The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Working Paper 74-1
THRIFT INSTITUTION COMPETITION WITH
COMMERCIAL BANI(S: EMPIRICALEVIDENCE
William Jackson
January 11, 1974
The views expressed here are solely those
of the author and do not reflect the views
of the Federal Reserve Bank of Richmond.
THRIFT INSTITUTION
COMPETITION WITH
COMMERCIAL BANKS: EMPIRICAL EVIDENCE
Introduction
Several banking law changes have been proposed
“to increase” com-
petition between commercial banks (“CB”), mutual savings banks (“MSB”),
and savings and loan
associations (“SLA”), This paper will examine the
empirical evidence concerning thrift institutions' effects on financial
conditions generally, and particularly on CB activity,to determine the
results of existing interindustry competition. If NSB and SLA are presently Important CB competitors, new analysis should reveal their influences, as an objective aid to proposed legislation. This issue is alb3
related to the geographical and functional nature of banking ‘products,’
for if nonbanks influence CB, then they should be considered as actual or
potential competitors when deciding “cases,” both by bank regulators and
by the courts.
Table 1, below, outlines these institutions' portfolio characteristics.
The important markets served by SLA and MSB are clearly time and savings
deposits ("TD") and mortgage loans. Thrift institutions generally do not
offer demand deposits ("DD') nor do they normally offer nonrealty 1oans.l
'The usual exceptions to-any such generalization exist.
For example,
Maryland MSB can offer DD, and Texas SLA can provide consumer loans. MSB
generally can hold corporate securities, Statistically, such legalities
have had little effect on financial competition.
Table 1,
Increase In
Mortgages/Increase
in TD, 1972
Total Loz;ls/Total
Deposits, 121X/72
Home Mortgage
Loans;TD. 12/31/?2
Other Mortgage
LoansfTD, 12131172
Mortgage Loans/TD
12/31/72
CB
17.84%
13.34%
31.18x*
39.23%
SLA
80.85
18.71
‘99.56
97.47
lOO.G?
HSB
45.46
28.28
73.74
54.84
76.99
Institutions
* 18.54% of DD and TD
Source :
--
Selected Portfolio Ratios, 1972.
Computed from [20].
71.58%
3
Theories of Nonbank Competition
MSB and SLA are assumed, in the literature, usually to, increase the
effective supply of both TD and mortgage loans (hence, total loans) where
they are present. Their rate or convenience competition may increase the
attractiveness of, and hence raise the total of, an area's savings. MSB
and SLA may also attract funds from locations outside of their service areas,
through effective deposit rate competition,*
If a downsloping demand for
financial services exists, their competition should thus raise TD rates
and lower mortgage loan rates for an area and thus for its CB.
Alternatively,
some have assumed that nonbanks divert funds away from CB through nonprice
competition without necessarily altering the volume of an area's financial
activity.
Changes in related "product markets" may occur, such as CB diversion of liabilities into DD, CB asset shifts into nonrealty loans and
investments, and CB asset size being smaller than otherwise, if nonbank
competition is effective. In behavioral terms, nonbank activity may cause
CB to "give up" competition in the TD and/or mortgage markets (e.g., California). Conversely, CB may believe that customers prefer to deal with a
"one-stop" financial company for all of their transactions. Thus, CB
managers may be forced to become more aggressive, seeking to maintain or
increase their firms' activity and market share relative to all of their
competitors, under nonbank competition.
*California SLA and Massachusetts MSB are examples of euch
interregional flows of TD funds.
4
The relationships between banks and nonbanks may also flow in a
reverse
direction. For example, Baxter, McFarland, and Shapiro [2]
suggest thatt
I
'Nonbank competition arises from the inability or
unwillingness of commercial banks in an area to
meet the needs of the consuming public adequately."
A more "rigorous" approach to this issue, that of Carson [4], is
extremely obscure. He suggests that nonbank TD competition diminishes
the TD/total deposit ratio of CB.
Since this ratio depresses CB earnings,
he believes that if nonbanks capture TD funds and lend them as mortgage
loans, whose proceeds return to CB as DD, CB asa
group will benefit. Yet
Carson's individual bank faced with nonbank competition may suffer a declining financial position,
Indeed, by substituting SLA and/or EiSBas competitors in differing
theories of the firm as a bank, it can be shown that almost any CB portfolio or price change can result from nonbank activity.
See [13] and its
bibliography. A brief examination of empirical studies of nonbank competition would thus appear to be highly relevant as a policy basis,
uregatad
Studies
Several historical studies of CB, MSB, and SLA interactions are
built upon a "demand for financial asset" foundation. They examine possible substitutability between various deposits and/or money, as extensions
of the Gurley-Shaw hypothesis, Baumol-Tobin inventory theories of transactions balances, or other externally-based concepts. These studies
generally find that thrift institutions' liabilities are highly liquid
substitutes for each other, DD, TD, and other possible near money assets.
15) is an example of, and (221 Is a summary of, such studies. This type
5
of indirect evidence for interindustry interaction is typified by Cohen
and Kaufman's [6] 1951-61 income elasticities of demand for DD (.84),
commercial bank TD (.80), and nonbank savings deposits (,98). Hence,
they conclude that great interindustry competition must exist.
12, 6, 7,
10, and 17) examine such measures as deposit growth, total savings, etc.,
within similar philosophical foundations, to arrive at similar conclusions,
Other studies, examining interindustry competition more closely,
also find that interindustry rivalry exists. Jung [15] finds that
Chicago bank and nonbank interest rates on similar mortgages tended to
converge in the five years after 1960. Bloch [3] examines several proxies
for competition at the national level. His CB and STA asset growth, market share, profit, and profit margin figures suggest that a "new competition"
between SLA and CB arose during the 1960's. Bloch attributes such results
to CB initiative, in contrast to part of Friend's [8] study emphasizing STA
aggressiveness. Dhrymes and Taubman's [7] analysis of SLA activity, including some CB effects, is unfortunately unable to shed any light on the
extent of interindustry competition.3
Micro Studies
The more relevant studies of interindustry competition attempt to
directly relate CB price and quantity variations to nonbank activity by
using microeconomic theory. Their initial assumption is that nonbanks may
be direct bank competitors. Such research is worth a closer examination.
3This is because a one-period lagged dependent variable seems to
capture the overwhelming majority of their equations' explanatory power.
6
Phillips (18) reports an earlier study finding that savings account
interest was slightly higher when SLA were 'large' relative to CB in the
area studied. Horvitz and Shull's 1121 regressions find that isolated
towns' unit banks have lower TD interest rates and lower TD/total deposit
ratios when other savings institutions are present (O-1 dummy variable).
But thrift institutions do not affect Horvitz and Shull's CB loan interest
rates or their loan/asset ratios.
Kaufman's [ld] Iowa county regressions
find that the ratio of SLA assets to CB deposits stimulates CB loan interest,
depresses the ratio of bank TD to total TD, but is not significantly related
to bank TD interest rates, CB earnings/assets, or to bank loans/assets.
Ware's [21] analyses use the number of SLA in Ohio nonSElSAcounties to
define nonbank competition, This measure does not influence CB DD service
charges, CB profitability, or CB thrift account interest. Conversely, SLA
numbers stimulate Ware's CB loan interest rates and price-spread monopoly
power proxy (loan interest rate minus TD interest rate). Aspinwall [l]
treats CB, MSB, and SLA as homogeneous sources of mortgage funds. His
analysis groups all three legal forms into numbers and concentration ratios
of firms in OMSA's.
These structural variations influence mortgage interest
rates in the expected direction. Friend's [8] first regressions seeking
nonbank influences on CB TD deposit rates, using "market area' figures,
are
so
"unimpressive" that these results were not published. His 1960 re-
gressions, however, using state-level figures, show that TD interest rates
are positively related to SLA and MSB competition. But such effects do not
appear in his regressions for 1966 or 1967.
mortgage
Neither do they appear in his
interest rate analyses at either the market or state level. Most
recently, Stuhr [19] uses
unity minus an
area "market share" of each of
25 New York banks as a competition variable, measured with respect to CB,
7
SLA, and MSB activity, Stuhr shows that while CB-only competition does not
affect TD interest, MSB and particularly SLA competition is positively
related to CB performance in this product market.
Similarly, SLA and MSB
competition in the mortgage market seems to lessen CB mortgage interest
rates,
Evaluation of Published Studies
Taken as a whole, both the micro and macro-oriented studies seem to
refute the hypothesis that thrift institutions do not alter either CB behavior or financial activity generally, But these analyses use 1950s and
1960s data, whose validity may have declined since the effective lmposition of interest ceilings on thrift institutions during the tight money
period of recent years* Moreover, new financial influences of a behavioral,
legal, or technological nature, extending both banks and nonbanks' product
and geographical markets in recent years, suggest that new analysis of
their competition is needed.
Empirical Analysis of Nonbank Competition
MSB and SLA competition with CB can be directly examined within the
framework of a larger study of the sources of American CB allocational
efficiency [13]. Honbank competition is defined as the ratio of time and
savings deposits held by MSA and SLA divided by these deposits plus CB time
and savings deposits in a state,4
This ratio directly measures the effectiveness
4MSB and SLA were combined to lessen the regional bias of MSB concentration in a few states, State levels were utilized for theoretical reasons, as
there appears to be a wide range of statewide linkages affecting the financial
structure in recent years, to a greater extent than in the past. Moreover, no
delineation of banking "markets" is available to cover 44 states.
8
of nonbank competition for funds with CB,
(The assumption is made that
recent historically high interest rates have lured as large a portion
of interest-sensitive Ml out of DD as would be likely to shift into
interest-bearing accounts in the near future.)s Moreover, the high
mortgage loan/TD ratios of MS,Band particularly SLA (Table 1) suggest
that this measure may also rather directly measure competition in the
mortgage loan market,
Less directly, as shown above and by [13], this
ratio may represent nonbank competition's effect on CB output or prices
generally, to the extent that CB portfolio decisions relate various
"single-product" deposit and loan markets to each other.
This study's dependent and many independent variables are balancesheet and income-statement data for 1644 CB in 44 states for the 1969-71
years, as provided by the FDIC,
These firm-level figures are combined
with appropriate state-level banking and socioeconomic traits in three
forms of analysis, whose results appear below.6
Correlation Analysis
Table 2, below, shows Pearsonian correlation coefficients between
nonbank competition and bank-firm variables, These correlations (except
for multibank holding company affiliation and adjusted loan interest) are
51,
NOW" accounts, which were not significant during the period
studied, may alter'the realism of this assumption.
he
data are group averages of three very highly similar banks
combined to prevent "disclosure'. No apparent selection or grouping biases
are present in this data,
Ray Cobble programmed these analyses, using
programs BMDO8M and BMDO2R.
Table 2.
Correlation Coefficients of Banking Variables
With Nonbank Competition
-0.17
Unit-type Bank, O.or 1 branch*
Small Branch Bank, 2-5 branches*
0.07
Large Branch Bank, 6 or more branches*
0.15
Multibank Holding Company Bank*
-0.01**
Time and Savings Deposits/Total Deposits
-0.28
"Investmentsu/Assets
-0.25
0.09
Cash Items/Assets
-0.30
Agricultural Loans/Total Loans
Commercial and Industrial Loans/Total Loans
0.27
Consumer and Individual Loans/Total Loans
0;10
Trust Revenue/Total Revenue
0.23
Equity/Assets
0.11
Labof Expense/Revenue
0.36
Occupancy Expense/Revenue
0131
Dividends/Net Income
0.08
Bank Asset Size
0.13
Operating Revenue Less Demand Deposit
Service Charges/Assets
0.08
-0.i3
Wet Income/Equity
Loan Interest Minus Loan Loss Provisions/Loans
Time and Savings Deposit Interest/Time and
Savings Deposits
0.03**
,
-0.09
"Monopoly Power" (Adjusted Loan Interest Rate
Uinus Time and Savings Account Interest Rate)
0.09
Loans and Discounts/Total Deposits
0.23
*Dummy Variable (0 if no, 1 if yes)
**Not highly significant.
10
highly significant (at
rather weak.
the 0.01 level of a two-tailed test). Yet they are
The strongest of these upward-biased relationships (0.36)
leaves over 87% of variation in labor cost unexplained, Some coefficients,
such as those of the TD ratio, the commercial loan ratio, the trust/total
revenue ratio, and the net income ratio, are very plausible. The positive
cost coefficients (labor expense/revenue, occupancy expense/revenue) suggest somewhat that nonbanks cause CB to raise costs, a presumptive sign of
nonprice competition, And nonbank competition does not seem to be associated
with smaller banks, as measured by either branches or assets. But the
hazards of generalizing from these coefficients may be illustrated by the
investments and agricultural loan ratios’ coefficients, which have the
opposite signs from those expected.7 Other influences such as regulation
may be at work to mask the true, separate, effects of nonbank competition
8
on CB.
Factor Analysis
Hence, factor analysis, combining nonbank competition with 52
regulatory, structural, managerial, demand, and time trend variables,
should reveal if nonbank competition is strongly related to other
forces.
(Factor relationships are not necessarily “causal.“) In fact,
it finds [14] that 13 patterns of common influence appear among the 53
variables analyzed.
7Competltion in realty loans should raise the relative profitability
of other earning assets, hence their ratios to loans or assets.
8This difficulty is the “multicollinearlty” that. infests financial
data generally;
[2, 13, 141,
11
Nonbank competition appears in two of these factors. One of them
groups it with statewide "demand" forces: financial employment, population density, urbanization, per capita income, and population growth.
This pattern also associates nonbank competition with limited holding company states, new bank entry, large (state) average bank size, :aqd CB deposit
inequality ("Gini coefficient"); but oppositely from CB agricultural
loans and agricultural employment. Additionally, this factor associates
nonbank activity with
many
households per CB office, as would be ex-
pected if interindustry "convenience" competition is effective,
This
pattern does not support the hypothesis that SLA and MSB activity causes
CB size to decline.
A second factor relates nonbank activitv to CB firm-level traits.
It groups SLA and MSB competition with equity capitalization, high labor
expense/revenue, and low CB time and savings deposit capitalization. This
factor again associates thrifts' activity with limited holding company
states.
It captures some agricultural loan activity, however, in contra-
diction to the first factor.
Thus, nonbank competition accompanies many exogenous 'demand' forces,
which should be associated with financial institutions' output generally.
But it is also associated with a few forces that
should depress individual
CB activity, such as new entry, equity capitalization, and labor expense.
Even factor analysis cannot completely resolve this issue's complexity.
Thus, nonbank competition is directly regressed on CB performance ratios,
controlling for other plausible causes of CB activi:y as suggested by a
similar factor analysis 1131. A stepwise regression process is used to
lessen any remaining multicollinearity. No nonlinearities appear in plots
of residuals.
12
Flow-Output Regression Ana.lysis
Table 3, below, explains a typical CB's flow-output proxy (operating
revenue less demand deposit service charges/bank assets). The dependent
variable measures a bank "activity" dimension which Is roughly comparable to
GNP and industrial firms output, relative to available CB resources*
!ChiS
equation contains both "production" and price-related variables.
Sixteen tiariablesinfluence this output proxy*
Nonbank competition
is a highly significant stimulant of adjusted revenue/assets. An increase
of 25% in the ratio of nonbank TD to total TD (relative to the sample's
mean value of about 0.41) in a state Would raise the typical bank's flowoutput by 0.76% (relative to output's mean value). This result seems to
reflect CB portfolio shifts into loans when forced with thrifts' competition (since nonbank competition is not slgnlflcantly related to loan
interest by either Table 1 or by later regression analysis).
Yet this results' absolute size is rather small. Other forces have
generally more significant effects in Table 3.
Their influences are worth
brief coverage.
The time trend of higher adjusted revenue/assets may reflect both
tight money and managerial improvements, A high cash ratio reduces output, reflecting financial "excess capacity.' Herfindahl concentration
(the sum of squared deposit shares and thus the oligopoly proxy) raises
adjusted revenue through its probable influence on average price.
Popula-
tion growth (demand) stimulates this measure, as would be expected. Both
unit-type and 2-5 branch banks have lower flow-output than extensively
branching firms.
Since the two other muttibank holding company laws bear
Table 3.
Variable
Stepwiee Regression for Flow-Output
Coefficient
Standard Error
3.58804
0.23272
237.71a
-42.59164
4.09341
108,26a
1971
2.07157
0.23551
77.37a
Herfindahl Concentration Index
1.79993
0.23253
59.92a
Population Growth
9.96616
1.42138
49.16a
- 1.94767
0.30403
41.04a
2.26478
0.39063
33.61a
17.39389
3.11130
31.25a
1.70967
0.36509
21.938
- 3.57043
0.76509
21.78a
1.31274
0.30784
18.18a
2-5 Branch Bank
- 1.36441
0.32155
18.01a
Per .CapitaIncome
- 9.38876
2.32187
16.35a
Time and Savings Deposit Ratio
- 6.50513
1.75102
13.80a
Nonbank Competition
4.31324
1.31665
10.73b
National Bank
0.64256
0.27702
5.38c
1970
Cash/Assets
O-l Branch Bank
Non-Multiback Holding Company State
Labor Cost/Revenue
Multibank Holding Company State
Equity/Assets
NonnernberState Bank
Adjusted Revenue/Assets Ratio (Mean)
Intercept
R2
F(16, 1627)
Standard Error of Estimate
Standard Error of Estimate/Mean
F Statistic
58.94061
63.36482
0.3628
57.90a
3.8388
0.0651
a Significant at the 0.001 level.
b Significant at the 0.01 level.
' Significant at the 0.025 level.
Variables are O-l dumy- variables, or cardinal values scaled to have means comparable to those of the dummy
The dependent variable is multiplied by 1,000 for technical reasons.
variables (0.10 to 0.90).
. ....I..' - -. -'.
.
14
positive coefficients, limited multibank holding company states seem to have
low output.g Labor cost is associated with high adjusted revenue/assets, as
high cost must be reflected in high average price.
The equity capitali-
zation ratio lowers output, behaving as a liability analogue of liquidity
in restraining CB activity [13], Nonmember state banks, and to a lesser
extent national banks, appear to generate larger output than state members
do.
Per capita income seems to lower output, perhaps reflecting regional
10
traits.
to
The TD capitalization ratio, finally, is negatively related
this measure, so that it alone may not indicate banking "competition."
Stock-Output Regression Analysis
Table 4, below, further explores nonbank influences on CB output,
this time defined as the traditional loan/deposit ratio, Nonbank competition
is the most significant influence on this stock-output proxy.
An Increase of 26% In nonbank competition would raise the typical
bank's loan/deposit figure by 4.23% relative to its mean (an 0.0245 absolute
increase). Hence, MSB and SLA again stimulate bank output generally.
Briefly, thirteen other influences also explain CB stock-output.
Both unit-type banks and 2-5 branch banks lend less of their resources than
do extensively branching ones.
The holding company laws' coefficients imply
that limited holding company states have lower output than other ones do.
Larger banks tend to lend more of their resources, partly supporting the
9
Changing CB holding company regulation has made these variables'
effects less important than they have been In the past.
"[l]
gives theoretical reasons why per capita
true "demand" proxy.
income may not be a
Table 4.
Stepwise Regression for Stock-Output
Variable
Coefficient
Standard Error
F Statistic
236.56052
24.30205
94.75"
- 55.68581
6.30470
78,01a
Multibank Holding Company State
54.21399
7.51935
51.98a
Non-Multibank Holding Company State
55.80005
8.08032
47.6ga
Bank Asset Size
34i64708
5.83978
35.20a
1971
- 19.33206
3.99363
23.43a
Cash/Assets
-274.41821
57.04234
23.14"
Labor CostjRevenue
-274.81299
58.87668
21.7ga
177.50931
44.74142
15.74a
Equity/Assets
- 57.85732
14.58771
15.73a
2-5 Branch Bank
- 23.37730
6.50762
12.90a
State Deposit Coefficient of Variation
50.28008
15.20710
10.93b
Rerfindahl Concentration Index
14.22313
4.58251
9.63b
9.55523
3.24929
8,65b
Nonbank Competition
O-l Branch Bank
Per Capita Income
5-Banking Organization Concentration Ratio Change
Loan/Deposit Ratio (Mean)
I tercept
RS
F(14.1629)
Standard Error of Estimate
Standard Error of Estimate/Mean.
See Table
580.98193
532.32544
0.2557
39.97a
75.5979
0.1301
3 '6 Footnotes.
.-
-
-
-
-
_
_
__
:
.-
._
_.
.
.-
-
-_
16
existence of "economies of scale." A time effect shows that 1971's loan/
deposit ratio was lower than that of previous years. As would be expected,
cash holdings depress output.
negative effect on output.
Similarly, high labor cost has a strong
In this equation, per
capita income behaves as a
demand proxy. Low leverage (high equity capitalization) again depresses output. A large coefficient of variation in CB deposit sizes stimulates this
ratio. Somewhat surprisingly, both Herfindahl concentration and a long-term
increase in the S-banking organization concentration ratio have positive,
not negative, effects on output. This result is consistent with CB lending
more of their assets when oligopolistic conditions allow higher gross loan
yields.
([13] finds several such effects.)
Price, Profit, and "Monopoly" Regression Analyses
Other regressions (reported by [13]) similarly consider the effects
of nonbank competition and eighteen banking and socioeconomic forces on CB
profitability, adjusted loan interest, TD interest, and "monopoly power."
Nonbank competition does not appear to significantly (0.05 level) influence
these banking ratios, when other traits' effects are included in the
11
regressions.
Rather, regulatory, managerial, etc., variations determine
these price-related CB characteristics. Nonprice competition, rather than
direct price competition, may thus characterize CB-MSB-SLA interaction, at
this study's level of abstraction [ll].
llIts near-zero partial correlations with these measures, controlling
for all significant variables, are:
Profitability
Adjusted Loan Interest
TD Interest
Monopoly Power
-0.023
-0.028
-0,033
-0,029
11
13
13
11
variables
variables
variables
variables
significant
significant
significant
significant
17
Conclusions
Efficient CB evidently need not fear SLA and MSB activity, for nonbank
competition does not affect CB firm-level profitability, contrary to [4],
In a broad sense, MSB and SLA seem to actually stimulate CB output, measured both as a flow ratio and as a stock ratio, Moreover, nonbank competition seems to be associated with high financial activity on the state level,,
(Although the causal relation may flow from demand to nonbank activity
rather than in reverse,) The sep,arateexistence of these thrift institutions thus seem to be desirable from a financial activity viewpoint.
Hence, this study suggests that SLA and MSB be considered as relevant competitors when regulatory decisions involving mergers and CB holding company
acquisitions are being made.
Finally, these results suggest that allowing
CB bank holding company acquisition of existing thrift institutions, in
states where the holding company is already represented, may have anti-competitive effects.
18
REFERENCES CITED
1.
Aspinwall, Richard. "Market Structure and Commercial Bank Mortgage
Interest Rates," Southern Economic Journal, XXX11 (April, 1970),
376-84.
2.
Baxter, Nevins, David McFarland, and Harold Shapiro. "Banking Structure
and Nonbank Financial Intermediaries," National Banking Review,
IV (March, 1967), 305-16.
3.
Bloch, Ernest. "The Setting of Standards of Supervision for Savings
and Loan Associations," in Friend (ed.), pp. 1629-36.
4.
Carson, Deane. "Bank Earnings and the Competition for Savings Deposits,"
Journal of Political Economy, LV (November, 1959), 580-88.
5.
Chetty, V. K. "On Measuring the Nearness of Near-Moneys," American
Economic Review, LIX (June, 1969), 1945-66.
6.
Cohen, Bruce, and George Kaufman. "Factors Determining Bank Deposit
Growth by State," Journal of Finance, XX (March, 1965), 59-70.
7.
Dhrymes, Phoebus, and Paul Taubman. "The Savings and Loan Business: An
Empirical Survey," in Friend (ed.), pp. 67-182.
8.
Friend, Irwin. "Changes in the Asset and Liability Structure of the
Savings and Loan Industry," in Friend (ed.), pp. 1355-1429.
9.. Friend, Irwin, (ed.). Study of the Savings and Loan Industry. Washington: Government Printing Office, 1970.
10. Goldfeld, Stephen, "Savings and Loan Associations and the Market for
Savings," in Friend (ed.), pp. 569-659.
11.
Grebler, Leo, and Eugene Brigham. Savings and Mortgage Markets in
California. Pasadena: California Savings and Loan League, 1963.
12.
Horvitz, Paul, and Bernard Shull. "The Impact of Branch Banking on Bank
Performance," Studies in Banking Competition and the Banking Structure. Washington: U.S. Treasury Deuartment, 1966, pp. 141-86.
13.
Jackson, William. "Commercial Bank Regulation, Structure, and Performance."
Unpublished doctoral dissertation, University of North Carolina, 1974.
14.
Jackson, William. Commercial Bank Performance and Structure: A Factor
Analysis Approach. Federal Reserve Bank of Richmond, Working Paper 74-5.
15.
Jung, Allen. "Terms on Conventional Mortgage Loans--l965 Versus 1960:'
National Banking Review, III (March, 1966), 379-84.
19
16.
Kaufman, George. "Bank Market Structure and Performance: The Evidence
from Iowa," Southern Economic Journal, XXX11 (April, 1966). 429-39.
18. Kaufman, George, and Cynthia Latta, "Near Banks and Local Savings,"
National Banking Review, III (June, 1966). 539-42,
18.
Phillips, Almarin, "Competition, Confusion, and Commercial Banking,"
Journal of Finance, XIX (March, 1964), 32-45.
19.
Stuhr, David, "Competition and Commercial Bank Behavior," Proceedings
of &Conference on Bank Structure and Competition. Chicago:
Federal Reserve Bank of Chicago, 1972, pp* 184-202.
20. U.S., Board of Governors of the Federal Reserve System, Flow of Funds
Accounts 1945-1972. Washington: Board of Governors of the Federal
Reserve System, 1973,
21. Ware, Robert. "Banking Structure and Performance: Some Evidence from
Ohio,"-Federal Reserve Bank of Cleveland Economic Review, March,
1972, pp. 3-14.
22.
Yesley, Joel. "Defining the Product Market in Commercial Banking,"
Federal Reserve Bank of Cleveland Economic Review, June, 1972,
pp* 17-31.
@FedFRASER