then we confront
and reject the weak form of the null hypothesis, using statistical evidence as well as information from bank examiner reports.
A. Comparisons of Bank Attributes
Leading Up to the Panic
We divide the Chicago banks in our sample
into three groups: panic failures (banks that
failed during the panic, June 2 0 - 2 7 ) , nonpanic failures (banks that failed at other
dates), and survivors (banks that did not fail
in the first seven months of 1932). We then
compare the ex ante attributes of these three
groups.7
In our analysis of survivors, panic failures,
and nonpanic failures, we focus on four ex
ante measures of bank condition: (1) the ratio
7
The dates we choose for the panic window are consistent with James's (1938) discussion, newspaper accounts of the beginning and end of the panic, and the daily
movements of the stock prices of the ten Loop banks reported in the Chicago Tribune, which reached their nadir
on June 27. The results are robust to reasonable alternative
definitions of the panic window.
DECEMBER 1997
of the market value of equity to the book value
of equity; (2) the estimated probability of failure of banks; (3) the rate of decline in bank
deposits; and (4) the interest paid on bank
debts. These various measures of bank risk are
available for different subsets of Chicago
banks, depending on the availability of data on
stock prices and interest paid on deposits.
Stock prices are not available for all Chicago
banks, and interest paid is only available for
Fed member banks. The data set for the study
consists of several components: balance sheet
data, income and expense data, and stock price
data. Balance sheet data from December 31,
1931 call reports were collected for all state
and national banks in Chicago, a total of 123
banks. Total assets and total deposits were also
collected for December 31, 1930, to permit
calculation of the changes in those variables
during 1931. Balance sheet data for the 22 national banks and 11 state banks that were
members of the Federal Reserve System come
from microfilm of the original Reports of Condition filed with the Office of the Comptroller
of the Currency (OCC) and the Board of Governors of the Federal Reserve System. State
nonmember bank balance sheet data are from
the compilation of Statements of State Banks
of Illinois issued by the Superintendent of
Banking of the State of Illinois. The disaggregated Reports of Condition of member banks
facilitated aggregation of balance sheet categories to reporting standards comparable with
the Statements of State Banks of Illinois. The
stock prices for Chicago banks are end-ofmonth observations published in the Bank and
Quotation Record (of the Commercial and Financial Chronicle). Interest payments are
available only for Fed member banks (from
the Reports of Condition).
1. Market-to-Book Value Ratios. Figure 3
plots the means and 50-percent inclusion
ranges for market-to-book value ratios for the
three separate groups of Chicago banks (survivors, panic failures, and nonpanic failures).
To adjust for survivorship bias in plotting
these trends, we retain failed banks in our sample after their date of failure, and assume that
their stock value after failure is zero.
The striking fact illustrated by Figure 3 is
that as early as January 1931 the banks that
VOL. 87 NO. 5
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
871
•Panic Failures
•Nonpanic Failures
Survivors
FIGURE 3. RATIOS OF MARKET-TO-BOOK VALUES OF EQUITY, END-OF-MONTH QUOTES, JANUARY 1931-JULY 1932
Notes: Survivorship bias can arise if banks that fail are excluded from subsequent means and standard deviations. We
correct for this bias by including failed banks and assuming they had zero market value of equity after the date of failure.
survived the June panic appeared to be a distinct group with higher average market-tobook ratios. The banks that failed during the
panic generally had slightly higher average ratios than those that failed at other times, but
throughout the prepanic period (January 1,
1931-June 20, 1932) the market-to-book
value ratios of panic failures were very close
on average to those of prepanic failures and
very different from those of panic survivors.
By January 1932, most of the panic failures
had market-to-book ratios less than unity. Figure 3 shows that all Chicago banks suffered
from capital decline during 1931 and 1932,
and that the banks that failed during the June
panic reached and maintained unusually low
market-to-book value ratios long before the
panic.
2. Failure Predictions. Next, we use ex ante
observable characteristics of Chicago banks
(based on bank data reported in December
1931) to compute scores predicting failure
during the first seven months of 1932. We
compare the probabilities of failure for the
three groups of banks (panic failures, nonpanic failures, and survivors). Our ex ante
scores indicate that panic failures and nonpanic failures on average were weaker banks
than survivors at least as early as the end of
1931.
We estimate the probability of failure using
a logit model of the links between bank characteristics (e.g., balance sheet ratios) and bank
failure. The danger of using ex post failures to
estimate failure risk, of course, is that special
events with low probabilities may have influenced actual failure experience during the panic
in ways that were unpredictable ex ante and
possibly unrelated to underlying insolvency.
For example, if a panic were a common shock
to all banks, then the level of reserves might do
an excellent job of forecasting panic failures
even if the banks that failed during the panic
872
THE AMERICAN ECONOMIC REVIEW
did not have higher ex ante probabilities of failing. Thus using ex post panic failures to construct a model of ex ante bank weakness may
bias toward identifying panic failures as inherently weak when in fact they were not. To avoid
(or at least minimize) this problem, we report
logit failure forecasts constructed from both insample and out-of-sample estimation.8 In the
out-of-sample forecasts, we exclude banks that
failed during the panic from the sample when
we estimate the coefficients relating bank characteristics to the probability of failure. This
constrains the panic failures to be predicted using model parameters that were constructed to
explain nonpanic failures, and thus prevents
special unpredictable events during the panic
from influencing predictions of failure.
Our logit results are reported in Table 1. We
include the following variables (all measured
at year-end 1931) in our specification: size
(log of total assets), the reserve-to-demand
deposit ratio (where reserve assets are defined
as cash and government securities), the real
estate loan share (defined as the ratio of loans
on real estate to total loans), the ratio of real
estate owned to illiquid assets (which mainly
includes repossessed real estate collateral, and
excludes bank premises), the ratio of last
year's retained earnings to net worth, and the
long-term debt ratio (bills payable plus rediscounts plus time deposits, divided by total
assets).
This combination of variables also forms the
basis of the logit models of White (1984),
David C. Wheelock (1992), Calomiris and
Wheelock (1995), and Mason (1995), all of
which are used to forecast bank failures for the
1920's and 1930's. This specification typically
is interpreted as capturing measures of each of
the following fundamental bank characteristics: bank size, asset liquidity, exposure to real
estate market risk, nonperforming loans (real
estate owned), recent bank performance (retained earnings/net worth), and bank liability
8
We also estimated logit models for prepanic failures
only (excluding the failures that occurred during or after
the panic). The results were essentially identical to those
we report below for nonpanic failures (which include failures that occurred after the panic), and so we do not report
them here.
DECEMBER 1997
composition. Bank liability composition is a
useful signal of weakness because—as White
(1984), Wheelock ( 1 9 9 2 ) , Calomiris and
Wheelock ( 1 9 9 5 ) , and Mason (1995)
argue—reliance on high-interest, borrowed
funds was an undesirable necessity only suffered by higher-risk banks (see also our discussion of debt composition below). While
not all variables prove significant in our logits,
we retain measures of the basic concepts that
previous studies have found to be important
even if they do not prove statistically significant in our sample. Excluding them would not
affect our results. We also experimented with
including two other variables (not reported in
Table 1): the ratio of book net worth to assets
and the percent changes in deposits or assets
of banks from December 1930 to December
1931. In neither case did the regressors add
predictive power to our models.
The results in Table 1 are quite similar for
the in-sample and out-of-sample specifications,
which is consistent with the view that failures
during panics were similar events to nonpanic
failures. The variable coefficients that are significant are of the expected signs. Banks with
higher reserve ratios, higher ratios of retained
earnings to net worth, and lower proportions of
long-term debt were less likely to fail.
Table 2 reports the mean and median predicted failure probabilities for the logit models
by category of bank (panic failure, nonpanic
failure, and survivor), and the significance levels for tests of differences across categories in
means and medians. These results indicate that
the banks that failed during the panic were less
risky ex ante than banks that failed outside the
panic and more risky than survivors. Comparisons using predicted values from in-sample
and out-of-sample regressions are similar. By
construction, the in-sample results show less of
a difference between panic and nonpanic failures. Also by construction, out-of-sample forecasts tend to have lower probabilities of failure.
Our results are consistent with the notions that
panic failures were much weaker banks than
panic survivors, and that failures during the
panic were a continuation of the same process
that underlay other failures.
3. Deposit Withdrawals and Debt Composition. If panic failures had been relatively weak
VOL. 87 NO. 5
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
873
TABLE 1—LOGIT MODEL RESULTS
Out-of-sample
In-sample
Dependent variable of all models: Bank failure N[0, 1] used for significance levels
Number of observations
Number of panic failures
Number of nonpanic failures
86
114
0
28
18
18
Log-likelihood
-18.95
-51.37
Restricted (slopes = 0) log-likelihood
-44.12
-76.88
50.34
51.03
3.38E-09
2.28E-09
Coefficient
Standard
error
Coefficient
Standard
error
3.31
7.40
1.49
3.73
Chi-squared statistic (k — 1 df)
Significance level
Variable name
Constant
Bank size (log of total assets)
-0.75*
0.54
-0.20
0.25
Ratio of reserves to demand deposits
-4.47***
1.81
-2.97***
0.83
Real estate loan share
Ratio of other real estate owned to illiquid assets
Ratio of net earnings to net worth
Long-term debt
* Statistical significance at
** Statistical significance at
*** Statistical significance at
-1.39
1.71
13.89
12.16
2.09
8.25
-25.33***
70.27
-15.20***
5.30
23.31***
6.48
12.05***
2.64
= 0.10.
= 0.05.
= 0.01.
institutions for months prior to the panic, then
they should have experienced larger rates of
depositor withdrawal before the panic. It is not
possible to obtain comparable records of deposit accounts for failed banks after the December 1931 call, but one can ask whether
panic failures experienced relatively large deposit withdrawals from December 1930 to December 1931. Table 3 reports data on the rate
of decline of deposits over that year. Clearly,
panic failures and nonpanic failures experienced much larger withdrawals than survivors
in 1931. Panic survivors experienced an average decline in deposits of 41 percent, com-
2.10
2.78
pared to 55 percent for nonpanic failures, and
33 percent for survivors.
The higher rate of decline in deposits for
panic failures and nonpanic failures during
1931 is reflected in the debt compositions of
those banks. Detailed data on the composition
of bank liabilities are available for all banks in
our sample, either from Federal Reserve or
state call reports. Table 3 presents data on the
liability composition of banks as of December
1931.
Interestingly, the shares of the various debt
categories vary systematically across the three
groups of banks. In particular, panic failures
874
THE AMERICAN ECONOMIC REVIEW
DECEMBER 1997
TABLE 2—MEANS AND MEDIANS OF FAILURE PROBABILITIES, BY CLASS OF BANK
In-sample logit
Mean
Median
Out-of-sample logit
Mean
Median
Nonpanic failures
Score
0.753
0.796
0.669
0.780
Standard error
0.042
0.033
0.068
0.053
18
18
Number
18
18
Panic failures
Score
0.556
0.625
0.269
0.166
Standard error
0.043
0.059
0.055
0.106
28
28
Score
0.248
Standard error
0.030
68
68
Panic, survivors
5.66***
4.67***
3.75***
2.25***
Panic, nonpanic
3.11***
2.19**
4.57***
4.42***
Nonpanic, survivors
8.13***
5.61***
10.78***
19.32***
Number
28
28
0.187
0.088
0.005
0.055
0.021
0.015
68
68
Survivors
Number
t-statistics for tests of differences
* Significant at
** Significant at
*** Significant at
=0.10.
= 0.05.
= 0.01.
(like nonpanic failures) tend to rely much
more on borrowed money (defined as bills
payable and rediscounts). As we have noted
above, the standard interpretation of this
finding—which is consistent with observed
differences in deposit withdrawal rates across
categories during 1931 reported in these
tables—is that when demandable debt is withdrawn from risky banks, those banks are
forced to rely on high-cost borrowed money,
typically collateralized by liquid bank assets.
As noted above, other studies have found that
the share of borrowed money is a reliable predictor of bank failure during the 1920's
and 1930's (White, 1984; Wheelock, 1992;
Calomiris and Wheelock, 1995; Mason,
1995). Moreover, examiners from the Office
of the Comptroller of the Currency used a reliance on borrowed money as a clear indication that a bank was having trouble. For
example, in referring to the trouble at the Hyde
Park-Kenwood National Bank, the examiner
wrote that: "... practically all of the bank's
bonds and securities are pledged [as collateral
for borrowed money] and the bank is now a
heavy borrower, the Chief Examiner advising
that the total borrowings on January 28
amounted to $684,000 due to the heavy decline in deposits."
In summary, panic failures and nonpanic failures experienced significantly larger withdrawals of deposits long before the panic. Consistent
with this deposit withdrawal pressure, panic
and nonpanic failures alike experienced fundamental debt financing reallocations characteristic of troubled institutions.
4. Interest Rates on Debt. Interest rates on
debt should indicate debtholders' perceptions
of the risk of bank failure. If panic failures and
VOL. 87 NO. 5
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
875
TABLE 3—DEPOSIT AND INTEREST RATE COMPOSITION, B Y CLASS OF BANK
Change
in
total
deposits
Change
in
total
assets
Demand
deposits
Due to
banks
Time
deposits
Borrowed
money
Interest
on total
debt
Survivors
-0.3251
-0.2145
0.5098
0.0301
0.4600
0.0197
0.0062
Standard error
0.0363
0.0491
0.0226
0.0078
0.0245
0.0070
0.0005
Number of obs.
62
63
Mean
68
68
68
68
18
Panic failures
-0.4115
-0.3397
0.4911
0.0216
0.4873
0.0831
0.0093
Standard error
0.0595
0.0299
0.0314
0.0058
0.0315
0.0159
0.0009
Number of obs.
28
28
Mean
28
28
28
28
11
Nonpanic failures
Mean
-0.5514
-0.3979
0.3835
0.0053
0.6113
0.1872
0.0116
Standard error
0.0316
0.0242
0.0336
0.0029
0.0325
0.0301
0.0014
Number of obs.
18
18
18
18
18
18
3
Tests of differences between means (t- statistics)
Nonpanic, panic
1.777**
1.381*
2.263***
2.131**
2.629***
3.338***
1.161
Panic, survivor
1.287*
1.637*
0.461
0.667
0.630
4.249***
3.386***
Nonpanic,
survivor
3.245***
1.971**
2.672***
1.618*
2.99***
8.192***
4.201***
Notes: Deposits are presented as a proportion of total deposits, equal to demand deposits, interbank deposits, time deposits,
and bills payable and rediscounts. Interest is reported as interest expense as a proportion of the relevant deposit category,
i.e., demand deposit interest expense/demand deposits. Interest is calculated as the amount of interest paid over the last
six months as a proportion of the total in each category of debt as of December 31, 1931. Changes in total assets and
deposits are from December 31, 1930 to December 31, 1931.
* Significant at =0.10.
** Significant at = 0.05.
*** Significant at =0.01.
nonpanic failures were more likely to fail ex
ante they should have been forced to pay
higher interest on their debt prior to the June
panic. For a small sample of Chicago banks
(31) that were Fed members, we have data on
the interest paid during the last six months of
1931 on each of the categories of debt discussed above (individual demand deposits,
bank deposits, time deposits, and borrowed
money). We report the aggregate amounts of
these in Table 3 as a fraction of the respective
outstanding debts shown on the December 31,
1931 balance sheets. The banks are grouped,
as before, according to their failure experi-
ence. It is important to keep in mind that our
reported interest rate differences capture the
experience of only a small sample of banks,
and are measured with error because we divide
interest flows over a six-month period by endof-year balances. Therefore, these data may
not provide an entirely accurate picture of interest rates paid as of December 1931.
In the column entitled interest paid on total
debt we compute the means for each of the
three categories of banks of the ratio of total
interest paid relative to total debt. We find that
panic failures and nonpanic failures paid significantly higher interest rates on debt than
876
THE AMERICAN ECONOMIC REVIEW
survivors. Panic failures paid an overall interest cost that was 50 percent higher than that of
survivors, and nonpanic failures paid nearly
double the interest rate paid by survivors.9
In summary, we find that panic failures and
nonpanic failures paid higher interest rates on
their debt than survivors. The higher interest
paid by those two classes of banks reflected
their relative reliance on high-cost funds (borrowed money and time deposits) rather than
higher interest costs on demand debt.
B. The Role of Declining Asset Values
in Chicago Bank Failures
The fact that panic failures appear to have
been stronger institutions on average than nonpanic failures, as measured by their publicly
available financial data as of the end of 1931,
has several possible explanations. One possibility is the weak form of the null hypothesis (that
a significant number of solvent banks failed during the panic). If one believed that bank characteristics (as measured in December 1931)
accurately reflect unchanging cross-sectional
differences in bank condition throughout the period January-July 1932, then panic failures consequently appear excessive. That is, under the
assumption of unchanging bank condition, the
fact that panic failures' characteristics lie between those of nonpanic failures and those of
9
We also examined the breakdown of interest paid according to each category of debt. Differences in total deposit risk show up in withdrawals of demand deposits
from banks, in changes in relative weights attached to various types of debt, and in overall debt costs, but not in
demand deposit interest rate differences. Other research
examining links between bank-failure risk and demand deposit interest rates during the Great Depression has also
failed to find a positive relationship between demand deposit interest rates and bank-failure risk. George J.
Benston (1964) analyzed banks during the period 19291935 and found no significant positive relationship between demand deposit interest rates and failure risk. As in
our sample, he sometimes found a negative (and insignificant) relationship between the two. One explanation for
these findings is provided by Gary B. Gorton and George
G. Pennacchi (1990), who argue that some bank depositors may be very unwilling to accept increasedriskon their
accounts. Risk-intolerant depositors may prefer to adjust
to changes in bank riskiness via changes in the quantity
of balances they hold with a bank rather than changes in
the interest paid on those balances.
DECEMBER 1997
survivors implies that the failure process was
less discriminating during the panic—that is,
that the traits of panic failures reflect a mix of
solvent and insolvent institutions.
But such an assumption is surely incorrect.
In Section I we presented evidence that local
asset values (and the value of bank portfolios)
declined dramatically in the first half of 1932.
This implies that the failure threshold for
banks was shifting over that period. In an environment of persistently declining asset
prices, the first banks to fail (nonpanic failures) will appear measurably weaker than
banks that fail only after asset values have
fallen much more (panic failures). To control
for changes in the probability of failure within
our period, we estimate a survival duration
model of bank failure. This model supports the
argument that declining fundamentals can explain the quality difference between (early)
nonpanic failures and (late) panic failures.
Our survival duration model is similar to our
logit model except that it forecasts the length
of time the bank will survive (measured in
days after December 31,1931), and allows for
changes in the underlying transition probabilities during our period, i.e., the conditional
probabilities of failure on any given day, via
a logistic hazard function. This hazard function effectively separates the effects of
changes in the probability of failure across individuals from shifts in the baseline probability of failure associated with diminishing
fundamental bank asset prices (Guido W.
Imbens, 1994 p. 703). The implied baseline
probability of failure estimated in our model
increases at a decreasing rate from January to
June, and declines during July. A technical
Appendix, available from the authors upon request, provides a formal discussion of our survival duration model.
The results of our survival duration model
are reported in Table 4. The results are qualitatively similar to those for the logit model in
Table 1, although, of course, coefficients in the
two models are of opposite sign. As shown in
Table 4, our survival duration model is capable
of approximating the gaps in time between
nonpanic failures and panic failures. Table 5
illustrates that the model overpredicts survival
duration on average for both panic failures and
nonpanic failures. That is, using the same scor-
VOL. 87 NO. 5
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
877
TABLE 4—SURVIVAL MODEL RESULTS
In-sample
Out-of-sample
Dependent variable: Log of time elapsed (in days) from December 31, 1931
Number of observations
86
114
0
28
18
18
Log-likelihood
-39.32
-84.36
Restricted (slopes = 0) log-likelihood
-67.08
-117.30
Number of panic failures
Number of nonpanic failures
Chi-squared statistic (k - 1 df)
55.51
Significance level
3.65E-10
Variable name
Coefficient
Standard error
65.88
2.85E-12
Coefficient
Standard error
Constant
1.77
5.21
4.53***
1.76
Bank size (log of total assets)
0.53
0.42
0.13
0.12
Ratio of reserves to demand deposits
2.69***
0.98
1.44***
0.39
Real estate loan share
-0.61
1.62
0.58
0.72
Ratio of other real estate owned to illiquid assets
—4.95
7.29
0.04
3.33
Ratio of net earnings to net worth
Long-term debt
* Statistical significance at
** Statistical significance at
*** Statistical significance at
-12.15***
3.63
7.41***
2.40
-5.80***
1.10
= 0.10.
= 0.05.
=0.01.
ing model (based on December 1931 characteristics), and allowing for time variation in
the hazard function, we estimate mean survival duration for nonpanic failures of 192
days, compared to 349 days for panic failures,
while the actual respective survival means
were 107 and 177 days. But the model accurately estimates the relative health of panic and
nonpanic failures. Our model predicts that
nonpanic failures survive (on average) 60 percent as long as panic failures, and in fact they
averaged 55 percent of the survival time of
panic failures. Thus our model does not underpredict panic failures relative to nonpanic
failures, as one would expect if panic failures
10.52**
6.07
were unwarranted and nonpanic failures were
warranted. In other words, when one accounts
for the time-varying probability of failure for
all banks, the model does as well estimating
cross-sectional hazards during the panic as
prior to the panic.
C. Further Evidence the Panic Entailed
Low Social Costs
The duration survival model results are
consistent with the view that only observably insolvent banks failed during the panic.
But these findings do not constitute a formal
rejection of the weak form of the null
878
THE AMERICAN ECONOMIC REVIEW
DECEMBER 1997
TABLE 5—MEANS AND MEDIANS OF DURATION PREDICTIONS
(IN DAYS FROM DECEMBER 31, 1931), BY CLASS OF BANK
In-sample duration
Mean
Median
Actual duration
Mean
Median
Nonpanic failures
192
168
107
131
Standard error
24
25
NA
NA
Number
18
18
18
18
349
253
177
177
Score
Panic failures
Score
Standard error
49
38
NA
NA
Number
28
28
28
28
1,482
688
NA
NA
308
99
NA
NA
68
68
68
68
Survivors
Score
Standard error
Number
t-statistics for tests of differences
Panic, survivors
2.35***
2.78***
NA
NA
Panic, nonpanic
2.46***
1.66**
NA
NA
Nonpanic, survivors
2.15**
2.69***
NA
NA
* Significant at
** Significant at
*** Significant at
=0.10.
= 0.05.
= 0.01.
hypothesis—that some solvent banks failed
during the panic, and that the social costs
from these failures were important. To investigate that possibility, we take a closer
look at panic failures, particularly at "outliers' ' that appear (on the basis of observable
traits in December 1931) to have been
healthy institutions. Examination reports on
the condition of these outliers reveal deep
problems in these institutions prior to the
panic, which were publicly known but not
captured by 1931 balance sheet ratios. In
many cases, bank fraud and accounting irregularities explain why banks that failed
during the panic appear stronger statistically
(on the basis of 1931 accounting data) than
they did to contemporaries, for whom their
problems were common knowledge by mid1932.
Table 6 presents data on the distributions of
logit scores for the three groups of banks using
in-sample and out-of-sample estimation. Note
that none of the panic failures has an out-ofsample score that is as low (that is, as good) as
the top quartile of survivors. The minimum
(best) out-of-sample score of the panic failures
is 0.00059, and the cutoff for the lowest (best)
quartile of survivors is 0.00025. Only six panic
failures had out-of-sample logit scores that were
lower than or equal to the median of survivors.
Were these six panic failures examples of
banks that were solvent but allowed to fail by
their fellow bankers? If so, were the social
costs of those failures high? It is easier to answer the second question. These six banks collectively represented a trivial proportion of the
bank assets of Chicago (1.2 percent of total
assets as of December 1931), and while it is
VOL. 87 NO. 5
879
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
TABLE 6—DISTRIBUTIONAL ANALYSIS OF LOGIT SCORES FOR EX POST PANIC FAILURES AND SURVIVORS
In-sample p(fail),
panic failures
Percent of survivors
estimated below
that probability
(in-sample logit)
Out-of-sample p(fail),
panic failures
Percent of survivors
estimated below
that probability
(out-of-sample logit)
Minimum
0.048
0.311
0.001
0.326
25th percentile
0.448
0.782
0.038
0.697
Median
0.626
0.910
0.175
0.832
75th percentile
0.701
0.932
0.469
0.929
Maximum
0.905
0.988
0.906
1.000
In-sample p(fail),
survivors
Percent of panic failures
estimated below
that probability
(in-sample logit)
Out-of-sample p(fail),
survivors
Percent of panic failures
estimated below
that probability
(out-of-sample logit)
Minimum
0.000
0.000
0.000
0.000
25th percentile
0.040
0.000
0.000
0.000
Median
0.162
0.038
0.006
0.139
75th percentile
0.403
0.222
0.074
0.300
Maximum
0.943
1.000
0.821
0.918
conceivable that some of them were solvent
banks, the social costs of their demise cannot
have been very large.
To answer the first question we explored the
specific circumstances of these six panic failures and their financial condition prior to failure. We searched the records of the OCC for
any relevant examinations and correspondence
with respect to the two national banks included
in the group of six (South Ashland National
Bank and Standard National Bank). We were
able to locate information about both of these
banks. Prior to the panic, both of these banks
had experienced large loan losses and were under investigation by the U.S. Attorney General
and the OCC for fraud.
Comptroller's office. That examination revealed that the bank had experienced large unaccounted losses that left its capital impaired,
and placed it in violation of its charter. The
Deputy Comptroller wrote that "the officers
and directors have been operating the bank
along unsound lines." In particular, the Deputy Comptroller criticized the bank's management for allowing a large loan to the bank's
Director/Manager that produced an enormous
loss for the bank. The examiner, in his May 5
letter to the OCC, wrote: "This bank is now
under the incompetent management of Director James G. Hodgkinson, who dominates the
policies; he is absolutely broke and the manner
in which he has furthered his own interests is
most reprehensible. A report on his operations
The records we discovered for South Ash- land clearly indicate that this bank was on the
has been made to the United States District
verge of failure at least two months prior to
Attorney." The details of the examination rethe panic, and possibly earlier. While the bank
veal fraudulent activities, including check kitwas closed on June 24 and placed in the hands
ing by Hodgkinson.
of a receiver on June 27, South Ashland's Vice
Standard National Bank was also involved
President, Guy Brown, had written to the OCC
in
a case of fraud. Its Vice President, who was
as early as June 2 to announce that the bank
also the Vice President of another bank that
had decided on May 25 to liquidate itself in
failed during the panic (People's National
response to an April 27 examination by the
Bank and Trust Co.) confessed to appropriating
880
THE AMERICAN ECONOMIC REVIEW
bank funds to finance his speculation in the
stock market. The individual and the two
banks were all under investigation by the
OCC and the U.S. Attorney General as early
as October 1931.
South Ashland and Standard are interesting
examples of banks whose accounts as of December 1931 do not provide pictures of their
true position prior to the panic. While their
scores in our models are very strong, their
condition according to the examiners was extremely weak. It is possible that some or all
of the four state bank outliers may be explicable in similar terms. After all, if the strength
of a six-month-old balance sheet were enough
to conclusively indicate a bank's strength,
asymmetric-information panics could never
occur.
The Comptroller's examination reports indicate that information aboutfraudand risk taking
by banks that failed during the panic surfaced
between December 1931 and April 1932.
Clearly, investors in bank stock detected special
problems in the banks that failed during the
panic in those same months. As Figure 1 shows,
the market-to-book value ratios for panic failures
fell sharply from January through April 1932.
The market-to-book ratio of surviving banks did
not decline nearly as precipitously.
The wealth of OCC file material we discovered led us to search for the examination records of the other national banks in our sample
that failed during the June panic. It is difficult
to quantify the statements of examiners (to
convert them to scores), but it is easy to summarize their content. We were able to locate
information for all but one of the other national
banks that failed during the panic. In every
case for which we have records, the bank examiners had indicated extreme problems at the
bank at least as early as the end of April 1932.
Following are excerpts from (or synopses of)
the examiner's remarks about each of these
banks:
(1) Jackson Park National Bank (Unpublished examination report, April 27,
1932): "... the condition of this institution remains highly unsatisfactory from
every angle. It will be noted criticized assets have increased materially since last
examination."
DECEMBER 1997
(2) National Bank of Woodlawn (Unpublished examination report, April 25,
1932): "The report of an examination ...
completed April 25 ... shows a bond depreciation ... which greatly exceeds the
bank's entire capital, surplus, and undivided profits ...."
(3) Bowmanville National Bank (Unpublished examination report, Letter of April
6,1932): "The report of the examination
... completed February 25 shows such an
unsatisfactory condition it is requested
that you hold a meeting with the directors
or a committee thereof and ascertain
whether or not something further can be
done to strengthen the bank, after which
this office would be advised fully of the
conclusions reached ... the solvency of
[the bank] is questioned in view of the
doubtful and loss items and the bond
depreciation."
(4) Midland National Bank (Unpublished examination report, April 27, 1932): "This
little institution is struggling along probably as best as could be expected under
the circumstances ... Loss of $1,000,000
in deposits within a year has just about
taken away earning capacity." "The report ... shows a bond depreciation of
$162,004 and losses of $16,041 in loans,
which consume the surplus fund, undivided profits and reserve for contingencies and impair the bank's capital to the
extent of $67,414."
(5) Hyde Park-Kenwood National Bank (Unpublished examination report, February
10,1932): "The report of an examination
of your bank completed December 28 ...
shows an exceedingly unsatisfactory condition and that you are confronted with a
serious situation. This conclusion is based
upon the slow and doubtful assets ...
shown in the report; the fact that practically all of the bank's bonds and securities
are pledged and the bank is now a heavy
borrower, the Chief Examiner advising
that the total borrowings on January 28
amounted to $684,000 due to the heavy
decline in deposits."
(6) Ravenswood National Bank (Unpublished examination report, April 18,
1932): "The report of an examination of
VOL. 87 NO. 5
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
your bank completed April 18 ... discloses several unsatisfactory conditions
as set out by the examiner throughout the
report and itemized on page 11, the correction of which should be effected as
rapidly as possible. The report shows a
high percentage of the bank's loans to be
in very unsatisfactory condition ...." By
May 9, accounting fraud was also discovered to have taken place by a former
employee.
These qualitative statements about national
banks that failed during the panic reinforce the
evidence from Tables 2 and 3 that panic failures were among the weakest banks in the system at the time of the panic.10
III. Conclusion
We have compared the attributes of banks
that failed during the Chicago panic of June
1932 to those of banks that failed at other
times in early 1932, and those of banks that
survived the period using a variety of standards of comparison. Comparisons of the
market-to-book value of equity, the estimated
probability of failure or duration of survival,
the rates of withdrawal of debt during 1931,
and the interest rates paid on borrowed money
lead to the same conclusion: failures of banks
during the panic reflected the continuation of
the same process that produced failures before
the panic. The special attributes of failing
banks are distinguishable months before the
panic and were reflected in stock prices, failure
10
The Chicago Bank of Commerce, the largest bank to
fail during the panic and the only Chicago Loop bank to
fail, had an estimated probability of failure of 0.00572 in
the out-of-sample logit, and an estimated probability of
failure of 0.448 in the in-sample logit. Four panic failures
had lower out-of-sample estimated probabilities of failure
than the Chicago Bank of Commerce, and seven panic
failures had lower in-sample estimated probabilities of
failure. The panic failures with lower estimated failure
probabilities (for which we have examination reports)
were viewed as severely troubled banks by examiners.
While it is not possible to determine whether the Bank of
Commerce was solvent or insolvent during the panic using
its logit or survival scores alone, we are able to say that
its scores were not unusual relative to panic failures that
were perceived by examiners as troubled institutions.
881
probabilities, the opinions of bank examiners,
debt composition, and interest rates.
We conclude that failures during the panic
reflected the relative weakness of failing banks
in the face of a common asset value shock
rather than contagion. The panic was precipitated by exogenous asset-price decline, and the
banks that failed during the panic were among
the weakest banks in the city. While asymmetric information between depositors and
banks precipitated a general run on banks, our
evidence suggests that this asymmetricinformation problem did not produce failures
of solvent banks.
If the risk of solvent banks failing during an
asymmetric-information panic is not high (as
the evidence from the Chicago panic suggests), that could have important implications
for bank regulatory policy. Deposit insurance
and government assistance to banks since the
Depression have been motivated in part by the
perception that bank failures during the Depression were a consequence of contagion,
rather than the insolvency of individual banks.
If private interbank cooperation, buttressed by
liquidity assistance from the monetary authority (like the assistance provided by the RFC
to the Chicago clearinghouse), are adequate to
preserve systemic stability, then a far less ambitious federal safety net might be desirable
(Calomiris, 1990).
Our findings lend support to James's (1938)
account of the role of interbank cooperation in
mitigating the costs of the banking crisis. The
limited duration and costs of contagion may
have reflected the cooperative intervention by
the Chicago clearinghouse, which used its liquid assets to protect at least one solvent bank
from unwarranted attack until the runs by uninformed depositors subsided. Absent such cooperation, the failure experience during the
panic of June 1932 could have been very different. Our evidence suggests, somewhat contrary to the portrayals in Friedman and
Schwartz (1963) and Bernanke (1983), that
clearinghouses continued to serve an important function during the Great Depression, and
did not see the Fed or the RFC as ''reliev [ing]
them of the responsibility of fighting runs"
(Bernanke, 1983 p. 260). How far can one
generalize from these conclusions? Was the
Chicago panic of June 1932 representative of
882
DECEMBER 1997
THE AMERICAN ECONOMIC REVIEW
other banking panics during the Great Depression? Because panics and waves of bank failure were scattered across time and location
during the Great Depression, we believe answering that question will require analysis of
other local panics, using detailed bank-level
data similar to those we have analyzed for the
Chicago panic. Defining and analyzing those
events is an important area for future research
on the causes of bank failures during the
Depression.
REFERENCES
Bank and Quotation Record (of the Commercial
and Financial Chronicle). New York: Dana,
various issues.
Benston, George J. "Interest Payments on Demand Deposits and Bank Investment Behavior." Journal of Political Economy,
October 1964, 72(5), pp. 431-49.
Bernanke, Ben S. ' 'Nonmonetary Effects of the
Financial Crisis in the Propagation of the
Great Depression." American Economic
Review, June 1983, 73(3), pp. 257-76.
Bhattacharya, Sudipto and Thakor, Anjan V.
"Contemporary Banking Theory." Journal
of Financial Intermediation, October 1993,
3(1), pp. 2-50.
Calomiris, Charles W. "Is Deposit Insurance
Necessary?" Journal of Economic History,
June 1990, 50(2), pp. 283-96.
"Regulation, Industrial Structure,
and Instability in U.S. Banking: An
Historical Perspective," in Michael
Klausner and Lawrence J. White, eds.,
Structural change in banking. Homewood, IL: Business One Irwin, 1993, pp.
19-116.
.. "Is the Discount Window Necessary?
A Penn Central Perspective." Review, Federal Reserve Bank of St. Louis, May/June
1994, pp. 31-55.
Calomiris, Charles W. and Gorton, Gary B.
"The Origins of Banking Panics: Models,
Facts, and Bank Regulation," in R. Glenn
Hubbard, ed., Financial markets and financial crises. Chicago: University of Chicago
Press, 1991, pp. 109-73.
Calomiris, Charles W. and Kahn, Charles M.
' 'The Role of Demandable Debt in Structuring Optimal Banking Arrangements."
American Economic Review, June 1991,
81(3), pp. 497-513.
Calomiris, Charles W. and Schweikart, Larry.
"The Panic of 1857: Origins, Transmission,
and Containment." Journal of Economic
History, December 1991, 57(4), pp. 80734.
Calomiris, Charles W. and Wheelock, David C.
"The Failures of Large Southern Banks
During the Great Depression." Working paper, Columbia University, November 1995.
Champ, Bruce; Smith, Bruce D. and Williamson,
Stephen D. "Currency Elasticity and Banking Panics: Theory and Evidence." Working paper, University of Western Ontario,
1991.
Chicago Tribune. Various issues, 1932.
Commercial and Financial Chronicle. New
York: Dana, July 2, 1932, 135(1), p. 71.
Diamond, Douglas W. and Dybvig, Philip H.
"Bank Runs, Deposit Insurance, and Liquidity." Journal of Political Economy,
June 1983, 97(3), pp. 401-19.
Donaldson, R. Glenn. "Sources of Panics: Evidence from the Weekly Data." Journal of
Monetary Economics, November 1992,
30(2), pp. 277-305.
"Financing Banking Crises: Lessons
from the Panic of 1907." Journal of Monetary Economics,
February 1993,
31(1),
pp. 69-95.
Esbitt, Milton.' 'Bank Portfolios and Bank Failures During the Great Depression: Chicago." Journal of Economic History, June
1986, 46(2), pp. 455-62.
Federal Reserve Bulletin. Washington, DC:
Board of Governors of the Federal Reserve
System, various issues.
Friedman, Milton and Schwartz, Anna J. A Monetary History of the United States, 18671960. Princeton, NJ: Princeton University
Press, 1963.
Gorton, Gary B. and Pennacchi, George G. ' 'Financial Intermediation and Liquidity Creation." Journal of Finance, December 1990,
45(1), pp. 49-72.
Imbens, Guido W. "Transition Models in a
Non-Stationary Environment." Review of
Economics and Statistics, November 1994,
96(4), pp. 703-20.
James, F. Cyril. The growth of Chicago banks,
volume 11. New York: Harper, 1938.
VOL. 87 NO. 5
CALOMIRIS AND MASON: BANK FAILURES DURING THE DEPRESSION
Kaufman, George G. "Bank Contagion: A Review of the Theory and Evidence." Journal
of Financial Services Research, April 1994,
8(2), pp. 123-50.
Mason, Joseph R. "Determinants and Effects of
RFC Lending to Banks During the Great Depression." Working paper, Office of the
Comptroller of the Currency, November 1995.
Mishkin, Frederic S. "Anatomy of a Financial
Crisis." National Bureau of Economic Research (Cambridge, MA) Working Paper
No. 3934, December 1991.
Office of the Comptroller of the Currency. Unpublished examination reports and letters
pertaining to national banks in Chicago,
1931 and 1932.
Annual report. Washington, DC:
U.S. Government Printing Office, various
issues.
Reports of Condition of (Federal Reserve Member)
Banks. Unpublished records, various calls.
Saunders, Anthony J. and Wilson, Berry. "Contagious Bank Runs: Evidence from the
1929-1933 Period." Working paper, New
York University, 1994.
Schwartz, Anna J. "The Misuse of the Fed's
Discount Window." Review, Federal Reserve Bank of St. Louis, September/October, 1992, 74(5), pp. 58-69.
883
Superintendent of Banking of the State of Illinois.
Statement showing total resources and liabilities of Illinois state banks at the close of
business. Springfield, IL: Journal Printing
Company, various issues.
Thies, Clifford and Gerlowski, Daniel. "Bank
Capital and Bank Failure, 1921 -1932: Testing the White Hypothesis." Journal of Economic History, December 1993, 53(4), pp.
908-14.
Wheelock, David C. "Deposit Insurance and
Bank Failures: New Evidence from the
1920s." Economic Inquiry, July 1992,
30(3), pp. 530-43.
White, Eugene N. "A Reinterpretation of the
Banking Crisis of 1930." Journal of Economic History,
M a r c h 1984, 44(1),
pp.
119-38.
Wicker, Elmus R. "A Reconsideration of the
Causes of the Banking Panic of 1930."
Journal of Economic History, September
1980, 40(3), pp. 571-83.
"A
Reconstruction of the Gold and
Banking Crises in 1931." Working paper,
Indiana University, October 1993.
Wigmore, Barry A. "Was the Bank Holiday of
1933 Caused by a Run on the Dollar?"
Journal of Economic History, September
1987, 47(3), pp. 739-56.