View original document

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

Federal Reserve Bank of Cleveland
Table 4 - Regression Results for Bank Earnings
Variable to be explained:
Average percent return on assets
Explanatory

Variable

Bank deposit size
Bank dummy (2 banks
1; 1 bank
0)
Number of thrifts in market
Deposit share held by thrifts
Deposit share held by credit unions
Market deposit growth (1982 to 1985)

=

=

First
Regression

Second
Regression

Positive"
Negative"
Negative"
NI
Positive
Positive

Positive
Negative"
NI
Negative"
Positive
Negative

NI - not included
a. Denotes statistical significance at the 5 percent level.
b. Denotes statistical significance at the 1 percent level.
SOURCE: Reports of Income and Condition, Board of Governors of the Federal Reserve System;
Data Book, FDIC; Summary of Savings Accounts by Geographic Area, Federal Home Loan Bank
Board, and the Rand McNally Credit Union Directory.

the deposit share held by thrifts for the
number of thrifts to determine if their
competitive strength was related
directly to their market position.
It has been claimed that smaller institutions may not be viable competitors in
a less-regulated environment. If larger
banks realize economies of scale or have
some other advantages over smaller
competitors, then bank size would have
a positive influence on earnings.
Credit unions are active competitors
for consumer deposits and loans, and
their presence and concentration in a
market may exert downward pressure
on bank earnings. Regardless of competitive conditions, banks would be
expected to earn higher profits in
markets with stronger demand. We used

market deposit growth from 1982 to
1985 to account for demand differences
across markets.
The statistical results, as determined
by regression analysis, are presented in
table 4. These findings are consistent
with the t-test results and support the
view that the thrift institutions exert
strong competitive pressures on banks.
When other factors, such as bank size,
market growth, and credit union competition, were taken into account, along
with the number of banks in the market, the presence of thrifts in the local
area still had a negative and significant
impact on bank earnings. We also found
that banks in markets where thrifts held
a larger share of the deposits earned
significantly lower returns on assets.

13. We found usable observations in the following states: Alabama, Arkansas, Florida, Georgia,
Illinois, Indiana, Kentucky, Michigan, Mississippi, Pennsylvania, South Carolina, Virginia,
West Virginia, and Wisconsin. Arkansas banks
were included in the sample because their earn-

ings were relatively stable, and many Arkansas
banks met our other sample criteria.

Federal Reserve Bank of Cleveland
Research Department
P.O. Box 6387
Cleveland, OH 44101

Material may be reprinted provided that the
source is credited. Please send copies of reprinted
materials to the editor.

14. Lynn A. Nejezchleb, "Declining Profitability
at Small Commercial Banks: A Temporary
Development or a Secular Trend?" Banking and
Economic Review, Federal Deposit Insurance
Corporation, June 1986, pp. 9-21.

Conclusion
This Economic Commentary looks at

the role of thrifts as competitors to
banks. However, we do not examine the
issue of banking market contestability.
Our findings suggest that regulators
could be underestimating the actual
competition for banking services in
local areas by not adequately accounting for competitors of banks when evaluating the competitive effects of bank
mergers. In some markets thrifts
might properly be regarded as full competitors of banks. While our bank sample is limited to markets with no more
than a few competitors, the presence
and the relative size of a thrift in those
markets had a strong and similar effect
on bank earnings as a competing bank
in those markets. Banks operating in
markets with at least one thrift or in
markets with another bank earned significantly lower profits than other
sample banks. Also, earnings were substantially lower at nonsample banks,
which typically operate in markets
with a greater number of thrifts and
banks.
We found no evidence to support the
view that thrift competition is inferior
to bank competition-at
least in terms
of impact on bank earnings. Although
the competitiveness of thrifts could
vary widely over the many products
and services provided by banks, the
presence of thrifts in local banking
markets is exerting downward pressure on bank earnings.
15. The competitive influence of banks located
outside of these markets is generally considered
weak compared to the threat of entry in larger
banking markets.
16. This earnings difference, however, was statistically insignificant.

BULK RATE
U.S. Postage Paid
Cleveland, OH
Permit No. 385

Address Correction Requested:
Please send
corrected mailing label to the Federal Reserve
Bank of Cleveland, Research Department,
P.O. Box 6387, Cleveland, OH 44101.

November 1, 1987
ISSN 0428-1276

ECONOMIC
COMMENTARY

Local Thrift
Competition and
Bank Earnings
by Paul R. Watro

Economic and legislative changes have
clearly eroded differences between commercial banks and thrift institutions.
Commercial banks traditionally offered
a unique cluster of products that were
not available from other institutions.
Because of the expanded powers of
thrift institutions in the early 1980s, however, thrifts have been operating more
like commercial banks, providing a wide
range of financial services, including
checking accounts and business loans.
As a result, banking regulatory agencies, such as the Federal Reserve System, the Federal Deposit Insurance
Corporation, and the Office of the
Comptroller of the Currency, are reevaluating some of their policies. For
example, the Federal Reserve Board is
currently reconsidering its policy of
prohibiting bank holding companies from
acquiring healthy thrift institutions
and has solicited public comments.
Some of the Federal Reserve's previous concerns about banking organizations operating thrifts have apparently
lessened with the removal of interstate
barriers to geographic expansions by
banks and with the elimination of statutory deposit-rate differentials between
banks and thrifts.'
Also, deregulation, along with technological innovations and increased
customer sophistication has fostered an
expanding number of firms providing
financial services and has intensified
competition among banks and thrift
institutions. These developments have
made competitive assessments of bank
mergers extremely difficult and open to
considerable debate.
Bank regulators rely largely upon a

structure-performance framework that
implies that market structure (the number and size distribution of competitors
in a market) is an important determinant of bank performance. With this
line of reasoning, the fewer the number
of competitors and the larger the share
of the market controlled by the largest
competitors, the greater the likelihood
that firms will be able to charge prices
above the costs of doing business and to
generate profits above the competitive
level. The traditional structuralist view
is incorporated in the Justice Department's merger guidelines, which are
used in various degrees by banking regulators when evaluating the possible
anticompetitive impact of bank mergers.
Some analysts downplay the importance of market structure, reasoning
that the structure itself may be a product of the competitive forces at work in
the market. In this view, market concentration is unlikely to lead to collusion and excessive profits when barriers
to entry are low. Supporters of this perspective argue that the threat of entry
by firms outside of the market will
exert competitive discipline on market
participants, regardless of the actual
number of competitors in the market."
Banking agencies and the Justice Department are required by law to prevent
bank mergers and acquisitions that
would have substantially adverse effects
on banking competition. The competitive assessment of regulators centers
upon three factors: the product market,
the geographical market, and the likely
anticompetitive effects of the transaction (if any) within those markets.'
For antitrust purposes, the product

market for banking has been defined by
the Supreme Court to be a separate and
distinct line of commerce. This landmark decision was made in 1963, reaffirmed in 1974, and was based upon the
view that commercial banks provide
local customers with an unique cluster
of services that are not available from
other depository institutions.' Since that
time, however, sweeping changes have
taken place in banking. For instance,
thrift institutions can now offer the
same services as commercial banks.
Other nondepository institutions, such
as securities brokerage firms, mortgage
banking firms, insurance companies,
and finance companies are providing a
variety of financial services that are
competing with bank products.
Regulators have reacted somewhat to
these major developments. To various
degrees, the existence of thrift institutions and other nonbank suppliers of
financial services is being considered in
regulatory assessments of competitive
factors. Moreover, the Justice Department is apparently more lenient toward
bank mergers than toward mergers in
other industries, because the department apparently assumes the competitive influence of thrifts and other nonbank institutions, as well as the
imprecise boundaries used to estimate
geographical banking markets."
This Economic Commentary discusses the importance of thrift competition and examines earnings differences
between banks operating in markets
with and without the presence of
thrifts." Bank earnings are used as a
measure of overall competition for
banking services in local markets. If

Paul R. Watro is an economist at the Federal
Reserve Bank of Cleveland. The author would like
to thank Mark Sniderman, James Thomson, and
Gary Whalen for their helpful comments, and
would like to thank John McElravey for his valuable research assistance.
The views stated herein are those of the author
and not necessarily those of the Federal Reserve
Bank of Cleveland or of the Board of Governors of
the Federal Reserve System.

1. For a discussion of Federal Reserve policy
regarding the acquisition and operating of thrift
institutions by bank holding companies, see Federal Register, Vol. 52, No. 186, September 1987,
pp.36041-36045.

3. For a discussion of the analyses used by regulators and Justice Department, see Robert E.
Hauberg, Jr. "Mergers and Acquisitions: Trends
in Competitive Analysis," Banking Expansion
Reporter, Golembe Associates, Inc., Volume 6,
Number 13, July 6, 1987, pp. 1,9-17.

2. See Gary Whalen, "Concentration and Profitability in Non-MSA Banking Markets," Economic
Review, Federal Reserve Bank of Cleveland,
Quarter 1, 1987, pp. 1-15.

4. United States v. Philadelphia National Bank,
374, U.S. 321 (1963); and United States v. Connecticut National Bank, 418 U.S. 666 (1974).
5. Robert E. Hauberg, Ir., p. 13. Under the revised

the presence of thrifts does not alter
bank earnings, it would be consistent
with the view that thrifts are not as
important as banks when gauging
banking competition.
On the other hand, if bank earnings
are significantly lower in markets with
thrifts, then the actual level of overall
competition for banking services in
such markets is likely to be underestimated by not including thrifts as full
competitors of banks. Such a finding,
however, would not imply that thrift
competition is equivalent to bank competition for all banking services.
The Strength of Thrift Competition
Thrift institutions compete aggressively
with commercial banks, particularly
since the early 1980s with the advent of
product and price deregulation for depository institutions. With their expanded
powers, thrifts can and often do offer
essentially the same banking services
as commercial banks. Yet banking
agencies and the Justice Department do
not generally treat thrifts as banks
when assessing the competitive effects
of proposed bank mergers.
For example, the Federal Reserve
Board usually counts thrift deposits
only at half their actual market level
whereas bank deposits are counted at
their full market level. In essence, the
Federal Reserve is assuming that the
competitive influence of a thrift is
much less than that of an equal-size
bank in the market.' This assumption
stems from the fact that many thrifts
are not active business lenders and that
business loans account for a very small
portion of their total loans,"
However, some banks also make relatively few commercial and industrial
loans, yet regulators consider all banks
equally according to their deposits in
the relevant market area." Moreover,
thrifts that are not presently extending
business loans still provide banks with
potential competition for these loans.
Other factors also support the inclusion of thrifts as full competitors of
banks. Thrifts often are deposit-rate
leaders in markets. The removal of
interest-rate ceilings on deposits gave
Department of Justice Merger Guidelines, any
market in which the postmerger HerfindahlHirschman Index (concentration measure) is over
1800, the department is likely to challenge a
merger that increases the Index by more than 50
points unless other factors indicate that the
merger would not substantially lessen competition. However, the Department of Justice has
informed regulators that a bank merger generally
will not be challenged in the absence of other Iactors indicating an anticompetitive effect unless

banks and thrifts freedom to compete
for deposits on the basis of rates.'?
Deposit-rate surveys, including a
detailed study by Mahoney, Mcl.aughlin, O'Brien, and White, show that
thrifts consistently offer higher average
rates than banks on retail deposit
accounts, especially on longer-term certificates of deposit (Clzs).!' Over the
period examined (October 1983 to
December 1985), thrifts paid nearly 70
basis points more than banks on retail
CDs with maturities of over six
months, which was much greater than
the 25 basis-point differentials allowed
under deposit-rate regulation.
The authors suggested several
explanations for the higher deposit
rates paid by thrifts. Compared to
banks, thrifts hold more longer-term
assets, so they have an incentive to pay
higher rates on longer-term liabilities in
order to match maturities and reduce
interest-rate risk.
Thrifts also have traditionally relied
much more heavily on retail time and
savings deposits and have tended to
cater to more interest-rate-sensitive
customers. In contrast, banks have
depended more on noninterest-bearing
demand deposits and on customers who
have been willing to accept more heavily subsidized services as compensation
for lower deposit rates.
In addition to these differences
between banks and thrifts, capital
requirements are lower at thrifts,
which may give them a cost advantage
over banks." Simple cartel theory suggests that cost differences between
firms reduce the incentive for collusion.
From this perspective, thrifts might
actually be a greater competitive force
in the market today than additional
banking organizations, particularly in
markets with only a few competitors.
Sample and Method Used
The impact of thrift competition on
bank earnings is examined using recent
data for a sample of 314 banks located
in 14 states." These banks are headquartered in the eastern half of the
country and operate in one- and twobank rural markets where they generate
the post merger HHI is at least 1800 and the
merger increases the HHI by at least 200 points.
The HHI is calculated by adding the squared
market share of each competing institution.
6. Researchers have generally found that thrift
competition does not significantly alter bank
earnings, but these studies were generally conducted prior to deregulation or used data prior to
1983. For the purpose of this article, thrifts refer
only to federally insured savings and loan associ-

all or most of their deposits. The sample was limited to those banks to minimize the effects of regional economic
conditions and nondepository competition and to isolate the effects of additional competitors on bank profitability.
Banks west of the Mississippi River
have generally been greatly affected by
the depressed energy and farm sectors
of the economy. These sectors have
been largely responsible for the record
number of bank failures in recent
years; the vast majority of the more
than 600 bank failures since 1981 have
taken place west of the Mississippi.
Moreover, a study on small-bank earnings by Nejezchleb found that profitability correlated strongly with regional
economic conditions.t! For example, in
1985 small banks east of the Mississippi had return on assets at about the
same level or modestly below their 1981
levels. In contrast, aggregate earnings
for banks west of the Mississippi were
at least 50 percent below 1981 levels.
The sample was restricted to one-and
two-bank markets that are approximated by counties. Thrifts do not operate in many of these areas, and other
nonbank competitors, such as mortgage
and finance companies, are often
absent in small rural markets. IS The
sample allows a direct earnings comparison between banks operating in
markets with and without the presence
of a thrift, as well as between banks
operating in markets with and without
another bank. The sample included
only those banks operating offices
within a single county, so that the
presence of a thrift in an area could be
related to profits earned in that market.
Bank earnings are measured by their
average return on assets for 1985 and
1986. A two-year average was used to
reduce the variability of earnings from
year to year. For each grouping, the
mean value of bank earnings was computed with a "t" test employed to
determine if differences were statistically significant between the various
groups. Earnings of sample banks were
also compared to earnings of nonsample banks. Although the sample was
designed to isolate the effects of local
thrift competition on bank profitability
ations and savings banks.
7. The Federal Reserve has given thrift competition more weight in approving a few recent proposed mergers. See, for example, the decision to
approve the acquisition of Chester Bank in Chester, Connecticut by the Hartford National Corporation of Hartford, Connecticut, Federal Reserve
Bulletin, September 1987, pp. 720·723.
8. Although virtually all commercial banks have

while accounting for the number of
banks in the market, other factors may
affect bank earnings. Consequently, we
also use regression analysis to account
for other factors, such as bank size and
market growth, when estimating the
effects of the presence of thrifts on
bank earnings.
Findings
The earnings of the sample banks and
various comparisons of those earnings
are presented in tables 1 through 4.
Table 1 shows the range and average
earnings of the total bank sample
broken down by the number and type
of competitors. Despite the use of a
two-year average, individual bank earnings varied widely. Nevertheless, average returns on assets were consistently
lower for the group of banks located in
markets with thrift competition. For
instance, banks earned 1.10 percent on
assets in markets with thrifts, regardless of the number of banks. In contrast, the sample banks earned 1.45
percent on assets in markets without
any thrifts or other banks.
We show average earnings differences
between groups and denote which ones
are statistically significant in table 2.
As expected, earning differences were
significantly lower for the group of
banks faced with thrift competition.
Banks in two-bank markets also earned
lower profits than banks in one-bank
markets. Further sample breakouts by
the number and type of competitors
also yielded anticipated results.
None of the findings suggest that a
thrift provides a bank with less competition than another bank. In fact, bank
earnings were actually lower in onebank markets with thrift competition
than in two-bank markets without thrift
competition. 16 Earnings differences between groups, however, could be due to
other factors rather than the number or
type of competitors. For example, bank
size could affect earnings if larger
banks have an advantage over smaller
banks.
We took "bank size" into account in
two ways. First, we compared bank
earnings of the sample banks without
business lending experience, one-fifth of all banks
had commercial and industrial loans accounting
for less than 5 percent of their total assets as of
year end 1986. Jim Burke, Stephen Rhoades and
John Wolken; "Thrift Institutions and Their
New Powers," The Journal of Commercial Bank
Lending; June 1987, p. 51.
9. One could argue that thrifts have a tax incentive to maintain a very large share of assets in

Table 1 - Bank Earnings and Market Competitors
Average Percent Return on Assets
Mean
Range

Total Sample
Thrifts
Thrifts, two banks
Thrifts, one bank
No thrifts
No thrifts, two banks
No thrifts, one bank
Two banks
One bank

1.18
1.10
1.10
1.10
1.30
1.23
1.45
1.15
1.37

Observations

-0.69 to 3.17
-0.69 to 2.12
-0.69 to 2.12
-0.02 to 1.63
-0.05 to 3.17
-0.05 to 3.17
0.45 to 2.73
-0.69 to 3.17
-0.02 to 2.73

314
185
174
11
129
89
40
263
51

Table 2 - Earnings Differences Across Markets
Difference in percent
return on average assets

Thrifts versus no thrifts
Thrifts, two banks versus no thrifts, two banks
Thrifts, one bank versus no thrifts, one bank
Thrifts, one bank versus no thrifts, two banks
Two banks versus one bank

-0.20"
-0.13c
-0.35b
-0.13
-0.22"

a. Denotes statistical significance at the 1 percent level.
b. Denotes statistical significance at the 5 percent level.
c. Denotes statistical significance at the 10 percent level.

Table 3 - Earnings Differences Between Sample and Nonsample
Banks
Average percent return on assets
Nonsample banks8

Differenceb

0.85
0.84
0.92
0.78

-0.34
-0.46
-0.31
-0.66

Total sample
No thrifts
No thrifts, two banks
No thrifts, one bank

a. Nonsample banks are aU other banks of similar size and located in the same states as the sample.
b. Difference equal to nonsample bank earnings minus sample bank earnings. AU differences are
statisticaUy significant at the one percent level.
SOURCES: Data for the above tables were derived from reports of Income and Condition, Board of
Governors of the Federal Reserve System; DataBook, FDIC; and Summary of Savings Accounts by
Geographic Area, Federal Home Loan Bank Board.

thrift competition to the earnings of all
other similar-size banks in the same
states. Table 3 shows that nonsample
banks earned substantially lower
returns than the sample banks. Average earnings of similar-size banks were
between 31 and 66 basis points below
the average earnings of the sample
banks operating in markets without
any thrifts. The earnings differences
between the nonsample banks and
sample banks may be due to many fac-

tors, including the fact that the nonsample banks typically operate in more
competitive markets.
We also took bank size, as well as
market growth and credit unions, into
account through regression analysis.
Bank profitability was regressed on
bank deposit size, the deposit share
held by credit unions, market growth, a
dummy variable indicating if the market
had one or two banks, and the number
of thrifts in the market. We substituted

qualifying assets such as mortgages, to receive
favorable tax treatment over banks. However, the
tax reform act of 1986 reduced the qualifying
level from 82 percent to 65 percent, which gave
thrifts more leeway to be active business lenders
and stiU maintain their favorable tax status.

O'Brien, and Alice White, "Responses to Deregulation: Retail Deposit Pricing from 1983 through
1985," Staff Study 15, Board of Governors of the
Federal Reserve System, Washington, D.C. Ianuary 1987.

10. Depository institutions are still prohibited
from paying interest on demand deposits.
11. Patrick Mahoney, Mary McLaughlin, Paul

12. In the early 1980s, thrift regulators reduced
capital requirements because of the financial
plight of the industry. On the other hand, thrifts
have been paying higher premiums for deposit
insurance than banks since 1985.

the presence of thrifts does not alter
bank earnings, it would be consistent
with the view that thrifts are not as
important as banks when gauging
banking competition.
On the other hand, if bank earnings
are significantly lower in markets with
thrifts, then the actual level of overall
competition for banking services in
such markets is likely to be underestimated by not including thrifts as full
competitors of banks. Such a finding,
however, would not imply that thrift
competition is equivalent to bank competition for all banking services.
The Strength of Thrift Competition
Thrift institutions compete aggressively
with commercial banks, particularly
since the early 1980s with the advent of
product and price deregulation for depository institutions. With their expanded
powers, thrifts can and often do offer
essentially the same banking services
as commercial banks. Yet banking
agencies and the Justice Department do
not generally treat thrifts as banks
when assessing the competitive effects
of proposed bank mergers.
For example, the Federal Reserve
Board usually counts thrift deposits
only at half their actual market level
whereas bank deposits are counted at
their full market level. In essence, the
Federal Reserve is assuming that the
competitive influence of a thrift is
much less than that of an equal-size
bank in the market.' This assumption
stems from the fact that many thrifts
are not active business lenders and that
business loans account for a very small
portion of their total loans,"
However, some banks also make relatively few commercial and industrial
loans, yet regulators consider all banks
equally according to their deposits in
the relevant market area." Moreover,
thrifts that are not presently extending
business loans still provide banks with
potential competition for these loans.
Other factors also support the inclusion of thrifts as full competitors of
banks. Thrifts often are deposit-rate
leaders in markets. The removal of
interest-rate ceilings on deposits gave
Department of Justice Merger Guidelines, any
market in which the postmerger HerfindahlHirschman Index (concentration measure) is over
1800, the department is likely to challenge a
merger that increases the Index by more than 50
points unless other factors indicate that the
merger would not substantially lessen competition. However, the Department of Justice has
informed regulators that a bank merger generally
will not be challenged in the absence of other Iactors indicating an anticompetitive effect unless

banks and thrifts freedom to compete
for deposits on the basis of rates.'?
Deposit-rate surveys, including a
detailed study by Mahoney, Mcl.aughlin, O'Brien, and White, show that
thrifts consistently offer higher average
rates than banks on retail deposit
accounts, especially on longer-term certificates of deposit (Clzs).!' Over the
period examined (October 1983 to
December 1985), thrifts paid nearly 70
basis points more than banks on retail
CDs with maturities of over six
months, which was much greater than
the 25 basis-point differentials allowed
under deposit-rate regulation.
The authors suggested several
explanations for the higher deposit
rates paid by thrifts. Compared to
banks, thrifts hold more longer-term
assets, so they have an incentive to pay
higher rates on longer-term liabilities in
order to match maturities and reduce
interest-rate risk.
Thrifts also have traditionally relied
much more heavily on retail time and
savings deposits and have tended to
cater to more interest-rate-sensitive
customers. In contrast, banks have
depended more on noninterest-bearing
demand deposits and on customers who
have been willing to accept more heavily subsidized services as compensation
for lower deposit rates.
In addition to these differences
between banks and thrifts, capital
requirements are lower at thrifts,
which may give them a cost advantage
over banks." Simple cartel theory suggests that cost differences between
firms reduce the incentive for collusion.
From this perspective, thrifts might
actually be a greater competitive force
in the market today than additional
banking organizations, particularly in
markets with only a few competitors.
Sample and Method Used
The impact of thrift competition on
bank earnings is examined using recent
data for a sample of 314 banks located
in 14 states." These banks are headquartered in the eastern half of the
country and operate in one- and twobank rural markets where they generate
the post merger HHI is at least 1800 and the
merger increases the HHI by at least 200 points.
The HHI is calculated by adding the squared
market share of each competing institution.
6. Researchers have generally found that thrift
competition does not significantly alter bank
earnings, but these studies were generally conducted prior to deregulation or used data prior to
1983. For the purpose of this article, thrifts refer
only to federally insured savings and loan associ-

all or most of their deposits. The sample was limited to those banks to minimize the effects of regional economic
conditions and nondepository competition and to isolate the effects of additional competitors on bank profitability.
Banks west of the Mississippi River
have generally been greatly affected by
the depressed energy and farm sectors
of the economy. These sectors have
been largely responsible for the record
number of bank failures in recent
years; the vast majority of the more
than 600 bank failures since 1981 have
taken place west of the Mississippi.
Moreover, a study on small-bank earnings by Nejezchleb found that profitability correlated strongly with regional
economic conditions.t! For example, in
1985 small banks east of the Mississippi had return on assets at about the
same level or modestly below their 1981
levels. In contrast, aggregate earnings
for banks west of the Mississippi were
at least 50 percent below 1981 levels.
The sample was restricted to one-and
two-bank markets that are approximated by counties. Thrifts do not operate in many of these areas, and other
nonbank competitors, such as mortgage
and finance companies, are often
absent in small rural markets. IS The
sample allows a direct earnings comparison between banks operating in
markets with and without the presence
of a thrift, as well as between banks
operating in markets with and without
another bank. The sample included
only those banks operating offices
within a single county, so that the
presence of a thrift in an area could be
related to profits earned in that market.
Bank earnings are measured by their
average return on assets for 1985 and
1986. A two-year average was used to
reduce the variability of earnings from
year to year. For each grouping, the
mean value of bank earnings was computed with a "t" test employed to
determine if differences were statistically significant between the various
groups. Earnings of sample banks were
also compared to earnings of nonsample banks. Although the sample was
designed to isolate the effects of local
thrift competition on bank profitability
ations and savings banks.
7. The Federal Reserve has given thrift competition more weight in approving a few recent proposed mergers. See, for example, the decision to
approve the acquisition of Chester Bank in Chester, Connecticut by the Hartford National Corporation of Hartford, Connecticut, Federal Reserve
Bulletin, September 1987, pp. 720·723.
8. Although virtually all commercial banks have

while accounting for the number of
banks in the market, other factors may
affect bank earnings. Consequently, we
also use regression analysis to account
for other factors, such as bank size and
market growth, when estimating the
effects of the presence of thrifts on
bank earnings.
Findings
The earnings of the sample banks and
various comparisons of those earnings
are presented in tables 1 through 4.
Table 1 shows the range and average
earnings of the total bank sample
broken down by the number and type
of competitors. Despite the use of a
two-year average, individual bank earnings varied widely. Nevertheless, average returns on assets were consistently
lower for the group of banks located in
markets with thrift competition. For
instance, banks earned 1.10 percent on
assets in markets with thrifts, regardless of the number of banks. In contrast, the sample banks earned 1.45
percent on assets in markets without
any thrifts or other banks.
We show average earnings differences
between groups and denote which ones
are statistically significant in table 2.
As expected, earning differences were
significantly lower for the group of
banks faced with thrift competition.
Banks in two-bank markets also earned
lower profits than banks in one-bank
markets. Further sample breakouts by
the number and type of competitors
also yielded anticipated results.
None of the findings suggest that a
thrift provides a bank with less competition than another bank. In fact, bank
earnings were actually lower in onebank markets with thrift competition
than in two-bank markets without thrift
competition. 16 Earnings differences between groups, however, could be due to
other factors rather than the number or
type of competitors. For example, bank
size could affect earnings if larger
banks have an advantage over smaller
banks.
We took "bank size" into account in
two ways. First, we compared bank
earnings of the sample banks without
business lending experience, one-fifth of all banks
had commercial and industrial loans accounting
for less than 5 percent of their total assets as of
year end 1986. Jim Burke, Stephen Rhoades and
John Wolken; "Thrift Institutions and Their
New Powers," The Journal of Commercial Bank
Lending; June 1987, p. 51.
9. One could argue that thrifts have a tax incentive to maintain a very large share of assets in

Table 1 - Bank Earnings and Market Competitors
Average Percent Return on Assets
Mean
Range

Total Sample
Thrifts
Thrifts, two banks
Thrifts, one bank
No thrifts
No thrifts, two banks
No thrifts, one bank
Two banks
One bank

1.18
1.10
1.10
1.10
1.30
1.23
1.45
1.15
1.37

Observations

-0.69 to 3.17
-0.69 to 2.12
-0.69 to 2.12
-0.02 to 1.63
-0.05 to 3.17
-0.05 to 3.17
0.45 to 2.73
-0.69 to 3.17
-0.02 to 2.73

314
185
174
11
129
89
40
263
51

Table 2 - Earnings Differences Across Markets
Difference in percent
return on average assets

Thrifts versus no thrifts
Thrifts, two banks versus no thrifts, two banks
Thrifts, one bank versus no thrifts, one bank
Thrifts, one bank versus no thrifts, two banks
Two banks versus one bank

-0.20"
-0.13c
-0.35b
-0.13
-0.22"

a. Denotes statistical significance at the 1 percent level.
b. Denotes statistical significance at the 5 percent level.
c. Denotes statistical significance at the 10 percent level.

Table 3 - Earnings Differences Between Sample and Nonsample
Banks
Average percent return on assets
Nonsample banks8

Differenceb

0.85
0.84
0.92
0.78

-0.34
-0.46
-0.31
-0.66

Total sample
No thrifts
No thrifts, two banks
No thrifts, one bank

a. Nonsample banks are aU other banks of similar size and located in the same states as the sample.
b. Difference equal to nonsample bank earnings minus sample bank earnings. AU differences are
statisticaUy significant at the one percent level.
SOURCES: Data for the above tables were derived from reports of Income and Condition, Board of
Governors of the Federal Reserve System; DataBook, FDIC; and Summary of Savings Accounts by
Geographic Area, Federal Home Loan Bank Board.

thrift competition to the earnings of all
other similar-size banks in the same
states. Table 3 shows that nonsample
banks earned substantially lower
returns than the sample banks. Average earnings of similar-size banks were
between 31 and 66 basis points below
the average earnings of the sample
banks operating in markets without
any thrifts. The earnings differences
between the nonsample banks and
sample banks may be due to many fac-

tors, including the fact that the nonsample banks typically operate in more
competitive markets.
We also took bank size, as well as
market growth and credit unions, into
account through regression analysis.
Bank profitability was regressed on
bank deposit size, the deposit share
held by credit unions, market growth, a
dummy variable indicating if the market
had one or two banks, and the number
of thrifts in the market. We substituted

qualifying assets such as mortgages, to receive
favorable tax treatment over banks. However, the
tax reform act of 1986 reduced the qualifying
level from 82 percent to 65 percent, which gave
thrifts more leeway to be active business lenders
and stiU maintain their favorable tax status.

O'Brien, and Alice White, "Responses to Deregulation: Retail Deposit Pricing from 1983 through
1985," Staff Study 15, Board of Governors of the
Federal Reserve System, Washington, D.C. Ianuary 1987.

10. Depository institutions are still prohibited
from paying interest on demand deposits.
11. Patrick Mahoney, Mary McLaughlin, Paul

12. In the early 1980s, thrift regulators reduced
capital requirements because of the financial
plight of the industry. On the other hand, thrifts
have been paying higher premiums for deposit
insurance than banks since 1985.

Federal Reserve Bank of Cleveland
Table 4 - Regression Results for Bank Earnings
Variable to be explained:
Average percent return on assets
Explanatory

Variable

Bank deposit size
Bank dummy (2 banks
1; 1 bank
0)
Number of thrifts in market
Deposit share held by thrifts
Deposit share held by credit unions
Market deposit growth (1982 to 1985)

=

=

First
Regression

Second
Regression

Positive"
Negative"
Negative"
NI
Positive
Positive

Positive
Negative"
NI
Negative"
Positive
Negative

NI - not included
a. Denotes statistical significance at the 5 percent level.
b. Denotes statistical significance at the 1 percent level.
SOURCE: Reports of Income and Condition, Board of Governors of the Federal Reserve System;
Data Book, FDIC; Summary of Savings Accounts by Geographic Area, Federal Home Loan Bank
Board, and the Rand McNally Credit Union Directory.

the deposit share held by thrifts for the
number of thrifts to determine if their
competitive strength was related
directly to their market position.
It has been claimed that smaller institutions may not be viable competitors in
a less-regulated environment. If larger
banks realize economies of scale or have
some other advantages over smaller
competitors, then bank size would have
a positive influence on earnings.
Credit unions are active competitors
for consumer deposits and loans, and
their presence and concentration in a
market may exert downward pressure
on bank earnings. Regardless of competitive conditions, banks would be
expected to earn higher profits in
markets with stronger demand. We used

market deposit growth from 1982 to
1985 to account for demand differences
across markets.
The statistical results, as determined
by regression analysis, are presented in
table 4. These findings are consistent
with the t-test results and support the
view that the thrift institutions exert
strong competitive pressures on banks.
When other factors, such as bank size,
market growth, and credit union competition, were taken into account, along
with the number of banks in the market, the presence of thrifts in the local
area still had a negative and significant
impact on bank earnings. We also found
that banks in markets where thrifts held
a larger share of the deposits earned
significantly lower returns on assets.

13. We found usable observations in the following states: Alabama, Arkansas, Florida, Georgia,
Illinois, Indiana, Kentucky, Michigan, Mississippi, Pennsylvania, South Carolina, Virginia,
West Virginia, and Wisconsin. Arkansas banks
were included in the sample because their earn-

ings were relatively stable, and many Arkansas
banks met our other sample criteria.

Federal Reserve Bank of Cleveland
Research Department
P.O. Box 6387
Cleveland, OH 44101

Material may be reprinted provided that the
source is credited. Please send copies of reprinted
materials to the editor.

14. Lynn A. Nejezchleb, "Declining Profitability
at Small Commercial Banks: A Temporary
Development or a Secular Trend?" Banking and
Economic Review, Federal Deposit Insurance
Corporation, June 1986, pp. 9-21.

Conclusion
This Economic Commentary looks at

the role of thrifts as competitors to
banks. However, we do not examine the
issue of banking market contestability.
Our findings suggest that regulators
could be underestimating the actual
competition for banking services in
local areas by not adequately accounting for competitors of banks when evaluating the competitive effects of bank
mergers. In some markets thrifts
might properly be regarded as full competitors of banks. While our bank sample is limited to markets with no more
than a few competitors, the presence
and the relative size of a thrift in those
markets had a strong and similar effect
on bank earnings as a competing bank
in those markets. Banks operating in
markets with at least one thrift or in
markets with another bank earned significantly lower profits than other
sample banks. Also, earnings were substantially lower at nonsample banks,
which typically operate in markets
with a greater number of thrifts and
banks.
We found no evidence to support the
view that thrift competition is inferior
to bank competition-at
least in terms
of impact on bank earnings. Although
the competitiveness of thrifts could
vary widely over the many products
and services provided by banks, the
presence of thrifts in local banking
markets is exerting downward pressure on bank earnings.
15. The competitive influence of banks located
outside of these markets is generally considered
weak compared to the threat of entry in larger
banking markets.
16. This earnings difference, however, was statistically insignificant.

BULK RATE
U.S. Postage Paid
Cleveland, OH
Permit No. 385

Address Correction Requested:
Please send
corrected mailing label to the Federal Reserve
Bank of Cleveland, Research Department,
P.O. Box 6387, Cleveland, OH 44101.

November 1, 1987
ISSN 0428-1276

ECONOMIC
COMMENTARY

Local Thrift
Competition and
Bank Earnings
by Paul R. Watro

Economic and legislative changes have
clearly eroded differences between commercial banks and thrift institutions.
Commercial banks traditionally offered
a unique cluster of products that were
not available from other institutions.
Because of the expanded powers of
thrift institutions in the early 1980s, however, thrifts have been operating more
like commercial banks, providing a wide
range of financial services, including
checking accounts and business loans.
As a result, banking regulatory agencies, such as the Federal Reserve System, the Federal Deposit Insurance
Corporation, and the Office of the
Comptroller of the Currency, are reevaluating some of their policies. For
example, the Federal Reserve Board is
currently reconsidering its policy of
prohibiting bank holding companies from
acquiring healthy thrift institutions
and has solicited public comments.
Some of the Federal Reserve's previous concerns about banking organizations operating thrifts have apparently
lessened with the removal of interstate
barriers to geographic expansions by
banks and with the elimination of statutory deposit-rate differentials between
banks and thrifts.'
Also, deregulation, along with technological innovations and increased
customer sophistication has fostered an
expanding number of firms providing
financial services and has intensified
competition among banks and thrift
institutions. These developments have
made competitive assessments of bank
mergers extremely difficult and open to
considerable debate.
Bank regulators rely largely upon a

structure-performance framework that
implies that market structure (the number and size distribution of competitors
in a market) is an important determinant of bank performance. With this
line of reasoning, the fewer the number
of competitors and the larger the share
of the market controlled by the largest
competitors, the greater the likelihood
that firms will be able to charge prices
above the costs of doing business and to
generate profits above the competitive
level. The traditional structuralist view
is incorporated in the Justice Department's merger guidelines, which are
used in various degrees by banking regulators when evaluating the possible
anticompetitive impact of bank mergers.
Some analysts downplay the importance of market structure, reasoning
that the structure itself may be a product of the competitive forces at work in
the market. In this view, market concentration is unlikely to lead to collusion and excessive profits when barriers
to entry are low. Supporters of this perspective argue that the threat of entry
by firms outside of the market will
exert competitive discipline on market
participants, regardless of the actual
number of competitors in the market."
Banking agencies and the Justice Department are required by law to prevent
bank mergers and acquisitions that
would have substantially adverse effects
on banking competition. The competitive assessment of regulators centers
upon three factors: the product market,
the geographical market, and the likely
anticompetitive effects of the transaction (if any) within those markets.'
For antitrust purposes, the product

market for banking has been defined by
the Supreme Court to be a separate and
distinct line of commerce. This landmark decision was made in 1963, reaffirmed in 1974, and was based upon the
view that commercial banks provide
local customers with an unique cluster
of services that are not available from
other depository institutions.' Since that
time, however, sweeping changes have
taken place in banking. For instance,
thrift institutions can now offer the
same services as commercial banks.
Other nondepository institutions, such
as securities brokerage firms, mortgage
banking firms, insurance companies,
and finance companies are providing a
variety of financial services that are
competing with bank products.
Regulators have reacted somewhat to
these major developments. To various
degrees, the existence of thrift institutions and other nonbank suppliers of
financial services is being considered in
regulatory assessments of competitive
factors. Moreover, the Justice Department is apparently more lenient toward
bank mergers than toward mergers in
other industries, because the department apparently assumes the competitive influence of thrifts and other nonbank institutions, as well as the
imprecise boundaries used to estimate
geographical banking markets."
This Economic Commentary discusses the importance of thrift competition and examines earnings differences
between banks operating in markets
with and without the presence of
thrifts." Bank earnings are used as a
measure of overall competition for
banking services in local markets. If

Paul R. Watro is an economist at the Federal
Reserve Bank of Cleveland. The author would like
to thank Mark Sniderman, James Thomson, and
Gary Whalen for their helpful comments, and
would like to thank John McElravey for his valuable research assistance.
The views stated herein are those of the author
and not necessarily those of the Federal Reserve
Bank of Cleveland or of the Board of Governors of
the Federal Reserve System.

1. For a discussion of Federal Reserve policy
regarding the acquisition and operating of thrift
institutions by bank holding companies, see Federal Register, Vol. 52, No. 186, September 1987,
pp.36041-36045.

3. For a discussion of the analyses used by regulators and Justice Department, see Robert E.
Hauberg, Jr. "Mergers and Acquisitions: Trends
in Competitive Analysis," Banking Expansion
Reporter, Golembe Associates, Inc., Volume 6,
Number 13, July 6, 1987, pp. 1,9-17.

2. See Gary Whalen, "Concentration and Profitability in Non-MSA Banking Markets," Economic
Review, Federal Reserve Bank of Cleveland,
Quarter 1, 1987, pp. 1-15.

4. United States v. Philadelphia National Bank,
374, U.S. 321 (1963); and United States v. Connecticut National Bank, 418 U.S. 666 (1974).
5. Robert E. Hauberg, Ir., p. 13. Under the revised