Full text of Business Review (Federal Reserve Bank of Philadelphia) : March/April 1986

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

IP
I
W

! ■

m W

ISSN 0007-7011

Federal Reserve Bank of Philadelphia

MARCH •APRIL 1986




Federal Reserve Bank of Philadelphia
Ten Independence Mall
Philadelphia, Pennsylvania 19106
MARCH/APRIL 1986

THE PRIVATE COSTS OF BANK FAILURES:
SOME HISTORICAL EV ID EN C E.......................................................................3
Brian C. Gendreau and Scott S. Prince
Bank failures impose costs not only on people with a direct financial stake in the bank, but
also on society — which is the main reason why banks today are so heavily regulated. While
the costs to society are inherently hard to measure, the direct private costs — fees, and so forth
— can be estimated by looking at data on bank failures from the pre-1930s, prior to much
regulation. An analysis of these costs suggests that they were, and probably remain, large
enough to discourage small banks from having a dangerously high ratio of debt to capital;
these costs appear to have been too small, however, to have such an effect on large banks.

WARM FEELINGS AND COLD CALCULATIONS:
ECONOMIC THEORIES OF PRIVATE TRANSFERS.................................... 15
Donald C. Cox and Robert H. DeFina
In the past decade, economists have broken ground in what traditionally has been the
territory of the m ore"social" of the social scientists. That territory is the family, and the aim has
been to understand how families allocate resources among their members. The research
branches into two lines of theory — "altruism" and "exchange." Surprisingly enough, each
theory also has implications for policymakers, especially regarding the effectiveness of
income redistribution programs. In the "altruism" framework, the effects of these policies are
offset by family behavior, and in the "exchange" framework, the effects are magnified.
The BUSIN ESS REVIEW is published by the
Department of Research every other month. It is edited
by Judith Farnbach. Artwork is directed by Ronald B.
Williams, with the assistance of Dianne Hallowell.
The views expressed herein are not necessarily those
of this Bank or of the Federal Reserve System. The
Review is available without charge.
Please send subscription orders and changes of
address to the Department of Research at the above
address or telephone (215) 574-6428. Editorial
communications also should be sent to the Department
of Research or telephone (215) 574-3805. Requests for
additional copies should be sent to the Department of
Public Services.
The Federal Reserve Bank of Philadelphia is part of
the Federal Reserve System—a System which includes



twelve regional banks located around the nation as well
as the Board of Governors in Washington. The Federal
Reserve System was established by Congress in 1913
primarily to manage the nation's monetary affairs.
Supporting functions include clearing checks, providing
coin and currency to the banking system, acting as
banker for the Federal government, supervising
commercial banks, and enforcing consumer credit
protection laws. In keeping with the Federal Reserve
Act, the System is an agency of the Congress, in­
dependent administratively of the Executive Branch,
and insulated from partisan political pressures. The
Federal Reserve is self-supporting and regularly
makes payments to the United States Treasury from
its operating surpluses.

The Private Costs of Bank Failures:
Some Historical Evidence
Brian C. Gendreau and Scott S. Prince*
Can unregulated financial markets be relied
upon to constrain bank leverage? How much
debt would banks choose to issue relative to
capital in the absence of regulation? The answers
to these questions depend upon the costs bank
owners and creditors can expect to bear in the
event of a bank failure. When a bank increases
its debt, it also increases the chances that its

"Brian C. G endreau is an Economist in the Research Depart­
ment of the Federal Reserve Bank of Philadelphia. Scott S.
Prince is with the Municipal Bond Department, Drexel Bumham Lambert, Inc. in New York.




future earnings may not be enough to cover the
principal and interest payments it has promised
on that debt. A rising level of debt relative to
capital, other things being equal, increases the
risk of bank failure. But if bank failures are costly,
bank stockholders and creditors will seek to
induce the bank to avoid excessive levels of debt
relative to capital. As a result, the market, by
acting to avoid bankruptcy costs, provides a
mechanism that is at least theoretically capable
of regulating bank leverage.
Though a great deal has been written about
bankruptcy costs and bank leverage decisions,
little is known about how much it costs for banks
to fail. In this article we present some estimates
3

BUSINESS REVIEW

of direct, private costs of bank failures. These
cost estimates are derived from the U.S. Comp­
troller of the Currency's published figures on
receiver and legal fees involved in national bank
failures between 1865 and 1934. Because these
figures are from the years before the existence of
the Federal Deposit Insurance Corporation
(FDIC), which now subsidizes bank failure costs,
they represent fairly accurate measures of the
out-of-pocket costs borne by stockholders and
creditors when banks fail. To anticipate this arti­
cle's findings, the direct costs of bank failures are
large enough to affect the capital decisions of
small banks, but are trivially small for large
banks.
THE COSTS OF BANK FAILURES
Banking today is one of the most heavily regu­
lated industries in the U.S. Banks are subject to
numerous restrictions on their asset choices, lia­
bility structures, locational decisions, and
participation in businesses not deemed closely
related to banking. Most of this regulation is de­
signed to prevent bank failures. Specifically, the
government is interested in protecting the public
from the social costs of bank failures. The social
costs of bank failures are those borne by third
parties: people other than the banks' owners and
creditors. If a spate of bank failures results in dis­
ruptions to credit markets, for example, the costs
of these disruptions are likely to be borne by the
public as well as by bank owners and creditors.
Borrowers, finding it difficult to obtain credit after
a series of bank failures, may respond by cutting
back on their production and consumption, lead­
ing to a reduction in spending and employment in
their communities, and possibly in other com­
munities as well. A widespread reduction in eco­
nomic activity could, in this way, affect large num­
bers of people wholly unconnected to the failed
banks.1

1See Ben S. Bemanke, "Nonmonetary Effects of the Finan­
cial Crisis in the Propagation of the Great Depression,"
American Economic Review, 73 (June 1983), pp. 257-276.

4



MARCH/APRIL 1986

Though the government and public at large are
likely to be concerned principally about the social
costs of bank failures, it is the private costs of bank
failures that are germane to bank liability structure
decisions. The private costs of bank failures are
those borne by bank shareholders and by indi­
viduals and firms with a contractual financial inter­
est in the bank—the banks' depositors, bondhold­
ers, and other creditors. These private costs are of
two kinds: the direct, out-of-pocket expenses in­
curred by stockholders and creditors during bank­
ruptcy proceedings following a bank failure, and
the indirect or hidden costs involved in operating
or doing business with a troubled or failed bank.
The Direct Private Costs of a Bank Failure.
When a firm defaults, that is, when it has been
unable to make principal or interest payments on
its debt, it is often declared bankrupt. Bankruptcy
is a legal procedure for resolving the claims on a
failed firm or for reorganizing a troubled firm. A
firm may file for bankruptcy voluntarily, or it may
be declared bankrupt by a court upon a petition by
the creditors. If there is no hope of saving the firm
as a going concern, the firm will cease operating,
will be discharged from some or all of its debts,
and will turn over its assets to a court-appointed
trustee for distribution to the creditors. If, instead,
there is a prospect of saving the firm, it will be
reorganized, and will obtain shelter from its debt
obligations.2
Banks are not covered by the U.S. bankruptcy
laws, but by the laws governing their chartering
agencies and the regulations issued by those agen­
cies. The procedures for resolving the claims on
failed banks, however, are similar to those govern­
ing the liquidation of nonbank firms, except that
banks are closed by their chartering agency, and it
is a receiver appointed by the chartering agency
who liquidates the bank. In this paper, the word
bankruptcy will be used in a generic sense to refer

2 See Michelle J. White, "Bankruptcy, Liquidation, and
Reorganization," Salomon Brothers Center for the Study of
Financial Institutions, Working Paper No. 304 (October 1983)
for a discussion of the law and economics of corporate
bankruptcy.

FEDERAL RESERVE BANK OF PHILADELPHIA

The Private Costs of Bank Failures

to both nonbank firms and banks.3
By the time a firm fails, the value of its assets will
usually have fallen dramatically. This fall in asset
values is commonly mistaken for the bankruptcy,
or is considered to be a cost of the bankruptcy. In
reality it is neither: bankruptcy is the legal state
which was triggered by the fall in asset values, and
not the fall in asset values themselves. Indeed, a
firm's asset values can fall to zero without the firm
being declared bankrupt Consider a firm financed
entirely by equity, in which each shareholder is
entitled to a proportionate share of the firm's earn­
ings, regardless of whether the firm prospers or
not. If the firm does so poorly that it has no current
or prospective earnings, the value of its assets will
fall to zero, and the firm's shares will be worthless.
The shareholders are likely to be unhappy about
this turn of events, but the firm did not promise to
make regular payments of any kind on its securi­
ties, and hence cannot be declared bankrupt.4
Going through bankruptcy proceedings, how­
ever, is not costless, and it is the expenses involved
in administering the liquidation of the firm's assets
or its reorganization that constitute the direct,
private costs of a firm's distress. These costs include
the fees paid to the trustee or receiver appointed
by the court or regulatory agency to oversee the
liquidation, and the legal expenses incurred by
those with claims on the firm's remaining assets.
In bankruptcies in the U.S., the receiver's and
lawyers' fees are paid by the court or bank charter­
ing agency from the proceedings of the liquidation.
The firm's shareholders and creditors ultimately
bear these costs because they receive only the
value of the assets remaining after the receiver's

3Before the creation of the FDIC in 1934, banks were
reorganized in much the same m anner as other firms. Since
the advent of the FDIC, however, troubled banks have been
merged into healthy banks rather than reorganized.
4To be more concrete, suppose you have bought shares in
a mutual fund. If the value of the shares drops precipitously
you would be unhappy, but you would not think that the
fund had become bankrupt. The distinction is that the mutual
fund has only ownership shares as liabilities, and not debt.
What you own is not worth as much as when you invested,
but no bankruptcy proceedings need be initiated.




Brian C. Gendreau & Scott S. Prince

and lawyers' fees have been paid (see BANK
LIQUIDATIONS — THEN AND NOW, p.6).
Indirect Bankruptcy Costs. In addition to the
administrative expenses of going bankrupt, failed
firms incur many indirect costs that a healthy firm
could avoid. Many of these costs arise out of the
confusion and inevitable inefficiency involved in
operating a bankrupt or nearly bankrupt firm.
Ailing firms, for example, are likely to experience
difficulty hiring or keeping employees. Similarly,
bankrupt firms are likely to find that their managers
are spending a considerable amount of time in
legal proceedings rather than managing the firm.
Other indirect costs reflect forgone opportuni­
ties. Whenever a firm fails, its customers are de­
prived of its product or service until they can locate
substitutes. Customers are likely to be reluctant to
deal with a troubled firm or a firm that is being
reorganized after being declared bankrupt because
they do not wish to incur the "shoeleather" costs
involved in establishing a new business relation­
ship if the firm ceases to operate as a going concern.
As a result, an ailing firm may find its business
falling off rapidly. In addition, a troubled firm is
likely to find it very costly if not impossible to
borrow in financial markets. Consequently, the
firm will be unable to take advantage of many
investment and sales opportunities open to sol­
vent firms.
Banks are particularly exposed to indirect
bankruptcy costs because their depositors are
both their customers and their creditors. In the
event of a bank closing, not only will the deposi­
tors have to find a new banking relationship, but
they also will likely be denied access to their
funds until their claim against the closed bank is
settled. For uninsured depositors, this delay
could be quite long. Although bank receivers
are required to liquidate a bank as quickly as
possible, it is not uncommon for a bank liquida­
tion to take over a decade.5

5The longest national bank liquidation on record is for the
Farmers and Drovers National Bank of Waynesboro, Pennsyl­
vania, which failed in 1906. This unfortunate bank was in
liquidation for the next 32 years.

5

BUSINESS REVIEW

MARCH/APRIL 1986

BANK LIQUIDATIONS— THEN AND NOW
Before the creation of the Federal Deposit Insurance Corporation (FDIC) in 1934, most deposits in
U.S. banks were uninsured. Deposits in some banks had been insured by state deposit guarantee funds,
but those funds had proven inadequate to deal with widespread bank failures, and all had ceased
operations one way or the other in the 1920s. In those days, a bank failure was treated much as any other
corporate bankruptcy. The bank was closed by the U.S. Comptroller of the Currency if it had a national
charter, or by its state bank supervisory agency if it had a state charter. The closing agency appointed a
receiver to sell the bank's assets, with instructions to pay out the proceeds to claimants in the following
order: the receiver's salary and all legal fees were paid first, depositors and other creditors were paid
next, and the bank's owners were paid last—but only if there were funds left over. Consequently, the
bank's creditors and owners bore the direct costs of administering and liquidating the bank after its
failure.
The FDIC has handled bank failures differently. Occasionally the FDIC has closed insured banks and
paid off the depositors directly. In these cases insured depositors have been paid in full (up to the
insurance limits per account), and the uninsured depositors have been paid out of the liquidation
proceeds of the bank's remaining assets, net of administrative and legal costs. Thus uninsured depositors
have remained exposed to bankruptcy costs, just as they were before the creation of the FDIC. Usually,
however, the FDIC has merged failing banks into sound financial institutions, with some financial
assistance from the FDIC, in transactions known as "purchase and assumptions." Because all deposits
are absorbed into the assuming bank in a purchase and assumption, depositors suffer no losses:
effectively, the FDIC has provided 100 percent deposit insurance. And because the failed bank is not
liquidated, there are no liquidation costs for the uninsured depositors and bank owners to bear.
Purchase and assumption transactions are not completely costless to depositors in that established
banking relationships are lost, but the costs are much lower than if the bank were closed and depositors
paid off.

By their very nature, indirect bankruptcy costs
are difficult to measure. It is hard to quantify the
costs of the disorder, inefficiency, and the lost
opportunities that accompany a business or bank
failure. Nonetheless, these indirect costs may be
large, and as such can be of substantial concern
to stockholders and creditors.
BANKRUPTCY COSTS AND THE
BANK LEVERAGE DECISION
Sometimes banks fail through events such as
bank runs or severe local economic downturns
that are no direct fault of their own. But bank
managers and owners can also make decisions
that affect the probability of going bankrupt. To
the extent that bankruptcy results in private
costs that would not have been incurred other­
wise, bank owners and uninsured creditors will
want to induce managers to take actions to avoid
those costs. (This is aside from any public-spir­
ited interest stockholders and creditors may have
6



in minimizing the social costs of bank failures.)
One important way bank managers can affect
the likelihood of bankruptcy is through the bank
leverage decision—the decision on how much
debt to issue relative to capital. Banks have at
least two motivations for issuing debt. First,
deposits—an important component of bank
debt—are closely linked to the provision of many
nondeposit services, such as checking, loans,
wire transfers, and cash management services. It
is possible to provide most bank services to
customers who are not depositors, but in many
cases it is convenient and economical for both
the bank and its customers if the services are
provided along with a deposit account.6 Second,

6See Thomas Gilligan, Michael Smirlock, and William
Marshall, "Scale and Scope Economies in the Multi-Product
Banking F i r m Journal of Monetary Economics, 13 (May 1984),
pp. 393-405, for evidence on joint economies in the produc­
tion of loans and deposits.

FEDERAL RESERVE BANK OF PHILADELPHIA

The Private Costs of Bank Failures

because banks can deduct interest payments
from federal income taxes but cannot deduct
dividends or retained earnings, it is cheaper for
a bank to issue interest-bearing debt than to
issue stock.
On the basis of these considerations, it would
appear that an unregulated bank should issue as
much debt as it can. But as a bank increases its
leverage, it increases the probability that it will
not be able to m eet its promised principal and
interest payments and thus be closed. With a
little foresight, uninsured creditors can see that
if the bank is closed they will bear bankruptcy
costs. Therefore, long before a closing becomes
imminent, uninsured creditors will demand
compensation in the form of higher interest
payments for bearing the anticipated bankruptcy
costs. But the higher interest payments will re­
duce the net earnings available to stockholders,
reducing the market value of their shares in the
bank. The result is that at some point the market
value of the bank will fall as managers increase
leverage. At that point, the bank's leverage may
be said to be excessive. To the extent that bank
managers are responsive to stockholders' wishes
(and there is every reason to believe that they
are in the long run) those managers will try to
avoid excessive leverage positions—those levels
of debt relative to capital that are so great that the
value of the bank declines.7

Brian C. Gendreau & Scott S. Prince

MEASURING BANK FAILURE COSTS
Precisely how leveraged a bank can become
before its value begins to decline depends criti­

cally on the size of expected bankruptcy costs.
Very few estimates of the cost of going bankrupt,
however, have been made; indirect bankruptcy
costs are inherently difficult to measure, and
data on direct bankruptcy costs are not collected
or published in any systematic manner for most
industries. It is particularly difficult to measure
the private costs of recent bank failures because
the FDIC effectively subsidizes bankruptcy costs.
In the past 35 years, the FDIC has typically dealt
with bank failures by merging moribund insured
banks with healthy financial institutions rather
than closing the banks and paying off their de­
positors.8 Because a merged bank does not go
through liquidation, its owners and creditors
are spared the direct costs of going through a
bankruptcy, and many of the indirect costs, too.
This does not mean that they were not con­
cerned about bankruptcy costs: no bank owner
or creditor could ever be sure that his or her
bank would be merged rather than closed by the
FDIC—and the FDIC has recently been closing
more banks than in past years. But there have
not been enough banks closed outright in recent
years to provide accurate and representative
data on bank failure costs.
Before the creation of the FDIC in 1934, how­
ever, the direct costs of bank failures were borne
explicitly by bank owners and creditors. Each
year between 1865 and 1934 the U.S. Comp­
troller of the Currency reported in his Annual
Report the legal expenses and receivers' fees for
every national bank failure. By looking at the
size of the legal and receivers' fees relative to the

7This discussion has assumed that bank leverage positions
are not constrained by regulation. Banks today are subject to
regulatory capital requirements specifying minimum ratios
of book capital to assets. These capital requirements are
intended to restrain bank risk-taking, and thus serve as a
substitute for the market's regulation of bank leverage.

ar*d Investor Risk Perceptions: Some Entailments for Capital
Adequacy Regulation, Journal of Bank Research, (Autumn,
1975), pp. 190-201, and Brian Gendreau and David Burras
Humphrey, "Feedback Effects in the Market Regulation of
Bank Leverage: A Time-Series and Cross-Section Analysis,"
Review of Economics and Statistics, (May 1980), pp. 276-280.

Empirical studies, however, have found that large bank
leverage positions in the 1970s were sensitive to measures
of debt or equity costs or both, suggesting that regulation
was not binding on large banks in those years. See H. Presscot
Beighley, John H. Boyd, and Donald P. Jacobs, "Bank Equities

8See Joseph F. Sinkey, Jr., in Problem and Failed Institutions
in the Commercial Banking Industry, Edward I. Altman and
Ingo Walter, eds. (Greenwich, CT: JAI Press, 1979), for a
description of the FDIC's procedures in handling troubled
and failed banks.




7

BUSINESS REVIEW

liabilities of failed banks it is possible to estimate
the direct bankruptcy costs incurred by holders
of financial claims on failed banks.
The Data on Bank Failure Costs. Up to 1904,
the Comptroller published the balance sheets of
all national banks each year, making it easy to
match failure costs to individual bank liability
figures. We collected data on every national bank
out of the 428 that failed between 1865 and 1904
for which figures were available. The result was
a sample of 200 bank failures, with the first bank
failing in 1872 and the last in 1904. This sample
is the largest compiled to date in a study of
bankruptcy costs.
Ideally, bankruptcy costs should be measured
relative to the market value of bank equity and
debt. Only book values, however, are available
for the sample of pre-1904 bank failures. To the
extent that accountants' book values measure
banks' economic fortunes imperfectly, using
book instead of market values will introduce
measurement error into the estimates. By 1920,
however, shares of many large banks were traded
actively on New York and regional stock ex­
changes, allowing the market value of bank equity
to be found for these banks. For this reason we
took a second sample, this time of large banks
that failed after 1920. The second sample was
limited to those banks with capital greater than
$750 million because they are the only banks for
which both stock prices and balance sheet data
are readily available. We examined the records
on every large national bank failure between
1920 and the creation of the FDIC in 1934, and
collected data on every failure for which com­
plete figures were available. The result was a
sample of 21 bank failures, the first of which
occurred in 1929 and the last in late 1933.9
In gathering the data for the sample of large

9This sample unfortunately omits three of the largest
national bank failures of the 1929-1933 period. Bank-specific
stock prices were not available for two banks that failed in
the Guardian group holding company of Detroit, MI, nor for
the failed Harriman National Bank and Trust Company of
New York City.

8



MARCH/APRIL 1986

bank failures, we took balance sheet figures from
Moody's Investors Manual whenever possible, and
from Polk's Bankers Encyclopaedia or RandMcNally's Bank Directory otherwise (the Comp­
troller had ceased publishing individual bank
balance sheets in his Annual Report in 1904). We
obtained stock prices from the Commercial and
Financial Chronicle if they were listed there, and if
not, from Moody's, Polk's or Rand-McNally's
publications. In most cases, we collected the
stock prices as averages of the high and low bid
prices for the year because the shares were not
traded actively enough to provide price quotes
on the financial report dates.
The Estimates. Table 1 summarizes the statis­
tics for the pre-1904 bank failures. This sample
contains observations from banks ranging in
size from only $88,000 to over $15 million in
total liabilities (measured the year prior to failure),
with an average size of about $843,000. As can
be seen in this table, liquidation was no quick
process for these banks; on average, it took almost
6 years to settle the affairs of a closed bank. Nor
was the bankruptcy process inexpensive: the
cumulative sum of all fees paid to the receivers
and claimants' lawyers during liquidation
averaged $30,140 per bank.
It would be misleading to compare these direct
costs of bankruptcy to the amount of funds in­
vested by shareholders and creditors by using
the banks' total liabilities on the date of failure.
In the period just before they failed the banks
may have experienced a run on their deposits or
a rapid fall in the value of their capital. Hence we
used total book liabilities figures from one year
before the failure date for each bank. For the
banks in the pre-1904 sample, total receiver and
legal fees averaged about 5.6 percent of total
liabilities. Though most of the banks in the sample
had ratios of direct bankruptcy costs to liabilities
that were close to the average, a few banks had
ratios that were considerably different, ranging
from a low ratio of 0.3 percent to a high ratio of
18.3 percent.
Table 2, p. 10, presents statistics for the sample
of large banks that failed between 1929 and
FEDERAL RESERVE BANK OF PHILADELPHIA

The Private Costs of Bank Failures

Brian C. Gendreau & Scott S. Prince

TABLE 1

SUMMARY STATISTICS FOR 200 NATIONAL BANK FAILURES
1872-1904
Mean
Total receiver and
legal fees3

Standard Deviation

Maximum

Minimum

$30.1

$34.3

$312.0

$2.0

5.9

4.2

32

0.3

Total book liabilities3

$843.3

$1,726.9

$15,002.0

$88.0

Ratio of debt to book
liabilities

66.8%

11.2%

91.8%

28.4% b

Ratio of receiver and
legal fees to book
liabilities

5.6%

2.9%

18.3%

0.2%

Number of years in
liquidation

NOTE:

All figures are book values. Liabilities are measured as of one year prior to the bank failure date.

aDollar figures are in thousands.
bThis implausibly low ratio of debt to book liabilities is not an error: it reflects the actual value of the notes and
deposits relative to total liabilities of the National Bank of Paola, Kansas, on October 5, 1897, the year prior to its
failure. The puzzle is how such a heavily-capitalized bank could have failed.

1933. Banks in this sample varied in size from
$4.7 to $62.3 million in liabilities, and averaged
$19.7 million in total liabilities as of a year before
failure. This seems like a wide range, but the
coverage in the sample is actually fairly narrow.
To put the sample in perspective, national banks
in 1929 ranged in size from less than $1 million
in liabilities per bank in Nebraska, North Dakota,
and South Dakota to over $221 million per bank
in New York City. Liquidation for the banks in
the later sample averaged a little over 9 years,
which was longer than it was for the banks in the
pre-1904 failures sample.
With more information available on the large
banks in the second sample, we were able to
compare direct bankruptcy costs to bank liabilities
in a variety of ways, providing a check on the
pre-1904 sample's estimates. First, we calculated
total (cumulative) receiver and legal fees as a



ratio of book liabilities a year prior to failure,
providing the same measure used for the pre1904 bank failures in Table 1. Second, because
the receiver and legal fees involved in each bank
failure were paid out over a number of years, we
computed the ratio of the discounted present
value of these costs to book liabilities for each
bank, discounting the fees back to the failure
date using the prevailing 1-year corporate bond
yield for each year. Third, we compared both the
undiscounted and discounted bankruptcy costs
to estimates of the market value of bank liabilities
as of one year prior to failure. We computed the
market value of bank equity by multiplying the
stock price by the number of shares outstanding.
For the market value of bank debt we simply
used the book value. Normally, the value of
fixed-rate debt will vary with the riskiness of a
firm's underlying assets, so using book values
9

MARCH/APRIL 1986

BUSINESS REVIEW

TABLE 2

SUM M ARY STATISTICS
FOR 21 LARGE NATIONAL
BANK FAILURES, 1 9 2 9 -1 9 3 3
Mean

Standard Deviation

Maximum

Minimum

Total receiver and
legal fees3

$676.7

$410.2

$1,684.0

$21.0

Receiver and legal fees
discounted back to
failure date3

$645.7

$402.9

$1,670.4

20.6

9.1

2.4

12.0

3.0

$19,689.1

$13,588.6

$62,323.0

$4,727.0

$2,548.6
$3,164.1

$1,579.2
$2,285.1

$7,918.0
$9,277.9

$1,188.0
$1,000.0

85.6%

4.7%

92.7%

74.9%

3.66%

1.75%

6.47%

0.44%

3.45%

1.65%

6.34%

0.41%

Number of years in
liquidation
Total book liabilities3
Total equity capital3
(a) Book value
(b) Market value
Ratio of debt to book
liabilities
Ratio of receiver and
legal fees to liabilities
(a) Ratio to book
liabilities
(b) Ratio to market
value of liabilities

aDollar figures are in thousands.

will be misleading. In the case of these banks,
however, the error introduced by using book
values is likely to be small: most of their debt was
in the form of short-term deposits or demand
deposits on which the banks changed the rate of
interest frequently. The way depositors would
have reacted to changing bank risk (barring an
imminent failure) would have been to demand a
higher level of interest payments per dollar
10



deposited—hence the value of the deposits would
usually have been close to their book value of
100 cents on the dollar. Our estimate of the
market value of bank liabilities, then, is the market
value of bank equity plus the book value of bank
debt.
From the figures presented in Table 2 it is
clear that neither discounting the bankruptcy
costs nor using the market value of bank equity
FEDERAL RESERVE BANK OF PHILADELPHIA

The Private Costs of Bank Failures

in measuring bank liabilities makes much of a
difference in the resulting ratios of bankruptcy
cost to liabilities. The discounted receiver and
legal fees are close to their undiscounted coun­
terparts—reflecting both the large fees typical in
the first years of liquidation and the low interest
rates of the 1930s and 1940s—and the market
values of bank equity happened on average to
be not very different from book values.10 How­
ever measured, direct bankruptcy costs for these
21 banks amounted on average to approximately
3.5 percent of their liabilities. Like the pre-1904
failures, most banks in this sample had ratios of
bankruptcy costs to liabilities that were close to
the average, but again there were some exceptions.
Bankruptcy costs consumed less than one-half
of 1 percent of the market value of one bank's
assets, while they accounted for over 15 percent
of the value of another bank.
JUDGING THE IMPORTANCE
OF BANKRUPTCY COSTS
The direct bankruptcy costs for the failed banks
reported in Tables 1 and 2 relative to the values
of the banks' liabilities are of the same order of
magnitude as those reported in other studies of
bankruptcies in nonfinancial firms. Jerold Warner
found that direct bankruptcy costs for 11 railroads
that failed between 1933 and 1955 amounted to
4 percent of the value of the firms' liabilities
(debt plus equity) one year prior to failure. In a
study of the recent failures of 19 retailing and
industrial firms, Edward Altman found that direct
bankruptcy costs were about 6 percent of the

10The simple correlation coefficient, r, between the book
and market values of equity is .57 a year before failure, and
.87 five years before failure. Surprisingly, the market values
are higher than the book values on average. This appears to
reflect the speculative run-up of stock prices in the late
1920s: one bank in the sample, for instance, had paid no
dividends since 1921 yet had a stock price in 1929 that was
twice its book value. In contrast to Warner's finding that the
market value of the securities of the firms in his sample
declined as the date of their failure approached, no such
movement was evident in the stock prices of these banks.




Brian C. Gendreau & Scott S. Prince

value of the firms' liabilities a year before fail­
ure.11
Are direct bankruptcy costs of this size large
enough to affect bank leverage decisions? Is the
prospect of losing one out of every 15 to 20
dollars of the banks' assets to the administrative
and legal costs of bankruptcy proceedings enough
to deter banks from issuing excessive quantities
of debt? Scholars differ on this point. Warner
pointed out that it is the expected cost of bankruptcy
that matters, so that bankruptcy costs must be
multiplied by the probability that the firm will
actually fail to give a reasonable estimate of the
costs the firm's investors could expect to bear.
He showed in an example that multiplying direct
bankruptcy costs of 5 percent of firm value by a
probability of going bankrupt of 5 to 10 percent
results in expected costs that are negligible. Alt­
man, however, judged that direct bankruptcy
costs of this size could not be dismissed as trivial,
and used a statistical failure prediction model to
show that the failure probabilities for the firms
in his sa m p le had b e e n fairly high, averaging 82
percent a year before failure and 58 percent two
years before failure.
While recognizing that our direct bankruptcy
cost estimates are not overwhelmingly large, we
believe that they are big enough to have an
impact on the leverage decisions of many banks.
In competitive debt markets, in which funds
flow from institution to institution in response
to interest rate differentials of only a few basis
points, a potential loss to bankruptcy costs of as
little as 1 to 2 percent of the value of an invest­
ment (100 to 200 basis points) is likely to be
viewed as a serious matter—serious enough,
certainly, to induce some investors to shy away
from such investments, and to induce bankers to
think about how to avoid those costs. In addition,1

11See Jerold B. Warner, "Bankruptcy Costs: Some Evi­
dence," Journal of Finance, 32 (May 1977), pp. 337-347, and
Edward I. Altman, "A Further Empirical Investigation of the
Bankruptcy Cost Question," Journal of Finance, 39 (Septem­
ber 1984), pp. 1067-1089. Altman's study estimates indirect
bankruptcy costs, too.

11

BUSINESS REVIEW

MARCH/APRIL 1986

to the extent that there are also indirect costs to
bankruptcies—and by Altman's estimates they
may be as much as twice as high as the direct
costs—they will serve to reinforce market parti­
cipants' concern with direct bankruptcy costs in
making leverage decisions.
Before coming to a final conclusion about the
significance of our bankruptcy cost estimates to
bank leverage decisions, it is important to point
out that our data suggest that it is proportionately
less expensive for a large bank to go through a
closing and liquidation than it is for a smaller
bank. In the sample of 200 bank failures from
1872-1904, which includes banks of a wide range
of sizes, direct bankruptcy costs rose with bank

size, but not on a one-for-one basis. Instead, for
every 1 percent increase in bank total liabilities,
bankruptcy costs rose by only .65 percent. (See
the Appendix: ECONOMIES OF SCALE IN
BANK FAILURES.) This relationship is illustra­
ted in a slightly different way in Figure 1, which
shows that the relationship between the ratio of
direct bankruptcy costs to liabilities and bank
size (measured by total book liabilities as of a
year prior to failure) is unmistakably negative.
For banks with total liabilities of less than
$500,000, the ratio of direct bankruptcy costs to
liabilities a year prior to failure averaged 6.3
percent; for banks with liabilities of from
$500,000 to $1 million, the ratio averaged 5.3

FIGURE 1

THE RELATIONSHIP BETWEEN DIRECT BANKRUPTCY COSTS
RELATIVE TO LIABILITIES AND BANK SIZE
Ratio of Direct Bankruptcy Costs
to Liabilities

8%

3%

0

1

2

3

4

5

6

7

8

9

10

11

12

13

14
15
Bank Size

(total liabilities in millions $)
NOTE:

Direct bankruptcy costs are total receiver and legal fees paid out during liquidation and liabilities are book

values as of a year before failure. The curve in the figure was fitted by a log-linear regression of the ratio of bankruptcy
costs relative to total liabilities on total liabilities.

12



FEDERAL RESERVE BANK OF PHILADELPHIA

Brian C. Gendreau & Scott S. Prince

The Private Costs of Bank Failures

percent, and for the banks with liabilities of over
$1 million the ratio fell to 3.3 percent. The largest
four banks in the over-$l million category, more­
over, all had ratios of only 1 to 2 percent.12
Given that direct bankruptcy costs do not rise
proportionately with bank size, those costs should
loom larger in the leverage decisions of small
banks than of larger banks. There is probably no
easy way to test this hypothesis, but it is interesting
to note that small banks tend to have compara­
tively low leverage ratios—that is, lower ratios
of debt to total liabilities—while large banks are
usually very highly leveraged. This was true
before 1934, and it is still true today.13 Though
many factors enter into bank leverage decisions,

12Would the bankruptcy of an uninsured bank today be
as costly as the bankruptcy of a comparable bank before
1934? Probably so. Many technological advances in infor­
mation collection and processing have occurred since the
1930s, which have doubtless restrained the rise in bankruptcy
costs. But the principal component of bankruptcy costs is
legal fees, and lawyers' salaries have more than kept pace
with salaries in other professions and with consumer prices
generally since the 1930s. Bankruptcy remains a long, tedious,
and expensive process.
13The ratio of equity capital to assets for insured banks in
1984 averaged 8.5 percent for banks with assets of $100
million or lower, 7.15 percent for banks with assets greater
than $100 million but less than $1 billion, and 4.6 percent for
money center banks with assets exceeding $1 billion. These
ratios are derived from return on equity and return on asset
figures reported by Deborah Danker and Mary McLaughlin,
"Profitability of Insured Commercial Banks in 1984," Federal
Reserve Bulletin, (November 1985), pp. 836-849.

the observed pattern of leverage by bank size is
at least consistent with small banks trying to
avoid incurring the bankruptcy costs that larger
banks can ignore with less peril.
CONCLUSIONS
In this paper we have presented estimates of
the direct costs to failed national banks of going
through bankruptcy before 1934. We used his­
torical data on bankruptcy costs because they
are not distorted by the indirect subsidies that
failing banks have enjoyed in recent years when
regulators have merged them into solvent banks.
Overall, our estimates show that the total admin­
istrative and legal costs of bank failures averaged
between 3 and 6 percent of the value of bank
liabilities. The estimates also show that direct
bankruptcy costs were fairly large relative to the
value of small banks, but negligible in relation to
the value of large banks.
Our estimates of the administrative and legal
costs of going through bankruptcy suggest that
these costs alone are high enough to affect the
leverage decisions of many banks. If regulators
ceased regulating bank leverage positions and
also allowed uninsured depositors to suffer losses,
the prospective direct costs of bankruptcy would
be high enough to affect the debt positions of
small and mid-sized banks. Direct bankruptcy
costs are so low relative to the value of large
banks' liabilities, however, that they are not likely
to have a material impact on large bank leverage
positions.

APPENDIX

ECO N O M IES OF SCALE IN BANK FAILURES
It is possible to think of several reasons why the administrative and legal costs of going through bank­
ruptcy might not vary in direct proportion to bank size. The affairs of a large failed bank might be consid­
erably more complicated than those of a small bank, and thus disproportionately more costly to unravel.
On the other hand, the fixed costs of a bankruptcy, such as the receiver's overhead, may be substantial
relative to the value of a small bank, but negligible relative to the value of a larger bank. Or the legal costs
of resolving a claim against the failed bank may be about the same for each account in the bank, but large



13

MARCH/APRIL 1986

BUSINESS REVIEW

banks may have more large balance accounts than small banks. The issue is at heart an empirical
one.
To see how bankruptcy costs varied with bank size, we used a statistical technique called regression
analysis. Specifically, we regressed the natural logarithm of total direct bankruptcy costs, ln(BC), on the
natural logarithm of bank total liabilities measured one year before failure, ln(711), using data from both
the 1872-1904 and 1929-1933 bank failure samples. The coefficient on the bank liabilities variable in a
regression specified this way measures the percentage change in the bankruptcy costs associated with a
1 percent change in bank size.3 A coefficient of 1.0 would indicate that bankruptcy costs rose on a onefor-one basis with bank size. A coefficient of less than 1.0 would indicate that a 1 percent increase in bank
size was associated with a less than 1 percent increase in bankruptcy costs, and econom ies of scale may
be said to have existed in bankruptcy. If the coefficient is greater than 1.0, it would indicate that
bankruptcy costs rose more than proportionately with bank size, indicating diseconomies of scale in
bankruptcy.
As can be seen in the regression results reported below, the coefficients on (the log of) bank liabilities
using data from the 200 banks that failed between 1872 and 1904 is .654. This is significantly different
from 1.0 at the 99 percent confidence level. Investors in these banks benefited from clear econom ies of
scale in bankruptcy.
For the 21 large national banks that failed between 1929 and 1933 the estimated coefficient is close to
unity, suggesting that no scale econom ies existed in bankruptcy for these banks. Unfortunately, we can
have little confidence in this conclusion because the sample is small and the standard error on the
estimated coefficient is large: we cannot, for instance, reject the hypothesis that the estimated coefficient
differs significantly from .65 at the 1 percent level. We are more comfortable with the scale econom ies
estimate from the earlier bank failures: it is based on many more observations—and from banks of all
sizes rather than just large banks.

Regression Results
A.

1872-1904 Bank Failure Sample:
ln(BC) = -.935 + .654 ln(TLl)
(.238) (.039)
R-squared: .591 Number of observations: 200
Standard error of the regression: .277

B. 1929-1933 Large Bank Failure Sample:
ln(BC) = -3.408 + .994 ln(711)
(2.42) (.249)
R-squared: .455 Number of observations: 21
Standard error of the regression: .682
NOTE: Standard errors are in parentheses. All liabilities are measured using book values. Results using discounted
bankruptcy costs or estimated market values for the liabilities were not materially different from the results reported
above.

aSee Charles R. Frank, Jr., Statistics and Econometrics (New York: Holt, Rinehart & Winston, 1971).

14



FEDERAL RESERVE BANK OF PHILADELPHIA

Warm Feelings and Cold Calculations:
Economic Theories of Private Transfers
Donald C. Cox and Robert H. DeFina*
Many economic policy initiatives redistribute
income and wealth among the populace, either
by accident or by design. Specific program
changes, such as a reduction in Social Security
benefits, as well as broader changes, such as the
suggested overhaul of our tax system, all funnel
resources from one group of people to another.

* Donald Cox is currently an Assistant Professor of Econom­
ics at Washington University in St. Louis, MO. Robert DeFina is
a Senior Economist and Research Advisor in the M acroeco­
nomics Section of the Philadelphia Fed's Research Depart­
ment.




Even monetary and fiscal policies that focus
strictly on the economy's overall performance
have implications for the distribution of income
and wealth.
Such redistributions can affect people's well­
being, or welfare, making certain individuals
better off at the expense of others. Not surpris­
ingly, these reallocations often spark heated de­
bates about whether the incidence of gains and
losses to people's welfare is desireable. Indeed,
the adoption of one policy over another often
turns on the associated income redistribution
and equity issues.
15

BUSINESS REVIEW

Government policies aren't the only reason
for income and wealth transfers, however. Private
individuals, acting on their own with no official
prodding, also redistribute significant amounts
of income, property, and personal services to
family members and friends. Conservative esti­
mates indicate that individuals make some $100
billion of private transfers annually (not counting
charitable contributions).1
These private transfers could prove a critical
element in debates over the welfare impact of
policy-induced redistributions. In particular,
recent economic analyses suggest that people
may take public redistributions into account
when they make their private transfer decisions.
For example, people may decrease their transfers
to a family member whose income rose due to a
government program change. As a result, private
transfers could alter or even erase the effects on
an individual's well-being that public policies
would otherwise have. To understand fully the
welfare implications of their actions, then,
policymakers must understand the interplay be­
tween public and private transfers.
EXPLAINING PRIVATE TRANSFERS:
WHAT GIVES?
The term "transfer payment" normally con­
jures up visions of vast federal bureaucracies
channeling money from the well-fed to the
hungry, from the landed to the homeless. And
while the government certainly reallocates a
substantial amount of funds, economists have
discovered that it does not monopolize that ac­
tivity. Private individuals, they now realize, also
redistribute significant amounts of resources

1-rhis article purposely excludes private contributions to
charitable organizations. We exclude these contributions
because the theories that we describe stress a family context
in which donor and donee are closely related. We mention in
passing, however, that these theories and their implications
might apply to some aspects of private charitable contribu­
tions. A recent study by Charles T. Clotfelter, Federal Tax
Policy and Charitable Giving, (Chicago: University of Chicago
Press, 1985), estimates that private charitable contributions
equaled almost $50 billion in 1980.

16



MARCH/APRIL 1986

amongst themselves. These individuals include
familial relations, such as parents and their adult
children, as well as friends who are not members
of the same family.
A recent study by Donald Cox and Fredric
Raines provides an interesting socioeconomic
snapshot of private transfer givers and receivers.2
Their analysis shows that individuals of all types
participate in the private transfer network. On
average, earnings among givers are almost twice
as high as earnings among receivers, while the
average value of assets among givers (financial
plus tangible wealth) is almost three times as
high as those among recipients. Givers tend to
be older than recipients, and fewer of them are
married. Average education levels for recipients
and givers are virtually identical.
Of the $100 billion of private transfers that
individuals make, about 40 percent comprise
bequests—the $5,000 Aunt Harriet wills to niece
Mary or the family heirlooms passed from gener­
ation to generation. The remaining 60 percent
constitute transfers between living family mem­
bers and friends, transfers that help defray the
recipients' cost of food, rent, education, and their
many other requirements and luxuries of daily
life.3
Most individuals wouldn't question the mo­
tives for those private transfers. Indeed, asking

2Donald Cox and Fredric Raines, "Interfamily Transfers
and Income Redistribution," in Martin David and Timothy
Smeeding, eds., Horizontal Equity, Uncertainty and Measures of
Economic Well-Being National Bureau of Economic Research
Conference Proceedings, forthcoming. Cox and Raines discuss
numerous socioeconomic characteristics of private transfer
givers and receivers, basing their analysis on the President's
Commission on Pension Policy Household Survey. That
survey was conducted in August, 1979. We mention only a
few of the many characteristics that they present, and direct
the interested reader to their study.
3The aggregate figure for bequests comes from Laurence J.
Kotlikoff and Lawrence H. Summers, "The Role of Intergenerational Transfers in Aggregate Capital Formation /'Journal
of Political Economy, (August 1981), pp. 706-732. The aggregate
figure for transfers among living individuals com es from
Mordecai Kurz, "Capital Accumulation and the Characteris­
tics of Private Intergenerational Transfers," Economica, (Feb­
ruary 1984), pp. 1-22.

FEDERAL RESERVE BANK OF PHILADELPHIA

Warm Feelings & Cold Calculations

the reason for someone's generosity would be
regarded widely as an affront. And yet, econo­
mists have begun to do just that. Economists ask
not because they are nosey or crass by nature
(although that is debatable in the minds of some),
but rather because of their broader interest in
understanding how individuals allocate scarce
resources among competing uses. Private trans­
fers represent, in actuality, one of many alterna­
tive uses for people's income, property, and
time. Not surprisingly, then, economists have
applied their analytical methods toward describ­
ing why people make those transfers instead of,
say, investing in real estate, buying a record, or
seeing a ballgame. Although economists' models
cannot capture all of the aspects of social interac­
tions among individuals and families, they have
attempted to examine some of the possible moti­
vations for private transfers. Thus far, they have
offered two complementary explanations.
The Altruism Model. One seemingly natural
motivation for people's giving is their caring for
others. People rarely remain unaffected by the
fortunes that befall their family and friends, but
stand ready to share the good times and mitigate
the bad. Economists recognize this, and have
adopted people's mutual concern as a basis for
explaining private transfers.
The particular version of "altruism" that econ­
omists have examined is one that is more rele­
vant to econom ic behavior than to philosophical
discussions of human motivation. In the eco­
nomic literature, this so-called "altruism" theory
begins, ironically, by presuming that self-interest
guides people's actions.4*That is, each individual
always tries to get the most satisfaction from his
or her own income. For some, happiness comes

4The theoretical underpinnings of the altruism model are
described in Gary Becker, "A Theory of Social Interactions,"
Journal of Political Economy, (December 1974), pp. 1063-1094,
Gary Becker, "Altruism in the Family and Selfishness in the
Marketplace," Economica, (February 1981), pp. 1-15, and Gary
Becker, A Treatise on the Family, (Cambridge, MA: Harvard
University Press, 1981). Becker notes that his use of "altruism"
is in a more limited sense than has been used generally. See
"Altruism in the Family . . . " p. 2.




Donald C. Cox & Robert H. DeFina

only from the goods and services that they pur­
chase for their own use. But for others, labeled
"altruists," pleasure also arises from the hap­
piness experienced by the people they know.
An altruist might, for example, get satisfaction
both from his recent vacation and from his sister's
exuberance at landing a new job. Altruists, then,
exhibit a split-personality, acting always in their
own interest yet caring about the well-being of
others in the process.
Altruism leads naturally to private transfers.
Because the altruist receives satisfaction both
from his own consumption and the happiness of
his family and friends, he is likely to use his
income to enjoy some of each. He can increase
the happiness of his close relations by giving
them money, no strings attached, to be used as
they prefer. Any satisfaction that they obtain
from the income transfer will also accrue to the
altruist.
How much income the altruist devotes to
private transfers and to whom he gives the
money will depend on the amount of extra satis­
faction that the additional income produces. The
altruist's self-interest tells him to spend his money
where it yields him greatest happiness. Thus, he
will make a particular transfer if the satisfaction
he gets exceeds the pleasure he could obtain
from some other use of his money.
The recipient's financial situation probably
helps determine the amount of satisfaction that
the transfer provides the giver. Presumably, a
family member with very little income will be
happier about the altruist's donation than will a
family member with a large income. An extra
$50 will probably mean more to someone who is
unemployed, for instance, than to someone who
has a lucrative investment banking job. Transfers
targeted toward more needy family members,
then, yield the altruist more satisfaction than
donations made to more affluent members.
Consequently, an altruist will more readily make
transfers to poorer individuals than to richer
ones. In this way, private transfers motivated by
altruism serve a compensatory role, muting dif­
ferences in family m embers' income.
17

BUSINESS REVIEW

The notion that private transfers play a com­
pensatory role accords well with the caring and
mutual concern that one often associates with
those transfers. But is that the only role for private
transfers? A moment's reflection suggests that
the richness and complexity of familial inter­
actions allows room for other roles as well. Cer­
tainly, numerous occasions arise when an indi­
vidual can freely pursue his or her own selfinterest, which may include improving the lot of
other family members. Yet, other occasions arise
when one's preferences becom e subordinate to
the family's collective choice. Deciding on an
acceptable vacation spot, what color drapes to
buy, and when to allow one's son to have his first
date, all represent instances when family mem­
bers can have different preferences but cannot
or will not impose their own choice on the others.
At such times, when m em bers' divergent inter­
ests conflict, the need for negotiation arises.
Families often resolve conflicts by bargaining.
One member might agree explicitly, for example,
to a vacation at the North Pole (despite her
dislike of the place), if the other family members
okay the purchase of paisley curtains (something
she likes but the others find in bad taste). Or
members might have an unspoken understand­
ing, whereby Junior's yardwork (something he
views with displeasure) “earns" him his parents'
permission to stay out late on Saturday night
(something that makes them a bit uneasy). These
types of family negotiations, in addition to altru­
istic feelings, can also give rise to private transfers.
Economists have described this type of behavior
in what may be called the “exchange m odel" of
private transfers.5

5Exchange models of private transfers are presented in B.
Douglas Bernheim, Andrei Shleifer, and Lawrence H. Sum­
mers, "Bequests as a Means of Paym ent," NBER Working
Paper No. 1303 (March 1984), and Donald C. Cox, "Motives
for Private Transfers," Hoover Institution Mimeo (July 1985).
Again, economists' models do not describe all aspects of
exchange behavior among family mem bers — children and
their parents do many things for each other because of
various motives. The bargaining or negotiations approach
discussed here is just one type of behavior.

Digitized 18 FRASER
for


MARCH/APRIL 1986

The Exchange Model. The premise of the socalled “exchange" theory of private transfers is
that transfers represent payments by one person
for the services rendered by another. Essentially,
the theory envisions a family-type setting, with
one person desiring services (call her the “par­
ent") that only another particular person can
provide but would rather not (call him the
“child"). These services might include, for in­
stance, companionship and conforming to par­
ental regulations.6*
The model depicts child and parent as rational,
self-interested people who “bargain" and reach
a voluntary agreement about the quantity of
services provided and the size of the transfer
paid for those services. The agreement might be
explicit, such as when a rich aunt controls a
wayward nephew with a promised inheritance.
Or it might be tacit, such as when college students
regularly call home in order to “earn" an allow­
ance from their parents.
Because the child may view the provision of
such services with displeasure, he will supply
services to the parents only if adequately com­
pensated for his troubles. "Adequate" here means
sufficient to leave him at least as well off as he
would be if he provided no services and received
no payment. If his expected payment is less than
adequate, then he'll consider some other use of
his time more valuable, and he will pursue that
activity rather than supplying the services to his
parents.
The parent obtains happiness from the child's
services in the same way that she does from
other goods and services. And, those factors that
determine her demands for other goods and
services also determine her demand for the child's

6The services associated with exchange relationships dis­
cussed here should not be interpreted in such a limited way
that they include only family actions involving "love" and
"affection." For example, Laurence J. Kotlikoff and Avia
Spivak, "The Family As An Incomplete Annuities Market,"
Journal of Political Economy, (April 1981), pp. 372-391, suggest
that such services might take the form of insurance arrange­
ments among family members aimed at protecting members
from the risks of uncertain lifetimes.

FEDERAL RESERVE BANK OF PHILADELPHIA

Warm Feelings & Cold Calculations

services. Thus, her income, the price of the ser­
vices, and how much she desires the services
will all influence the quantity of services for
which she bargains.
The size of the transfer that the parent makes
to the child and the amount of services that the
child supplies to the parent emerges from the
bargaining process.7 Based on their respective
situations, parent and child will make offers and
counter-offers in an attempt to reach mutually
agreeable terms. The agreement ultimately
reached will reflect the initial bargaining posi­
tions of the parent and child, as well as their
negotiating savvy. The stronger the bargaining
position of an individual, the more concessions
he or she will be able to extract from the other
party. A key determinant of a person's bargain­
ing strength is the degree of well-being that he
or she can achieve in the absence of any agree­
ment. Chances are that a person who is very
unhappy in his current state will be more anxious
to strike a deal that improves his condition than
one who feels quite contented with the way
things stand. The better off an individual is from
the start, the stronger his or her bargaining posi­
tion will be, and the more advantageous the deal
he or she will be able to negotiate.
Empirical Evidence on Private Transfer Mo­
tives. Although introspection suggests that altru­
ism and exchange are plausible motives for some
types of private transfers, economists like to
gather more formal, statistical support for these
explanations before accepting their validity.
Empirical research on private transfers remains
at an early stage, so existing information is prelim­
inary and incomplete. Nonetheless, available
studies do confirm that both altruism and ex­
change motives are at work.8

7The outcome of the bargaining process envisioned by
the exchange model is the so-called "Nash solution," which
is commonly used in econom ic models of bargaining. An in­
structive discussion of the Nash solution is contained in R.
Duncan Luce and Howard Raiffa, Games and Decisions, (New
York: John Wiley and Sons, 1957).
8The following discussion of empirical results is not exhaus­
tive. A more complete catalogue of relevant studies is found




Donald C. Cox & Robert H. DeFina

Two studies that focus on bequests illustrate
the point. One, completed by Nigel Tomes,
examined the relation between the size of inheri­
tance that people receive and their income level.9
He found an inverse relation, and concluded
that bequests perform a compensatory role con­
sistent with the altruism model. Another study
of bequests, conducted by B. Douglas Bernheim,
Andrei Shleifer, and Lawrence Summers, found
evidence of exchange motives. Using different
data from Tomes', they analyzed whether parents
use bequests to influence their children's behav­
ior.101They determined that the frequency of a
child's visits and phone calls increases the larger
his or her expected inheritance becomes. This
finding leads them to conclude that parents use
transfers, at least in part, to obtain services from
their children. Research focusing on transfers
between living individuals, rather than on be­
quests, also yields evidence that supports both
the altruistic and exchange motives models.11
The current theoretical and empirical devel­
opments in the economic analysis of private
transfers represent first steps in what will likely
be an exciting area of research for some time to
come. As it turns out, the forthcoming insights
may be of more than academic interest. Under­
standing why people make private transfers has
important implications for the impact of eco­
nomic policies.
PRIVATE TRANSFERS
AND PUBLIC REDISTRIBUTION
Virtually all changes in social and economic
in Bernheim, et al., "Bequests as a Means of Payment," and
Robert A. Poliak, "A Transaction Cost Approach to Families
and H o u s e h o ld s Journal of Economic Literature, (June 1985),
pp. 581-608.
9Nigel Tomes, 'The Family, Inheritance, and the Intergenerational Transmission of Inequality," Journal of Political Economy,
(October 1981), pp. 928-958.
10B. Douglas Bernheim, Andrei Shleifer, and Lawrence
H. Summers, "Bequests As a Means of Payment."
11 See, for example, Jere R. Behrman, Robert A. Poliak,
and Paul Taubman, "Parental Preferences and Provision for
P ro g en y "Journal of Political Economy, (February 1982), pp.
52-73, and Donald C. Cox, "Motives for Private Transfers."

19

BUSINESS REVIEW

policy redistribute income and wealth, either
intentionally or unintentionally and, hence, alter
people's financial position. A reduction in Social
Security benefits combined with a cut in payroll
taxes, for example, channels income from retirees
to workers. Cuts in educational loan guarantee
programs also redirect money, from student
borrowers (or their parents) to taxpayers in gen­
eral. And generally, policymakers agonize over
the effect that these sorts of redistributions have
on people's financial positions and well-being.
Indeed, such concerns often occupy center stage
in discussions regarding the underlying policy's
worth.
Private transfers may complicate these issues
significantly. For by changing people's financial
status, policymakers may spark adjustments in
people's private transfers, whether those trans­
fers are motivated by altruism or exchange. And
because those adjustments might carry their
own welfare implications, they might alter the
welfare impact that public redistributions would
otherwise have. Policymakers, then, must under­
stand these potential responses if they are to
assess accurately the ultimate impact of their
actions on people's welfare.
No single, summary statement can describe
how private transfers respond to public redis­
tributions. Obviously, the precise way that pri­
vate transfers respond and what the welfare
consequences of those responses are will depend
on the many particulars of the situation. Those
particulars include whom the policy change af­
fects, what the motives underlying the private
transfers are, and so forth. But to get a flavor of
the possible responses and their consequences,
consider the hypothetical cases of the Donors
and the Barters.
The Cases of the Donors and the Barters. Ellie
Donor and Andy Barter are sailing enthusiasts,
and each owns a sailboat. Ellie has a nephew,
Eldon, and Andy has a niece, Ann. Both Eldon
and Ann are unemployed and each receives a
monthly unemployment check of $100. In addi­
tion, Ellie sends Eldon $50 every month based
on her altruistic feelings toward him. Andy like­
20



MARCH/APRIL 1986

wise sends Ann $50 every month, but his trans­
fers are motivated by exchange, not altruism.
Specifically, he sends Ann money to induce her
to come talk about windlasses and bowdecks,
something Ann dreads.
Recently, Congressman Newbill has recom­
mended increasing monthly unemployment
benefits to $150 from $100. He suggests that the
needed revenue for the increase come from a
monthly tax of $50 per sailboat ownerT2 This
policy initiative redistributes income away from
sailboat owners, like Ellie and Andy, and toward
unemployed persons, like Eldon and Ann, pre­
sumably to improve the well-being of the un­
employed at the expense of sailboat owners.
Before supporting Newbill's proposal, however,
the Ad Hoc Committee on Maritime Taxation
wanted to be sure that the welfare implications
of this redistribution for typical families, like the
Donors and Barters, would be the intended ones.
First, consider what happens to the altruistic
family—the Donors. On the surface, the program
change appears to have the anticipated and de­
sired consequences. The change transferred $50
from Ellie to Eldon, making him better off at her
expense.
But the story doesn't end there. Prior to the
boost in unemployment benefits, Elbe's altruistic
feelings compelled her to send Eldon $50. The
program change altered both her and Eldon's
financial situation, however, and this caused
Ellie to reevaluate her previous decisions. Ellie
realizes that the government is now doing what
she had intended to do anyway—transferring
$50 from herself to Eldon. Consequently, she
now sees no need to send him the monthly
check for $50 as she did voluntarily before. Ellie's
altruistic feelings still compel her to send him
$50. But she cares little how she sends it. Whether
she sends it directly, or through the government,l

l 2In order to concentrate solely on the influence of private
transfers, we assume that neither the unemployment program
nor the tax on sailboat owners has any impact on any aspect
of individuals' resource allocation decisions other than
(possibly) their private transfer decisions.

FEDERAL RESERVE BANK OF PHILADELPHIA

Warm Feelings & Cold Calculations

she receives the same satisfaction from Eldon's
increased well-being. And because the govern­
ment now performs the task, she cuts back her
private contribution to zero.
Elbe's response to the government program
change has striking implications: by adjusting
her private transfers, Elbe will erase any welfare
effects that the program change might have had.
Due to Elbe's response, Eldon's total income is
$150 both before and after the increase in un­
employment benefits. Moreover, Elbe still gives
Eldon only $50 as before, although her contribu­
tion is more roundabout. She pays the $50 in
taxes, and the government then gives the $50 to
her nephew. Ultimately, Elbe and Eldon remain
in precisely the same situation as before and,
hence, neither experiences any change in well­
being.
The altruistic link between Elbe and Eldon,
and the associated private transfers, are a barrier
to the success of Congressman Newbill's propo­
sal. On the surface, his plan seemed reasonable
and likely to achieve the desired welfare effects.
Indeed, the proposal's initial impact yielded just
those results. But the program change also caused
Elbe to rethink her private transfer decisions.
And, her subsequent response produced unin­
tended welfare effects that completely neutralized
the redistribution program.
The program change precipitates different
added welfare effects when exchange motives
underhe private transfers, as they do in the Barter
household. Initially, the change enhances Ann's
welfare. She, like Eldon, receives additional income
which allows her to buy more of the things that
give her pleasure. Uncle Andy is less fortunate;
his now higher taxes reduce his spendable income
and, hence, diminish his well-being.
Congressman Newbill's proposal, then, ap­
parently yields the desired results, as it did with
the Donors. That is, Ann benefits at Andy's ex­
pense. But as with the Donors, the story doesn't
end there. The reason is that the program change
alters the relative bargaining positions of the
two. By increasing Ann's income and, hence, her
well-being, the program change improves her



Donald C. Cox

S' Robert H. DeFina

initial bargaining position. And by decreasing
Andy's income and, hence, his well-being, the
program change erodes his initial bargaining
position.
This shift in relative bargaining positions has
its own welfare implications. Because the redis­
tribution places Ann in a relatively more advan­
tageous bargaining position than before, she can
now obtain a more favorable agreement than
before. For instance, Ann might now visit Andy
less, but receive the same size transfer as before.
This leaves her somewhat better off, and Andy
somewhat worse off, than the initial effect im­
plied.
The exchange-motivated private transfer ar­
rangement thus magnifies the program change's
initial impact. The program change improves
Ann's welfare not only by allowing her to spend
more but also by allowing her to strike a more
favorable deal with her Uncle. And, the program
change diminishes Andy's well-being not only
by forcing him to spend less but also by causing
him to accept a less favorable deal from his
niece.
Congressman Newbill's proposal seems des­
tined to produce unwanted side effects which
ultimately alter its intended results. When altru­
istic motives underlay private transfers, as they
did for the Donor's, adjustments in private trans­
fer decisions may completely erase the intended
impact of the program change. When exchange
motives underlay private transfers, as they did
for the Barter's, adjustments in private transfer
decisions may magnify the intended impact of
the change. These effects, moreover, would not
be taken into account if people's private transfer
decisions were ignored.
While focusing on a specific redistribution,
the hypothetical cases of the Donors and the
Barters illustrate the more general point that
private transfer decisions might respond to public
redistributions, and thereby alter the impact that
public redistributions would otherwise have. As
the example shows, the response will differ de­
pending on what motivates the private trans­
fers. When altruism motivates private transfers,
21

BUSINESS REVIEW

those transfers will play a compensatory role,
dampening any welfare impacts that public redis­
tributions might have on either the giver or the
beneficiary. Private transfers play no such com­
pensatory role when they stem from exchange
considerations. In that case, private transfers
will respond not to changes in people's well­
being per se, but to shifts in their relative bar­
gaining positions. As a result, exchange-motiva­
ted private transfers will reinforce any welfare
effects that public transfers might have. Currently
available information is insufficient to assess
accurately the practical significance of private
transfers for the many policy changes that redis­
tribute income and wealth. But given the poten­
tially critical role that private transfers play, poli­
cymakers would do well to examine their likely
importance when contemplating policy changes.
PRIVATE TRANSFERS:
A FLY IN THE POLICY OINTMENT?
Economic research on people's motives for
private transfers has uncovered a potentially
important policy issue. In particular, people are
likely to adjust their private transfer behavior in
response to public redistributions, and thereby
alter the welfare implications that public trans­
fers would otherwise have. This conclusion flows
from two economic models that provide plausi­
ble and complementary, though not exhaustive,
descriptions of why individuals make private
transfers. According to the altruism model, peo­
ple make transfers because they share in the
recipient's subsequent happiness. And, if the
altruism model accurately describes some aspects
of our society, then policy redistributions may
ultimately be undone by changes in private

22



MARCH/APRIL 1986

transfers, for altruistic private transfers serve a
compensatory role. According to the exchange
model, people make transfers as a quid pro quo,
for services rendered by the recipient. And, if
the exchange model accurately describes some
types of behavior, then policy redistributions
may ultimately be magnified by changes in private
transfers. In that case, private transfers compound
the gains and losses of individuals as their bar­
gaining positions change.
Preliminary empirical evidence has been
found to support both explanations, which is
not at all surprising. It's easy enough, indeed, to
imagine that a single individual could embody
both kinds of behavior—for example, altruistic
motives toward close family members, and ex­
change motives toward more distant relations.
Research on this issue is at an early stage,
however, and economists are just beginning to
grasp the complexities of private transfers. More
sophisticated analyses that account not only for
altruism and exchange, but also other economic
factors, will probably be required to understand
fully the motives for private transfers and their
implications for policy. These might include, for
instance, attempts by families to insure against
the risks associated with uncertain lifetimes. In
addition, noneconomic factors, such as the roles
of tradition and religious custom, might also
prove crucial influences in private transfer deci­
sions. And while economists cannot yet provide a
firm ground from which to survey the interplay
between public redistributions and private trans­
fers, policymakers will need to keep a watchful
eye on developments in the analysis of private
transfer behavior.

FEDERAL RESERVE BANK OF PHILADELPHIA

Business Review
Federal Reserve Bank of Philadelphia
Ten Independence Mall
Philadelphia, PA 19106
Address Correction Requested