i "\W-v‘x-v-vu-g -.

STATE BRANCH BANKING LAws¥4IIIEIR 'EFFECT I
0N ECONOMIC ACTIVITY

Thesis for the Degree of Ph. D.
MICHIGAN STATE UNIVERSITY
RANDALL H. HOEMKE
1969

f «A‘- A.”-
35cm:

 
     

Michigan State
UniverSIty

. 5W“ “fl '

This is to certify that the

thesis entitled

STATE BRANCH BANKING LAWS——THEIR EFFECT
ON ECONOMIC ACTIVITY
presented by

 

Randall H. Hoemke

has been accepted towards fulfillment

of the requirements for

Ph.D. degree in Economics

MIKE -

: Z :r
Major professor
Dateflm} / §(7’

0-169

 

ABSTRACT

STATE BRANCH BANKING LAWS—~THEIR IMPACT
ON ECONOMIC ACTIVITY

By
Randall H. Hoemke

In 196A, when discussing the rapidly expanding
economy of the State of California, J. Fred Weston
raised the following question:

Have the characteristics of commercial

banking in California contributed to that growth?

Or, has the growth taken place in spite of the

characteristics of commercial banking in Cali-

fornia?

His query is the basis for this attempt to deter—
mine which of the three systems of banking, as delineated
by varying limitations on branching specified by individual
state law, is most successful in stimulating economic
activity in the state in which it is located.

The literature investigating branching as an alter-
native to Unit Banking is reviewed. Only the study by

Cohen and Reid specifically addresses itself to the public

interest and it is inconclusive in this regard. The

Randall H. Hoemke

remainder concentrates primarily on analyzing cost struc-
tures suggesting that lower marginal costs imply lower
price and larger quantity of investment funds devoted to
specific loan categories if the applicable demand curve is
downward sloping. Recent literature recognizes the impor-
tance of the demand side of the market in determining the
proportion of the portfolio devoted to Loans and his-
counts.

Information regarding portfolio income limited the
portfolio breakdown to (a) Cash Items, on which no income
is earned; and (b) United States Government Securities,
(c) Other Securities, and (d) Loans and Discounts on which
income is earned. The primary source of data was the year—
end "Call Report" and the Annual Report of the Federal
Deposit Insurance Corporation for the years 1946 through
1966. (Hue proporticn1<3f the total tunn;flalio devoted to
each of these components is designated the Y (or dependent)
vardjnile. lfliis is Ill acccnfliance Ivith Nuxrk by (hilbraitfli,
Alhadeff, and others.

Russell's study was instrumental in establishing the
independent profit variable [X3 in Equation (1) and Xu in
Equations (2) through (A)]. Inclusion of the per capita
real income variable (X2) and the variable reflecting
liquidity needs (X1) is in accordance with principles

<9Xpressed by Robinson.

Randall R. Hoemke

Review of the elements of demand for each of the
portfolio components led to the decision that the demand
curves facing bankers in each of the banking systems are
similar. The supply structure was analyzed and it was
decided that there is no suggestion of difference in the
willingness of bankers in varying market structures to
supply a quantity of'fhnnhs to a specific component at a
given price. Consequently, the following hypothesis is
adopted:

It is hypothesized that no difference exists
among the systems as to how they determine the pro—
portionate share of their portfolio to devote to any
one of the four components. Stated more specif-
ically in terms of the statistical computation per-
formed, it is not expected that the Statewide
Branching system or the Limited Branching system is
significantly different (as revealed by F tests)
from the Unit Banking system in the proportion of
its funds devoted to any one of the components anal-
yzed (the Y variable) given the same motivations (as
designated by the X variables) for having made that
determination.

The use of Dummy Variables in the regression equa—
tions allows the variance attributable to the S,atewide
Branching system (D1) or to the Limited Branching system
(D2) to be isolated from the variance present in the entire
banking system. The technique makes possible, through the
summing of the Constant and the Regression Coefficient of
the Dummy Variable or through observation of the F signif-
icance, an evaluation of the extent to which the system

indicated by a Dummy variable is different from the Unit

Banking system.

Randall H. Hurmkw

1k) detiwnnine ii‘ any ciiffereruxe BXiLHLT aHKNLg tin? var-

ious baimfirur systtnma, four setmuzific twang; are iwm'flarmcd
(l) Cross-section regression for each of the twenty—one
years of the study (1946 through 1966); (P) Time-series
regressions across the entire time period of the study;
(3) Eknyhnated Y valiuns at t M?ih€3n5 (m‘tmk: X variables
bascml on tin: CPQSgF—SficthWl estirmitingjtniuaticnhi; arni (U)
tkfl,imat«”i'{ ViltbuT at timsrmeans (if tiwi X vardJniles inSPd
on the time—series estimating equations.

'Phe {druihigs oi‘iAh# study iihiicate tine followirnt:

(l) The Statewide Branching system is significant

(differern: from tin? Unit [kniking symflxnn in tin»

lv

manner in which it responds to similar motiva—

ticans as rewrarmis tJie 13r6u3o1%,iori of‘ its: pt)PtTV)llF)

(l

dtflthFBd t;o:

(a) (jfifiil Itern3:

 

Y'
'.

In tim?(7ross-secti(u1 {egrwusinuh:, the
Statewide Branchinrjsystmnnzwr; found to

;;iitni-fi-cau1t'ly di ffX3r<n1t ffiioni tlic Hillt

i3arfl<ing§ sygflzeni at ‘the> 10 guermnent. le\nbl i

(H‘

i 1

12 out of 21 regressions—~in su.5 percent

of the examples tested. in the Time—Series

Regressions, the Statewide Branching sys

was found to be significantly different
from the Unit Banking system at the

cern; lcnnel. th(‘3iStlIH3tCWl Valvk~s (if Y

5 pr

tum

i"

(b)

Rah drill i1. fine Irzk.‘

(Test 3) for the Statewide Branching system
were less than the Estimated Values of Y
for the Unit Banking system for each of the
El years of the studv——in 100.0 percent of
the examples tested. The average of the
swttirnatthj 3’ vafllue.; ftn" tin? St ithVldl‘
Firminrfiiiiig .iy;ittwn is 7. 3 [Dei'ccuii, lt‘fitl tliaii
tin“ avcrmnye for tin: Unit [kinkiint.systeih
Tin) est;imaiw9d ‘{ valara ill ttna Tiime-stm"ios
Regression for the Statewide Branching
svstem is 3.U percent lower than the esti—

Ir

mated Y value for the Unit Brnking system.

C!)

lhiited tates Government CeCLudifivnu

 

The Cross—section Regressions revealed that
tin? Statevflxfie Branchirnt systeni Ls signifl-
icnritly (liffeiwnit iian thrrllnit,l%ankirnf
systaun at tin} 10 perwnnn; level iiiES out IW'
21 regressions——in 22.7 percent of the
examples ested. The Time—series Regres—
sion indicates no difference between the
Statewide Branching and the Unit Banking
system. The Estimated Values of Y in the
Cross-section Regressions for the Statewide
Brcnching system is found to be lower than
those in thETtht Banking system in all

years except one. The average of the

(a)

Riindf‘ll 1 H . ”003114? *‘

Custirmite<l Y \F11U6H1 is; 10.?35 gh‘PC(Hit liuver'
than tdué averenya for tim‘lhiit Bankiin: sys—
tenn ‘The Ekdxhnated \Hrhre of Y iriiflua'fime—
serdeas Regressitnn3 is 4.0 peimmnn; lower
tiHNi that;.for tine Uniizilankirn; systcnn.

Loans and Discounts, Net:

 

The Cross—section Regressions indicate that
the Statewide Branching system is signif-
i(vint;ly (iif‘FeIW3nt; fd”0fll tin: ihiit Baiikirig
;qi:ttwn ai. tiu> 1(1 permnfini, le\nel iri U (gut (if
21 repressions——in 18.2 percent of the
réxanniles iréstcmi. llie Tfirne-ansritma Hegiwcs—

sion indicates that the Statewide

U

Branching syst’m is not significantly

(1

different from the Unit Banking system.
The Estimated Values of Y in the Cross—
section Regressions for the Statewide
Branching system are higher than those for
the Unit Ranking system for each of the Pl
years of‘ the study. The average’.) of‘ tinese
estimmfimwi { values for the Statewide
Branching system is 12.1 percent higher
than the average for the Unit Banking sys—
tem. The estimated value of Y in the Time—
series Regressions is 2.1 percent higher

that that of the Unit Banking system.

Fandall H. Hoemke

(2) The Statewide branching system is not different

fronttflne Knit Finn inn sgxmnnn as Iwnyirds tlmygvro-

portion of its portfolio devoted to "Other

Securities.”

(3) The Limited Branching system is similar to the
Vnit Rankine system in all components.

It is concluded that the Statewide Branching system
is different from the Unit Banking system as stipulated
above and, in addition, that it devotes less of its invest—
ment portfolio to liquidity needs and United States Govern-
ment Securities (components not contrihutinr to local
economic activity) while devoting a larger proportion of
its portfolio to Loans and Discounts (a component con—
tributing to local economic activity) than does the Unit
Banking system. The Statewide Branching system, therefore,
does contrihute more significantly to economic activity

than do the alternate systems.

STATE BRANCH BANKING LAWS-—THEIR EFFECT

ON ECONOMIC ACTIVITY
By

Randall H: Hoemke

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Economics

1969

PLEASE NOTE:

Not original copy.
Some.pages have very
light type. Filmed
as received.

University Microfilms

PREFACE

Branching and its limits are of present concern to
legislators, economists, and bankers. in many states,
bills are being considered for changing the limits within
which banks are permitted to branch. lt is hoped that
this study will assist by providing clearer insight into
the relationship between the operation of the industry and
branch banking laws.

Appreciation is extended to Robert F. Lanzillotti,
Chairman of my Advisory Committee, who first introduced me
to the study of the banking industry. He has demonstrated
almost endless patience throughout the project.

To Thomas R. Saving, my major professor, 1 offer my
gratitude for his understanding, participation, and inspir-
ation not only in the research but throughout my graduate
study prwwnfiun. Witimnn.}iis enC(nu%nynmnit, it MHHJld not
have bcmni complctrmi.

My thanks to John Henderson, also a member of the
Committee, for establishing an excellence in scholarship
Which set the standard to which I aspired throughout the

undertaking.

ii

lecoenition is also extended to Leo Dohagne and Terry
Ellis, who assisted in preparing the computer input; to
Arthur Havenner and Donald Brown, who did the programming;
and to Roy Gilbert who desivned the regression equation.

\
-4

Any errors, of course, are my own reimzponsihility.

Rainiall l1. Hcmnnkc

iii

PREFACE.

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

Chapter
I. INTRODUCTION
II. THE STATE LAWS
III. REVIEWING THE LITERATURE
IV. THE THEORETICAL STRUCTURE.
V. GATHERING THE DATA
'VI. THE REGRESSION RESULTS.
XIII. EVALUATION AND CONCLUSIONS
BIIBLIOGRAPHY . . . . . .

iv

Page

ii

vii

18
6O
86
95
lU2

léu

Table
IV-1

VI-la

VI—lb

VI-lc
VI—1d
VI—2

VI—3
VI—u

VI—Sa
'VIE—Sb
VI—6a

VI—6b
VI‘7a
VI—7b
v1.8a

VI... 8b

LIST OF TABLES

Reduced Form Regression Equations

Regression Results, Equation (1):
Cash Items, Cross Section.

R.)

Regression Results, Equation (
US Gov Sec, Cross Section.

)1
Regression Results, Equation (3):
Other Sec, Cross Section

Regression Results, Equation (A):
Loan & Discounts, Cross Section.

Regression Results, All Equations,
Time Series . .

Y Variable Time Trends, A11 Equations.

Correlation of Insured and Uninsured
Banks, 1950-1966. .

Intra-system Comparison, Cross Section
Intra-system Comparison, Time Series

Significance of Regression Coefficients,
Cross Section. .

Significance of Regression Coefficients,
Time Series

Analysis of Signs, Cross Section,
Regression Coefficients

Analysis of Signs, Time Series,
Regression Coefficients

Estimated Values of Y*, Equation (1):
Cash Items (Cross Section) .

Estimated Values of Y*, Equation (2):
US Gov Sec (Cross Section) .

Page
85

118

I22

124

126

128

129
130

131

I32

134

137

Table

VI-8c

VI—8d

VI-8e

VI—9

Estimated Values of Y*, Equation (3):
Other Sec (Cross Section). .

Estimated Values of Y*, Equation (A):
Lns & Disc (Cross Section) .

Estimated Values of Y*, A11 Equations
(Time Series). . . . . .

Isolated Intercepts and Slopes,

Y Variable Time Trends, A11 Equations.

vi

Page

138

139

1A0

1A1

LIST OF FIGURES

Figure Page

II-l States classified according to State
Banking Law as regards limitation on
branching--Federal Reserve

Classification . . . . . . . . . l6
II—2 States classified according to State

Banking Law as regards limitation on

branching--as revised . . . . . . . 1?

vii

(NIAPTWHI I

IPVTWOINJCTIIDN

The market structure of commercial banking in the

United States is closely controlled by various overlapping

jurisdictions at both state and federal levels, making

banking the most regulated industry in the country. The
industry would reflect "the market structure that is most

ideal in terms of public need"1 through the interaction of

ITree—markct crmuyfi3itive Iknwnxt if these fkukwnt\vere allowed

tLQ exercise their influence.

Because the market structure of banking 's determined

buy legislative and executive decision rather than the exi—

gyencies of the market—place, discussion regarding the

nkerits of varying market structure has filled the pages of

beankers' association publications, scholarly journals, and

PEEports of hearings in the halls of Congress. The discus—

531CN1 has expanded rapidly since 1960 as a result of two

dCfVelopmentsz (l) Th3(kmumrollerafllthe Curmnuw shan-

CI‘Jrlewi tin: (HJflgflffWHiLI\flé {K)ll(ll€:3 prwevi<\u;fly {JPHIWZI:h?d lny iiis

 

 

\

.D 1Comptroller of the Currency, Annual Report, 106M,
IDepartment of the Treasury (Washington, D. C.: Government
‘I‘lJnting Office, 196“), p. l.

l

ofdlice anal (2) ‘fhrJV? was :1 rerunved_cnn1eeru11vitiiidlahniinf

anci stiuictiire {L1 nu his igo aittziiniiig ancontunic (EXprhisi«ni.

The Comptroller's determined policy to use the influence

Of his office to advance eXpansion has made timely any dis—

,_/

cussion Iwugirding tiwéimalative nmmfiix (if unit (:nui small)

banking versus branched (and large) banking. Analitical

difficultiex; are cwujounterwmi, not tin: ieast cfi‘iwiich is

that largeness and branching do not necessarily exist

together. Many branched banks are smaller than the largest

of the unit bunks.

‘Ww? New Orientation

{Eur zanti agwiirntt iiiwniciiirug }NiVC? ciiarnieci

J

Tin? arwmimeings

(over the years. Some have become more applicable while

cathers have lost their significance. For example, earlier

Eirguments against branching were offered within a techno—

lmagical framework much more primitive than that which

€?Xists today. The argument previously advanced regarding
tide capability of a branch to operate with less liquidity

huecause of its facility in transferring funds from its home

Ckffficc is not relevant. With well—developed lines of com—
”NJriicatixni incliuiing dLHMI—phone enni renkux: comgnmxné input

Strations, branches become extensions of their home office.
injfs loan capability of each branch is equal to the loan

Capability of its home office. Branches operating with

r: ‘ .. ._ 0 V _ V o
qC3dern tnmfiinriues are run; separate tunn<s an igzznmzumed by

a . . . .
Galbraith“ who writes from the Vieprint of the Canadian

nationwide branched system——a system somewhat different as

to geographical removal and remoteness from that observed

in the United States where geographical areas as regards

branching extend only to state borde s. Also, the United
States does not have large undeveloped territorial areas.
Consequently, his fine treatise does not readily apply to
the domestic market structure.

Branching in the early days of the nation was not a

necessary solution to the need for financial resources.

In the East where there was a long history of settlement,

:1 sufficient number of wealthy citizens were available to

{Drovide the necessary capitalization to organize a bank.

IBecause communication was poor, branch banking did not pre—

seent economies of scale. Branches of early banks were

aaatonomous and often had individual boards of directors.

Chae of the most successful of the early branched banks, the

IBEink of Indiana, operated thirteen branches. Each was

1C>cally organized, had its capital subscribed by its own

43t<3ckholders, and paid its own dividends subject only to

tia<3 zipgwwival (of twie :3utw~rvi.uary luaarmi of‘<3orn;rol {it tin)

home office.3

 

\
t . 2John A. Galbraith, The Economics of banking Opera—
143T1s (Montreal: McGill University Press, 1963;, pp. 105—

\

153-

 

N

Ck} 3Comptroller of the Currency, Annual Report, 1965—66
.pIQEDELrtment of the Treasury (Washington, D. C.: Government
‘ alrdting Office, 1965-1966), p. 5.

Today, the Bank of America operates branches through—
out the State of California, some of which are more than

one thousand miles from its home office. Each branch is

an extension of the home office with almost immediate com—

munication. Under these modern conditions, branching takes

on a new and different aspect.

The Literature
After some years of little or no concern given to

twink market structure by members of the academic or govern-
nuental community, Alhadeff, in 1954, published his basic

st:udy.u It drew contrasts between unit and branched banks

iri California. The work inspired a resurgence of interest

iri whether an extension of the branching privilege repre—
SEflited a problem or an opportunity to the public or the

bariker. These and other milestones in the literature will

be :Peviewed in Chapter III.

J. Fred Weston of the University of California at
[“33 .Angeles, in a statement prepared as part of the econ—
(”niCB analysis of the proposed merger of the Crocker—Angelo
Nat iCnaal Bank and the Citizens National Bank, stated the
fOllOwing after having cited seven specific characteristics
State of Cali—

0 . . .
f 'tlle developing economic structure in the

fOPnia:
\

 

Ba “David A. Alhadeff, Monopoly and Competition in
.ialfiiéglg (Berkeley: University of California Press, 1954).

The inter-connection between the seven char—
acteristics may now be considered. California
presents an impressive record of growth. Have the
characteristics of commercial banking in California
contributed to that growth? Or, has the growth
taken place in spite of the characteristics of com—
mercial banking in California? In my judgment a
persiasive case can he made that the characteristics
of commercial banking in California have made sub-
stantial contributions to the outstanding growth and
achievment of the California economy.5

The discussion on branching continues because of the

very unique structure of the banking industry in the United

States. Each state establishes its own banking system.

Some states subscribe to the branching technique while

others are determined not to allow branch.s at all.

IBecause the national banking system is restricted to the

tDranching privilege extended to state banks by state laws,

tflae national system is the center or nucleus for any con-

tiroversy regarding the relative mertis of one system over

ariother. The action of state legislatures has generated

axztivity on the part of the Controller of the national

bariking system, but his actions, in turn, have influenced

6 It is evident that even more

L)

iStéite banking SUPePViSOFS°

”Karwging and branching would take place if the regulatory

 

\

5J. Fred Weston, Economic Analysis of the Proposed

M , . ,. .
fiEZEgigr of the Crocker-Angelo dational Bank and the Litizens
L£2E3;9nal Bank, Affidavit submitted to Comptroller of the

Cur??ency, leU.

"Com etition Confusion and Com—
) )

6Almarin Phillips, ‘
(March, 106A), XIX,

m . . ..
#irn31al Banking," Journal of Finance
' (IWarCh, 196U), pp. 32—35.

6

agencies werwiinflmived completely ifiknntflma_hidustry. Crosse

states the following:

Without regulatory controls there would be a
great many more mergers and a great many more new
branctww; :stabliinnai. This trwnni PGpPOSCfMfiS'ChB
judgment of bank management and represents the
long-range interests of the stockholders. These
are the views of management both in the acquiring
banks and those that they acquired. it takes two
to tango! It is clear that banke's believe sin—

t there are efficiencies of scale, even

cerely thav
economies. And I think they are right.7

'The PanrnewOIWC for timélitudy

it was the stand taken by Weston cited ahove together

with a similar general position he adOpted in reviewing a

book8 on California banking that gave rise to an interest

in banking market structure and its role in stimulating or,

.at least, supporting economic growth.

Recognizing that the United S,ates is divided into

.fifty states each with its own individual banking organi—

Zéation and its own specific law governing branching and

rwecognizing, also, that the states subscribing to a general

Classification of (a) Statewide Branching, (b) limited

. . . . ‘
}3I“tu1chirnx, or. (c) Ihiit [hankirnt arwé sonknfliat (Xflltlglu)u3),

(

{i

x

"Bank Structure and Competition,"

Ziltrwzir°d. I). Circis:se*,
35”

lliliignal of Finance (May, 1965), XX, #3 (May, 1965): 9°

 

Review of California Banking in a

8J. Fred Weston,
Bankirnt

£1£1§§1ing Economy, ed. by Hyman P. Minsky, National
E9View, Vol. 3, #U, pp. 586-590.

1953—1952," Federal

________.

Federal Reserve Sys—

 

SEEEESIEVe Bulletin, Board of Governo*s,
EPrn (Washington, D. C.: Government Printing Office), p.
ES

9"Changes in Banking Structure,

lZLE;

study which would contribute to a resolution of the ques—

tion as to which of the three banking systems as regards

branching was best for the economy of the state would be

one which would attempt to determine which of the three

systems was most successful in meeting the demands made

g)

upon it. The theory under which the study was structured

's covered in Chapter IV.
The concltnfi(Ni that "the»;n¢hicipal impact cfi‘tflns com—

mercial ban's' operations on the economy '3 through their
lending activities rather than through their deposit func—

tion"10 led to structuring the study as an analysis of the
investment portfolio of the banking systems. Russell's

study11 of the bankers' portfolio through the medium of his

IJtility function and his conclusion that a banker responds

t;o profit motivations in maximizing his utility was crucial

111 encouraging an analysis of how banking groups reacted to

.iracome motivations. Robinson's instructional work regard—

‘irig management of bank funds was also important in

portfolio.12

approaching the analysis through the bankers'

 

 

10"The Commercial Banks and Economic Growth," The

Devine Institute of Finance, New

§3L111etin (New York: C. J.
fOrk University, December, 1960), #12, p. Cl.

. 11William R. Ressell, "Commercial Bank Portfolio
‘ACiJ ustments,” American Economic Review (May, 196U), Supple-
rn€?Y1t, pp. SUM-553.

’)
l‘Roland I. Robinson, The Management of Bank Funds

(New York: McGraw—Hill, 19627.

The importance of banking to economic growth is also

‘

asserted by the Institute of International Finance.13 It

concluded that "there is no shortage of bank capital at the

present time, that none can be envisaged in the near

future, and that inadequacy of capital resources of the

commercial banking system is not likely to he an obstacle

" It “ets fknfidi the

k)\

to long—range economic progress.

assurance that, while it may not be possible to state that

banks are specifically causal factors to economic growth,

they are, at least, very instrumental in facilitating or

assisting that growth.

A Word of Caution

A study which seeks to contrast one system or thing

ivith another cannot help but express its arguments and con—

czlusions in terms of one such entity in contrast with the

cather entity considered. As a result, authors are some-

t imes ascribed sympathetic positions when, in fact, such

“Has not their position. For example, in this study, it is

'fory easy to put the discussion in terms positive toward

br‘anch banking. This is simply because branch banking

cuss something which does

' L

e><ists——and it is difficult to dis

YICTC exist.
\

 

" The I. I. F.

13”Bank Resources and Economic Growth,
Busi-

£3;1;L].7t1n (ptwq York: hknv York [Hflflfl?rfiity, {Hohool (if
“3:::;, February 2U, 10MB), Bulletin #IQR, p. 3.

CHAPTER II

THE STATE LAWS

The banking system in the United States developed into
system composed of banks chartered by the individual

imposed a system of federally-
the

a dual

states over which has been

Further complicating understanding is

chartered banks.
federally—chartered banks

current limitation on branching by
placed on state-chartered banks.
subject to the

In

to the restrictions
of the system are

effiect, the national banks

liaws of the fifty states in the extent to which they can

ezxpand by branching.
In viewing the derivation of the industry structure

furom a national standpoint, we have actually described the

Eitmnosphere iiixwiich each stinks.individually'tnnxfl lished the
ZLEUN which defines the industry structure in that state. It

‘VCNJld have been revealing to trace the develOpmcnt of the

151%! in each individual state. However, sheer magnitude pre—

Cilleed its being undertaken.

Differences in State Laws

The state laws regarding branching are many and varied.
to tMéiiroadly

(‘6‘. . .
’rriCPPally reVIewed, those laws permit states
0

J

Id

classified into three distinct and separate groups. Some
states allow branches to be established anywhere in tne

state, subject, of course, to the approval of the State

L)

K1 as 'Wivatcu-

K

Banking Authority. These states are classific

wide Branching" States.
do not allow hanks

(‘f‘

at the other extreme, some sta es

.L/
WA”

to operate branches at all. These states are classi
"Unit Banking" states. In some of these states, limited—
service offices are allowed, but the establishment of branch

offices offering the full range of st;.'

home office of the bank is specifically p

State of Iowa has the largest number of such limi

(affices. It bad 161 such branches immediately after World

Vkir II. By the middle of the 60's, the number had eXpanded
t() 212. There are some full—service branches in this group
avtiich were in Operation at the time the state law was
ckianged. There are also some states in this group which
have allowed branches to shed as exceptions to the

14391. The total number of branches Operated in states which
IDYTDhibit branching amounted to 568 in the mid-sixties,
ineluding, of course, some states other than Iowa which
39‘3rfinitted limited-service branches. Even though brahcdih
(1CDENS exist to some extent, the states are classified as
be-’10ngjing to this group because of a :eneral practice of

J ' o o y ‘ o o :, __\ ~ _. - ,
FPOhibitinp; branching or of allowing the e tanlisnment

O f - o v u n 1 o
f‘lCes whicn offer only limited serVices.

11

There is another group of states which allow branches
on some limited bas's. These limits are set in many ways.

The most common is a stipulation that branches may be estab—

lished only within the county limits of the home office

or within a specific number of miles from the

location,
home office location, or within neighboring counties, etc.
These states are classified as "Limited Branching" states.

The classifications have remained quite constant

throughout tflua period cm7ifliis study (ljflfiiiflirouch 1966).

Changes which have taken place in the state laws during the

twenty—one post—World War ll years have not been such as to

realign the states. For example, New York relaxed its
lvranching restriction: for banks located in the New York

The State of Virginia (in 1998 and 1952)

”metropolitan area.
it formerly imposed on mer-

e>xpanded the limited provisions
In 1963, New Hampshire changed from the "Unit

Exsrs.
I361nking" to the "Statewide Branching" group. For pur—
in the Unit

IDCDSes of‘the study, however, the state was kept

£3aliking category for the entire twenty-one year period of

category for the greater

tliea study because it fell into tiis

and because the 8 ate—

Ijllnfloer of years under consideration
“’lcie branching system in that state did not develop to any

(Z)
“'thent by 1966.
Extracts of the state laws regulating financial

i 1 w . o . . o 0 N o
r), "t’l tilt lini:: zipgncrir 1r: {1 [JUilli Gilt lrni (uf i.hi3 liaiik lint iniw

12

Journal.1 In this publication, existing state laws of all
re extracted by recognized legal

states except Vermont a

firms (sources are given) expressing the essence of the law

in lay tt’"l"m£:3.

The States Classified

'fication of the states into the three groups

cited above was accomplished by the Board of Governors and
appeared in the Federal Reserve Bulletin.2
In an article which appeared in the National Banking

Review, Shull and Horvitz3 analyzed banking market structure

in accordance with the Federal Reserve Bank's classification.
South Dakota and Wisconsin, where

Ilowever, tlmgzzrrmue that

EDranching is prohibited, actually have a significant number
/

(If branches in existence—~69 in South Dakota and 162 in Wis—

19 time of their study——and that this justifies

(3C3n8lll at ti”

tljeir being classified with the states in the "Limited

IBranching" category. In like manner, they argue that Maine,

While quite generally restricting branches to neighboring

C3CH1nties of the home office location, actually, because of

\
1Banking Law Journal, Encyclopedia of Banking Laws

(Ifilrtford: Lamont—Cross, 1963).
in Commercial Banking Structure," Federal
C ° Federal Pcserve Bank,

1) 2”(Niarnges
ma'gerve Bulletin (WQShington’ D'

m), p. 1195.

3 3Bernard Shull and Paul M. Horvitz, "Branch Ranking

Eij’i the Efinhicturm-()f Competititni," Natiruvil Hankirw:lh*view
xfl:11“ch, 170”), Vol. 1, #3, pp. 301—3H1-

13

the small number of counties and the large number of excep-
in the Statewide Branching

tions to the limitation, belongs

category.

The District of Columbia

Neither the Federal Reserve Bank nor Shull and horvitz

include the District of Columbia in their classifications.

the District of Columbia are under the direct

The banks in
supervision of the Comptroller of the Currency and are,

therefore, national banks. They are allowed to branch

throughout the District area, but only within the limits of

District of Columbia is bordered on three

that area. The
is a Statewide

Itides by the State of Maryland, which
firanching state. The portion of the metropolitan area lying
iri Arlington County, Virginia, is outside the geographical
1 imits of the District itself. The State of Virginia is a

Ilinfited Branching state.
The classification of the District of Columbia pre—
fsenated a problem in logic because banks are permitted to
br‘anch District—wide, but the District of Columbia is only
(Drle‘ city. It was finally somewhat arbitrarily determined to
C:léiisify the District of Columbia in the "Limited Branching”

Ceitlegpry because (a) its branches are permitted only within
tk7E? limits of the District, (b) banks are not allowed to

b . .
Par'lCh even throughout the metropolitan area of the City,

'ar 0 o
1‘1 ((3) it is "X{H?Ctn?(l tiiat tine g2r€n1t6?r <30rlterfiiA: (if 1Jh(‘
:3 t ' o . .

LJ'CJqJ W] ll hr: rlrfiiwri t>ei;w(>er1 tiie Uriit; kirn<irig Vikiini {in(l tiie

 

f3tainewixie ii? nichiiig strotu).

these two groups structured as

1”

it was deemed desirable to keep

closely as possible to the

structure used in other studies. This decision is not con-

sidered to be of crucial importance as the presence of the

District of Columbia in any one group would not be distort—

ing because of the relative size of the District to the

total of the entire number of states making up a group.

The Classification Revised

The states have been rcgrouped in accordance with the

classification used by economists Shull and Horvitz with the

addition of the District of Columbia to the Limited

EBranching group, as follows:

 

St3atewide Branching

‘

lilzaska
Axeizona
CEilifornia
(3CHanecticut
Delaware
”EUNaii
ICialio
Maine
{VIa Py land
PREVada
”O rt h Carolina
be£{on
§FNO<fle Island
QCDthfl Carolina
Ut ah
SQEITUNDnt
Jaisklington

\

Limited Branching

Al ab ama
District of Columbia
Georgia
lndiana
Kentucky
Louisiana
Massachusetts
Michigan
Mississippi
New Jersey
New Mexico
New York

Ohio
Pennsylvania
South Dakota
Tennessee
Virginia
Wisconsin

ihiit Pkwakirn;

Arkansas
Colorado
Florida
Illinois
Iowa
Kansas
Minnesota
Missouri
FMnitana
Nebraska
New
Hampshire
Pharth Ikn<ota
Oklahoma
Texas
West
Virminia
Wyoming

 

Line maps of the United States depicting these cate—
gories are included herein. The first (Figure ll—l ) sets
forfifli the ififideral.lhéserwma Boaiwi classil‘haaticui and tlk>
second (Pagan?) 11-? ) sets forth the classification out—
lined immediately above.

Attention is invited to the extent to which the groups
are cohesive entities--that the states which are of the same
classification are, generally, contiguous. The states mak—
ing up the Statewide Branching group are found primarily in
the West, with a scattering on the eastern seaboard. The
states allowing no branching are primarily in the mid-
western section of the country. Only Florida, West Vir—
ginia, and New Hampshire are separated from the main group.
The cohesiveness of the Limited Branching Group is violated
only by the removal of New Mexico and South Dakota from the
main body of the states lying in the eastern portion of the
mid—west to the east of the Mississippi River.

Because banking markets are primarily local in that
the greater proportion of a bank's business comes from its
immediately surrounding area, the fact that these groups
are contiguous makes a comparison of banking systems by
grouping states according to their branching laws somewhat
more meaningful than it would be if the three s stems were
thoroughly and totally homogenized. Economic analysis of
the effects of varied branching prohibition is aided by the

fiCOVfiTflUliCQl (H;H;ribution.

.QoeprHMHmmmHo m>gmmmm Hmpmemmnnmcfigocmpn so COHpmpHEHH mesmmmg
mm seq mcflxcmm mpmpm 0p wsflcpooom emanammmao mmpmpm "HIHH mmzwfim

 

l6

 

 

 

II

 

x wcfixcmm use: u A V
. ./.I. 9 wcficocmam smudge u A v
_ . frufiw wcficocwum mcfismpmum u v

 

J

 

(8.2.Q_l_

 

 

 

 

    

. ~9..u!.‘.l
fix / I!

 

 

.eomfi>mh maulmsficocmpn co coapmpfiefifi mehmwwh
we add mafixzmm opmum on mcfiesooow UmHMfimmmao mmpmpm "NIHH mLSMfim

 

l7

 

     

III

a ./)\.,
:c <

V ecfixcmm ss::

mcflcoszm copes“;

wCHLech: oUH3;»spu

 

 

 

 

 

 

 

 

 

CHAPTER III
REVIEWING THE LITERATURE

Bank market structure as regards the extent to which
banks are permitted to branch has been under close scrutiny
of late in an attempt to determine which system is best for
those (hnnanCHrhT anti Hipplyirn: bank'iflarvices. 'The aiwn1 of
investigation is difficult, not because data are unavail—
able——the regulatory agencies have seen to that——but
because the data made available do not make for easy answers
to most evident questions.

The most frustrating difficulty encountered in review-
ing the literature concerning banking market structure is
the confusing of economies of scale and of structure. It is
frequently stated that branch banking is more efficient than
uriit ljarflilrUj EUld it is oineri irnplyieti trust a liranicfii tuinL; czn1
be operated at less eXpense than a unit bank of the same
size. In the Horvitz study,1 a specific distinction is made
between economies brought by becoming larger versus the econ—

omies, if any, brought by expanding a multi—plant operation.

 

1Paul M. Horvitz, Economies of Scale in Banking,"
Egivate Financial Institutions, Commission on Money and
Credit (Englewood Cliffs, New Jersey: Prentice—Hall,

1963), p- 8.
18

19

This distinction is not always clear and one finds that
"big" and "branched" become synonyms or, at least, that the
same results are attributed to "bigness" as are attributed
to "branching." Some of this is understandable as expansion
in the banking industry, because of the local nature of its
market, is often accomplished through branching and many
large banks are branched.2 A serious error is committed by
taking the next and seemingly logical step that everything
said about large banking being able to operate at less cost
also assumes these cost savings are attributable to expan-
sion by branching. Actually, deterioration of control and
increased occupancy costs in a multi—plant operation add

to Operating cost.

Another difficulty encountered in the study of banking
markets is that the regulatory agencies, in asserting their
influence, have created a market situation to which no
regularly accepted economic market models apply.

Concentration statistics are effective in analyzing
markets. Because banking markets are local, however, their
usefulness is weakened. An area of low concentration may be
experiencing many local markets with monopolistic suppliers.
A low concentration statistic does not indicate the absence

of monOpoly influence as could be presumed in other industries.

 

2Irving Schweiger and John S. McGee, "Chicago Banking,
the Structure and Performance of Banks and Related Financial
Institutions in Chicago and Other Areas," Journal of Busi—
ness (Chicago, Illinois: University of Chicago, Graduate
School of Business, 1961).

 

20

This difficulty is recognized in the Edwards study3 when the
author states:

'Large metrwuxxlitan areas :fiunild represewfla:umne
of the most competitive banking markets in the nation.
They offer more banking alternatives and more aggres—
sive management than do smaller communities . . .
communities with less than 50,000 people are not
likely to have more than three or four hanks. This
study examines areas with no less than 20,000 people.
The presence of monopolistic practices in metropol-
itan markets would indicate greater departures from
competitive conditions in smaller population centers.

Further difficulty is indicated by Schweiger and

1!

McGee when they find that markets with concentrated banking
charge the highest rates for new car loans and cash install-
ment loans. The rates charged varied inversely with the
number of banks in each town.

The problem is revealed to be of wider ramification
when it is realized that a market with only one supplier is
as much a monopoly in a statewide branching state as it is
in a unit banking state. To further complicate the argu-
ment, the market serviced by one bank in a state allowing
branching would enjoy more complete services and policies
determined in more competitive areas if that monopoly bank
was a branch of a larger bank with its home office located

in a competitive metropolitan center. The local market

served by such a monopoly bank would be more competiively

 

5Franklin H. Edwards, "Concentration and Competition
in Commercial Banking: A Statistical Study," Research
E<e£gyrt [10. 26 (Bruytoriz lfledrnwil Fies(¢rvréliarfl< olrfsostiin,
lgbfl).

11.. . .
Schweiger and McGee, on. Cit.

--L——__..__

 

oriented than if it was served by a unit bank in which all
decisions were made in the local market and not in consider-
ation of the competitive structure of the home office

location.

An Overview

 

Phillips,5 when recognizing the desireability of com—
petition hecause of its effects on market performance and
the allocation of resources, admits that the individual
bankers are not persuaded by academic arguments that indus—
try competition is lacking. As in other industries, bankers
confuse market rivalry with competition and insist that even
in a monopoly market competition exists because of bankers'
concern with profits, market shares (with banks in other
towns and other financial institutions), and growth. As
evidence of the lack of competition in the banking industry,
he cites:

(1) the low failure rate and general stability
in the structure of the industry, (2) the persist—
ence of firms of less than optimal scale, and (3)
price performance which is inconsistent with_the
results of multi—lateral market competition.5

He reviews the highly skewed distribution existing in
the banking industry: of the 12,937 insured commercial

banks at the end of 1963, 7,370 had total deposits of less

than five million dollars. At the other extreme, only 31H

 

5Phillips, op. cit., pp. 32—u5.
6;bid., p. 3M.

banks had deposits of more than one hundred million dollars
and sixty—four had deposits of over five hundred million
dollars.7 If competition existed, large banks would not
allow smaller banks to exist in their markets——the smaller
banks would be absorbed. Advantages of s'ze are not
CXplOlth as they would be if the market were more freely
competitive.

Alhadeff8 suggests that economies of scale are sub-
stantial up to at least five million dollars in deposits,
then reach a plateau until the very largest banks are
reached. If economies reach to five million dollars in
deposits, fifty-seven percent of all commercial banks are
of less than optimal size. Of the 12,933 insured commercial
banks in 1962, 7,370 had deposits of less than five million
dollars, 7,705 Operated in one—bank towns——towns where the
market was not of sufficient size to allow the attainment
of optimum size. This argument is not entirely convincing
because 1,800 banks operated in two-bank towns and 362 banks
operated in three—bank towns. Many of these banks are
smaller than optimum size. Competitive forces, if they
existed, would have caused these banks to join torether to

attain opitimum size.

 

7Federal Deposit Insurance Corporation, Annual Report,
1962 (Washington, D. C.), p. 13A.

8

 

Alhadeff, ob. cit., p. 106.

23

Price behaviour in banking is especially difficult to
analyze because of lack of published information. Informa—
tion which is published is known not to apply to all sit-
uations. For example, an advertised rate will apply only
to a very small segment of the total credit risk. Banks,
in contrast to suppliers in other industries, charge dif— {—m
ferent rates to different customers for the same type and
size of loan. Some of the difference is attributed to var-

iation in risk, type of collateral, time of repayment,

 

amount of compensating balances, etc. But these factors do _J
not account for all of the difference. Differences arise

because the customers and the bank are in bilateral rather

than in openly competitive bargaining positions in the nego—

tiating process for loans. The c stomer with the greater

number of alternatives (including not borrowing at all

because of his excellent financial situation) will tend to

get loan rates which are lower than will a customer with

fewer or no alternatives. The bank operates as a discrim-

inating monopolist. The willingness of the customer to

"shOp" around for better loan rates is precluded by the

nature of the transaction. The borrower, because he finds

the role of "supplicant" unpleasant and because he does not

readily understand the intricacies of the banking process, 1
allows the banker to act as a discriminating monopolist even 9
in situations where the classical definitions of monopoly do

not exist. Banks also discourage customer "shopping" for

better price by stressing the advantages of "one-stop"
banking. A strong influence restricting competitive price
determination is the desire by customers to minimize the
disutility of a situation which, at best, seems to interfere
with their individual sense of propriety as regards reveal-
ing information they deem to be confidential.

One could eXpect within the classical analytical
framework to observe that interest rates would be lower and
possess less variance in markets containine many suppliers
than in markets having few or one supplier. Unit banks
might be eXpected to offer more attractive rates than
branched systems or holding company banks, but there is
little to support such conclusions. The Edwards study0
suggested that market structure variables are not associated
with interest rates in any significant way.

After reviewing the influence of public regulation and
the private rationalization of competition, Phillips sug-
gests that commercial bank performance could be improved
primarily by changing its market organization through the
medium of changing the intent and purpose of the regulatory
agencies. Most important from the standpoint of this study,
he recommends that the current "needs and convenience" cri—
terion be abandoned in favor of a less restrictive basis,

that arbitrary limitations on de novo branching and

 

9Edwards, op. cit., pp. 71-76.

25

branching by merger be relaxed, and also that the prohibi—
tion against branching over state lines he removed. He also
recommends that restrictions on interest rates paid on
deposits be eliminated.

Phillips recognizes that his suggestions for change
would establish a national banking industry and spell the
end of the dual banking system as it now exists. He does
not expect that such changes will be accomplished very

rapidly.

A "Seminal” Study10

 

The work that formed a basis or beginning for many
others was the one by David Alhadeff of the University of
California.11 In 195“, Alhadeff published a study of Cal-
ifornia banking in which he directed his attention pri—
marily to differences in load factor (the percent of Total
Assets represented by notes and investments) in banks of
varying size and of varying organizational structure as
regards branching. Prior to the Alhadeff study:

claims and counter-claims about relative operating
efficiency usually had been supported by descriptive
reasoning. For example, opponents of branch banking
claim that branch banks have higher operating costs

than unit banks that possess the same output because
it is costly to coordinate the operations of more

 

Stuart I. Greenbaum, "Competition and Efficiency in
the Banking Industry—~Empirical Research and its Policy
Implications," Journal of Political Economy, Supplement
(August, 1967), p. U56.

 

llAlhadeff, o . cit.

than one banking office. Branch banking proponents
counter with the assertion that branch banks tend to
employ more 'progressive' branch managers who Oper—
ate their branches more efficiently than the depart-
ment managers of unit banks operate their depart—
ments. Economies of scale Operations are claimed
for branch banks because they usually are larger
than unit banks. That economies of scale do exist
has been generally assumed rather than demon-
strated.12
Alhadeff's study was the first to make extensive use
of empirical data.13 He concluded that the "load factor"
showed:
no clear tendency to vary with the size of unit bank
except that the smallest unit banks invariably have
the lowest load factorslu
In contrast, the load factor tended to:
vary directly with the size of a branched bank with
the pattern of variation most sharply defined in the
extremes of the different sizes of branched banks.
Furthermore, the load factor for the average of the
branched banks was higher in every year than the load factor
of the largest unit banks. Even branched banks with the
lowest load factor among all branched banks had a higher
load factor than the average of the group of the largest
unit banks. Alhadeff analyzed the differences which he

observed by recognizing an important element which affects

the load factor-—the ratio of time deposits to total

 

2George J. Benston, "Branch Banking and Economies of
Scale," Journal of Finance (May, 1965), XX, #2, p. 312.

 

13Horvitz, o . cit., p. 10.

l“Alhadeff, op. cit., p. 57.
15Ibid., p. 57.

27

deposits. Again, although no strong pattern exists among
unit banks between size of bank and the ratio of time
deposit to total deposits, the smallest category of unit
banks appeared to have comparitively low time deposit
ratios. This relationship is not fortuitous but causal as
the ratio of time deposits to total deposits affects the r—m
load factor in either or both of two ways. Because time
deposits are subject to lower reserve requirements than
demand deposits, a bank with high time deposit to total

deposit ratio is able to "produce" more loans than another

 

H'.q< "

i

bank without that high ratio. Also, high time deposits

imply less pressure on the bank for liquidity.16
Alhadeff goes on to state that bank size can be

expected to affect liquidity requirements in somewhat the

same manner. A large bank with a large number of depos-

itors will not be unstabilized if even a large depositor

makes a substantial withdrawal. A small bank, on the other

hand, is more subject to drastic variation in liquidity

demand because the accounts do not vary in size with the

size of the bank. Another factor cited as contributing to

the difference in load factors is the general practice in

larger banks of carefully investing excess reserves in the

Open market. Another is the necessity of small banks to I

carry correspondent balances with larger banks in metropol— {

itan centers on a continuing basis. These balances, of

16Ibid., pp. 58-59.

course, are required in part by having to provide for the
unstable liquidity demand mentioned above. As a result, the
author concludes that smaller unit banks actually carry
higher balances than do larger banks and the demand for
liquidity is sufficient to overcome the generally higher
reserve requirements to which larger banks are subjected.17
The load factor of branched banks is aided because
inter—branch mobility of funds reduces liquidity require-
ments established by high time deposit ratios. Internal
mobility permits a high turnover of resources as excessive
demand in one area is met by shifting funds from areas
where demand is not so strong. Thus, it is possible for
branched banks to have a higher average load factor than
the ayerage of unit banks even when the time deposit ratios
of the two groups is very similar. In fact, branch banks
reflect higher load factors than unit banks even in those
cases where banks have lower time deposit ratios. Thus,
the inter—branch mobility of funds is instrumental in deter—

18

mining load factor. The proportion of a bank's portfolio
devoted to loans and investments is a function both of size
of bank and of bank structure, but the influence of bank
structure is more important. He concludes:

In the net interplay of these factors, branch

banks emerge with a clear superiority in load

17lbid., p. 16.

 

18Ibid., p. 60.

2?

factor over unit banks. Quite specifically, this

conclusion means that, with given resources, branch

banks have an inherent superiority over unit banks

in their ability to produce all kinds of bank

credit.19

Alhadeff finds that unit costs for different size

banks decline fairly sharply in the lower ranges (up to
approximately five million dollars in total deposits) and
then remain quite constant up to a size of about fifty
million dollars, then decline again in the range of the

’7)
’0 Greenbaum21 argues that Alhadeff used

larger banks.
loans and investments s a measure of bank output, relating
the ratio of operating expenses to earning assets to still
another variable——Total Deposits——in order to derive his
cost curves. Greenbaum recognizes that the literature is
replete with discussions as to the meas ring of bank output.
The use of earning assets to measure output implies that
lending is the primary productive activity of commercial
banking. This, of course, is what Alhadeff argued when he

0 o 9’)
established the framework for his study.e¢

Greenbaum con—
tends that the use of earning assets as an output measure
implies that all forms of bank credit are perfect substi—

tutes to the community. This is in violation of the study

 

19Ibid., p. 60.

201bid., p. 106.

 

21Greenbaum, loc. cit., p. 466.

 

22Alhadeff, loc. cit., pp. 0—19.

/

 

30

performed by him in 1961423 and the study by Hester and
Zoellner in 1966.214 Greenbaum states that the reasons for
using earning assets as a stock variable to represent flow
concepts, while not uncommon, should be explicit and reason—
ably compelling as flow measures are readily available.

He also suggests that current operating earnings would be
preferable as measures of bank output.25 In answer to those
who would argue "that earnings assets, deposits, total
assets, and current Operating earnings are so highly corre—
lated that the issue is more apparent than real," he found
adjusted coefficients of determination which ranged from
0.015 to 0.701 for these items. Further, the functional
relationships were linear, quadratic, or cubic depending on

the measure used.26

 

23Stuart 1. Greenbaum, ”Bank Structure and Costs: A
Statistical Study of the Cost-Output Relationship in Commer—
cial Banking," Ph.D. Dissertation (Baltimore: Johns Hopkins
University, 196“).

2“Donald D. Hester and John F. Zoellner, ”The Relation
Between Bank Portfolios and Earnings: An Econometric Anal-
ysis," Review of Economics and Statistics (November, 1966),
XLVIII, pp. 372—386.

 

25Greenbaum, ”Competition andEfficiency in the
Banking Industry,” 100. cit., p. 467.

 

26Greenbaum, "Banking Structure ani Costs," op. cit.,
p. 247.

A Similar Rut Rroader Studv

 

In 1963, Paul M. Horvitz27 completed a study sin 1'

4r

in form to that of Alhadeff, using i;a covering a bri

geographical area. His results regtrdin; “cram ies of

are very similar to Aliadeff. In contrast to Al lideff’:
confident case for the superiority of branched hanks, 'ur—
vitz concludes the t Operating eXpenses of branched hawk
greater than for unit ban's. As h3 stat it:

This means, surprisingly enough, that four fif-
teen million dollar unit banks can be operated at
lower cos t than a sixty million dollar branched

q
bank.¢8
Even the mzilles size gH10‘p of unit banks had lb or

J

"3 {1,5 m

'f‘ ‘ ‘. if)
(. J. i

cost, on the average,than the largest size branched hanks.

This was true in all cases except where banks had 1233

t;iif1r1

twenty- five percent of total deposits in derknd deposits.

The Gramley Research

 

q

‘ ‘ ' " Q
Lyle Gramley, in his well—known stucy,6' used rev

1 "
C l x \}

sion analysis on data secured from Earnings and Conditian

Reports to which he had access through the Federal Reserve

Bank of Kansas City. He researched 270 unit banks in
Kansas City District and averaged the annual data for

years 1956 through 1959. His regressions used as a

 

27Horvitz, o . cit., pp. 1-5U.

281bid., p. 37.

f. .L). If“

29Lyle Gramley, Study of Scale E ooncmies in Ranking

 

(Kansas City: Federal Reserve Bank, 1962).

dependent variable the ratio of current operating expenses
to total assets while his main independent variable was
total assets. Gramley considers the problem of defining an
output measure very extensively and, in order to avoid this
problem, he restricts his results. He states the following
as part of his introduction:
This study, then, will have to be content with
a discussion of scale economies that is narrower in
its scope and its implications. Since it is con—
cerned with costs of operation and earnings at banks
small and large, it necessarily deals with the ques—
tion of efficiency in the performance of banking
functions and its relation to bank size. The concept
of efficiency that is dealt with, however, is a nar—
row one. The study is concerned with efficiency at
individual banks, and does not seek to determine
whether, from the standpoint of achieving maximum
social efficiency, the bank system should be composed
of small or large units.30
Because of this limitation, Gramley's results are of
little application to the purpose and intent of this study
as we intend here to throw some light on the very problem
which he has eliminated—-to aid in determining which banking
system as delineated by its varying capacity to branch is
best for the economy of the state in which it exists.
It will be sufficient for the record to observe that

Gramley's study of member banks in the Tenth Federal Reserve

District, most of which were unit banks, "did not lead to

 

3Olbid., p. U.

33

definitive conclusions supported by precise qualitative
measLU%as."31
The study revealed that pre-tax rates of return on
bank assets rise with increasing bank size over the full
range of bank sizes present in the sample. The author con-
cludes with the following:
It is evident, nevertheless, that rates of
return on invested capital, both before and after

taxes, rise quite sharply with increasing bank
size among Tenth District Member Banks.34

Output Redefined

 

Greenbaum33 attempted a study somewhat similar to that
done by Gramley except that he did not avoid the "output"
definitional problem. He divided output into two compo—
nents: (a) lending, and (b) all other. Lending output was
defined as the gross yield—weighted sum of the diverse
earnings assets in each bank's portfolio. Experimentation
indicated that sixteen classifications of earning assets
provided ample disaggregation and satisfactorily homogeneous
earnings—asset groupings.3u He argues that his output meas-

ures recognize (l) the multi—product nature of commercial

 

31£bid., p. 58.
321bid., p. 59.
33Greenbaum, "Banking Structure and Costs," op. cit.

3“Greenbaum, "Competition and Efficiency in the
Banking Industry," loc. cit., p. U69-

3U

bank output, (2) interbank price differences resulting from
imperfect markets, and (3) production as a flow process.35
Greenbaum's study investigates the costs of banking
and looks especially at costs in groups of banks differing
in their market structure as regards branching. The study
covers Federal Reserve member banks in two Federal Reserve
Districts. In the Tenth District, only New Mexico allows
branching. In the Fifth District, West Virginia was the
only state that prohibited branching. In the Tenth District
all but four of the 7&5 sample banks were classified unit
banks, while in the Fifth District, of the “13 sample banks,
131 operated branches. The technique was to determine the
difference, if any, in the earnings banks experienced on
their assets as the quantity of funds invested in the spe-
cific type of assets varied. Current Operating expense was
used as a measure of cost. Cost was introduced into a
regression equation by setting operating expense as a per—
cent of Total Assets equal to a series of variables meas—
uring output (type of loans) as a percent of total assets.
In the Kansas City District (where unit banking is
prevelent) the study revealed a wide range of size of firm.
The optimum size of firm was approximately 700 times as
large as the smallest observed firm. Expansion from the

smallest to the optimum size brought with it an estimated

¥

 

LU
\fl

thirty-three percent reduction in average cost. Moreover,
the elasticity of the cost function was —0.00l at the min-
imum observed output--i.e., a one hundred percent increase
in output would bring about a 0.1 percent reduction in
average cost. Thus, there were only very modest economies
of scale over a large range of output. The author concluded
that it was not to be expected that cost considerations
would bring either output expansion or contraction.

In the Richmond District, the banks were divided into
unit and branch banking groups. In the branched group, a
variable was introduced relating to the number of branches
the banks operated. In these tests, holding either the
average plant size or the output constant and allowing the
other to increase, resulted in lower average cost. In
branch banking, therefore, economies could be derived that
are impossible in a unit banking environment.

When unit banks are merged into multi-office
firms, costs will be reduced, even if the output of
each office is left unchanged.36

The unit banking sub—sample of the Richmond District
revealed a downward sloping cost curve suggesting that
where branch banking is permitted, rising average costs may
be forestalled by acquiring additional plants instead of

increasing the output of existing facilities.

 

36mm. , p. 328.

36

In effect, the institution of unit banking
imposes a low-level optimum output rate upon firms
in the banking industry.37
The cost performance of branched systems were compared

with collections of unit banks having equal total output and
number of offices. The results indicated that branch banks
operated at lower average costs at all comparable levels of
output than do the collections of unit banks. Surprisingly,
the same result was not obtained when Richmond District
banks were compared with similar groups of Kansas City Dis—

trict Banks.

Output Defined as Number of Accounts

38

George Benston performed an interesting study of

s
bank cos s in which each bank was divided into six separate
units: (1) Demand Deposits, (2) Time Deposits, (3) Mort-
gage Loans, (A) Installment Loans, (5) Business Loans, and
(6) Securities. Each type of output is measured in terms
of number of loans or deposit accounts with the average
total value of the transactions included among the para—
meters of the cost functions.

A confusing aspect of Benston's study is that both

liabilities of the bank (Deposits, for example) and Assets

 

37Ibid., p. 328.

38George J. Benston, "Economies of Scale and Marginal
Costs in Banking Operations," National Banking Review, II
(June, 1965), 507—5U9.

 

37

of the bank (Loans and Securities) are considered measures
of output. The assumption of a liability such as deposits
would give rise to the production of output in the manner
of loans or the purchasing of securities. A portion of the
deposit balance exists because loans were produced--but the
deposit balance so generated is transferred to other banks
as loan proceeds are expended. The purchasing of deposits
represents acquisition of factor input.

Benston provides for the evaluation of cost structures
in banks which are branched through the use of dummy vari-
ables. He determines through regression technique that
slight economies of scale exist as regards the level of cost
experienced as the number of demand deposit accounts are
increased. The cost curve as regards time deposits is very
flat as the geometric mean number of accounts is reached.
The mortgage loan marginal cost curve slopes upward at a
decreasing rate; the installment loan marginal cost curve
is consistently "U" s aped; and the business loan marginal
cost curve is rather sharply downward sloping. The econ-
omies of scale in the handling of securities were not as
great as expected. The elasticities of the cost curves were
not as low as were expected. Apparently there are some
costs that rise as the size of the portfolio increases.

In another article, Benston reviews his results as

regards the extent to which economies of scale exist in

38

branched banks. It is here that he attempts to answer the
following questions:

(1) Is a unit bank of a given size more effi—
cient than a branch bank of the same size, ceteris
parabus, and, if so, how much are these costs and
in which specific bank services are they found?

(2) If banks should increase their demand deposits,
installment loans, or other banking services by,
say, ten percent, by what percentage will its costs
increase, irrespective of its being a unit or a
branched bank? (3) If branch banks, per 33, are
more eXpensive to operate than unit banks, but if
there are economies of scale, will merging several
unit banks into one branched bank result in lower
or higher operating costs, and in which banking
service will these net costs or cost savings be
found?39

 

len ton recognises that he cannot draw conclusions
that may be applied to the entire banking system because the
population from which his samples are drawn was limited to
medium—sized banks. The study showed that branch banking
does entail additional costs that are not offset by econ—
omies of scale:
Branch banks with one or two branches do not
have costs that are very different from those of
unit banks; approximately half of the additional
costs are due to occupancy expenses; and the mar-
ginal cost of additional branches beyond eight is
probably very small.”0
His analysis of occupancy exnenses indicates that it

is not evident that total costs expended on occupancy by a

group of unit banks would be greater than those expended by

 

39Benston, ”Branch Banking and Economies of Scale,"
100. cit., p. 313-

uOIbid., p. 331.

39

a branched bank with the same number of units. Contrary to
other studies, he finds that branching can be accomplished
only at higher costs. He states:
It does not necessarily follow from these
findings that charges would be higher to depositors
and borrowers of branch banks. Other differences
between branch and unit banks, such as ability and
desire to take risks, may offset or reinforce dif—
ferences in operating expense.”1

Thus, he suggests that additional research be con-

ducted to isolate further some of these differences.

An Analysis of Expense Structure

 

Using tabular statistical examination technique, Hor—
vitzu2 achieves results quite similar to those of Alhadeff.
He evaluated aggregate data published by Federal Deposit
Insurance Corporation and found, when looking at expense
classifications as a percent of loans and investments, that
expenses tended to decline as bank size increased. He con—
ducted an analysis of the expense structures which allowed
comparison among the largest banks in the country (New York
Federal Reserve District), the largest unit banks (Chicago
Federal Reserve District), the smallest unit banks (Kansas
City Federal Reserve District), and the largest branched
systems in the country (San Francisco Federal Reserve Dis—

trict).

¥

ulIbid., p. 331.

 

u2Horvitz, op. cit.

UO

One conclusion of importance to this study is tnat a
clear pattern results:
when branch and unit banks were compared. For every
size and time deposit ratio the branch banks had
higher loan ratios.u3
This was in accordance with the findings of Cchweiger
and McGee who also found a strong tendency for branched
banks to make more loans than unit hanks. Forvitz found,
further, that branched banks had more mortgage loans than
unit banks of the same size and time deposit ratio. T
results were again in accordance with the Cchweiger and
McGee study which found that:
lar e branched banks . . . relative to assets
3 3
. . . lend about five times as much on home mort-

gages and one and one-half times as much on con-
sumer installment loans as very large unit banks.

uu

Total earnings expressed as a percent of loans and
investments showed a substantial advantage for branched
banks over unit banks—-something which could be expected
from the high loan ratios and the higher preportion of high
income installment loans experienced by branched banks. in:
only were branch bank costs higher but there was also con—
siderable spread between branch and unit banking costs. as
the author states:

This means, surprisingly enough, that four

fifteen million dollar unit banks can he operited

u3lbid., p. 3M.
UM

Schweiger and McGee, loc. cit., p. 226.

Ml

at a lower cost than a sixty million dollar
branched bank.

In summary, Horvitz states:

While it seems clear that economies of scale
in banking do exist, they are relatively small and
are actually less important than the diseconomies
of branch structure. These diseconomies are such
that the per unit costs of a large branched bank
are as high or higher than the costs of a small
unit bank. This means that branching cannot be A”
solely supported on the basis of cost aivantages. 0

He states, further, that the argument that branch
banking must inevitably lead to monopoly is untenable

because a small bank is more efficient in making small

loans than is a large bank. In competition with a large

branched bank, the small bank finds some disadvantages but

has some advantages. He also states that the argument that

branched banks would mean more facilities for some towns
not dependent on cost advantages for branched banks. He

continues:

Branches can exist in communities where unit
banks cannot operate profitably even without any
cost advantages. A community which provides a large
deposit volume but little or rm) loan demand may
prove a suitable location for a branched but not for
a unit bank. The same is true of a communitv with a
large loan demand but little deposit volume.h7

. "
1 b

_..

Branching is a welcome means of expanding onc's oper—

ation when that branching will take place into markets which

 

uSHorvitz, op. cit., p. 37.
“51bid., p. 52.

u7lbid., p. 52.

U2

are not evenly balanced as to the availability of deposit
volume in comparison with the volume of loan demand which
is available. These imbalances can be overcome by the
inter—relationship which exists in a branched operation——
an advantage this form or structure provides over the more
rigidly separate structure of unit hanking.

In conclusion, Horvitz states that nation—wide
branched banks are very unlikely to be seen in the near
future not only because of the general suspicion of the cit—
izenry of bigness in this field and the opposition it would
face as a violation of states' rights but because the small
unit bank could compete on nearly equal terms with the large
bank. Large banks have no advantage from the cost side and
this would preclude the unit banks' being superceded by

branches of a large bank.

An Expansion of the Benston Approach

 

Y

In a very extensive study, Frederick w. Bell and Neil

B. Murphyl48

examine costs in commercial banking and econ-
omies of scale inherent in expansion whether within existinr
facilities or by acquiring additional banking offices. The

authors criticize Alhadeff, Horvitz, Greenbaum, and Gramley

because they used as a basic independent variable a stock

 

148Frederick W. Bell and Neil B. Murphy, gosts in
Commercial Banking, Research Report #Ul (Boston: Federal
Reserve Bank, 1968).

 

“3

variable which was related to a flow of costs. This is
inconsistent with production theory. Other criticisms were
made, especially of the Gramley and Greenbaum studies. The
Benston study was thought to be the most fruitful to the
measurement of bank costs. The authors see their study as
an expansion on the Benston approach when they state:

This study on returns to scale in commercial

banking seeks to extend and enhance the Benstonian
approach to the measurement of bank costs.u9

tionship in which the authors stipulate as a measure of out-
put the number of accounts serviced by the bank under con—
sideration. Inputs to this output were Total Direct Costs,
Labor Input, Capital Input, Materials Input, Wage Rate,
Rental Rate on Capital, and Price of Materials. Branching
was introduced with dummy variables allowing distinction to
be made by number of branches operated. Variables intro-
duced the consideration that more risky loans were serviced
at higher cost than were business loans.

Using ordinary least squares technique after logar-
ithmic transformation, increasing returns to scale were
found for many bank functions:

Demand deposits, business and real estate
loans, securities, trust departments, and busi-
ness development showed economies of scale that
are statistically different from constant returns

to scale at the five percent level. All functions
except the safe deposit and time deposit functions

 

“91mm, p. 10.

 

AA
exhibited point estimates below unit for the scale
parameter.50
Economies of scale were found for expansion of size
of firm——for expansion within existing facilities. No econ—
omies of scale for expansion through increasing the number
of branches were found. As a result, the authors stated:

Expansion via merging and branching may sub-
stantially offset any economies which might accrue.51

The authors also investigated the varying productivity
of labor as scale increased and found that the dis-aggrega-
tion of the cost function revealed a persistent tendency to
economies with increases in scale. As regards branching,
some of the economizing of labor was found to be impossible
because of the necessity to duplicate labor services in the
various branch locations. The additional cost of branching
is mainly attributable to this problem and to the necessity
for additional Capital investment required in a branched
operation. In fact, in a separate part of the study, the
authors determined that:

if bank expansion comes largely through branching,
economies of scale may be offset by increased
branching costs. Hence, firm expansion may take
place under constant returns to scale.52

The study attempts to evaluate the impact of bank

structure on inter- and intra-regional variation in pro-

duction costs:

 

501bid., p. U7.
511bid., p. 68.

52Ibid., p. 181.

“5

It was demonstrated that a substantial inter—

regional dispersion in production costs exists due

to bank size, organization structure and relative

wage levels. This may produce differential service

charges between market areas. Intra—regional var—

iation in cost was much less.53

The results give some justification for a regulatory

policy in which new bank entry is encouraged to fOStGP com—
petition. Also, it is suggested that the recent growth in

branched systems of branched banks acquiring unit banks is

not so much to get efficiency but to achieve monopoly power.

The Effects of Concentration

 

A study quite different from those reviewed as to
intent and content and which has implications for a study
of branch banking is one by Franklin R. Edwards.54

Edwards attempted to determine whether or not the
existence of monopoly power in commercial banking had a sig—
nificant effect on mortgage rates. In determining the
extent of monopoly power, he included in his analysis not
only the large statewide branched banks but also bank
groups——banks which are members of a holding company regis—
tered persuant to the Bank Holding Company Act of 1956.

Especially interesting from the standpoint of this
study is the part of the paper which is concerned with bank
organization. To accomplish this, a dummy variable indi-
cating the type of market organization is added to the

regression equation. The author concluded that the form of

 

53Ibid., p. 215.

51‘Edward-.-_‘., op. cit.

M6

bank organization had no significant effect on differences
in rates charged on loans.

As to an evaluation of the economic effect of mergers,
Edwards concludes that many other relevant factors must be
considered. At the heart of the merger question is the con—
troversy over branch or unit banking. He states that branch
banking provides more banking offices and a wider range of
banking services than does unit banking. Branch banking
also offers a lesser risk of failure because it can spread
the risk over a widely-spread geographical area. Branch
banking, because it is big banking, is solving the manage—
ment succession problems faced by small unit banks. It
also enhances credit mobility in that branch banking permits
quick and convenient transfers from surplus areas to deficit
areas.

But, in spite of all these advantages, branch banking,
Edwards reminds the reader, does not offer the personal and

informal services that unit banking offers.

Interests Served by Merger

 

A study which has implications for branch banking
because of its recommendations for changes in regulatory
policy is the one by Kallman Cohen and Samuel Reid.55 The

authors set the framework for their concern with the

 

55Kallman J. Cohen and Samuel R. Reid, "Effects of
Regulation, Branching, and Mergers on Banking Structure and
Performance," Southern Economic Journal, XXXIV, #2 (October,
1967), 231—2u9.

 

“7

problems of the effects of regulation, branching, and mer—
gers on banking structure and performance as follows:

It is obvious that the vast majority of mer—
gers, 92 1/2 percent, were consummated in States
which permit branching. The absorption of banks
through merger in unit states was more than off—
set by the 687 new entrants during this decade.

It is also interesting to note that both in abso-
lute and in relative terms there was less entry

of new banks in the thirty—two states permitting
branching than in the eighteen unit banking states
in this period.

For those who believe in the preservation and
promotion of banking alternatives, the future is
indeed bleak unless there is legislative and regulae
tory agency recognition of the problem. In the
absence of this recognition, it appears that banking
alternatives will decline at an accelerating rate in
the years ahead. This pessimistic outlook is based
on the following factors: First, the majority of
bank mergers occur in states which permit branch
banking, and Second, it is highly probable that
branching laws will be liberalized in many states
in the future. This prediction is based upon the
effective redistricting which will eventually cause
a shift in the balance of power away from the rural
areas toward the metropolitan areas.

A curious phenomenon exists in that a pro—
merger attitude has such wide-spread acceptance
among legislators and regulatory agencies without
any clear-cut demonstratable evidence to support
this position. The assumed benefits of bank mergers
seem to be almost universally accepted as a fact of
economic life.5
In 1960, the State of New York changed its banking law
to permit New York City commercial banks to open branches in
counties surrounding cities which counties have a pOpulation
of at least 700,000 people. At the present time, this

includes Nassau and Westchester Counties and provides for

the merger or purchase of assets of banking institutions

 

56lbid., pp. 231-232.

H8

involving New York City and Westchester or Nassau County
Banks.

The State of Virginia also changed its banking law.
Prior to l9U0, Virginia permitted branching of both types,
that is, general full branching and merger acquisition of
another bank to be Operated as a branch. in 19U8 and
1952, the act was amended to expand the geographical area
of branching. Branching in Virginia is now permitted
anywhere in the State but only by merger. De novo branches
continue to be restricted to the immediate area of the
exis ing bank.

The study sets forth that there are three interested
groups in bank behaviour: bank management, bank stock—
holders, and the public. A seri s of equations is used to
represent the management interest. First, management is
interested in the percentage change in total assets from
one period to another. A similar equation is constructed
which sets forth the percentage change in total deposits,
for the change in loans and discounts, and for the change
in employees.

Stockholder interests are set forth similarly in the
percentage change in Operating earnings per share adjusted
for stock splits and stock dividends.

The interests of the public are represented by equa—

tions setting forth the percentage change in total assets

“9

for the entire network of banks which had been combined into
a single bank and for the percentage change in loans for the
entire network of banks which had been combined into a
single bank.

The study was first run on a nation-wide sample of
165 large commercial banks. It was run the second time in
an attempt to see if the conclusions previously reached
about these 165 large banks would remain valid when consid—
erably more homogeneous groups of banks were examined for
relatively brief periods of time. The data for the study
was taken from a three-year period determined by the date
of passage of the recent changes in banking legislation:
1959 to 1961 was used for New York while 1961 through 1963
was used for Virginia. Samples consisted of all commercial
banks in the respective states from which it was possible
to obtain information concerning all the relevant variables.

It would appear that management interests were well—
served by the merger activity in each of these states. All
four of the variables representing_management interests are
positively and significantly associated with merger activity
in both New York and Virginia. Stockholder interes s do not
appear to be served by the bank mergers which have taken
place. The public interest variables are not significantly
associated with merger activity. Since mergers, per se,
lead to a reduction in the number of banking alternatives

available to customers, it would have been necessary that

there be a significant positive association between the
public interest variables defined and merger activity to
have made a prima facie case for bank mergers to be in the
public interest. Therefore, the statistical findings in
New York and Virginia indicate that most bank customers at
least in the short run have not benefited from bank expan—
sion by the merger route.

The following policy considerations grew out of the
study: (1) Liberalize branching and encourage merger activ-
ity. As a result, one would expect an accelerated pace
toward a concentrated banking structure in the state. Major
benefits accrue to the owners of the acquired bank since a
premium is paid for his franchise. (2) Liberalize branching
and promote de novo branching instead of mergers. In this

(3)

instance, the public will gain banking alternatives

a
0

It is the opinion of the authors that the relaxing of
branching restrictions, provided the citizens of the state
demonstrate their approval preferably through the voting
privilege, will help stimulate banking competition if pri-
ority is given to de novo branching rather than the merger
alternative. Possibly, it should be provided that a public
announcement be made concerning a proposed branch with pri-
ority for that location being given to qualified new
entrants. This would increase the number of independent
decision making units. Ideally, the citizens of a state

would have a more direct voice in the type of banking law

51

which is passed in the state. The possibilities of this
action materializing, however, are slim since the various
bank trade associations undertake extensive lobbying activ-
ities. In the absence of the direct voting privilege, it

is recommended that state legislatures concerned about opti-
mum banking structure should give serious consideration to
the various feasible alternatives. The welfare of special
interest groups, the authors reiterate, should be subor—

dinate to the public interest?7

Banking in Chicago

 

The Schweiger and McGee report on Chicago Banking is
an extensive study prepared for the Chicago Association of
Commerce and Industry in an attempt to "evaluate the ade-
quacy of the financial structure of Chicago and Illinois."58
Because of its complexity and length, it cannot be reviewed
in detail. A summary of findings which appeared in the
Journal of Business in 1961, indicated that small banks
place a smaller proportion of their assets into loans than
would larger banks of the same type whether they were unit
or branched banks. Unit banks averaged smaller loan ratios
than did branched banks of comparable size in each of the
types of community tested. In fact, the scale of lending

by branch banks was so much above that of unit banks that

 

57Ibid., p. 2H9.

588chweiger and McGee, loc. cit., p. 203.

 

52

small branched banks averaged a larger ratio of loans to
assets than did unit banks of much larger size in comparable
communities. This pattern was also revealed in the Chicago
area when it was compared with comparable banks in similar-
sized citifh; in otlmnizireas. [hilt and brmnuui banks MHWW?
found to have different lending patterns. Very large unit
banks loaned heavily in the commercial and industrial areas
while branched banks had a more diversified loan portfolio.
Small branched banks provided more business financing
relative to assets as well as more mortgage and consumer
financing than did comparable unit banks. Unit banking
systems granted considerably less residential mortgage
credit than did branched bank syste.s. As a result, the
authors concluded that branched banks do not behave as

mere combinations of smaller banks.

As to facilities, a unit banking system provides a
larger choice of banking firms in a large city. In con—
trast, a branched system provides a wider choice of facil-
ities in smaller areas. Unit banking areas have relatively
more one-bank towns and towns with no banks than do branched
bank areas. A branched bank system provides three to four
times as many banking offices relative to population in
large cities as do unit systems and, roughly, twice as many
in suburban areas. Unit banking responds slowly to shifts
in pepulation and industry in providing new or additional

banking facilities.

53

Branched bank areas experience a larger volume of
long-term savings in local institutions relative to income
tha11(ha unit;tiankirhg areas. In urdl: hankiih: areas, :Mivers
put more if their resources into postal savings, savings
bonds, and cash, thus limiting the lending capacity of the
banks. Multiple banking areas also revealed that total
savings were higher on the average for both the banks and
the savings and loan associations.

Apparently the potential for attracting
savings into local institutions are so inadequately
tapped in unit banking areas that both banks and
savings and loan associations can increase their
deposits with the change in structure and the more
vigorous competition for savings that often accom-
panies it.59

Regarding costs and earnings, the authors state:

Branch banks tend to have higher costs and
lower net current earnings on capital than do unit
banks of the same size. But branched banks are of
larger average size. As a consequence, a greater
proliferation of banking offices can be achieved
more cheaply, and with higher net current earnings
by branch banks than by many small unit banks. No
doubt, some of the present disadvantages of unit
banking stem from blocked entry; but not all. Even
in periods of essentially free entry into banking——
as in the twenties——branch ba king provided more
facilities than unit banking.

Two Studies by Shull and Horvitz

 

Bernard Shull and Paul Horvitz, in two articles,61

analyzed the influence of branch banking on the structure

 

59Ibid., f). 211.
60Ibid., p. 215.
61Shull and Horvitz, loc. cit., pp. 30l-3U1; Shull and

Horvitz, "The Impact Of Branch Banking on Bank Performance,"
National Banking_Review (December, 196“), pp. lU3—188.

 

 

7i
i—

of the market and also on bank performance. In the first,
the authors make the observation that there are fewer banks
in states with branch banking——and that the decline in the

number of banks over the last decade has been greater in

(‘1‘

ates with branch banking. The downward trend shows a

U)

slackening in recent years. This would appear to be a very
striking and significant conclusion in view of what one
would expect to have taken place in view of the recent

trends toward increased branching through th

(D

merger and de
novo routes. In the second, the authors find that areas
with pepulations over 7,500 eXperience larger increases in
number of banking offices under branch banking than under
unit banking. In smaller communities, there are mire
banking offices in unit banking states but the differences
are quite small. It becomes evident that the first article
does not discuss banking offices. It focuses its attention
on banks as entities——on banks, regardless of the number of
branches. The conclusions of the first study are not sur—
prising within this framework. One is easily led to forget

that bank markets are primarily local in nature and a com~
munity served by a branch of a larger bank or by a single
unit bank is being served within a monopoly framework. The
authors recognize this and determine that their data is
inadequate to ana yze the local market structare in th;
industry. They then gathered data on the number of banks
in each town in the United States not a part of a metropol—

itan area. Within this structure, they found the highest

55

average number of banks in branch banking states. While
the difference was not large between unit banking and branch
banking states, it was significant enough:
to refute the implication of overall state compar—
isons; i.e., that the extension of branch banking
leads to smaller numbers of competing banks in
local markets.62
Then the authors looked at Standard Metropolitan Sta-
tistical Areas and found that smaller metropolitan areas
had about the same number of banks whether in branch or
unit banking states. In the larger areas, there were con-
siderably more banks under unit banking than under branch
banking. Again one must remember that the authors are
talking about banks——and not about banking offices. Con-
centration statistics reveal that unit bank areas are less
concentrated than branched areas but the difference does not
appear to be significant from an economic point of view.
A final summary of the conclusions of the first arti-
cle states the following:
Our analysis suggests that neither in terms of
number of competitors, nor concentration (measures
of actual competition), nor in terms of conditions
of entry (potential competition) have the structures
of local banking markets been adversely affected by
branch banking in the United States. The weight of
evidence suggests that, to the contrary, market

structures are adversely affected by restrictions on
branch banks,63

 

62Shull and Horvitz, "Branch Banking and the Structure
of Competition," loc. cit., p. 3&0.

63Ibid., p. 3M1.

56

In the second article, the authors conclude that there
are systematic differences in the performance of branch sys-
tems and unit systems. Other studies are cited as having
indicated that:

Branched systems tend to have high loan-asset
ratios (and low liquid-asset ratios), a high pro—
portion of retail—type loans in their loan port-
folio, lower rates and lower maturity on retail-
type loans, high rates on time and saving deposits,
and high service charges on demand deposits.

Branch banks are typically larger but for any given
size, tend to have higher costs than unit banks.
The loan mix at branch banks and the higher loan—
asset ratios would tend to make branch bank profits
higher, but this is offset to some extent by their
higher costs.“4

They point out that the important difference between
branched and unit systems is in the range of services pro-
vided. Branched systems offer a large variety of services.
They also offer greater convenience because they have mul-
tiple offices. The additional service is, at least in
part, a function of their larger size. The authors note
that it must be remembered that the size of banks in unit
banking communities is severely restricted by their inabil—
ity to branch. The capability to branch, then, is a sig-
nificant influence on bank performance. The threat of
potential entry is not as serious to a branched bank as it
is to a unit bank and the threat of increased competition

by branching causes unit banks in branch banking states to

operate differently than do unit banks in unit banking

 

6“Shull and Horvitz, "The Impact of Branch Banking on
Bank Performance," 100. cit., p. 176.

57

rn

states. ihey operate more nearly like the branched banks
ifilth which) they cxmnpete. Tfliis, of‘cxnirse, ii; not BJILHM?X-
pected conclusion.

The authors note that there is a general tendency for
branches of banks to follow pricing policies similar to
those of their main offices and to offer, as far as is
possible, the same services as are provided at the main
office. Since the main office is typically located in a
large city with substantial numbers of banks in direct com-
petition with each other, the branch brings into what would
otherwise be a monopoly market the polici 3 determined in

the competitive market of the home office.

Bank Entry and the Public Interest

 

Motter and Carson65 evaluated the extent to which bank
entry served the public interest.

When the State of New York, in 1960, passed its
Omnibus Banking Bill, it opened the suburban counties of
Westchester and Nassau to entry to New York City Banks. The
expansion into these counties was rapid. An economic sit-
uation was created which made possible the evaluation of
the effect of bank eXpansion under relatively closely con-
trolled conditions. The study, in large part, was devoted:

to an ex—post evaluation of the effects of these pol—
icies in so far as this can be accomplished by

 

65David C. Hotter and Deane Carson, "Hank Entry and the
Public Interest: A Case Study,".National Hanking_kevivw
(June, 1964), pp. 469-512.

 

58

examination of bank costs and rates of return, the
structure dimension of bank performance, and the
benefits that have redounded to consumers of bank
services.66

Two interrelated tests were adopted to determine if
the county was overbanked:

Are the rates of return to Capital for Nassau
banks in the aggregate, and for most individual
banks, at a level which threatens the ability of
the system to provide an adequate quantity and qual—
ity of banking output over the long term . . . Is
the N.ssau banking system currently performing ade—
quately from the standpoint of its customers, and
has there been an improvement or a worsening of
this performance in recent years?

The authors determined that overbanking did not exist
in the county because rates of return were at levels which
allowed the continued supply of a broad line of banking ser-
vices and products and that the system materially improved
in performance in a number of dimensions in recent years.
Specifically, the study revealed that the increase in the
number of banking offices in the counties did not result in
a deterioration of the aggregate rate of return to capital
in comparison with all insured banks in the United States.
The expansion of banking offices in the counties provided
additional banking service to the customers. The evidence
also indicated that the increase in banking competition

lowered installment loan rates. Analyzing the expansion by

submarkets of the counties showed high correlation between

 

66Ibid., p. 511.

 

67Ibid., pp. 511—512.

the distribution of banking offices and dis ributien of
deposits with various income and population measures
The authors concluded overall:

Nassau County has a viable dynamic and bal—
anced banking system which compares very favorably
with the banking systems of other area" in the
extent to which it meets the needs s cus-
tomers. he recent sharp increase e number
of banking offices represented a r s of
catching up with the growth of demand for banking
services in the country.

"’9

There are in the literature many examples of empirical
research concerned with the demand side of the market. This
literature often directs its attention to the problem from
the standpoint of the demand for credit from the banking

community by the business or household sector. A typical

’

69

which he

Ho
:5

example of such an approach is that by Frazer
questions the effectivity of the monetary controls of the
Federal Reserve System at various phases of the business
cycle. He is concerned with an important aspect of the
entire problem under consideration in this study but it is
felt that the field is too broad to se taken into consid—

eration.

 

68Ibid., p. 512.

69William J. Frazer, Jr., Camgen'11m en “enetary
Policy Guidelines and Federal Reserve Structure Subcom—
mittee on Domes tic Finance of the Commi toe on flanking and
Currency, House of Representatives (Washington, D. C.:
Government Printing Office, 1968), pp. 191—138.

 

 

 

THE THEORETICAL STnUCTWRE

The purpose of this study is to determine the differ—
ences, if any, which exist among the various banking systems
as to restrictions on branching and their impact on the
economic activity of the state in which those systems are
located. Various studies have touched upon this subject,
mostly tangentially. One study approached the problem in
a discussion of the overall role of commercial banks in
economic growth, and concluded:

The principal impact of the commercial banks'
operations on the economy is through their lending
activity rather than through their deposit func-
tions.1

Another states that ”the characteristic function of the

’7
a.

banking system is provision of business loans." Later,
arguing that the market for bank loans can be examined in
isolation, the author states:

Trade credit is the only important example of

alternative sources of supply for short—term bank
business loans which can meet, at least, some of

 

1”Commercial Banks and Economic Growth," The Bulletin,

op. cit., p. 21.

.2A1hadeff, Monopoly and Competition in Banking, op.
01 ., p. 9.

 

 

6O

61

the tests of a reasonable substitute. At best,
however, even trade credit is a highly imperfect
substitute for bank loans and under most con—
ditions does not constitute a competitive alter—
native source of supply. Hence, the market for
bank loans can be examined by itself."3

Finally, Galbraith makes the following statement which
eXpresses the essential features of the major hypothesis of
this study:

The banking system that maximizes its cash reserves
and which keeps these reserves most fully utilized
does a better job of supplying funds to the com—
munity than banking systems that are less efficient
in these tasks.“

Many of the studies reviewed in Chapter III generally
attempt to pinpoint the differences among banks located in
different market structures especially with respect to their
changing capacity to make loans as they expand either "firm"
or "plant." While those questions are similar to the one
being posed here, the studies only suggest or imply that
differences noted among the systems result in a greater or
lesser quantity of loans being made with the resulting
impact on economic activity being left to the speculation
of the reader. Only Cohen and Reid5 and Motter and Carson0

attempt to evaluate the extent to which public interest is

served. The first study was inconclusive in this regard while

 

3Ibid., p. 19.

”Galbraith, The Economics of Banking Operations, op.
cit., p. 173.

 

5Cohen and Reid, loc. cit., pp. 23l-2M9.
6

 

Motter and Carson, loc. cit., pp. 469-5l2.

 

62

the second revealed that increasing the branching privilege
provided additional banking sucess.

In this study, an attempt is made to determine the
extent to which the economic activity in a state is influ—
enced by the bank market structure specified by state law.
As commercial banks supply investment funds to the existing
demand, economic activity is generated. The alternatives
available to the banker vary as to their direct impact on
local economic activity. Because the Federal Government
does not spend in specific geographic areas in accordance
with the volume of federal indebtedness marketed in the
area, no local impact can be directly attributed to invest-
ment funds devoted to this alternative. Somewhat similarly,
investment funds devoted to "Other Securities" find local
market impact only incidentally. But, because most of the
Loan and Discount business is found in a commercial bank's
local market, it follows that if banks in one system tend
to make more Loans and Discounts than banks in another sys—
tem, given the same internal motivations for making Loans
and Discounts, the first system would be making a more sig-
nificant contribution to local economic activity than would
the second. A system able to operate with less of the Total
Assets devoted to satisfying liquidity needs would also aid
economic activity as funds released from this employment are
devoted to income—generating activity.

Assuming that bank loans contribute to economic

activity as they facilitate income—generating spending, this

63

study measures the proportion of a banker's portfolio
devoted to certain components of investment opportunity,
given the motivations for placing investment funds in alter—
native components. It is assumed that bankers are profit
maximizing businessmen who are motivated to invest available
funds in their various portfolio components in accordance
with the varying demand for those funds. Stated somewhat
differently, it is proposed that the relative extent be
investigated to which bankers in varying market structures
allocate investment funds to the components of their port-
folios given the internal motivations for doing so. Any
differences observed would be revealing as to which system,
given the internal motivations, devotes alarger proportion
of its portfolio to investment opportunities which generate
economic activity directly in the bank's local area and

less to satisfying its liquidity demand--thus fulfilling

the purpose and intent of the study.

Availability of Income Statistics

 

Consideration of the motivations influencing alloca-
tions to certain portfolio components leads one, within the
profit maximizing assumption, to an examination of the
extent to which bankers are motivated by profits or income-—
not only on the specific portfolio component under consid-

eration but also on alternative sources available. The

6A

recognition that bankers respond to changes in profit oppor~
tunities is examined in a recent study by Russell.7

The tests to be performed are limited by the fact that
banks' income statistics are available only for (a) United
States Government Bonds, (b) Other Securities, and (c) Loans
and Discounts, Net. While income information is available
on other activity (trust departments, etc.) of commercial
banks, these income sources are not directly attributable
to specific portfolio components and, therefore, are not
related to the investment process. A significant part of
the portfolio is one on which no income is earned and yet
is important in the determination attempted. Should one
banking system differ from another in the extent to which
it devoted part of its investment funds to liquidity demand,
it would also differ in the quantities of investment funds
it allocated to income producing components. Accordingly,
allowance was made for indicating the varying extent to
which banks provided for liquidity needs.

Another portfolio component on which no income is
directly earned is the portion of the Total Assets devoted
to Bank Premises, Furniture and Fixtur,s, Real Estate, and
Miscellaneous Assets. While these are worthy of investment
because of their contribution to over—all efficiency of the

banking operation, they do not represent a portfolio

 

7Russell, op. cit., pp. BAA-553.

65

component to which income can be directly attributed. This
portion is excluded from the consideration.

The four portfolio components taken into consider-
ation, therefore, are the following:

1. "Cash Items" or "Liquidity."

2. United States Government Securities.
3. Other Securities.
A. Loans and Discounts.

These four investment opportunities are considered to
be substitutes to the banker. They are not perfect sub-
stitutes, however, because they vary as to (a) their
capacity to satisfy liquidity needs, (b) as to term, (c) as
to risk, and (d) as to costs the banker will undertake to
service the component. The relative importance of these
considerations is reflected in the income earned on each
specific component.

The extent to which a banker provides for liquidity
(in excess of his reserve requirements) is influenced by
the income which is earned in the remaining components of
his portfolio. As income on the remaining components
rises, the banker is motivated to reduce his desire for
liquidity; as income on alternative components falls,
liquidity would not represent so great a sacrifice,
resulting in his being more willing to place investment
funds into this non-income generating component.

The isolation of the portfolio components which are

applicable to the study allows a consideration of the

66

demand and supply framework in which the analysis can be
accomplished. First, consider the demand curves which a
banker faces for each of the portfolio components under

consideration.

The Demand Curves

 

8

The Price Variable

 

As to United States Government securities, the demand
curve facing any individual bank in the country, irrespec-
tive of state branching restrictuions, can be assumed to
be absolutely elastic. Each of the banks, regardless of its
firm size or the extent to which it is branched, finds that
demand for its investment funds exists at the prevailing
price in quantities far beyond its capacity to buy. Each
bank can supply investment funds to this demand up to its
full capability without affecting market price. In effect,
each bank is in the nature of a "purely competitive" seller
in the market for United States Government Bonds. Here, in
effect, is the classical case of a monopsonistic buyer (the
United States Government) buying investment funds in a
purely competitive factor market. All banks are in the
same relative relationship to the demand for United States
Government securities regardless of the particular market

9

structure in which they are located.

 

8Robinson, op. cit., pp. 201-210.

91bid., pp. 376-393.

67

The classification of "Other Securities" includes
(a) Obligations of States and Subdivisions, (b) Securities
of Federal Agencies and Corporations, (0) Other bonds,
notes, and debentures, and (d) Corporate Stocks (including
Federal Reserve Bank Stock). The demand facing the indi—
vidual bank for investment funds represented by "Obligations
of States and Subdivisions” and "Securities of Federal Agen—
cies and Corporations" is similar to that for United States
Government Bonds. These securities are offered in a
national market and, therefore, all banks in the country
are in a purely competitive relationship to the demand.10

Some lessening of elasticity is introduced into the
demand curve by the presence in this category of "Other
Bonds, notes, and debentures" and "Corporate Securities."
These securities are offered into the market on a more atom—
ized basis. In each of these cases, the interest rate which
must be paid on the security is the price for having her-
rowed the funds. The price is a restricting influence on
the quantity of investment funds purchased. More investment
funds would be borrowed and more investment eXpenditure made
if the price at which these funds could be borrowed was
lower. Consequently, the demand curve for this portion of
this component of the portfolio is downward sloping. It

does not, however, cause the demand curve for the entire

component to deviate very far from being absolutely elastic

 

lolbid., pp. 39u—u12.

68

because the elastic components of this category are com~
pletely overwhelmed by the relative magnitude of "Obliga—
tions of States and Sub—divisions." There is no reason
that the demand curve for this portfolio component would be
significantly different whatever the structure of the
banking market as to the geographical limits to which banks
are allowed to branch. Size might be of importance in
determining the extent to which a bank could affect market
price so that such large banks might face a downward sleping
market demand curve much as a monopolist does. Since there
are unit banks as large or even larger than the largest of
some branched banks, this possibility would not seem to be
meaningful in a geographical delineation based on branching
restrictions in state banking laws.

Constructing a balanced portfolio by evaluating rela-
tive risk, hedging against eXpectations, and balancing term
and liquidity considerations precludes buying sufficient
quantities of any one security to affect market price. The
banker will buy relatively small quantities of many secur-
ities rather than specializing in one. Good portfolio
management would induce a banker to act more as a purely
competitive supplier than as a monopolist.ll

in summary, it seems safe to assume that the demand
curve facing banks in the country for investment funds

represented by the "Other Securities" category would be

 

llihid., pp. 133-1uu, 323—3u2.

69

similar regardless of the market structure as regards
branching prescribed by the individual state law.

The "Loans and Discounts"

component records the extent

to which a bank services its local market. It is generally

assumed that large business loans are serviced in a regional

or national market making the larger banks competitors for

these loans. But the bulk of the Loans and Discount busi—

ness is found in a bank's local area.12
The demand curve for Loans and Discounts is assumed

to be downward sloping. The price charged by bankers for

a loan is a rationing device to limit the quantity of

loans granted to the quantity of fUnds available for

making loans as well as to satisfy the bank's profit

objectives. Because banks practice price discrimination

in granting loans based, at least in part, on relative

risk, it must be assumed that loans of some given risk

would not have been denied had the general structure of

interest charges been lower. The general structure of '

interest charges might have been lower had there been a

greater volume of investment funds available for the pur4

pose of making loans. Loans are often denied because the

investment cannot support the interest rate required to make

the loan profitable enough for the bank to undertake the

risk involved. Stated differently, it might be assumed that

investment opportunities not undertaken because loan rates

 

12Ibid., pp. 129—32u.

70

are too high to make the investment profitable, would be
undertaken if the cost of borrowed capital had been less.
All of this implies a downward sloping demand curve for
investment funds secured by Loans and Discounts.

There is reason to assume that bankers generally
cooperate with each other through trade and professional
associations. These practices are evidenced by the close
similarity among competing banks, the establishment of sim-
ilar and relatively inconvenient banking hours, and that
"shOpping" for price in a local market does not often yield
advantage. It is recognized that exceptions to this gener—.
alization can be cited, but sufficient observations of these
characteristics of the industry warrant the generalization.
In summary, while there may be many banks in a single mar—
ket, a truly competitive relationship does not exist among
them. ‘

According to Greenbauml3 bank practices, in fact, pre—
clude the establishment of a fully competitive market. This
exists even in markets which appear to be.very competitive
on the basis of concentration statistics.lu Because of
these practices, branches of banks in metrOpolitan centers
may be similar to unit banks in their relationship to their

demanders. A branch of a bank in a one—bank market is no

 

l3Greenbaum, "Competition and Efficiency in the
Banking Industry—-Empirical Research and its Policy Impli—
cations," loc. cit., pp. U6l—U8l.

l“Edwards, Op. cit.

 

71

more or less a monopolist in that market than is a unit
bank. In fact, a monopoly bank which is a branch of a bank
located in a large metropolitan center brings to its monOp—
oly market a method of operation determined by competitive
relationships as it brings to its local area the policies
and practices of its home office.

In any event, both branches of banks and unit banks
can be assumed to be facing a downward sloping demand
curve for Loans and Discounts available from their local
market. Accordingly, no difference is assumed to exist in
the demand curve for Loans and Discounts facing banks in
separate banking systems classified according to state
restrictions on their Opportunity to branch. It is assumed
that the demand curve for Loans and Discounts is similar no

matter what the market organization.

The Income Variablels

 

In addition to the price variable represented here by
the income earned on the portfolio component under consid-
eration, it could be expected that demand for Loans and
Securities would be affected by the level of income in the
market in which the bank is located. In the case of the
demand for funds by the United States Government, the rela—
tionship of income to quantity demanded is somewhat unclear.
It could be predicted that the influence of this variable

would be very small and then only to the extent that local

4*

15Robinson, 0p. cit., pp. 20-3“, 201-210.

72

income per capita represented a similar income level used

in determining fiscal policy at the national level. Should
income be low, national policy might call for deficit finan—
cing to bring about economic growth. On the other hand,
during certain periods of relative economic affluence, def-
icit financing has also taken place.

As to the "Other Securities" component, the influence
of income is again quite unpredictable because this com—
ponent is, for the most part, made up of "Obligations of
States and Subdivisions." While it might be desirable that
deficit spending by these subdivisions of government take
place to a greater extent in time of declining or low
national income than during inflationary periods, it is
entirely pos~ible that local governmental units are more
eXpansion minded in times of economic prosperity than in
times of recession. Of course, in the portion of this com—
ponent more closely related to the private industrial stock
and bond market, one would expect that changes in quantity
demanded of investment funds would be positively related to
changes in income in order to satisfy the increased demand
for goods and services which accompanies higher income.

In the case of the Loans and Discounts component, the
role of income is much more clear. As income rises, it can
be expected that demand for loans and discounts also would
Piiué. Businesses anild demand.irnnmitment capital ir1<nwhxr
t”) meet the increased demand for goods and services, and

Consumers, in attempting to satisfy their desire for goods

73

and services, would demand additional quantities of loans
and discounts for consumption purposes (given their expecta-
tion of continued earnings and willingness to undertake
increased indebtedness). However, a precise projection of
the relationship of the income variable to the quantity of
investment funds devoted to Loans and Discounts is not pos—
sible because one is unable to isolate whether the change

in income brought about the change in demand for Loans and
Discounts or if the increase in Loans and Discount demand
acted as a causal factor to the change in income.

The demand curve for liquidity,l6

or "cash items," is
difficult to define since it is a demand the banker faces
as a result of market conditions and supervisory regulation.
The shape of the demand curve becomesdifficult to speculate
about because it cannot be assumed to have a price variable
comparable to that found in the demand curves already dis-
cussed. A banker will face an increase in quantity demanded
of liquidity as his demand deposits increase as a proportion
of total deposits. Some of this demand is created by super—
visory authority in the imposition of reserverequirements
which, because they are higher on Demand Deposits than on
Time Deposits, would rise as this ratio rises.

Another determination of quantity of liquidity

demanded is the level of personal income in the banker‘s

market with the expectation that this indication of economic

 

16Robinson, op. cit., pp. 51-68.

7“

activity is expected to decrease the demand for liquidity
as demand for loans increased. The elasticity of this
demand curve is quite unpredictable because the relative
magnitude of these influences in determining demand for
liquidity is unknown. However, it might be assumed that
the demand is positively sloping——reflecting that the
influence of the Demand Deposits as a percent of Total
Deposits ratio is the most important determinant of this
relationship.

In summary, it is postulated that no difference
exists in the demand curve facing banks in any of the port-
folio components under consideration no matter what the
market in which the bank is located as determined by state
law.

Before giving consideration to the supply curves rele-
vant to each of the demand curves already discussed, it is
necessary to recognize that we assume that the principles of
profit maximization are observed in that the total quantity
of output of investment funds is determined by the inter-
section of the marginal revenue and marginal cost curves for
each bank in the consideration. The demand curve for the
bank's entire operation can be assumed to be the sum of the
demand curves for each of the portfolio components discussed
together with those not included in the study (Bank Prem-
ises, Furniture and Fixtures, and MisCellaneous Assets).

The Supply Curve for the bank's entire operation can be

75

considered to he reflecting, in classical form, the mar»
ginal costs of producing the output and placing it into the
market. Consequently, recognition is given to the deter-
mination of total quantity of output in light of the var-
iable eXpense structure discussed in the literature
reviewed in Chapter III. Should variable costs be lower

in one system as compared to another, the marginal cost
curve for that system would be expected to be lower. And
the system having lower variable costs would, of course,
experience a profit maximizing quantity of output which
would be larger than a system with a higher variable cost
structure. While this determination is important in under—
standing the working of this very inter—related and com-
plicated market, it does not constitute the major purpose
of this study. We are here concerned with a determination
of which banking system, classified according to branching
restrictions of various state laws, will allocate the avail-
able output in such a manner as to be more instrumental in
generating economic activity in its local area than will
some alternate system. Consequently, the quantity variable
in this study is not an absolute quantity of dollars or a
number of loans. It is, rather, a percent of the total
portfolio which is offered in each of the four Specific com—
ponents under consideration. We are, therefore, concerned
with the decision-making process which a banker faces not

as regards the total quantity of investment funds to offer

/

76

into the market in light of the profit maximizing deter-
mination of his total marginal revenue and marginal cost
curves, but the proportion of his total investment funds
or output which would be devoted to each of the profit-

making Options open to him, given his need for liquidity.

The Supply Curves

 

The decisions on portfolio mix examined in this study
are, of course, part of the decision—making process under-
lying profit-maximizing output determination. Several
basic factors are instrumental in determining this alloca-
tion, all reflecting the motivations a banker takes into
consideration in making the allocation to the four portfolio
alternatives open to him. These four portfolio components
are considered to be substitutes to the banker. They are
not perfect substitutes because they vary as to (a) the
income which they can produce, (b) the risk they represent,
(0) the length of time they require that funds be so
employed, and (d) the ready facility with which they can
be converted to cash. If they were perfect substitutes,
this study would not be possible since the entire output
would be devoted to the option representing the most impor—
tant consideration at any one time. It is expected that,
if they were perfect substitutes, the banker would spe—

cialize.

77

The Price Variable17

 

One factor a banker considers in making a decision as
to how much of his portfolio should be devoted to any one
component of the existing demand is the income he expects
to earn on the investment funds offered. The income or rev—
enue expected on investment devoted to a specific component
is, of course, the price variable. Since no income is
available on investment funds devoted to liquidity, this
component does not have a price variable in the same form as

does each of the remaining components.

The Alternative Income Variable

 

Part of the price consideration for any one component
is the income which can be earned on alternative options
available. Consequently, each of the alternatives available
to a banker must be weighed not only as regards the income
which can be earned on that specific component, but also as
regards the income which must be foregone when output is

invested in one of the components in preference to another.

The Liquidity Restraintl8

 

Another fact which the banker must take into consider-
ation is the extent to which he is required to provide for
his liquidity needs. Since reserve requirements are sub-

stantially higher on Demand Deposits than on Time Deposits,

 

l7Robinson, Op. cit., pp. 201-210.

181b1d., pp. 55-59.

78

and since a banker's reserve requirement will vary as these
components of Total Deposits vary, his liquidity requirement
is indicated by the percent of Total Deposits represented by
Demand Deposits.

In the case of the three income—producing components
of a banker's portfolio, we would expect that the quantity
of funds supplied to a specific component would be (a) pos—
itively related to the price variable, (b) negatively
related to the variable indicating the income which is sac-
rificed, and (c) negatively related to the variable speaking
for the restraint imposed by liquidity requirements.

In the case of the Liquidity Component (herein enti—
tled "Cash Items"), only two determinants are observed-—the
(Demand Deposits)/(Total Deposits) ratio and the variable
indicating the income which is sacrificed by having to sup-
ply the liquidity demand. The first of these is expected to
be positively related to the quantity of funds supplied
while the second is expected to be negatively related to

the quantity supplied.

The Reduced Form Regression Equations

 

The proportion of the portfolio observed from the
aggregate figures of bankers' financial statements are
equilibrium proportions, i.e., proportions determined in
accordance with the principles outlined above. Conse-
quently, reduced form equations are constructed which will

recognize the factors instrumental in determining the

79

proportion of the total portfolio devoted to any one spe-
cific component from both the demand and supply sides of
the decision. These equations are set forth in Table IV—l
in the classical form of Y = f(X1, X2, X3, Xu).

To eliminate the multicollinearity which exists
between the last two variables in Equations (2), (3), and
(A), these concepts will be combined. The price variable
will be expressed as a ratio: (Income Earned, this Com—
ponent)/(Income Earned, alternate Components).

As outlined above, there is no reason to believe that
the decision to supply investment funds to any one of the
components in view of the existing demand for the output is
different in one banking system (according to branching

restrictions) as against another.

The Hypothesis

 

It is hypothesized that no difference exists among
the systems as to how they determine the proportionate share
of their portfolio to devote to any one of the four compon-
ents. Stated more specifically in terms of the statistical
computation performed, it is not expected that the Statewide
Banking System or the Limited Branching System is signifi-
cantly different (as revealed by F tests) from the Unit
Banking System in the proportion of its funds devoted to any
one of the components analyzed given the same motivations
for having made that determination. The regression equa—

tions are framed in such a fashion (using Dummy variables)

80

to indicate the difference, if any, which might exist among
the banking systems in the extent to which either the State-
wide or Limited Branching System might differ from the Unit
Banking System through the differences in the alpha term

(or constant) of the regression equation. The hypothesis
can also be stated that the alphas (or constants) in each

of the regression equations are not expected to be signif-

icantly different from each other.

Reasons for Rejecting the Hypothesis

 

No attempt will be made in this study to determine
the specific causes for any differences which may be
observed. The purpose of this study is to determine if
some differences exist and to leave to others the isolation
of the specific reasons for those differences. It is pre—
sumed that any differences observed are caused by the dif-

fering limitations on branching.

On the Demand Side

 

Should differences be observed, causing the hypothesis
to be rejected, such differences can be speculated to have
been caused by varying structural characteristics inherent
in the different market organizations. These differences
may be attributed to differences in concentration, to the
presence of varying degrees of monopoly, to economies of
scale brought by expansion through expanding the multi—

plant operation of a branched system.

81

The system experiencing market conditions which
more closely approximate those of pure competition would
find that its demand curve for Loans and Discounts would
be more elastic than that system which is more nearly
monopolistic. Any deviation in the demand curve from
being absolutely elastic would cause the quantity of out-
put devoted to this specific component to be significantly
reduced as the marginal revenue curve deviated from the
average revenue curve. The further the market deviated
from the purely competitive ideal, the further would be
the equilibrium quantity reduced.

Deviation of the demand curve from being absolutely
elastic can be caused by concentration as is argued by
Edwards.19 However, because of restriction on entry and
on expansion and the fact that a branch of a bank can be
as much a monopolist as a unit bank in a similar market
situation, concentration in the banking industry is not
analogous to concentration in other industries and such
deviation of the demand curve could be caused by the pres—

ence of monopoly power regardless of the concentration

statistics with similar allocative results.

¥

1 x O
9ndwards, (q). Cit.

82

On the Supply Side

Another possible cause for difference might be found
in the manner in which bankers in different branching struc-
tures respond to the motivations causing them to devote out-
put to each specific component of their portfolios. For
example, it can be argued that branched banks are able to
operate with less potential investment capital tied up in
liquidity requirements than are unit bankers.2O Any such
difference observed would allow a greater prOportion to be
,devoted to the income-producing portion of the portfolio.
Should it be observed that branched systems devote less of
their output to liquidity needs, this argument will tend
to be confirmed.

It is also argued that branch bankers will tend to
devote a greater proportion of their portfolio to Loans and
Discounts and, therefore, make a more significant contribu-
tion to the local economy than would bankers in alternative
systems. Should it be determined that branched systems do
actually devote a larger proportion of their output to
Loans and Discounts given similar internally determined
motivations as other systems, the argument would be~
strengthened. Such excesses of one system over another
can be deemed to be throwing some light on a determination

of the superiority of one system Over another.

—_

20Chapman and Westerfield, op. cit., pp. 18ll-l88.

83

Should one type of state banking system be observed
to be devoting a larger prOportion of its portfolio to the
"United States Government Securities" component or to the
"Other Securities" component, it could be concluded that
this system was demonstrating a decidedly conservative
investment policy because it would be electing to sacrifice
income for purposes of taking little risk. In this con—
nection, it should be noted that about half of the states
allow their state-chartered banks to invest required
reserves in government securities. Consequently, there
exists some inter-relationship between the liquidity com-
ponent and the United States Government Securities compon—
ent. The states allowing this investment of reserve funds
are quite evenly spread among the three systems. Because
it was impossible to isolate the extent to which this was
taking place, it was necessary to ignore it. In any event,
it is not believed to be of significant importance because
of the small proportion of the entire system (State banks'
which are not members of the Federal Reserve System) able
to take advantage of this option. In 1966, for example,
banks in this classification represented only 8.8 percent
of the Total Assets of the entire Statewide Branching sys—
tem, only 6.5 percent of the Limited Branching system, and

only 9.9 percent of the Unit Banking system.

8H

Anticipated Analyses for Policnyurposes

In addition to the above speculations on the causes
for any differences which might be observed, it is expected
that certain policy implications will be examined based upon
an analysis of the measured responses of different state
banking systems to the variables determining the allocation
to the various portfolio components. This analysis will be
based on the relative significance of the motivational var-
iables in determining the percent of the total portfolio
devoted to any one specific component.

If some light is cast on the variables which are most
significant in motivating bankers to take certain specified
action, regulatory authority would be assisted in deciding
which policy action should be undertaken to accomplish cer-

tain ends.

85

”wzoaaom mm mmsHm> mcfimppmo coHpMSUm some 0p poops mam moHQMHLm> mEESU 039

wcfixcmo msofipm> one coozpmo .mcm we .mocmLmMMflp esp oesofiocfi on com: ops moaomfipm> mEESQ

odomflsm> oopmcmfimmo we maomwpm> wcflpasmop one
m we can ope moanmfism> omega .:x was mx mmaowfism> coozpmo zpflsmmcflaaoofipfise opmcfieflam 09a

II
(\J
O

weazocmsm empHEuS emaufimmmfio soc mu spasm ea 0
mascocmtm assuage emauummmflo mu spasm ea H
mcfinocmsm mefizmpmpm emfiufimmmfio soc ma spasm ea 0
mauzocmsm meazmpmsm emaeummmflo ma spasm an H

II II II
r—h—4m
DOD

.mEopm>m

.mx

.sx magmata> s9 emea>ue mx mupmusm>IIOHpms

”mopoz

 

mpcsoowflm new mcmoq 0cm
mempH :mmo mscfle moomm< Hmpoe
wwwm segue «com >00 m: flosoocH

 

smug s was Hmsoe
Human w mcq mEoocH

coapmazqom
mEoocH Hmom

mpHmOQmQ Hmpoe
mnemomoo pcmEmm

 

 

 

mpomm< proe
pesoomfla
w memoq

 

poz .mpcsoomfia pew memoq ”sz coflpmswm

 

oom pmcuo use
meopfl memo mscfie muomm< Hmpoe
Howfio w mad «cow >oo m: «mEoocH

 

com Lmnpo awooe soapmHSQom muHmOQmo Hmpoe

 

 

 

Howm gmzpo mEoocH mEoocH mem muamomma pcmEoQ

mummm< Hmpoe
0mm smcpo

 

mmfipfipsomm pmzuo "Amv :ofipmsqm

 

omm >00 m: use
mEopH cmmo mscfle mpomm< Hmuoe

 

Homfio w mcqiwomm Locpo .mEoocH

omm >00 m: Hmpoe coapmH3dom mpHmoqoo Hmuoe

 

 

 

Hoom >00 m: mEoocH oEoocH Hmom mpfimomoa pcmsoo

mummm< Hmpoe
0mm >00 m:

 

 

mmapfipsoom ucmECLm>oo mmpmum omens:

"Amv coflpmSUm

 

wEmpH zmmo mscflE wpowm< Hapoe
mcwoq w oom co mEoocH

 

coapwHSQom mpamogoa HapoB

 

 

 

maomOfiHQQm poz mEoocH Hmmm mpflmomea passes

mpmmm< Hapoe
mepH cmmo

 

 

mEopH ammo "AHV cofipmsqm

 

ax

mx mx fix

.mcofiumzom :ofimmmsmmm Epom pmospmm

CHAPTER V
GATHERING THE DATA

It was evident early in the theory development that
the objectives of the study would require an analysis of
the investment portfolio of banks over a lengthy period of
time in order to compare the relative performance of the
three banking systems to each other.

Because different banking systems exist as a result
of differing individual state laws, it was evident that the
analysis should be approached from the basis of state totals
with states classified according to the limitation placed
on branching by those individual state laws. Unit banks do
exist in states permitting branching. However, research
has established that banks not operating branches but
located in states which do permit branching actually Operate
differently from unit banks located in states which do not
permit branching.1 The study showed that unit banks located
in branch banking states Operate more nearly like branched

banks than do unit banks located in unit banking states.

 

lShull and Horvitz, "The Impact of Branch Banking on
Bank Performance," loc. cit., pp. 162-171.

86

87

Therefore, it was decided not to attempt to gather data on
the basis of individual banks, branched or unit, but to find
a source of data revealing detailed information on the com-
posite of bankers' portfolios already summed to state

totals.

Classifying the States

 

The Federal Reserve Bank had already classified states
according to their individual state laws as regards branch—
ing2 and this classification was revised by Shull and Ror-
vitz3 which reclassification was deemed to be applicable to
the objectives of this study. Line maps setting forth the
states which are classified as belonging to either of the

three groups are included in Chapter II.

Period of the Study

 

The period selected for the study was the years fol—
lowing World War II--l9u6 to the present-—a period of suf—
ficient length with none of the distorting characteristics
(World War II and the Great Depression) existing immediately

prior to this interval.

 

2"Changes in Commercial Banking Structure," OD. cit.
p. 1195.

3

3Shull and Horvitz, "Branch Banking and the Structure
of Competition," loc. cit., pp. 3Al-3A2.

88

Sources of the Data

 

The Balance Sheet data selected for the study was
obtained from the "Call Report" for the end of each year
published by the Federal Deposit Insurance Corporation.“
Bank income statistics were taken from the Annual Report of
the Federal Deposit Insurance Corporation.5 State Income
and Population statistics were taken from the publications
of the United States Department of Commerce.6

The data sources provide information in great detail
for insured banks which group comprises the larger propor-
tion of the entire national system. For example, in June,
196A, of the 1U,189 Commercial Banks in the United States,
13,728 were insured. Insured banks constitute 97.6 percent
of the Total Assets of all Commercial Banks. Consequently,
it was decided to use the insured Commercial Banks' statis-
tics to represent the entire banking system in the inter-
system comparison being planned.

To strengthen the argument to use insured bank statis—

tics to represent all banks, it was planned to accumulate

 

u"Assets, Liabilities, and Capital Accounts; Commer-
cial and Mutual Savings Banks," op. cit.

5"Annual Report," Federal Deposit Insurance Corpor-

ation, op. cit.

6"Statistical Abstract of the United States" (Washing-
ton, D. C.: United States Department of Commerce, various
volumes providing most recent data, 19A6 through 1966);
"Survey of Current Business" (Washington, D. C.: United
States Department of Commerce).

89

data for both insured and uninsured banks with the intention
of correlating these groups to indicate the extent of their
similarity. The results of the correlations will be
reviewed later.

Certain problems presented themselves in the compila—
tion of the data. For example, for many of the earlier
years, the States of Alaska and Hawaii were not reported as
separate entities because they were not, as yet, separate
states. Figures for these areas were included with statis-
tics for other territories. It was necessary to estimate
the statistics for these areas from those figures which were
given. This was done by assuming that all the banks in the
group were of the same size. It meant, simply, that if
Alaska had two banks out of a total of ten for the terri—
tories, Alaska received twenty percent of the total statis—
tic reported. The same assumption was made for Hawaii.

For the earlier years of the study (19U6 through
19u9), statistics as to income were not available for
Alaska. The Governor General of the Territory estimated
that per capita income in Alaska was twenty-five percent
greater than that in the United States continental area.7
Based on this information, income for Alaska was computed
by first determining the per capita income for the contin-

ental United States for a specific year, multiplying it by

¥

7Governor General, Alaska Territory, Postwar Alaska
(Washington, D. C.: United States Department of the Inter-
ior, Division of Territories and Island Possessions, l9u9).

 

90

125 percent, and then multiplying this figure by the popula-
tion of the territory for that year.
NO such problem existed for Hawaii as separate figures

were available for each year for which they were desired.

Processing the Data

 

Completing the procedure outlined so far did not pre-
sent the data in proper form. It was necessary to compute a
set of data which would lead to the construction of the
regression equations desired. This was accomplished by
designing a computer program which would sum certain of the
figures recognizing the various sub-divisions necessary to
make the proper comparisons. The summed statistics were
called "Regression Statistics" and were constructed as
indicated below.

In the first of these regression statistics, the com—
puter was instructed to compute the sum of the following
items for each year of the study separately for each state
and to punch into cards the information so computed:

Currency & Coin

Reserves with the Federal Reserve

United States Banks Demand Balances

United States Banks Other Balances

Balances with Foreign Banks

Cash in'Process of Collection

This statistic was designated "Regression Statistic

LHJO, Cash Items."

 

91

Similar regression statistics were prepared for each

of the following concepts by combining the specific items

indicated:

Regression Statistic

 

A01, US Gov Securities

A02, Other Securities

A03, Valuation Reserve
AOA, Real Estate Loans

A05, Lns to Fin Snsts

A06, Lns to Brokers $ Dlrs

A07, Loans to Farmers

A08, Commercial & Ind Lns
A09, Pass Auto Loans
A10, Other Retail Cons Lns

A11, Res Repair & Modern
Lns

A12, Personal Expenditure
Lns

A13, Single Pmt Personal
Lns

AlA, All Other Loans

A15, Other Assets

A16, Demand Deposits, IPC

Items
US Gov Obligations
Obgl of Sts & Sub rm:
Sec of Fed Agen "
Other Bnds Nts Deb
Corp Stock

Val Res

Real Est Lns

 

Lns to Fn Banks
Lns Other Fin Inst

Lns Brokers Dlrs in See
Other Lns P&C Sec

CC Guaran Lns to Farm
Other Lns to Farmers

Comm & Ind Lns
Pass Auto Lns
Other Ret Cons Lns

Res Rep & Mod Lns
Prsl Exp Lns

Single Pmt Prsl Lns
All Other Loans

Fur Fixt Real Est !
Misc Assets

—-a-_

Ind Ft Corp Dem Dep

 

Regression Statistic Items
A17, Demand Deposits, Cert & Off Checks
Other S Gov Demand Deposits

Sts & Sub Dem Deposits
US Comm Bnks Dem Dep
US Mut Sv Dem Dep

Fn Govts Dem Dep

Bnks Fn Cty Dem Dep

A18, Time Deposits, IPC Ind Pt Corp Time Dep 4

A19, Time Deposits, Other US Gov Time Dep i
Sts & Sub Time Dep
US Comm Bnks Time Dep
US Mut Sv Time Dep
Postal Saving
Fn Govts Time Dep
Bnks Fn Cty Time Dep

LLZ_ .3131,“ .L

 

A20, Real Income Personal Income
(Actual dollar figures were converted to real
income by dividing Personal Income by Con-
sumer's Price Index.)

A21, Population Population
A22, Income on US Gov Sec Int US Gov Oblgns
A23, Income on Other Sec Int & Div Other Sec

Profit Sec Sold or Redeemed
Recoveries, Sec

Trfs from Res, Sec

Losses on Sec Sold
Charge—offs, Sec

Trfs to Res, Sec

A2A, Income on Lns & Disc Int & Disc, Loans
Sv Chg Fees Loans
Recoveries, Loans
Trfs fr Res, Loans
Losses & Charge-offs, Loans
Trf to Res, Loans

To provide totals of certain Regression Statistics 1

established above, certain of them were summed and desig-

‘ H.>*

nated as follows:

93

Regression Statistic

 

A25,

A26,

A27,

A28,

A29,

A30,

A31,

A32,

A33,

A3A,

A35,

A36,

A37,

Having reduced the basic data into sums,

Total Lns & Disc, Net

Total Assets

Total Demand Deposits

Total Time Deposits

Total Deposits

Income on US Gov Sec,
Other Sec, Lns & Disc

Income on Other Sec &
Lns & Disc Net

Income on US Gov Sec
& Lns & Disc Net

Income on US Gov Sec
& Other Securities

Total Assets minus
Cash Items

Total Assets minus
Cash Items and US Gov
Sec

Total Assets minus
Cash Items & Other Sec

Total Assets minus
Cash Items minus Lns &
Disc

Source

Regression Statistic

A03,

A07, A08, A09, A10, A11,

A12, A13, AlA
Regression Statistic
A01, A02, A15, A25

Regression Statistic
A17

Regression Statistic
A19
Regression Statistic
A29
Regression Statistic
A23, A2A
Regression Statistic
A2A

Regression Statistic

A2A

Regression Statistic

A23

Statistic
A25

Regression
A02, A15,

Regression Statistic
A15, A25

Regression Statistic

A15, A25

Regression Statistic
A02, A15

it was

A00,

A16,

A18,

A27,

A22,

A23,

A22,

A22,

A01,

A02,

A01,

A01,

HOW

gaossible to compute the specific elements of the regression

 

-H

9A

equations as outlined in Table IV-l, page 85. These com-
puted statistics were recorded on punched cards which repre—
sent the body of data used as input to the computer for pur—

poses of accomplishing the objectives of the study.

CHAPTER VI
THE REGRESSION RESULTS

In order to accomplish the purpose of this study as
stated in Chapter IV, The Theoretical Structure, the fol—

lowing regression equation was designed:

Yij = fIXlO , x2 , X3 , X“ , D1, 01(X10

) D (x )
Jojij 13' ’120’

J J

D (X , D X , D , D X , D X , D x ,
1~ 313') 1< “11) 2 2<1O,> 2< 20,) 2< 313)

J J
D (x )1.
2 Aij
Subscript 'o' = Applicable to all components.
Subscript 'i' = Specific component as indicated below.
Subscript 'j' = Specific year-~l9A6 through 1966.
D1 = 1 if state is classified Statewide
Branching.
0 if state is classified Limited
Branching or Unit Banking.
D2 = 1 if state is classified Limited

Branching.
0 if state is classified Statewide
Branching or Unit Banking

95

The

Ylj

The

(AYl

96

Equation (1): Cash Items

 

variables are defined as follows:

(Cash Items)/(Total Assets).

(Demand Deposits)/(Total Deposits).

(Real Income)/(Population).

Not applicable.

(Income, US Gov Sec, Other Sec, Lns & Disc)
(Total Assets minus Cash Items).

expected signs are as follows:

j)/(Axloj) > O.

(AYlj)/(AX20 ) 2 0.

J

(AYlj)/(Axulj) < o.

The

The

Equation (2): US Gov Sec

 

variables are defined as follows:

(US Gov Sec)/(Total Assets).

(Demand Deposits)/(Total Deposits).

(Real Income)/(Population).

(Income us Gov Sec)/(Total US Gov Sec)

 

 

7.4.. .0... .
.
‘1

 

(Income Other Sec & Lns & Disc)/(Total Assets
minus Cash Items and US Gov Sec)

Not applicable.

expected signs are as follows:

97

(Ang)/(AXle) < 0.

(AY2J)/(AX2OJ) § 0.

(AY2j)/(AX323) > 0.

The

X“3J

The

Equation (3): Other Sec

 

variables are definded as follows:

(Other Sec)/(Total Assets).

(Demand Deposits)/(Total Deposits).

 

(Real Income)/(Population).

(Income Other Sec)/(Total Other Sec)
(Income US Gov Sec & Lns & Disc)/(Total Assets
minus Cash Items and Other Sec)

Not applicable.

expected signs are as follows:

(AY3j)/(AX10J) < O.

(AY3j)/(AXQOJ) § 0.

(AY3J)/(AX33J) > O.

The

Equation (A): Lns & Disc

 

variables are defined as follows:

(Lns & Disc)/(Total Assets). |

(Demand Deposits)/(Total Deposits).

(Real Income)/(Population).

-l

1.‘

I“ V‘F
1:. 'I 1

98

(Income Lns & Disc)/(Tota1 Lns & Disc)

 

X3 . (Income US Gov Sec & Other Sec)/(Total Assets
“J minus Cash Items and Lns & Disc).
quJ = Not applicable.

The expected signs are as follows:

(AYuj)/(Axloj) < o. r

V
O

(AYuj)/(AX2OJ)

V
O

(AYu3)/(Ax3uj)

 

Cross Section Regressions fl:

 

Cross section regressions were run for each of the
twenty—one years of the study (19A6 through 1966) and for
the difference between the last year of the study (1966) and
the first year (19A6). The results of the regressions are
set forth in tables as follows:

Table VI-la: Regression Results, Equation (1): Cash
Items.

Table VI-lb: Regression Results, Equation (2): US
Gov Sec.

Table VI-lc: Regression Results, Equation (3): Other
Sec.

Table VI-ld: Regression Results, Equation (A): Lns &
Disc.

Each cross section regression was based on a total of
fifty-one observations (the fifty states plus District of (

Columbia). ,

99

Time Series Regressions

Time series regressions were run for each of the basic
equations for all years of the study (19A6 through 1966).

The results of the regressions are attached hereto as fol—

lows:
Table VI-2: Regression Results, All Equations, Time
Series.

Each time series regression was based on a total of
one thousand seventy-one observations (the fifty states

plus District of Columbia for twenty-one years).

Time Trend Regressions

Time trend regressions on the Y variables were run for

each of the basic equations for all the years of the study

(19A6 through 1966). ‘The results of the regressions are

attached hereto as follows:

Y Variable Time Trends, All Equations.

Table VI—3:

Each of these time trend regressions was based on a
total of one thousand and seventy—one observations (the

fifty states plus District of Columbia for twenty—one

years).

Correlation of Insured with Uninsured Banks,
Table VI-A

In Chapter V, Gathering the Data, arguments were pre—

sented for using the insured group of banks to represent the

In June, 196A, only 3.2 percent of

entire banking system.
were not insured with

the total number of commercial banks

 

-V'D—

100

the Federal Deposit Insurance Corporation. These uninsured
banks constituted 2.A percent of the Total Assets of Commer—
cial Banks. To strengthen the argument for using insured
banks to represent all banks, it was planned to run correla—
tions between these two groups for the years of the study on
the dependent variable to be used in the study. Such corre-
lations for the absolute dollar amounts in each of the four
components under consideration were run. Please see Table

VI-A for the results of these correlations.

 

[w
., ‘
“.

It is evident that these two groups are similar to
each other as to changes in the quantity of dollars invested
in Loans and Discounts only. In the case of this component,
the coefficient of covariance is positive and high——0.9119.
The two groups do not correlate well in the "cash items"
component. In the other two components, the similarity is
not strong.

That these two groups of banks did not correlate well
did not strengthen the argument to use insured banks to
represent all banks. However, because the main area of
analysis in testing the hypothesis was expected to be in
the Loans and Discounts component, it was decided to proceed
with the original plan of using the insured group of banks

to represent the entire banking system. l

Intra-system Comparison, Table VI-5

 

The use of dummy variables in the regressions makes

130ssib1e the determination of the extent, if any, to which

101

the Statewide Branching and Limited Branching systems are
significantly different from the Unit Banking system.

As to Equation (1) Cash Items, in the cross-section
regressions, (Table VI-5a), the Statewide Branching system
(represented by Dummy 1) is revealed to be significantly
different from the Unit Banking system in twelve out of the
twenty—two separate cases—-or in 5A.5 percent of the exam—
ples tested. Because the hypothesis of this study stated
that there was no difference between the three systems, in
the case of the Statewide Branching system versus the Unit
Banking system as regards the extent to which they devote
investment funds to the liquidity component of their port—
folios given the same motivation to do so, the hypothesis is
rejected--there is a significant difference between these
two systems in the manner in which they respond to similar
motivations for devoting investment funds to the "Cash
Items" component.

In the "US Gov Sec” and "Lns & Disc" components, the
Statewide Branching systems is significantly different from
the Unit Banking system only five (22.7%) and four (18.2%)
times respectively out of a total of twenty-two examples.
While this is still greater than would be expected within
the limits of chance (1.1 times out of 22 if significant at
5.0 percent or 2.2 times out of 22 if significant at 10.0

percent), the argument that the Statewide Branching system

 

102

is different from the Unit Banking system is not over-
whelmingly conclusive.

In the "Other Sec" component, there is no year in
which the Statewide Branching system is significantly dif-
ferent from the Unit Banking system. Thus, in this category,
the hypothesis of this study is accepted.

As to the Limited Branching system, no significant
difference exists between this system and the Unit Banking
system in any of the four categories under consideration.
Any difference observed is due purely to chance.

In the time series regressions (Table VI—5b), the
Statewide Branching system is significantly different from
the Unit Banking system as to Equation (1), Cash Items, and
as to Equation (3), Other Sec. It is not significantly dif-
ferent as to Equation (2), US Gov Sec, or as to Equation
(A), Lns & Disc. The Limited Branching system is signifi-
cantly different from the Unit Banking system as to Equation
(1), Cash Items; Equation (3), Other Sec; and Equation (A),
Lns & Disc. The Limited Branching system is not signifi-
cantly different from the Unit Banking system as to Equation
(2), US Gov Sec.

Significance of Regression Coefficients,
Table VI-6

 

In Tables VI—6a and VI-6b is set forth an analysis of
the significance of the regression coefficients in the cross

Section and time series regressions respectively. In the

 

103

cross section analysis, it is noted that, except in the case
of Equation (1), Cash Items, very few of the regression
coefficients are significant at the ten percent level a suf—
ficient number of times to allow any serious analytical work
to be done on the interpretation of the relative magnitude
of those coefficients in their relationship to the deter—
mined variable.

1. In the case of Equation (1), Cash Items, variable
X1, (Demand Deposits)/(Total Deposits), is significant in
determining the proportion of a unit banker's portfolio
devoted to cash items twenty—one times out of twenty-two at
the five percent level. The influence of variable X2, (Real
Income)/(Population), is not important in this respect as it
was significant at the ten percent level in only two out of
twenty—two regressions. Variable X4, (Total Portfolio
Income)/(Total Assets less Cash Items), was significant at
the ten percent level nine out of the twenty—two times indi—
cating that unit bankers are motivated by profit considera—
tions in the prOportion of their portfolios devoted to
liquidity requirements.

In the Statewide Branching system, variable X1,
(Demand Deposits)/(Total Deposits), was significant at the
ten percent level fourteen times out of twenty-two; variable
X2, (Real Income)/(Population), was significant five times
out of twenty-two; and variable XA’ (Total Portfolio Income)

/(Total Assets less Cash Items), was significant at the ten

10A

percent level eight times out of twenty—two. In this cate-
gory, it would appear that the Statewide Branching system

is significantly motivated by the selected variables and
that it is motivated significantly differently than is the
Unit Banking system in the extent to which these variables
determine the proportion of the portfolio devoted to liquid-
ity considerations.

None of the variables is significant at the ten per—
cent level in the case of the Limited Branching system indi-
cating that this system is not significantly different from
the Unit Banking system in the extent to which these vari—
ables determine the proportion of the portfolio devoted to
this category.

2. In Equation (2), US Gov Sec, in the Unit Banking
system, none of the variables is significant a sufficient
number of times to indicate that these selected variables
are instrumental in the determination of the proportion of
the investment portfolio devoted to this component. In the
case of the Statewide Branching system, only variable X1,
(Demand Deposits)/(Total Deposits), is significant at the
ten percent level and it is significant at this level nine
out of the twenty-two times. In the Limited Branching sys-
tem, the profit variable X3 is significant at the ten per-
cent level four times out of twenty-two.

3. In Equation (3), Other Sec, the only variable

which is at all significant in the determination of the

105

proportion of the portfolio devoted to Other Securities is
in the Statewide Branching system and is the profit vari—
able (X3). It is significant at the ten percent level only
three times out of twenty—two.

A. In Equation (A), Lns & Disc, in the Unit Banking
system, only the profit variable (X3) is significant in the
determination of the proportion of the portfolio devoted to
this component. It proved to be significant at the ten per—
cent level only four times out of twenty-two. In the State—
wide Branching system, only variable (Xl), (Demand Deposits)
/(Total Deposits), proved to be significant a sufficient
number of times to be considered meaningful. It was signif-
icant at the ten percent level only three times out of
twenty—two. There were no variables in the Limited Branch-
ing system which proved to be significant determinators of
the proportion of the portfolio devoted to Loans and Dis-
counts.

5. The significance of the regression coefficients
in the time series regressions (Table VI-6b) is somewhat
more consistent. The only surprising elements are that the
Unit Banking system is not significantly motivated by profit
considerations (Variable X4) in the determination of the
proportion of its investment portfolio devoted to Cash
Items, that the Statewide Branching system is not signifi—
cantly motivated to the determination of the proportion of

its portfolio devoted to United States Government Securities

106

or to Loans and Discounts by liquidity considerations (Var-
iable X1), and that the Limited Branching system is not
motivated by liquidity considerations (Variable X1) in
Equation (2), US Gov Sec, nor by profit considerations (Var—
iable XA) in Equation (3), Other Sec, or in Equation (A),
Lns & Disc. The variable providing for real income per
capita (X2) is erratic throughout the study.

The Signs of the Regression Coefficients,
Table VI—7

 

Herein is set forth an analysis of the number of
times that the signs of the regression coefficients appear
as expected. In the case of variable X2, (Real Income/
(POpulation), a specific sign could not be anticipated.
Therefore, both signs and the number of times they appear
are given, Also, the percentage that each represents of a
total of twenty-two examples is also given.

1. In the cross section regressions (Table VI—7a),
it can be noted that the Unit Banking system responds con-
sistently positively to changing reserve requirements (Var—
iable X1 in the proportion of its portfolio devoted to Cash
Items. Neither the Statewide Branching nor the Limited
Branching systems are similarly motivated. The relationship
of changing real income per capita (Variable X2) to changing
liquidity needs is shown to be significantly positive in the
case of the Unit Banking system and in the case of the

Statewide Branching system. The Unit Banking system

107

responds positively to changes in real income per capita
seventeen times out of twenty-two while the Statewide
Branching system responds positively twenty-one times out
of twenty—two. In contrast, the Limited Branching system
responds negatively twenty-one times out of twenty-two to
such forces in the proportion of its portfolio devoted to
liquidity considerations.

Surprisingly, the Unit Banking system does not respond
as expected (negatively) to changes in profit available on
alternative components. In only two out of twenty-two exam-
ples did it indicate the expected relationship. In con-
trast, the Statewide Branching and Limited Branching systems
respond as expected nineteen and twenty times, respectively,
out of a total of twenty-two examples.

2. In Equation (2), the Unit Banking system adjusts
the proportion of its portfolio devoted to United States
Government Securities negatively as its reserve needs change
in twenty-one out of twenty—two examples. The Statewide
Branching system does not respond to such changes as expec-
ted--in fact, in only one example out of twenty-two is it
indicated that these variables are negatively related. The
Limited Branching system reacts somewhat more as expected
in that it does respond negatively eight out of a possible
twenty-two times.

All three systems indicated a preponderance of posi-

tive relationships as to variable (X2), (Real Income)/

108

(Population). The Unit Banking system responded positively
thirteen times out of twenty-two; the Statewide Branching
system responded positively twenty times out of twenty—two;
and the Limited Branching system responded positively four-
teen times out of twenty—two.

Variable X3, (Income this Component)/(Income alterna—
tive components), is most often negatively related in all
three systems as regards the proportion of the portfolio
devoted to United States Government Securities. This is
not as expected in that, as the return on this component
increases (the numerator), the value of this ratio increases
and it is expected that the quantity of investment funds
devoted to this component would increase. As the return on
alternative components increases (the denominator), the
value of this ratio decreases and it is expected that the
quantity of investment funds devoted to this component
would decrease. In other words, the Y variable is posi-
tively related to this profit variable.

3. As to Equation (3), Other Sec, the Unit Banking
system and the Limited Branching system respond as expected
(negatively) to variable X1, (Demand Deposits)/(Total Depos—
its), thirteen and twenty-one times, respectively, out of a
possible twenty—two. The Statewide Branching system indi-
cates a negative relationship only nine times out of twenty-

two.

109

The Statewide Branching system increases the prOpor-
tion of its portfolio devoted to the "Other Securities" com-
ponent as Real Income per capita increases in nineteen out
of twenty—two examples while the Limited Branching system
responds similarly in seven out of twenty—two examples. In
contrast, the Unit Banking system responds in this manner
only one time out of the twenty-two examples.

The response to the profit variable (X3) is quite
erratic in the case of Other Securities. As the Unit
Banking system responds positively only thirteen times out
of twenty-two, the Statewide Branching system responds posi-
tively only nine times out of twenty—two, and the Limited
Branching system responds positively only ten times out of
twenty-two. Only the Unit Banking system responds as
expected a majority of times.

A. In Equation (A), Lns & Disc, the Unit Banking sys-
tem and the Statewide Branching system respond as expected
(negatively) to changes in liquidity needs (X1). They
respond negatively seventeen and nineteen times out of
twenty-two, respectively. The Limited Branching system
indicates such a negative relationship only three times out
of twenty-two examples.

The real income per capita variable (X2) reflects a
positive relationship thirteen times in the case of the
Unit Banking system and eleven times in the case of the

Limited Branching system. This variable is positively

110

related only four times out of twenty-two in the case of the
Statewide Branching system.

The profit variable (X3) in the case of Equation (A),
Lns & Disc, reflects the expected relationship (positive)
only three times out of twenty—two in the case of the Unit
Banking system. This variable is positive eleven times out
of twenty-two in the case of the Statewide Branching system
and seventeen times out of twenty-two in the case of the
Limited Branching system.

5. In the time series regressions (Table VI-7b), the
signs of all the coefficients in Equation (1), Cash Items,
are as expected except for Variable X1 (Demand Deposits)/
(Total Deposits), in the cases of the Statewide and Limited
Branching systems.

In Equation (2), US Gov Sec, the sign of variable (X1),
(Demand Deposits)/(Total Deposits), is not as expected in
the case of the Unit Banking system and in the Statewide
Branching system. As to the income variable (X2), the Unit
Banking system reflects a negative relationship while both
isranched systems reflect a positive relationship. Only the
Lindted Branching system reflects the expected sign (posi—
tive) as regards the profit variable.

The "Other Securities" component [Equation (3)]
reflects somewhat the same erratic pattern as does Equation
(2). The Statewide Branching system does not respond as

emeicted (negatively) to the need for liquidity (Variable

111

X1 while the Limited Branching system does respond nega—
tively to changes in this variable. The income per capita
variable (X2) elicits a positive response in the case of
the Unit Banking system and the Statewide Branching system
while bringing forth a negative response in the Limited
Branching system. Only the Limited Branching system
responds positively to the profit variable (X3).

In Equation (A), Lns & Disc, the Unit Banking system
and the Statewide Branching system reflect the expected
sign (negative) as to the relationship of the liquidity var—
iable (X1) to this component of the portfolio. The income
variable (X2) reflects a positive relationship in the case
of the Unit Banking system and a negative relationship in
the case of the Statewide Branching system and the Limited
Branching system. The profit variable (X3) has the expected
sign (positive) in the case of the Statewide Branching sys-
tem and the Limited Branching system while reflecting a
negative sign in the case of the Unit Banking system.

It is evident that the regression equations, being
reduced form equations, are reflecting problems of identi-
fication which give rise to the erratic nature of the signs
and to the low number of variables which are found to be

Significant determinators of the dependent variable.

Estimated Values of the Y Variable, Tables VI-8
Tables VI-8a through 8e set forth estimated values of

the Y variable for each of the Equations under consideration.

112

These are first computed for the cross section equations for
each year of the study. They are then computed using the
time series regressions as a basis for the estimation.

1. Table VI-8a sets forth the estimated values of Y,
the proportion of the portfolio devoted to cash items, as
computed from regression Equation (1), Cash Items. The
estimated values of Y are computed at the mean values of the
X variables. A significant consistency is revealed in this
computation. The estimated value of the Y variable is less
for the Statewide Branching system than it is for the Unit
Banking system for each of the twenty-one years. The aver-
age of the estimated Y values for these twenty-one years for
the Statewide Branching system is 0.1956 as compared with
0.2075 for the Unit Banking system or is 5.73 percent lower.
The average for the Statewide Branching system (0.1956) is
7.30 percent lower than the average of the Limited Branching
system (0.2110). The Limited Branching system had as esti-
mated Y value higher than the Unit Banking system for all
but four of the twenty-one years.

The fact that the Statewide Branching system is so
consistently less than either of the other two systems, and
especially less than the Unit Banking system, is strong evi-
dence that the Statewide Branching system is significantly
different from the Unit Banking system in the proportion of
its portfolio it would devote to liquidity needs as a result

of the motivations represented in the variables selected for

113

the study. Not only is the Statewide Branching system dif-
ferent from the Unit Banking system in this most signifi-
cant manner, but it is different in a manner which enables
it to make a more significant contribution to economic
activity as it devotes the portion of its investment funds
not committed to liquidity needs to the remaining components
of its portfolio.

2. As to Equation (2), US Gov Sec, (Table VI-8b) the
Statewide Branching system reflects a lower estimated Y
value than the Unit Banking system in all years but one
(19A9). Similarly, the Limited Branching system reflects
a lower estimated Y value than does the Unit Banking system
in all years but one (19A9). The average estimated Y value
for the Statewide Branching system (0.3003) is 10.55 percent
lower than the Unit Banking system's average estimated Y
value (0.3357). The Limited Branching system reflects an
average estimated Y value (0.3196) which is A.80 percent
lower than the Unit Banking system's average.

3. In the "Other Securities" component of the port—
folio (Table VI-8c), the Statewide Branching system is
experiencing an estimated Y value which is higher or equal
to that of the Unit Banking system for fourteen of the
twenty-one years. The Limited Branching system experienced
a higher’ estimated Y value than did the Unit Banking system
for thirteen of the twenty-one years. The average of the Y

values for the Statewide Branching system (0.0807) was 3.86

11A

percent higher than that of the Unit Banking system
(0.0777). It was higher by 2.02 percent than the average

Y value of the Limited Branching system (0.0791). The aver-
age of the Y values of the Limited Branching system was 1.80
percent higher than the average of the estimated Y values
for the Unit Banking system.

A. In the most significant component of the portfolio
in terms of the intent and purpose of this study, Equation
(A), Loans and Discounts, the Statewide Branching system
reflects a higher estimated Y value than the Unit Banking
system for each of the twenty-one years of the study. Its
average estimated Y value (0.A06A) was 12.1 percent higher
than the average estimated Y value of the Unit Banking sys-
tem (0.3625) and was 7.88 percent higher than the average
estimated Y value for the Limited Branching system (0.3767).

5. Table VI—8e presents the estimated Y values for
all equations based on the time series regressions. Sur—
prisingly, both of the Branched systems reflect a lower
estimated Y value than does the Unit Banking system in
Equation (3), Other Sec, and (A), Lns & Disc. The differ-

ences, however, are very slight.

Y Variable Time Trends, Table VI—9

 

Time trend regressions were run for the Y variable of
each of the four equations for the twenty—one years of the
Study. The regression coefficients, standard errors, and

Significances are presented in Table VI—3 above.

115

A most revealing way of presenting the relative rela-
tionship of each of the banking systems to each of the other
systems is to compute separate or isolated intercepts and
slopes for these time trend lines. The isolate intercepts
and slopes are presented in Table VI-9.

1. As to Equation (1), Cash Items, it is revealed
that the time trend line for the Statewide Branching system
lies significantly lower on the Y axis (Proportion of port—
folio devoted to Cash Items) than does either of the other
two systems. The slopes of all three time trend lines are
negative with the Unit Banking system having the steepest
slope. The estimated values of the Y variable at the mean

of the T variable are as follows:

Statewide Branching system: 0.1856
Limited Branching system: 0.2lAA
Unit Banking system: 0.2235

2. As to Equation (2), US Gov Sec, the intercept for
the Statewide Branching system is only slightly higher than
that for the Unit Banking system. The Statewide Branching
system's time trend line is declining with a steeper slope
than is that of the Unit Banking system. The Limited
Branching system has a lower intercept than the other two
systems and its slope is declining at a rate only slightly
less than that of the Unit Banking system. The estimated
values of the Y variable at the mean of the T variable are

as follows:

116

Statewide Branching system: 0.3002
Limited Branching system: 0.3125
Unit Banking system: 0.32A5

3. Equation (3), Other Sec, shows intercepts and
lepes very similar in all systems. The slopes are positive
in all cases with the Limited Branching system having the
highest intercept and a slope less steep than the Unit
Banking system. The slopes of the Statewide Branching sys—
tem and the Unit Banking system are so similar as to make
the time trend lines, to all practical purposes, parallel.
The estimated values of the Y variable at the mean of the T

variable are as follows:

Statewide Branching system: 0.0820
Limited Branching system: 0.081A
Unit Banking system: 0.080A

A. In Equation (A), Lns & Disc, the Statewide Branching
system has the highest intercept of the three systems with
the Unit Banking system having the lowest intercept. The
slope of the time trend line is largest in the case of the
Statewide Branching system and is lowest in the case of the
Limited Branching system. It would appear, in accordance
with these time trend lines that the Statewide Branching
system is devoting a larger proportion of its portfolio to
the Loans and Discount component and that the proportion of

its portfolio so allocated is increasing at a faster rate

117

over time than is that proportion increasing in the other
two systems. The estimated values of the Y variable at the
mean of the T variable are as follows:

Statewide Branching system: 0.A1A6

Limited Branching system: 0.3771

Unit Banking system: 0.3595

TABLE VI-la:

Regression Results,

Equation (1):

11.8

.Cash Items, Cross Section

 

 

 

 

Period Variables
Constant X1 X2 X1; (D1) (D1)(X1)
19A6 -0.18163161 0.A2A0689Aa 0.00001690 3.A202AAA5 0.5A56980Aa —0.A2699859a
(0.1532863?) (0.107A1A20) (0.0000323?) (2.50057785) (0.22283779) (0.1A62AA65)
19A? —0.15818309 0.3?208616a 0.00000325 5.29513511a 0.5A3A2203a -0.3A523355a
(0.1A93A675) (0.0967111?) (0.0000339A) (2.5285350?) (0.23838282) (0.1A2281AA)
19A8 -0.205A0283 0.352998883 0.00002195 6.501527538 0.A3A?A390a -o.2598A855a
(0.1532939?) (0.09A02058) (0.00003311) (2.8389AA?1) (0.1961880?) (0.12051215)
19A9 —0.35065311a 0.396758A9a 0.0000A656 8.327A2723a 0.699531613 -0.363113?98
(0.09A30291) (0.08985005) (0.000028A2) (1.2?605593) (0.179029AA) (0.11898955)
1950 -0.199A6321 0.37A13290a 0.00001A88 A.98839863b 0.20A029A8 -0.15853119
(0.1A093728) (0.08867699) (0.000031?9) (2.8782196A) (0.15909333) (0 10110608)
1951 -0.26A605AA 0.A0095827a 0.00002138 6.231615053 0.33090A77b -0.20798857
(0.1582A063) (0.10906230) (0.00002960) (3.01887A75) (0.18375695) (0 12A87A90)
1952 —0.22598580b 0.A15A9816a 0.00000868 A.36A58601b 0.287392A3b -0.229837298
(0.12119021) (0.08600A8A) (0.0000232A) (2.AA651316) (0.1AA830A6) (0.100898A2)
1953 -o.AA93?825a 0.A71A8091a 0.0000A193b 8.0983A88Aa 0.6791575?a -0.3A251560a
(0.15175300) (0.09027186) (0.00002A6A) (2.959388A8) (0.16691696) (0.10A29677)
195A —o.23526610 0.38289295a 0.00001065 A.53A53A82 0.AA660151a -0.22810759a
(0.1588602?) (0.09190906) (0.000025A0) (2.8653936?) (0.1732A807) (0.10588A98)
1955 -o.26771505 0.A201A516a 0.00002761 3.75A8199A 0.A1A61538b -0.270A3A69a
(0.181A829?) (0.11051353) (0.00002936) (3.25296992) (0.208AA261) (0.12802285)
1956 -0.36123377a 0.AAA068??a 0.00003991b 5.2966910?a 0.57331985a —o.2?282679a
(0.1A5216AA) (0.09652378) (0.0000233A) (2.51859953) (0.1?3A8202) (0.11159213)
1957 -o.300A38?8a 0.39109201a 0.0000206? 5.2A2A535Aa 0.AA80653Aa -0.23629292a
(0.1A0986AA) (0.09055765) (0.00002206) (2.56366738) (0.159003A?) (0.10A61220)
1958 —o.18860611 0.36185856a -0.0000033A 1.81061AA3 0.27573051 -o.1?619356
(0.17637915) (0.10168270) (0.00002192) (2 75002719) (0.19313002) (0.1131033?)
1959 -0.0A668583 0.35967A73a -0.00000A59 -0.0101AO?6 0.12681856 -0.15968266
(0.13091995) (0.0807A739) (0.00002052) (2.39153A3A) (0 15756215) (0.09505082)
1960 -0.11121818 0.392258808 -0.0DOOOAA? 1.0889A8A9 0.1AA15676 -0.1?081A17
(0.160220?8) (0.0907967?) (0.00002055) (2.58107596) (0.18A52332) (0.1035821A)
1961 0.125A8953 0288914522a 0.00000388 -2.9779AAA8 0 0366631? —0.112?5?23
(0.2A87001A) (0.11650221) (0.0000206A) (A.0159?923) (0.27679906) (0.130A62A3)
1962 -o.05A06051 0.3100851?a -0.000013A6 1.302A2180 0.0A950762 -0.16667708
(0 1653A789) (0.10AA5935) (0.000021A7) (2.67529A35) (0.19625550) (0.11821298)
1963 -o.20271872 0.36566910a 0.00001A16 2.A1973293 0.3626A789a —o.1855??92b
(0.13807A2A) (0.09171556) (0.00001853) (2 A3360291) (0.16A52190) (0.1032A085)
196A -0.0851??25 0.A0173385a 0.00000505 0.053219A1 0.159130A5 -0.2126031?b
(0.1582157A) (0 10172213) (0.00002152) (2.2235A3A3) (0.1881809A) (0.112913A9)
1965 —0.069231?9 0.381A3502a —0.00000296 0 38355635 0.213695A8 -0.2285637Aa
(0.1A23221A) (0.08921120) (0 00001600) (2.1988A1A1) (0.16096009) (0.099716A3)
1966 -0.19253?95 0.A07AA273a 0.00000362 2.A0599A01 0.26520755 -0.21119300b
(0.1?869235) (0.10111A1A) (0.0000153A) (2.66A20356) (0.201A60A8) (0.11199718)
66 less —0.0?639536 0.23569000 0.00302935 0.2A063792 0.09052196 -0.2369AA59
A6 (0.0739938?) (0.1581A753) (0.0000A31?) (2.12218275) (0.19A61201) (0.1?A8389A)
Note: Superscript '"' indicates variable is significant at 5.0 per cent.
Superscript '0' indicates variable is significant at 10.0 per cent.

Standard errors of coefficients

are indicated in

brackets.

119

 

Variables

“I- ---33 .'!‘I

Regression Adv!c(:

 

 

(01)(X2) (01)(xu) (02) (02)(x1) (Dg)(X2) (Dg)(Xu) R BAR 2 Std Error
-0.0000196? -9.AA838A12a 0.32807319 -0.223AAA63b -0.000053?9 -3.21316102 0.5A38 0.028637A6
(0.00003A35) (A.A902A799) (0.19838966) (0.12900611) (0.00003638) (A.A9812A52)

-0.00001639 -11.?3583A90a 0.21171279 -0.15A6A699 -0.00003010 -2.21520817 0.6027 0.02566933
(0.00003818) (A.9312165?) (0.186177A3) (0.115679A8) (0.0000A001) (3.7350A119)

-0.00002506 —8.89286AA5a 0.2092196? -0.111A9302 -0.00003762 -2.A?61756A 0.5725 0.02615632
(0.00003665) (3.99522326) (0.19871195) (0.11372238) (0.0000396?) (A.39773986)
-0.0000?171a-12.8AA?090Aa 0.136823A0 -0.1A2A87A3 -0.00001597 -0.09A91536 0.7573 0.02658168
(0.00003A36) (3.A9AA3290) (0.11585178) (0.1112A369) (0.000032A7) (1.63011376)

-0.00002139 -2.60A38A01 0.1879602? -0.13638A59 -0.00002A96 ~1.61588613 0.6690 0.02AA3268
(0.00003A83) (3.3A858106) (0.1??65663) (0.11095368) (0.0000375?) (3.?6652066)

-0.00001163 -5.87068869 0 2530A987 -0.122879A3 -0.00003519 —3.38A27109 0.55A8 0.027760A0
(0.00003311) (3.?1565326) (0.20179773) (0.13A261A8) (0.000036A2) (A.08931285)

-0.000013A3 -3.A31?0A10 0.16102875 -0.1?627282 -0.000006?7 -0.1258320A 0.6390 0 0228726A
(0.0000266A) (3.02A65792) (0.1631A381) (0.1095019?) (0 00002906) (3.A9A05882)
-0.00006397a-10.2A326883a 0.A3313389a -0.22660228b —0.0000A97? -5.A001661? 0.6A79 0.02276208
(0.0000276A) (3.26A10A90) (0.1901A791) (0.11395502) (0.00003002) (3.81629366)

-0.0000A?2A -6.10505008b 0.2058708? -0.1092?196 -0.00002891 -1.9769A596 0.6A22 0.023078A0
(0.00002881) (3.1286A039) (0.21201579) (0.11833??A) (0.00003106) (3.86561915)
-0.00005855b -3.7A867A8A 0.15727588 -0.11600?31 -0.00002?91 -0.22525A61 0.5588 0.02619269
(0.00003285) (3.8?289106) (0.2386A680) (0.13839110) (0.00003A82) (A.563676A6)

«0.000067563 -7.A08A22A3a 0.3611A8A6b -0.1879A683 -0.0000A958b -3.603398A0 0.6185 0.0216231A
(0.00002603) (3.20A91A98) (0.19A60606) (0.11699088) (0.00002756) (3.75956730)

-0.0000A708b -5.A5A09287b 0.2?276AA1 -0 1A1580A5 -0.000031A3 -2.9125?212 0.6185 0.0212360A
(0.00002502) (2.9226550A) (0.183680A9) (0.11036806) (0.00002592) (3.A9990779)

—0.00002A13 —2.89360163 0.109A79A6 -0.11A76312 -0.00000985 —0.15193965 0.6000 0.021A5069
(0.00002537) (3.20697A78) (0.2118587?) (0.1191982A) (0.00002568) (3.5881335?)

-0.0000186A 0.A1A23293 0.08901200 -0.109836A9 —0.00001000) 0.3A0025A3 0.6025 0.02089203
(0.00002375) (2.91392275) (0.173086A0) (0.09993620) (0.00002371) (3.A12366?1)
—0.0000106A —0.39175813 0.1A200255 -O.12805666 -0.000009A8 -0.?3600115 0.6216 0.0219A399
(0.00002386) (3 0AAA0011) (0.192A29AA) (0.11063965) (0.00002AA8) (3.25033811)

-0.00001209 1.11562383 -0.1533A383 0.00332220 -0.00000239 3.65082958 0.5879 0.022A3227
(0.0000230?) (A.5A3598A6) (0.30588856) (0.1A15333A) (0.0000251?) (A.95865359)

0.00000160 0.80229957 0.177A893A -0.0A898898 -0.00000238 -3.11630553 0.A952 0.0218A1A5
(0.00002A67) (3.18111015) (0.2103603?) (0.12265A65) (0.00002A7A) (3 771563A2)

-0.00002286 —A.618A3013 0.26617666 -0.1A79153A —0.00002661 -2.59890108 0.586A 0.01903955
(0.00002119) (2.85A60288) (0.20076765) (0.1071678?) (0.00002189) (3.70359991)

-0.00000010 -0.836A9122 0.23069230 -0.1A01A395 -0.00001A82 —2.52567589 0.5210 0.0201A691
(0.00002A0?) (2.93676090) (0.19629086) (0.11838576) (0.00002385) (3.190A5166)

-0.00000502 —1.?A02917A 0 283A3662 -0.1387A7A3 -0.000011A3 —3.871?5200 0.6162 0.01787191
(0.00001860) (2.55A73350) (0.18959733) (0.10532522) (0.00001879) (3.08735165)

-0.00000386 -2.9?920?51 0 23191631 -0.129A2956 —0.00000062 -A.18531590 0.5573 0.113A623?
(0.00001810) (3.1719551?) (0.2139022A) (0.11686671) (0.000017A2) (3.36280320)

-0.00001868 -3.55767985 0.13921713 —0.06595577 0.00000315 -5.51802835 0.3A39 0.02A5A1f5
(0.0000A650) (2.86A93338) (0.11197865) (0.2080A262) (0.00006569) (3.95A56725)

 

1.2(3

 

 

 

 

TABLE VI-lb: Regression Results, Equation (2): US Gov Sec, Cross Section
Period Variables
Constant x1 x2 x3 (01) (Dl)(x1)
19A6 0.75535386a -0.300A6A08 0.00005803 -0.22A5?899 —0.A355936?b 0.A8?388?8a
(0.19768675) (0.181865A7) (0.000060A?) (0.187A8091) (0.2A653A96) (0.21051720)
19A7 0.52887A01a —0.18A?922A 0.0001A7AAa -0.28861510 -0.372068A0 0.3765A21A
(0.2AA3878A) (0.21609392) (0.00005976) (0.18631358) (0.30399918) (0.2A522513)
19A8 0.3880A05Aa —0.1A?30321 0 00005106 0.13A26736 -o.2??23123 0.29A01560
(0.19071709) (0.1767A3l7) (0.00005622) (0.13790727) (0.22578865) (0.2050A969)
19A9 0.5101A301a —0.20501A67 0 00006525 -0.060522A3a -0.321?0193 0.31609869
(0.182573A7) (0.17569222) (0.00005535) (0.01A539A9) (0.22650905) (0.20A72093)
1950 0.6023A03?a -0.28875973 0.00003A53 -0.07375373 -0.3A8A?293 0.3A213A19
(0.21200752) (0.1990A629) (0.00006066) (0.21589271) (0.25038570) (0.221A8137)
1951 0.A9996623a -0.19A09?11 -0.00001258 0.01363A5A -0.1963?580 0.2A8A2A23
(0.20615956) (0.19105AA9) (0.0000AA51) (0.16677986) (0.2A283853) (0.2165A556)
1952 0.63658265a -o.329A3085b 0 00002083 0.0?5A813? -0.2998A113 0.39933660b
(0.18235150) (0.16909112) (0.0000A501) (0.22026162) (0.22236A15) (0.198650A?)
1953 0.AA773359a —0.22A29AA3 0.0000301? 0.0718036? -0.15888217 0.32592800
(0.1927657?) (0.1678A623) (0.0000A21A) (0.17022778) (0.22835915) (0.19A63707)
195A 0.5A935671a -0.25635272 0 00002918 -o.09339666 -0.19886355 0.28A25A36
(0.1650616A) (0.1675A933) (0.0000A258) (0.1993981A (0.193956A8) (0.19A5A866)
1955 0.50669021a -0.25106605 0.00001360 -0 02329089 -0.15163688 0 31162A91
(0.187738A0) (0.1821083A) (0.0000A671) (0.1A186618) (0.21968182) (0.20981515)
1956 0.A786?32Aa -0.1)1225A0 0.00000635 -0.05398A59 -O.1A658310 0.26050265
(0.226A7780) (0.2065673A) (0.0000A8A8) (0.16809691) (0.26819703) (0.23668172)
1957 0.A803A69Aa —0.20A8?AA3 0.00000232 —0.0AA672A8 -0.1652865A 0.27576AOA
(0.21351605) (0.19A13718) (0.0000AA87) (0.20723631) (0.2A180001) (0.22373636)
1958 0.A2153592a -0.225258?7 0.000027A0 0.00A15963 -0.09A9302? 0.26035085
(0.1?082803) (0.1?611821) (0.0000A213) (0.1983338?) (0.2000A310) (0.20050936)
1959 0.A?7838A9a -0.219307AA —0.00001552 -0.003A9963 —0.20781006 0.27779A07
(0.1?0AA650) (0 16503853) (0.0000A175) (0.1A113829) (0.20338199) (0.19A92558)
1960 0.A2283002a —0.22A5A02A 0.0000053A -0.006203A3 -0.195869A1 0.3069A109b
(0.1A99A766) (0.15131331) (0.00003659) (0 110A2A72) (0.18091273) (0.1?A153A7)
1961 0.557559398 —0.28239A03 -0.00001863 —0.09191868 -0.36610318a 0.35091035b
(0.12701582) (0.1?23A059) (0.00003011) (0.1568288A) (0.16792519) (0.1926159?)
1962 0.A955951?a —0.2917A?0? -0.000010?3 -0.030793?1 .0.23008017 0.A18A3899a
(0.1525A79A) (0.1?A00513) (0.000035A?) (0.119033AA) (0.1?566631) (0.19293505)
1963 0.510392913 -0.2731A29A -0.00002??2 -0.051A8763 -0.3171?011b 0.A0931?A1b
(0.1669365A) (0.18320635) (0.00003695) (0.1A002A5A) (0.18598990) (0.206975A0)
196A 0.A510AA1Aa -0 237A7117 -0.00001808 —0.0579339A —0.1195371A 0.AAA3A87Ba
(0.1769667?) (0.1811A069) (0.0000368A) (0.13010588) (0.22A9A10?) (0 20A9119A)
1965 0.A15526A0a -0.2?059389 —0.00002822 0.0185775A -0.32051?96b 0.3903AAA9b
(0.1A928965) (0.18A252A8) (0.00003AA9) (0 1663AA36) (0.1?506939) (0.21387998)
1966 0.A1A22721a -0.27)79)72 —0.00093126 0.0115639A —0.31633001b 0.361A000Ab
(0.1591953?) (0.185AA931) (0.0000292?) (0.12206252) (0.1665A806) (0.21192A73)
66 less -0.30?52A0)a 0.0?266625 —0.00000118 0 00758556 -0.02702805 0.10526850
A6 (0.1A253386) (0.32712A6A) (0.00010010) (0.18398275) (0.1A75695A) (0.3?0106AA)
Note: Superscript 'a' indicates variable is significant at 5.0 per cent.

Superscript 'r

Standard errors of coefficients

J

indicates variable is significant at 10.0 per cent.

are indicated in brackets.

121

 

 

 

Variables Regression Advice:
(D1)(X2) (Dl)(Xu) (Dz) (D2)(X1) (Dg)(X2) (D2)(Xu) R BAR 2 Std Error
-0.00003158 0.251300A2 -0.11119556 0.09317989 0.000019A5 -o.03529622 0.2A0A 0.051A1A05
(0.0000631A) (0.2?011358) (0.2A077132) (0.22081696) (0.00006723) (0.26577732)
-0.00008A58 0.A191268? -0.003A78?6 0 02A051A8 -0.00002831 0.01965900 0.2330 0.0566A285
(0.00006971) (0.30196900) (0 28265320) (0.25A65285) (0.00007562) (0.251109A5)
0.0000086A 0.05568876 0.111563A3 -0.0?A65218 0.00003915 -0.26392915 0.2515 0.050262A3
(0.00006395) (0.19799613) (0.232A6366) (0.21398635) (0.00006869) (0.23502911)
0.00000315 0.16627905 0.00059A1A -0 02989780 -0.000000?0 0.0370A553a 0.6315 0.0508A522
(0.00006A31) (0.22185672) (0.21739693) (0.2157A922) (0.00006A06) (0.01509623)
0.0000A3A5 -0.03623555 -0.08??8959 0 05861035 0.0000A002 -0.0912815? 0.22A0 0.05016131
(0.00006833) (0.2?581786 (0.25067A31) (0.2A201775) (0.0000695A) (0.265AA926)
0.00003579 -0.20A29968 0.03727173 -0.03A59978 0.0000A639 -0.23910575 0.1176 0.0A596832
(0.00005053) (0.2A8A90A7) (0.2A532206) (0.23092150) (0.00005537) (0.23181392)
0.00006526 -0.3A38A076 -0.012A70?3 0.0?353288 0.00006685 -0.A0502628 0.1?A8 0.0AA88536
(0.00005137) (0.28560938) (0.22526350) (0.215593Al) (0.00005600) (0.27996130)
0.00000593 -o.26753805 0.13370833 -0.037A3390 -0.00000A78 —0.2A638953 0.1302 0.0A30A310
(0.0000A833) (0.23067873) (0.2A6A1178) (0.21376A91) (0.00005118) (0.259A5186)
0.00001693 -0.1815?627 0.01522651 0.0A756969 0.00000503 —0.18390589 0.179A 0.0A30682?
(0.0000A901) (0.22901663) (0.21705183) (0.213A2955) (0.0000517?) (0.30761770)
0.00001721 —0.2513A705 0.12778A8A -0.011A991A 0.00002516 -0.32?70363 0.1A23 0.0AA50518
(0.0000530A) (0.19536790) (0.229AAA71) (0.2309523A) (0.000059AA) (0.23520338)
0.00001292 -0.18726020 0.1?950A9? -0.09163181 —0.00090915 -0.19675769 0.068A 0.0A65876?
(0.00005A9A) (0.23070669) (0.27055A02) (0.25038152) (0.000056AA) (0.2A283338)
0.00002002 —0.19906982 0.12A76151 —0.01689597 —0.00000875 -0.213?82A5 0.0705 0.0A558852
(0.0000516A) (0.2A555769) (0.26A77A21) (0.238101A0) (0.00005181) (0.317A2979)
0.0000075A —0.2991910A 0.115330A3 0.0770685A —0.00002139 —0.30329502 0.1139 0.0A137753
(0.0000A86?) (0.2A0A937A) (0.21?2A221) (0.21A02513) (0.0000A826) (0.3077069?)
0.00003295 -0.1A15683A 0.05675161 0.09882225 0.0000203A -0.2?121000 0.1295 0.0A250323
(0.0000A8A8) (0.19916396) (0.20256999) (0.20A93088) (0.0000A83?) (0.18339875)
0.00000921 —0.11718126 -0.0A503998 0.083989A3 —0.00901051 -3.01507795 —9.9009 0.9382511»
(0.0000A291) (0.15828005) (0.1991A562) (0.1861A636) (0.0000A175) (0.17831503)
0.0000A359 0.01192A13 -0.03977A60 0.232519AA 0.00003369 —0.3AA28266 0.18A2 0.03A69238
(0.00003AA8) (0.23112620) (0.17773381) (0.20665770) (0.00003A87) (0.23968670)
0.00002689 —0.218086A5 0 0662082A 0.1682087? 0.00002260 -0.365A8925b 0.2598 0.0359A00)
(0.0000A0A?) (0.15718522) (0.19996090) (0 20316103) (0.0000A0AA) (0.20A77315)
0.0000A502 —0.120A876? 0.09A81162 0.1A295131 0.0000AA02 -0.A1171326b 0 26A5 0.037A2507
(0.0000A332) (0.19A18A06) (0.23093A09) (0.21606629) (0.0000A300) (0.238AA721)
0.00003688 -0.A0668882 0.01917882 0.10937678 0.00000751 -0.15557253 0.2A98 0.0369151?
(0.0000A17A) (0.2519A09?) (0.23A0A8A8) (0.2138933?) (0.0000A101) (0.20992986)
0.0000A0A1 -0.0?A308AA 0.2550172? 0.11590189 0.00002709 .0.51680505b 0.279? 0.035220A5
(0.00003920) (0.219638A8) (0.23597998) (0.2169057?) (0.00003866) (0.26327900)
0.00003573 ~0.0360?20A -0.385513A1 0.199A5338 0.00011601 0.25732316 0.2183 6,636 saga
(0.00003A68) (0.1238A326) (0.2?838A92) (0.21700086) (0.000033A5) (0.2688822?)
0.00000292 -0.06591350 -0.11178A33 —0.A03AAAA5 -0.0000372A 0.2059170? 9.1699 0.9677757?
(0.0001057A) (0.19380863) (0.2183A579) (0.A2988586) (0.0001A938) (0.2761537?)

 

.2...‘".-'o.|.

.f!

 

star

122

 

 

 

 

TABLE VI-lc: Regression Results, Equation (3): Other Sec, Cross Section
Period Variables
Constant X1 X2 X0 (DI) (Dl)(X1)
19A6 0.0358A611 0.02550832 —0.00000951 0 00631892 0.059290A8 -0.080935A3
(0.07A27AA6) (0.06723828) (0.00001865) (0 00532275) (0.0809095?) (0 07619739)
19A? 0.0?575586 0.003092A7 -0 00001788 0.00550211 0.03139223 -0.06059085
(0.0786679A) (0.0?1A2602) (0.0000218?) (0.01165889) (0.08772323) (0.08157099)
19A8 0.0?180602 —0.002A96?5 —0.00000919 0.00A2281? -0.006891A5 -0.009579A3
(0.07A85696) (0.067A85A3) (0.00001926) (0.01990763) (0.0838025?) (0.0?8A5357)
19A9 0.076A789A 0 00751111 —0.00000902 —0.011576A2 -0.00321998 0 00961689
(0.0666336?) (0.06A08169) (0.0000198?) (0.0088A135) (0.07637A09) (0.07A55389)
1950 0.07628798 0 00026623 -0 00000919 0.0032A681 0.096A9A18 -0.06395018
(0.08789511) (0.08790520) (0.00002AA3) (0.01630296) (0.09671282) (0.100101A1)
1951 0.09167AA7 -0.01366A2? —0.000015AA 0.0259A95? -0.00A25AA3 0 00680003
(0 08888777) (0.08671575) (0.00002153) (0.03803AA6) (0.09883030) (0.098779A8)
1952 0.07897826 0.01002010 —0.00001505 0.01550119 —0.01209A8A 0.00316067
(0.07817016) (0.080829A5 (0.00002095) (0.030027A9) (0.0895915?) (0.09A6978A)
1953 0.05A1250? 0.0317155A -0.0000101? 0.020A6880 0.0A316079 -0.053A7125
(0.09630208) (0.09798612) (0.00002A8?) (0.03995352) (0.1070823?) (0.113256A3)
195A 0.058839A7 0.05135613 0.0000026A -0.02A389A2b 0,0519u065 -0.02906518
(0.0?108365) (0.0?A5A91?) (0.000018A6) (0.01275289) (0.0800A723 (0.0866A9AA)
1955 0.057A63A0 0.0A857730 -0.00001283 0.01A29521 0.05771015 —0.0A309088
(0.08701979) (0.08566558) (0.000021A6) (0.05302260) (0.09598790) (0.0986A1A1)
1956 0.06825989 0.0?A16039 -0.00002631 -0.019598A6 0.03778990 -0.06081131
(0.09827370) (0.10323833) (0.00002190) 0.03265302) (0.1066296A) (0.1160A815)
1957 0.20535A22a —0.10129A52 -0 00003583 0.08093282 —0 1020369A 0 11702128
(0.095699A6) (0.113697A1) (0.000023A7) (0.063A9A95) (0.1051A235) (0.12735789)
1958 0.16161680 -0.009?9537 -0.00003252 -0.005052A3 -0.0A011022 0.038537A2
(0.09782388) (0.10691AA3) (0.0000275?) (0.02898371) (0 10658752) (0.11986692)
1959 0.15720A71b -0 02372162 -0.00002777 -0.033?3000 —0.06915538 0.06217111
(0.08022371) (0.08263069) (0.00002AA6) (0.03225870) (0.08985500) (0.09601A8?)
1960 0.2260185Aa —0.06916650 -0.00003193b —0.0A300585 -0.13180882 0.12350706
(0.07375556) (0.0??99986) (0.00001666) (0.02793198) (0.08062033) (0.08822396)
1961 0.19A08200a -0.032?0815 -0.00001563 —0.06A22931 -0.09070156 0.02787753
(0.0875958?) (0.10119A95) (0.00002381) (0.0A800933) (0.09871526) (0.11555A59)
1962 0.187623A6 -0.0A221708 -0.00000511 -0.0731A177 -0.088006?9 0.11556002
(0.12668A19) (0 12172522) (0.00001983) (0.0678A606) (0.13A36275) (0.1306906A)
1963 0.2005A111 -0.031196A1 -0.0000161A -0.0706127A —0 0585383A 0.06893310
(0.12665970) (0.1311A978) (0.00002622) (0 077A6561) (0.1A320355) (0.1A590131)
196A 0.1386A165 -0.00836023 —0.00001AA0 0.00016031 -0.03827063 0.05A89905
(0.1A016256) (0.1A0873A3) (0.00002665) (0 10067000) (0.1A915726) (0.1573A539)
1965 0.200397A0b -0.02880207 -0.00000063 —0.115152AA -0.115AA8A0 0.0765A916
(0.10162563) (0.11710781) (0.00002363) (0.10518202) (0.1129970?) (0.12996866)
1966 0.13250659 -0.0A935A86 -0.00000929 0.10A29066 0.01A13025 0.1073103?
(0.1073091?) (0.10926896) (0.000019A2) (0.12A89353) (0.1195592A) (0.12296511)
66 less 0.10017155b —0.0513023A -0.00003A51 0.0023A555 -0.05851A53 _0.09962989
A6 (0.056690A2) (0.15002107; (0.0000398A) (0.00822605) (0.053562A0) (0.16368600)
Note: Superscript 'a' indicates variable is significant at 5.0 percent.
Superscript 'b' indicates variable is significant at 10.0 percent.

Standard errors of coefficients are indicated in brackets.

123

Variables Regression AdVice:

 

 

(Dl)(X2) (Dl)(X3) (Dz) (D2)(X1) (D2)(X2) (D2)(X3) R BAR 2 Std Error

0.0000150A —0.01592181a 0.11322103 -0.090927?5 -0.00001213 -0.00968A03 0.1?2A 0.0179716A
(0.0000196A) (0.00707826) (0.08558696) (0.08031801) (0.00002333) (0.00965850)

0.00001160 -0.00928339 0.08871800 -0.07802399 -0.00000216 -0.023A3500 0.0257 0.0196968?
(0.00002A98) (0.01A773A8) (0.091A5752) (0.08597355) (0.00002720) (0.01982199)

0.00000592 0.0067822A 0.09767870 ~0.o7905832 —0.0000119A —0.01895901 0.0183 0.0192655A
(0.000022A6) (0.02111930) (0.0896036?) (0.08628063) (0 00002A77) (0 02819823)

0.00000523 —0 00820632 0.07925500 -0.11215998 0.00000098 0.007A0952 0.0881 0.01859581
(0.00002291) (0.01828535) (0.07936136) (0.0?879993) (0.00002300) (0.00903372)

0.00000068 —0.0528076Aa 0.09301905 -0.11683869 -0.0000220A 0.0A270769 0.2AA0 0.02A2A72A
(0.00002819) (0.02122280) (0.10776559) (0 11036A82) (0.00002926) (0.0312276A)

0.00000A0? —0.00203818 0.1108A3A9 -0.135AA272 -0.00000937 0.0166335? 0.0A65 0.0222223A
(0.00002A61) (0.0A303731) (0.10778382) (0.10753158) (0.00002525) (0.0AA78681)

0.00000A60 0.0295921? 0.16199A73 -0.2180?5A?b -0.00001282 0.05802A61 0.0358 0.021512A5
(0.00002A08) (0.03805210) (0.10292196) (0.11A69123) (0.00002502) (0.0508A092)

0.00000589 -0.010712A2 0.1A2068A1 -0.15986AAA -0 00000809 -0.0108123A -0.1560 0.02500A93
(0.00002839) (0.0AA6882?) (0.12333358) (0.13293A70) (0.00002913) (0.056A6381)
-0.00001395 0.00295138 0.15509839b -0.1572A136 -0.00000396 —0.0199A980 0.25A5 0.0192A226
(0.000021A0) (0.01A3A5A3) (0.09058951) (0.0950A990) (0.00002223) (0.0201A715)
-0.0000083A 0.003A68A3 0.13302A50 -0.1720A193 -0.00000675 0.0236A782 —0.0025 0.02093925
(0.00002A88) (0.0538008?) (0.106A822?) (0.10962388) (0.00002A57) (0.05879576)

0.00000731 0.0A179289 0.12607515 —0.18556698 0.000005A1 0.0565682? 0.0883 0.021A3572
(0.00002A90) (0.03507513) (0.11A65021) (0.1220A525) (0.00002A71) (0.03783719)

0.00002012 -0.062023A2 0.0?1615A3 —0.06A135?3 0.00000180 -0.12199A05 0.0636 0.02351A07
(0.00002725) (0 067A901A) (0.12533502) (0.13838178) (0.00002769) (0.0?387661)

0.00002A07 -0.02A9376A 0.07250291 —0.08188882 0.00002185 —0.0517367A 0.2233 0.02310386
(0.0000307A) (0.03091025) (0.1130A93A) (0.12775011) (0.00003088) (0.036566A3)

0.00001355 -0.0283A3A1 —0.00A61605 -0.0397330? 0.0000186A 0 00875203 0.1953 0.021A2271
(0.000027A1) (0.03329529) (0.099A9115) (0.10752866) (0.00002739) (0.03662113)

0.00001912 0.0153122? 0.031150A? -0 005A1805 —0.000001?5 -0.03519A57 0.3900 0.01773A35
(0.00001931) (0.02953719) (0.08A978A2) (0.09A86656) (0.0000191?) (0.03603383)

0.00000668 0.06753136 0.0?96A930 -0.16309077 —0 00000072 0.03391A95 0.0533 0.02281360
(0.00002639) (0.05785AA9) (0 10997373) (0.12557391) (0.000026A6) (0.058A3081)

-0.00000227 0.01505050 0.08760865 -0.0A663518 —0.00000511 —0.06A10977 0 2127 0.02033A19
(0.00002288) (0.0?6201A8) (0.13517788) (0 1A122325) (0.00002281) (0.09103333)

0.00001171 -0.0236?938 0.0989A888 -0.152A1088 0.0000001A —0.00A12091 —0.0002 0.02712552
(0.00003011) (0.10186001) (0 15582175) (0.15A63600) (0.00002961) (0.1A905350)

0.00001702 —0.055?8211 0.09738710 -0.19?8167A -0.000002A6 0.05153201 -0.0673 0.02635099
(0.00003012) (0.12278218) (0.1521A773) (0.16A89868) (0.0000293A) (0.11277633)

0.0000035? 0.12793519 _0.029A2700 ~0.095873A7 -0.000017A9 0.2376896Ab—0.0000 0.0238069;
(0.00002702) (0.1156A39?) (0.13009955) (0.13927676) (0.0000259A) (0.13789511)

0.0000075A —0.20A29821 -0.11222030 -0.09AAA9A2 —0.00000716 -0.096555A3 0.1102 0.021013A6
(0.00002229) (0.13A25213) (0.12A02520) (0.12870732) (0.000021A8) (0.1376920A)

0.00002079 —0.018980A2b —0.0300081A 0.0?8A9522 0.00002650 -0.00872869 0.1839 0.122 479
(0.0000A169) (0.0100AA22) (0.0807AA03) (0.18691769) (0.0000588A) (0.01133609)

 

1.214

 

 

 

 

TABLE VI-ld: Regression Results, Equation (A): Loan 8 Discounts, Cross Section
Period Variables
Constant X1 X2 X3 (01) (Dl)(X1)
19A6 0.266560558 0 05233676 -0.00001205 -0.0A6909A2a 0.20A0805A -0.29385?52a
(0.12375805) (0.12293872) (0 00003388) (0.01217352) (0.13969919) (0.1A286000)
19A? 0.A195?931a -0.009A952A -0.0000A120 -0.05110?56a 0.16A6622A -0.2607882A
(0.15A39663) (0 16672302) (0.0000A533) (0.0169A186) (0.18031119) (0.18850780)
19A8 0.A0A99A66a -0.09671133 -0 0000A099 -0.00506092 0.11595506 —0.1807131A
(0.16723565) (0.1536A691) (0 0000A501) (0 02195683) (0.1998851A) (0 17783859)
19A9 0.?A670325a (0.0?033213 -0.00010A18 —0.1A1A025Aa -0.1A158981 -0.3256066?
(0.2A028502) (0.25A8A852) (0.00007538) (0.02581179) (0.281A521A) (0.29331781)
1950 0.392267968 -0.06208686 -0.00000A35 -0.020A3853 0.25869896 -0.155A0?6?
(0.15780878) (0.1597668?) (0.0000A50A) (0.02A13102) (0.181A72A7) (0.181605A5)
1951 0.A2339769a —0.10091132 0.00001101 -0.02290100 0.2216086A -0.1187?259
(0.155A289A) (0.1658038A) (0.00003761) (0.02010912) (0.17313612) (0.1879A836)
1952 0.387561A?a —0.07621605 0.00002A09 -0.02152563 0.26A63622 -0.1610A790
(0.1?900265) (0 157A8758) (0.0000A2A1) (0.02919672) (0.20086783) (0.187015A2)
1953 0.6071159Aa -0.235?7023 -0.00003222 —0.01??0505 0.09230183 0.01A39353
(0.16379A50) (0.150A7273) (0 00003760) (0.027A2502) (0.18361512) (0 17A10662)
195A 0.58533052a -o.165A7159 -0.00001919 ~0.0AA82829 0.0557567A -0.0A7A8633
(0.17153753) (0.152A7375) (0.00003803) (0.032530A6) (0.190A5630) (0.1765980A)
1955 0.59A62659a -0.18170980 -0.00001677 -0.026291A2 -0.0003A083 0.0007236?
(0.18A6257A) (0 17022832) (0.0000A309) (0.03165755) (0.2090599A) (0.1959060?)
1956 0.5889A356a -0.28989?97 0.0000008? 0.00078326 0.1A327238 0.0A620652)
(0.20690083) (0.1733900) (0 0000A206) (0.03230939) (0.2275358?) (0.19912000)
1957 0.AA000155b -0.15A87109 0.00003668 -0.00772288 -0.02835567 -0.0?819127
(0.22030561) (0.16691633) (0.00003856) (0.0A21A976) (0.25359961) (0.19053168)
1958 0.AA8957638 -0.0923A072 0.00001A96 —0.01651731 0.23922816 -0.1?256883
(0.16662658) (0.1A328210) (0.00003701) (0.0A651768) (0 1826266A) (0.16507268)
1959 0.A58AA602a -0.073923A6 0.00006298b -0.0A697912 0.15887018 —O.1801?9A6
(0.15137281) (0.1A1712AA) (0.00003368) (0.0A186057) (0.16691623) (0.16353986)
1960 0.5A115216a -0.10310330 0.0000A2A8 -0.06783921 0.15736570 -0.21899?29
(0.13320505) (0.13068832) (0.0000308A) (0.0A6A9751) (0.15607A32) (0.152A5026)
1961 0.323A1765 0.012AA300 0.0000AA63 0.00092675 0.A2650073 -0.28503291
(0.20629131) (0.166620A3) (0.0000318?) (0.0A853168) (0.22816382) (0.18726631)
1962 0.A330581?a -0.01802915 0.00003562 -0.02520590 0.52A69009a -0.32192A0Ab
(0.200AA722) (0.16617959) (0.0000363A) (0.0532085A) (0.231727A6) (0 18629902)
1963 0.5A365639a —0.073A8683 0.00003191 —0.05219175 0.1?128823 -0.2AA50A51
(0.20236629) (0 170068A1) (0.00003A05) (0.05582220) (0.23368173) (0.18938005)
196A 0.A99-92773a -0.12288569 0.00002906 -0 00391678 0.A5A00268a —0.2?29?1?0
(0.185A3938) (0.17059765) (0.00003501) (0.06335810) (0.211929A8) (0.1922A137)
1965 0.8AA060A7a -0.17679A87 -0.00000185 -0.13782801 -0.13?72137 —0.10333150
(0.33932970) (0 18A07901) (0.0000A52A) (0.10223A96) (0.36011869) (0.20271652)
1966 0.28267031 0.00529259 0.00005063b 0.0A66A391 0.55919732a -0.35299271b
(0 18709628) (0.18032708) (0.00002683) (0 05339A81) (0 2061736A) (0.201A3A23)
66 less 0.273259A9a 0.20008A5A 0.0000816? -0.0A60612Ab 0.07588510 —0 30351608
A6 (0.110A7003) (0.32119A76) (0.000087A?) (0 02378965) (0.113A2525) (0.3A991376)
Note: Superscript 'a' indicates variable is significant at 5.0 per cent.
Superscript '0' indicates variable is significant at 10.0 per cent.

Standard errors of coefficients are indicated in brackets.

£1235

 

Variables

Regression Advice:

 

(011(x2)

(011(x3)

(D2)

(D2)(X1)

(02)(x2)

(02><x3)

R BAR 2 Std Error

 

.00002A50
.00003578)

.000003A8
.00005272)

.00001A93
.00005232)

.00006732
.00008938)

.00002818
.0000521A)

.OOOOA6A8
.0030A328)

.0000A8A5
.0000A823)

.00002A85
.0000A330)

.00002118
.0000AAA8)

.00000697
.0000A932)

.00000265
.0000A69A)

.000016A3
.0000AA75)

.00000753
.0000A26l

.00005185
.00003963)

.00003A72
.00003689)

.000057Al
.00003568)

.00001509
.0000A166)

.00003202
.00003878)

.0000508A
.00003953)

.000006A7
.0000A875)

.00005A73b
.0000320A)

.00010275
.00009102)

.03115A72

.039A3053)

.11632077a

.0A258836)

.02205078
.03186759)

.005A203A
.02A80A60)

.00729021

.03762725)

.03710270

.03301976)

.0102A751
.03973630)

.01071618
.0355002A)

.0A903273
.03856303)

.05728326
.0537A6A2)

.02972587

.0A9A9506)

.0A601891
.0A195AA2)

.O5A30391
.053A130A)

.03507578
.055A3A12)

.12068169
.0739A006)

.05788197
.06931220)

.O6A08Al6
.0707618A)

.15075635
.10903A2A)

.08692590

.05802836)

.0A329908
.03318767)

-0.

(O

—O

-0

O.
(O.

-O.
(O.

-0.
(0.

-0

-0

-O.
(O.

-0.
(0.

-0.
(O.

-O

-0.
(0.

—O

-O

0

-0

-O

-0

—0

0

—O.
(0.

03393238

.1510936A)

.09A21005
(0.

19106611)

.09853903
(0.

233A3295)

08252A69
29550A01)

0855A55A
20270550)

120A7576
20366312)

.05123056
(0.

25A89628)

.38969997
(0.

23676255)

255A6281
22871052)

33515607
2737l3u9)

5A50A0A5a
2502A58A)

.3552373
(0.

3
2A973885)

1038A212
25607792)

.21517593
(0.

17A33A10)

.23386955
(0.

18957755)

.08633886
(0.

2A806726)

.19319132
(0.

.281A76A8
(O.

2916A088)

26637182)

.2663A032
(O.

26A16331)

.59030A96
(0.

37035380)

.1A9A0772
(O.

20673A21)

00233192
161AA79?)

—0

-O

.06A52739
.1A76A567)

.08367565
.19660A2A)

.17636522
.187036A0)

.071A7679
.308289A9)

.16575A6O
.19662701)

.15771935
.20060775)

.15966269
.2051A66A)

.33731272b
.19A92A39)

.23926333
.193922A2)

.2856222A
.21866058)

.A2601605a
.21061996)

.18017133
.203738A9)

.13A60759
.179890A6)

.0151A918
.17807132)

.1080A556
.1629135A)

.0A10A213
.201515A7)

.O2A7081?
.19533A01)

.0529A39A
.19791707)

.13563970
.20077865)

.05292865
.21506780)

.11980933
(0.

21163777)

.0869A710
(0.

39406199)

-0
(0

-0

(0.

O.
(O.

—0
(0

—O
(0

—0
(0

-O
(0

- 0

(O.

-0
(O

.00001A38
.OOOO39A1)

.00000039
00005576)

00000820
00005723)

.0000593“
.0000861A)

.00001988
.00005272)

.00001052
.0000A630)

.00002371
.00005262)

.00005165
.OOOOASSI)

.00002983
.0000A65A)

.0000A751
.00005A37)

.00007275
.0000A820)

.00001001
.0000A363)

.00002001
.0000AA39)

.00001A96
.000038A3)

.00001115
.00003638)

.00003382
.00003636)

.00000669
.0000A150)

.00001068
.00003897)

.00001198
.00003898)
.0000A873
.0000A799)

.00002571
00003076)

.000013A0
.00012503)

.01073790
.019872A?)

.01579A29
.02382807)

.0168090A
.0A03ll26)

.01891051
.03A79059)

.000A0623
.03150676)

.00657028
.02608900)

.008A66A7
.038A167A)

.01977912
.0A028A10)

.01383590
.0A899667)

.01820319
.0A698861)

.03922673
.0A015170)

.0938212A
.0580A683)

.00626919
.07689396)

.09A28371a
.0A561681)

.10505539
.0751A339)

.01219688
.0690A581)

.09680907
.10687A28)

.13725899
.10303A92)

.0986AA39
.10518098)

.25762527a
.127A3108)

.000A0902
.06297868)

.03682573
.03000775)

0

O

O.

«0

.A562

.A067

.2383

.6668

.2795

.3082

.2767

.3A6A

.2973

.1612

.A066

.A112

.3918

5235

.A950

.38A9

.527A

.A905

.A992

.0011

0

O

0.

.03351011

.0A331A17

.0A358275

.0705A558

.0A2OA716

.0A002861

.0Al61399

.03861986

.03922739

.0A1A77A3

.0393192A

.O3765A06

.035AA628

.03577058

.03307613

.03A8875A

.03A98813

.03A202l3

.03A63753

0335505“

.03388280

,(1A6‘361‘)?

 

 

TABLE VI-2:

Regression

126

Results, All Equations, Time Series

 

. . '3"..:|=.—‘ “12.22——

 

 

 

Period
Constant X1 X2 Xg-Xu. (DI) (DIHXI)
Equation (1): Cash Items
All -0.0580A2889 0.392316933 —0.00000301 -0.095917A7 0.23697597a —0.2A10A?88a
Years (0.02AA735A) (0.02105692) (0.00000A78) (0.22122969) (0.02772632) (0.02A6539?)
Equation (2): US Gov Sec
All 0.3617817‘1a 0.201A5956a -0.0000680Aa -0.10577285a 0.03575331 0.02536293
Years (0.06A3o)7A) (0.05701A26) (0.00001A7A) (0.01973231) (0.0731A917) (0.06807611)
Equation (3): Other Sec
All 0.16800918a -0.12367281a 0.00000A60 -0.00738A02a .0.069956A6a 0.0731A7A9a
Years (0.02029AA8) (0.0179059?) (0.00000A73) (0.00316363) (0.022786AA) (0.021A1515)
Equation (A): Loans & Discounts
All 0.66951A05a -0.3826019Aa 0.00005558a -0.0561102la -0.00286997 -0.07513A2A
Years (0.063A693A) (0.06052508) (0.00001510) (0.009A2677) (0.07210932) (0.0709511A)
Note: Superscript 'a' indicates variable is significant at 5.0 per cent.
Superscript 'b' indicates variable is significant at 10.0 per cent.
Standard errors of coefficients are indicated in brackets.
(*) In Equation (1), this variable is X“.

In Equations (2), (3), and (A), this variable is (X3)/(Xu), and is designated X3.

1.2'7

m

Regression Coefficients, Standard Errors, Significance: Regression Advice:

(Dl)(x2) (D1)(X3-Xu)* (D2) (D2)(Xl) (D2)(X2) (D2)(X3-Xu) H BAR 2 Std EPI‘OI‘

 

 

Equation (1): Cash Items
-0.00000872 —1.65829300a 0.1350u119a —o.12333180a —0.0000133ua —0.51692u62b 0.7108 0.02590105
(0.00000509) (0.27751281) (0.0293uu67) (0.02596159) (0.0000055u) (0.2784091u)

Equation (2): US Gov Sec

0.0000688].a —0.343;
03

’3 -0.00587700 -0 11170001 0.00002213 0.070514252a 0.3682 0.080909u9
(0.00001698) (0.’ )

(0.077002u6) (0.07225201) (0.00001712) (0.0205uu32)
Equation (3): Other See

0.00000620 -o,102555778 0.0597u517a -0.050u6277a -0.00001u79a 0.000u7152 0.2423 0.026036u9
(0.00000540) (0.00378926) (0.02uu0189) (0.022909u7) (0.00000550) (0.00362791)

Equation (u): Loans & Discounts

-0.00002080 0.0U36U8U6a -O.1822968Sa 0.21977010a —0.00000102 0.01287982 0.u252 0.0818999“
(0.00001720) (0.0lOU2621) (0.07791255) (0.076051U2) (0.00001762) (0.01256912)

 

1228

.mpmxompn Cw bmpmowncfi mpm mpcmHOHummoo ho mpoppm Utmocmpm

.ocmopmo 0.0H um ucmofiuficwfim ma manmficm> mmowoflwcfi .p. pqfiLowLqum
.pcmopmo o.m um pcmofimficwfim ma mapwfipm> mmpmoflvcfi .m. poflpomcmozm umpoz

 

 

 

Amamfioooo.ov Ammmmmsoo.ov Aflmmfioooo.ov Asafiamaoo.ov Asmwooooo.ov Ammoasmoo.ov ”omflo s was
maoooomo.o omm~.o mmwooooo.on ammeOHmo.o mssmmocoo.o mmzssmmmo.o mmsmmmooo.o «mamflmmom.o coflpmsom
Awamooooo.ov Amaammmoo.ov Awwmooooo.ov Asmmmwmoo.ov Afimsooooo.ov Awmmoomoo.ov ”omm Lasso
magmmmmo.o asom.o mceoocoo.o- mooomsoo.o macooooo.o mmmmsfioo.o mfiwoeoooo.o maflflfloasm.o coflpmzom
Aommfioooo.ov Assmamaoo.ov Awmmfioooo.ov Amsmoaaoo.ov Amsmmswoo.ov Asmasmmoo.ov H8m sou m
mmmeHmo.o Haze.o amsmmoooo.ou msmemofio.on «mammoooo.on mmmmsmmoo.o mmwwsomoo.ou amsosmmos.o censuses
Asmmooooo.ov Ammsmamoo.ov Amsmooooo.ov Aommmmmoo.ov Aaowooooo.ov Ammsmamoo.ov ”mepH ammo
wowsmsmo.o moms.o anesfioooo.o mmmammao.ou mommmoooo.o «mammmoso.ou «osmosooo.ou mammmmmmm.o scavenge
tettm cum m m<m m AEVANQV Ammo ABVAHQV AHQV mefie ucmpmcoo
uw0H>o< cofimmmLmom "mochHaflcme ampoppm chapcwum .mpcmfiofimmooo coflmmmsmom

 

mcofipmsum HH< .mecoge mess wfiomfipm> » um-H> mamae

129

TABLE VI—U: Correlation of Insured and Uninsured Banks,

 

 

 

 

1 9 5 0 - 1 9 6 6 f“
m” -_-__*‘_ _ _‘ __ 3....-- YI
._._.-i..--...__ _—._._... ___.__..._. .- .. .—__..._ - -__.... I
Coefficient of Coefficient of
Equation Covariance — R Uetermination - R2
Equation (1), + ,
Cash Items: —0.0703 0.00U9 :4fi
Equation (2), 7 .
US Gov Sec: 0.36%1 0.1290
Equation (3),
Ottner Shea:
State & Hun.
Bonds ~0.1510 0.0225
Other See. -0.5814 0.336“
Equation (4),
Lns & Disc: 0.91l9 0.8281

 

 

130

 

 

 

 

 

 

 

 

 

0. ..
. 0
.mm 00 HopOp m 00 pcmo Lon 0.0H pm pcmoHuHcmHm Hxv
m.z H H m.: H 0 0.0 0 0 H.m m . H m0
m.mH : m 0.0 0 0 w.mm m H m.:m mH 0 H0
93 om 3 mém 0 m 0.03 mm mm 58 m z 2388
:proe mo ROH p< um p< *HwQOB no u0H p< mm p< *HwoOB 00 ROH p< Rm u¢ *HMpOB mo ROH p< “m p<
cm to m cm L
p 0 m "mHm mmEHp .oz ucmo L m umHm mmEHp .02 0:00 pom "mHm mmEHu .02 u 0 mm ”mHm mmEHu .oz mmHomHLw>
”.omHQ a mcq “H20 coHpmsvm ".omm mepo ”Amv :oHuwscm ”.omm .>00 m0 ”Amv coHumzdm ”mEopH :mwo "HHV coHpmsvm

 

coHpoom mmopo

.COmequoo EmpmzwlmpucH

uwmIH> mqm<9

I

Lquatior

H

T“

Intra-system Comparison, T

 

 

TABLE VI-Sb

 

 

 

b0

"—1

 

ea
in

(:3

r“!

13".
LO

1“7
‘v ’0

1'0

Y‘U’

’\ 07
x.) [9

l ;

5%

 

131

Q.)
’3 4

‘
\

(I)
V K—-

Y

i 93

1v

J «~'
6 a

3“.

ff)

TH

i.)

4,)
r!)

(U

Ye

N ’7‘)

.1

' )

\.
l.

,‘V‘.

(2)

U.‘
L; )

‘b‘

("II

(V

‘H

(“J

Lt'

up! Qantl‘VIc‘ -

 

 

lfi32

 

.mm 00 H0000 0 00 0:00 :00 0.0H 00 00000000000 H00

 

 

 

 

 

 

 

 

0H000HH000 002 0H000HH000 002 0H000HH000 002 0.0 0 0 stvaav
H.0 m m m.0 H 0 0.0H 0 H 0H000HH000 002 Amvamav
0.0 0 0 0.0 0 0 0.0 0 0 m.0 H 0 ANXVANQV
H.0 m H m.0 H 0 0.0 0 0 H.0 m 0 AHXVHNQV
0H000HH000 002 0H000HH000 002 0H000HH000 002 0.0m 0 0 stVHHov
m.0 H H 0.MH m m 0.0 0 0 0H000HH000 002 AMXVHHQV
m.0 H 0 0.0 0 0 0.0 0 0 0.mm m m AmvaHov
0.MH m H 0.0 0 0 0.00 0 m 0.m0 2H HH HHvaHov
0H0000H000 002 0H0000H000 002 0H000HH000 002 0.00 0 0 2020
m.0H 0 m 0.0 H 0 0.0 H H 0H000HH000 002 mev
H.0 m 0 0.0 H 0 m.0 H H H.0 m 0 H020
0.0 0 0 0.0 0 0 0.0 H 0 0.00 H0 H0 HHxv
0H0000 00 00H 02 um 00 0H0000 00 00H 00 00 00 *H0000 00 00H 00 um 00 0H0000 00 00H 0< 00 00
0:00 :00 “me m0EH0 .oz 0:00 :00 "000 m0EH0 .oz 0:00 L00 ”0H0 m0EHO .02 0:00 000 ”0H0 m0EH0 .oz m0HpmHhm>
“.0000 a 0:0 “H00 :0000300 ".000 :0:00 ”HMV :oH0mzcm .>00 0: “Amv :oH0msom "mE00H :000 ”HHV :oH0mSUm

 

:oH0o0m 00000 .m0C0HoH000o0 :onm0am0m 00 00:00H0H:0Hm

”wwIH> mqm<8

‘ . . ...rW0.

 

133

 

 

 

 

 

0H000HH000 002 0H000HH000 0oz 0H000HH0000 0oz 00% 0: Axxvnmmv
ox o: oz 02 00% 00% 0H000HH000 0o: meVHmQV
0: oz 00% 00% oz oz 00H 00% AmeHmmv
00% 00% 00% 00% oz oz 00% 00% AHXVHNQV
0H000HH000 0oz 0H000HH000 0oz 0H000HH000 002 00% 00% AJXVAHQV
00% 00% 02 oz 00% 00% 0H000HH000 00: meVAHQV
02 02 02 02 000 000 02 02 ANXVHHQV
02 oz 00» 00% oz oz 00% 00% HHxVAHQV
0H000HH000 0oz 0H000HH000 0oz 0H000HH000 002 oz oz Hzxv
00% 00% 00% 00% 00% 00% 0H000HH000 0oz Amxv
00m 00% 02 oz 00% 00% 02 oz Amxv
00% 00% 00% 00% 00% 00% 00% 00% AHXV
00H mm 00H 0m 00H 00 00H 00
.00 0:00H002000 H00 0000H0H0000 “00 000000000H0 “00 0000H0H00H0
.00H0 0 000 ”000 .000 00000 ”Hmv .000 .>00 00 ”H00 0500H 0000 "0H0 00H00000>

 

:oH0mzvm

I’ll‘illl‘lil‘lllill' ‘11- II.l.1|lv!||n"l'l ll‘lhl

- ’ l|| Ill)‘ I'll"

00HL0m 0EHE .00C0H0000000 Cofimmmammm ho 00:00HQHCWHW unmlH> mqm<E

IL3M

.mm 00 H0000 w 00 Axv

 

 

 

 

 

 

0H0000H000 002 0H0000H000 002 0H000HH000 002 0.00 00 - stvAmev
0.00 0H + 0.00 0H + 0.00 0 + 0H000HH000 002 meVHmav
0.00 HH- 0.00 0H: 0.00 0 n 0.00 Hm-
0.00 HH+ :\+ 0.Hm 0 + -\+ 0.00 0H+ -\+ 0.0 H + u\+ meVAmav
0.0H m - 0.00 Hm . 0.00 0 u 0.0 H + HHvamov
0H000HH000 0oz 0H000HH000 0oz 0HomoHHamm 0oz 2.00 0H I HzxvaHQV
0.00 HH + 0.00 0 + 0.00 0 + 0H0000H000 002 meVHHov
0.H0 0H- 0.0H m . H.0 m u 0.0 H :
m.0H 0 + u\+ 0.00 0H+ |\+ 0.00 0m+ n\+ 0.00 Hm+ u\+ meVHHQV
0.00 0H - 0.00 0 - 0.0 H - 0.0 0 + HH202H00
0H000HH000 002 0H000HH000 002 0H000HH000 002 H.0 m . H020
0.0H 0 + H.00 0H + 0.00 0 + 0H000HH000 002 H020
0.00 m n 0.00 Hm: 0.0: m n 0.mm 0 u
H.00 0H+ -\+ 0.0 H + s\+ H.00 mH+ -\+ 0.00 0H+ -\+ 0020
0.00 0H - H.00 0H - 0.00 Hm . 0.00H mm + AHxV
H0009 00 *0000000 :mHm H0000 00 *0000000 :me H0000 00 *000:000 :me H0009 no *0000000 :me
0:00 :00 LO0Ezz 00020 0:00 :00 000832 U0qu 0:00 000 000552 00me 0:00 000 000852 00qu 00HO0H00>
".00H0 0 0:0 ”H00 :oH0wzcm "000 :0:00 ”HMV :oH0wzam H.000 .>00 0: ”Amv :oH0msdm ”0E00H :000 ”AHV :oH0msdm

 

 

00:0H0Hmm0oo :oH000Lw0m .coH0o0m 00000 .0:me 00 0H0>H0:¢ "0%: H> mqm<0

Analysis of Signs, Time Series, Regressi

E VI-7b:

TABL

 

 

 

‘QC.

09.1"

\
"\

Qt

(3)

C.)

(If)

C)
(T3

(1):

 

 

 

 

r4
0‘. a:
:1 M.
4) -y{
C) ("D
Q)-
J—‘- S-
(7‘, i_|
"1, -r1
DJ ‘2
>4
{1-1
r-o
OJ C
:1, i1".
4.) 0r!
C) (0
“if:
T%
(L)
.0.) "‘
C) ‘0‘
G} -[~1
0.0)
M
[I]
H
m 51”.
'3 If.
4.) 'r‘1
(LL ('3
"I:
(I)
+3 (:1
c) ‘0
(ll "-1
£1.03
>4
[1]
H
m C
:5 (LU
$.) "'1
C) 0)
"2'5
(1)
+3
C) C
0) E10
Can—J
>< L”)
[:1

 

(x1)

/_

+/"-

(x2)

,
(J

#4

Orv.

\11

'(‘i
C)

at ’1

/

135

i .-

/‘\
,0

5 J
LU

C)

"*1

Q".
. *4

C)

19
\'

cat

".4

appl

'Ot

I
H

e

cabl

O
1
.1.

A

A

2‘01

‘
, \
\_I

.l

applicable

,‘\

 

 

136

TABLE Vl—Ba: Estimated Values of Y*, Equation (1): Cash
Items (Cross Section)

.——..—.H- .~. .- . __.--”_ -_-__---.— .- H-H-__ -_-

JtCMxnvide inw1nahimn; Lindtmwi Hrnrnwxinfi Urflt; Bankirnfi

 

 

YQQP Ctmtfia State; States
19““ ”-5337 0.2330 0.2u20

1947 0.23“) 3.2M60 0.2U67

1948 0.2360 0.25MB 0.2M98
19M) 0 221M 0.2386 1 2755
1950 0.2215 0.2 1“ 0.211u
1951 0.2396 0.2552 0.2501
1952 0.2206 0.?M23 0.9309
1953 3.2182 0.2366 0.2310
1956 6.1865 0.2266 0.2172
1055 e) 111% O 226M 0 21§3
1936 d Jlfiu 0.2’6, O 2193
1957 0.118) 3.2169 0.2167
1958 0.189U 0.2023 0.1972

1959 0.1871 0.1975 0.1901

1960 6.1833 0.1966 0.192
1961 0.1737 6.1895 0.18am
1962 0.1505 0.1680 6.1631
1963 0.1505 0.1609 0.1661
196M 0.161u 0.1657 0.1679
1965 0.1u82 0.163: 0.1593
1966 0.1515 0.1611 0.1608

Mean of

21 Years 0.1956 0.2110 0.2075

66 less
“6 0.0170 0.1087 0.1387

 

(*) At the means of the X variables.

TABLE VI—8b:

137

Sec (Cross Section)

Estimated Values of Y*, Equation (2):

-._._._.

.__._

US

__ —__ -—._~'—...__

Gov

Statewide Branching Limited Branching Unit Banking

 

Year States States States
1996 0.5197 0.5100 0.5329
1997 0.9562 0.9610 0.9805
1998 0.9150 0.9179 0.9309
1999 0.9607 0.9201 0.9193
1950 0.3668 0.3828 0.9050
1951 0.3976 0.3698 0.3798
1952 0.3982 0.365 0.3850
1953 0.3395 0.3993 0.3702
1959 0.3920 0.3538 0 3733
1955 0.3056 0.3202 0.3395
1956 0.2798 0.3010 0 3195
1957 0.2696 0.2896 0 3195
1958 0.2761 0.3021 0 3206
1959 0.2979 0.2689 0 2998
1960 0.2385 0.2600 0 2790
1961 0.2909 0.2689 0.2822
1962 0.2208 0.2560 0.?692
1963 0.1937 0.2395 0.2993
1969 0.1588 0.2155 0 2276
1965 0.1986 0.1923 0 2016
1966 0.1399 0.1681 0.1850
Mean of ’
21 Years 0.3003 0.3196 0.3357
66 less
96 —0.3759 —0.6100 —0.3291

 

(*) At the means of the X variables.

 

TABLE

1952
1953
1 9 {J 9

1955

1961
1062
1963
1969
1065

1966

V1

Mean of

—8c:

21 Years

66

~ ('1 xi
1 ‘2.) L.) x‘

96

_

:1," 15 Li, 6 {13'1“} 1

0.
0.

i).

—0.

(*) At the means

lit.“

‘- r
.JLI(‘1 .'

Estimated

(V

L. ,
[0.87

1011913
".1 r)1 0

“13611

0861

Of the

sec (Cr

1217111c1111112 111:0: t2>d

States

0.

K‘
4 1
‘ 0
1‘2
1
L; ,

_O.

0179

I] f— '9'“
o ‘. ‘a’ £1”,

1" J L ;
0897

r v’ r I
.1870

f) 1"; 1) )l

0197

X variables.

C.)

Q
0

0959

1'61 8 1.1 7

r ‘7) " .
j ‘ ) v.) I

f ‘ (:1) ' I?

1.} ‘J 77 (J
r J (J 1 j

 

 

TABLE

V1-8d:

 

imatad

.a
(Jr

139

17.1
OBS

ues of Y*,
Section)

6:63111111‘11311

(9):

1..) 11 L1 8C

 

 

 

(*) At

the

NH"- U 1‘] :3

(Q1 f

1.116

X ‘Jarfi:ublx s.

itatwvldm DPEHCnlflf pim1ted Franc inn Unit sanxln;
9*.r’ tat“: ‘ State: States
1396 51.311? 3.1w“! 0.1812
1997 a. 978 1.1979 0.211:
1990 1.95,8 i1.1‘«‘ 1.25;
1099 0.‘928 0.1«9~ 6.271
11:30 0 , 1369 T1 . 27:1 at“ .. 8G1
19191 11 . Q (111:), 1') . :7 «r17 {7} _[ {/4 If '9
195? 0.3918 0.6“? 0.110’
1003 1. '>; 0.54«9 0.<BU)
1959 0.3“ 0 0.119 0.3317
1055 0.9025 0.37‘6 0.1v13
ljiflfi 0.’tlfi7 {1.31-fi+ 3. {325
1957 0 9109 0.933, 0.33:7
1058 0.9296 0.3091 0.1631
1050 1.9623 0.9987 0.3818
196) 1.970» (rJdn ~1J7x,
1061 0.9710 0.9 1 1.9912
1816:.) ‘J.r!1'pfl) ‘1). 11.1111. 9.1.": ,1 "I;
1963 0.6}18 0.9». 1.17::
1069 0.6966 0.9101 0.98.
1065 0 5986 0.t17 1.9?
1966 0.1597 0.53 0.:191
Mean of
21 Years 0.9069 0.3767 0.308)
66 less
96 0.1970 0.105% 1.11.00:
‘1. ¥“__p ,fi__“~fl3__f-___

FJ
k
C)

5 ‘ Vl-8e: Estimated Values of Y*,
(Time Series)

 

 

TAELE
Statewide 11; 1mg
Equatinn firan‘hlng States Brafiching thte 0101103 States
Equation (1):
_/\ ' -'x I "‘ J" 2‘ ." r f _’ Q
11.11/17] L) .<.?].L.J‘.’) 3 .21j1'1' .1
K.

00.01 ltmflns:
1' 8.3090

2):
0.2171

Equatien (

 

US GOV SOC:
hq1U1titni (3):
\ 1 (‘ F ‘\ v /‘ (1 I“..’* A /\ (1" -
Utner Sec: f).N/73 0.‘036 {Jujgba
h] ”J 7'7") 0 “I'Kp'
L--,)(:t ‘-.,I J!

Equation (9):
Lns & Disc: 0.
” variable:

n t 1\/:)
L1 . U. A

 

mQEHS

(*) At the

141

an

 

 

 

m:mmmooo.o mmoammom.o wmmmmooo.o Hooommmm.o mmfimmooo.o ofizmomzm.o ”omfla a med
.sz coflpmzdm
Hwomoooo.o maaamw:o.o mm:moooo.o maammmmo.o mwowoooo.o mwommmzo.o ”.omm Logpo
.Amv coflpmscm
mommmooo.oi mqozmmmz.o wwammooo.on mmfismmmz.o o:mamooo.ol momwmmw:.o ”.omm >00 mm
.Amv coflpmsvm
oumoaooo.on smmmmmmm.o wommoooo.ou Hamammwm.o omnwoooo.ou mmmmmmmm.o ”mEmpH zmmo
.AHV coflpwsvm
ARV pcmpmcoo AEVANQV Amav + pampmcoo “EVAHQV AHQV + pampmcoo coflpmsam
“macaw ”uqmogmch ”macaw “pamopmch ”macaw "pamopmch

 

”mmpwpm mcflxcmm pHc:

”mmuwpm wCHcochm UmpHqu

”mwpwpm mcflcocmpm muflzmpmpm

 

mCOHpmzvm HH¢

.mvcmhb QEHB mapmapm> w

.meOHm ucm mpgmopmch UmpmfiomH ”m-H> mqm¢e

CHAPTER VII

EVALUATION AND CONCLUSIONS

A Restatement of the Intent and Purpose
of the Study

 

 

Chapter IV, The Theoretical Structure, opened with the
following statement:
The basic purpose of this study is to deter—
mine the difference, if any, which exists among the
various banking systems as to restrictions on
branching and its impact on the economic level of
the state in which they are located.
The chapter goes on to state that "an attempt is made
to determine the extent to which the economic activity in a
state is influenced by the bank market structure specified
by state law." Recognizing that the major income—producing
activity of banks is making loans--loans which generate
economic activity--it follows that an analysis of the rela-
tive extent to which bankers allocate investment funds to
the various components of their portfolios, given the same
motivations for making such an allocation, would reveal if
any difference exists among the systems of banking delin-
eated by their individual state laws as regards the limits
to which they are permitted to branch. Of course, it is

necessary that assumptions be made regarding the relative

lU2

1U3

impact on the local economy of the alternative opportunities
available to the banker in the allocation of his investment
funds.

Limits on availability of income statistics, consid-
ered to be a major determinant of the quantity of invest—
ment funds devoted to each of the portfolio components,
required that the portfolio be divided into three major
parts:

(1) United States Government Securities.
(2) Other Securities.
(3) Loans and Discounts.

It was also recognized that the need for being liquid
presented a restriction on the quantity of investment a
banker could make in the income-producing components of his
portfolio so it was decided to include a component entitled
"Cash Items" as one of the alternatives available to him.

These four components represented substitutes avail-
able to the individual banker for the investment of his cap-
ital. The motivations influencing the banker's decision in
making an allocation of investment funds among the four
substitute components were prescribed to be the following:

(1) The need for liquidity represented by the
ratio of Demand Deposits to Total Deposits.
(2) The real income per capita in the state

in which the bank is located.

ff

iuu

(3) The rate of income earned on the funds
invested in a specific component. Of
course, no income is earned on the cash
component.

(u) The rate of income earned on funds invested
in alternative components.

These four motivations were designated the indepen-
dent variables leading to determination of the proportion
of the Total Portfolio invested in each of the substitute
components. After a somewhat detailed analysis of each
determinant in its relationship to the Y variable (Total,
this Component)/(Total Portfolio), reduced form equations
were constructed reflecting the considerations outlined
above. In order to eliminate multi—collinearity between
the last two X variables, they were included in the
regression equations as a ratio: (Rate of return, This
Component)/(Rate of return, Alternate Components).

The theoretical considerations reviewed in Chapter
IV led to a statement of the hypothesis as follows:

It is hypothesized that no difference exists
among the systems as to how they determine the pro—
portionate share of their portfolio to devote to
any one of the four components. Stated more spe-
cifically in terms of the statistical computation per-
formed, it is not expected that the Statewide
Branching system or the Limited Branching system
is significantly different (as revealed by F tests)
from the Unit Banking system in the proportion of its

funds devoted to any one of the components analyzed
given the same motivations for having made that

lUS

determination . . . The hypothesis can also be
stated that the alphas (or constants) in each
of the regression equations are not expected to
be significantly different from each other.

Also, as stated in Chapter IV, no attempt will be
made to determine the specific causes for any difference
which may be observed. The purpose of the study is to
determine if such differences exist and to leave to future
research the isolation of the specific reasons for those
differences. The study, as it is structured, will suggest
that any differences which might be observed will be due
to the varying market structure. Possibilities for such

differences in banking markets of varying branch structure

are reviewed in Chapter IV.

The Statistical Test

 

As outlined in Chapter VI, The Regression Results,
the statistical test was formulated in such a way as to
make possible a comparison of Statewide Branching or of
Limited Branching with the Unit Banking systems. Having
eliminated the variance attributable to the Statewide
Branching system and to the Limited Branching system from
the entire banking system, only the variance attributable
to the Unit Banking system remains. Consequently, the
comparison can be made on the basis of the extent to which
either of the branched systems is different from the Unit

Banking system.

IU6

The Statewide Branching System

 

It can be confidently concluded that the Statewide
Branching system is significantly different from the Unit
Banking system in certain specific areas of the study as
follows:

As to Equation (1), Cash Items

 

The Cross Section Regressions

 

The Statewide Branching system is revealed to be sig-
nigicantly different from the Unit Banking system at
the ten percent level of significance in twelve out

of twenty-two separate cross section regressions
(Table VI-Sa). It was significantly different from
the Unit Banking system in 5h.5 percent of the exam-
ples tested.

The average of the twelve regression coefficients for
the D1 variable which were significant at the ten per-
cent level is 0.H8050836 (Statewide Branching system)
in contrast with the average value of the Constant for
the same years of -0.26693A38 (Unit Banking system).

The Time Series Regressions

 

The time series regressions (Table VI—Sb) indicate
that the Statewide Branching system is significantly
different from the Unit Banking system with a signif-
icance greater than 0.0005. It is recognized that

the presence of time bias in this regression weakens

147
its argument. It is presented here for purposes of
confirming the more convincing argument appearing
immediately above.

The Estimated Y Values

 

A most convincing argument for the difference between
the Statewide Branching system and the Unit Banking
system is presented in Table VI—8a. Herein are pre—
sented estimates of the Y values for each year of the
study based on the cross section regressions. They
are computed at the means of the X variables. Every
year for the twenty-one years of the study (l9u6
through 1966) the estimated Y values for the Statewide
Branching system are lower than those of the Unit
Banking system; in fact, the average for the Statewide
Branching system is 0.1956 as compared with the aver—
age for the Unit Banking system of 0.2075. The State-
wide Branching system is 5.73 percent lower. This
consistent pattern presents a very strong confirming
argument for the difference between the Statewide
Branching system, given the motivations assumed to be
relevant to all three systems, and the Unit Banking
system. The direction of difference indicates that
the Statewide Branching system consistently operates
with a lesser proportion of its portfolio devoted to
the liquidity component thus releasing investment
funds for use in the economic-activity generating

components.

. ma]

1&8

Y Variable Time Trends

 

An analysis of the twenty-one observations of the Y
variable over time removes the influence of the
selected independent variables. This analysis, as
set forth in Table VI—9, reveals that the Statewide
Branching system has consistently over the twenty—one
year period operated with a smaller proportion of its
Total Assets devoted to the Cash component than has
the Unit Banking system. Its time—trend line lies
significantly lower on the Y axis than does that of
the Unit Banking system. The estimated Y value based
on the time-trend regression and computed at the mean
of the time variable for the Statewide Branching sys-
tem is 0.1856 as compared with the estimated Y value
for the Unit Banking system of 0.2235. The estimated
Y value for the Statewide Branching system is 16.9
percent lower than that of the Unit Banking system.
It would seem that the study has conclusively stated that
the Statewide Branching system is significantly different
from the Unit Banking system as regards the manner in which
it responds to the stipulated motivations for devoting a
certain prOportion of its investment funds to the ”Cash
Items" component. Furthermore, it is indicated that the
Statewide Branching system is capable of and has actually
Operated with a lesser proportion of its investment funds

devoted to liquidity needs than has the Unit Banking system.

1&9

This is deemed tremendously significant as the investment
funds not devoted to liquidity needs may be devoted to
alternate components some of which are more directly con—
cerned with the generation of economic activity in the
economy of the state in which the bank is located. To
counter those who would argue that the proportion of the
investment portfolio not devoted to liquidity needs is, in
the case of the Statewide Branching system, actually
invested in Other Assets such as Real Estate, Furniture

and Fixtures, etc., it is pointed out that, while the State—
wide Branching system does indeed invest more of its Total
Assets in this alternative, there is not a sufficient
quantity of investment funds so invested to obviate the
advantage which is indicated. For example, in l9u6, the
Statewide Branching system invested 0.0102 of its Total
Assets in the "Other Assets" component as compared with
0.0068 so invested in the Unit Banking system. By 1966,

the prOportion so invested in the Statewide Branching system
had increased to 0.03U6 while the Unit Banking system had
increased its preportion so invested to 0.0231. The average
proportion of the portfolio so invested over the twenty—one
year period of the study was, for the Statewide Branching
system, 0.0222 and, for the Unit Banking system, 0.0150.
While the Statewide Branching system devoted half again as
much more (“8.0%) to this component than did the Unit

Banking system, the magnitude of difference is not

150

sufficient to absorb the quantities of investment funds

released from unproductive employment as outlined above.

As to Equation (2), US Gov Sec

 

The argument for the significance of the difference existing
between the Statewide Branching system and the Unit Banking

system is not so strong in the case of this component.

TY“? Groin;(Secticniliegrmnniions

 

The Statewide branching system is significantly
different from the Unit Banking system five out of
twenty-two examples, or 22.7 percent. While this is
still greater than would be expected within the limits
of chance, the difference was not overwhelmingly con—
clusive. See Table VI-5a.

The Time Series Regressions

 

The Statewide Branching system was not indicated to

be significantly different from the Unit Banking sys—
tem as regards the United States Government Securities
component in the time series regressions (Tatle
VI-Sb).

The Estimated Y Values: Table VI-8b

 

In this analysis, the argument for the difference
between the Statewide Branching system and the Unit
Banking system is very strong. In all years except
one (1949) a lower estimated Y value is indicated for

the Statewide Branching system (0.3003) than is

151

indicated for the Unit Banking system (0.3357). It‘
is, in fact, 10.55 percent lower.
The Y-Variable Time Trends
An evaluation of the actual proportion of the port-
folio devoted to United States Government Securities
over time indicates that, while the intercept for the
Statewide Branching system is only slightly higher
than the Unit Banking system-—O.U6928907 (Statewide);
O.U636UOU2 (Unit)-—the time trend line is declining
somewhat more steeply. The estimated Y values at the
mean of the time variable indicates more signifi-
cantly the extent of the difference between these two
systems: Statewide Branching system, 0.3002; Unit
Banking system, 0.32U5. See Table VI-9.
While not as strongly indicated as in the "Cash Items" com—
ponent, the study, nevertheless, indicates that the State—
wide Branching system is significantly different from the
Unit Banking system as regards the manner in which it
reacts to similar motivations in allocating a proportionate
share of its investment funds to the United States Govern-
ment Securities component. Even more important, because
Government Securities Investment does not represent, neces—
sarily, a generation of economic activity in the local mar—

ket of the bank, the difference reflected is in the

152

direction of the Statewide Branching system consistently
investing less in this component than has the Unit Banking
system, thereby releasing yet more quantities of investment
funds to the portfolio components which are more local mar—

ket oriented.

As to Eguation (3)J Other Securities

 

)

The .ross Section Regressions: Table VI—Ga

(-

In this portfolio component, there was no indication
in the cross section regressions that the Statewide

Branching system was different from the Unit Banking
system in the manner in which it makes allocation to
Other Securities as a result of the motivations pre—
sented.

The Time Series Regressions: Table VI—Sb

Strangely, the Statewide Branching system was found

to be siimiificantly'ciifferent fiknn the ihiit Bankirw:

system in this regr;ssion when the hias of time was

yncesznit. It \MQS :signiifdxzard;ly (iiITferwent zit Tile f‘ive
percent level.

The Estimated Y Values: Table Vln8c

 

In fourteen of the twenty—one years, the Statewide

Branching system reflected an estimated Y value whicn
was higher than that of the Unit Banking system. The
average of the Y values for the twenty—one ye rs com—

puted at the mean of the X variables indicates that

153

the Statewide Branching system (0.0807) was 3.86 per-
cent higher than the Unit Banking system (0.0777).

The Y-Variable Time Trends: Table VI-9

 

The estimated Y values reflect the similarity of the
two systems as regards the "Other Securities" com—
ponent. The Statewide Branching system has an esti—
mated Y value, computed at the mean of the Time var-
iable, of 0.0820 as compared with the estimated Y
value for the Unit Banking system of 0.080U.
It can be concluded that the Statewide Branching system is
not significantly different from the Unit Banking system as
regards its response to the selected variables in deter—
mining the proportion of its investment funds devoted to
the "Other Securities" component. In addition, it has not
actually performed significantly differently from the Unit

Banking system over the last twenty—one year period.

As to Equation (Ml, Loans & Discounts

 

The Cross Section Regressions: Table VI-5a

 

The Statewide Branching system is significantly dif—
ferent from the Unit Banking system only four times
out of twenty-two in the Cross Section regressions.
While this is greater than would be expected within
the limits of chance, the argument as presented by
the cross section regressions is not overwhelmingly

conclusive.

154

r11

he Time Series Regressions: Table Vl~5b

 

The Time Series regressions did not indicate that the
Statewide Branching s'stem was significantly different
from the Unit Banking system in the manner in which it
responded to the selected variables in determining the
preportion of its investment funds devoted to Loans
and Discounts.

The Estimated Y Values: Table Vl-8d

 

in contrast to the above findings and most important
for purposes of this study, the estimated Y values
computed at means of the X variables indicates that
the Statewide Branching system reflects a higher
estimated Y value for each of the twenty~one years of
the study than does the Unit Banking system. The
average estimated Y value for this twenty~one year
period for the Statewide Branching system (0.“06U) is
12.1 percent higher than the Unit Banking system
(0.3625).

Y Variable Time Trends: Table VI—9

 

AWhen the influence of the determining variables is
removed and the trend over time of changes in the Y
variables is analyzed, it is evident that the State—
wide Branching system is consistently able and actu—
ally does devote a larger proportion of its portfolio
to Loans and Discounts than does the Unit Banking

system. Moreover, it is increasing the preportion of

155

its portfolio devoted to this very important component
at a faster rate than is the Unit Banking system. The
estimated Y values computed at the mean of the time
variable are consistent with those estimated in the
cross section regressions in that the Statewide
Branching system is devoting a significantly larger
preportion of its portfolio (0.Alb6) to the Loans and
Discounts component than is the Unit Banking system
(0.3595). The estimated Y value was, in fact, 15.3

percent higher.

The Limited Branching System

 

As might be expected, the Limited Branching system,
being somewhat of a composite of both the Statewide
Branching system and the Unit Banking system as it allows
branching only within some specifically designated limita—

tion, is not so consistently different from the Unit Banking

U4

ystem as is the Statewide Branching system. It is indi—
cated to be different in some specific areas of the study

as follows:

As to Equation (1), Cash Items

The Cross Section Regressions: Table VI—5a

 

There is no significant difference between the Limited

Branching system and the Unit Banking system in this

category.

156

jiug'rime Serdrxsihqgressions: YVdflxa‘Jl-Sb

 

The time series regressions indicated that the Limited

Branching system was significantly different from the

H

Unit Bankin ystem as regards the "Cash Items com-

0Q
m

ponent.

The Estimated Y Values: Table VI—Ba

 

The estimated Y vlaues computed at the mean of the X
variables indicates that the Limited Branching system
has estimated Y values which are higher than those of
the Unit Ehuiking systtnn for all tun; four of'tjha
twenty-one years. The average of the estimated Y
values over the twenty-one year period for the Limited
Branching system was 0.2110 as compared with the aver—
age for the Unit Banking system of 0.2075. This indi-
cates that the Limited Branching system finds it nec—
essary to Operate consistently with a larger propor—
tion of its total portfolio devoted to satisfying
liquidity needs than does the Unit Banking system. in
this respect, the Limited Branching system is the
least desirable system of the three in that it finds
it necessary to devote the largest proportion of its
portfolio to this unproductive employment.

Y Variable Time Trends: Table VI—9

 

The Y variable time trends indicate that the Limited
Branching system has an intercept lying closer to that

of the Unit Banking system than does the Statewide

157

Branching system. While the slope of this time trend
line is negative, the steepness of its slope is very

similar to that of the Unit Banking system. Th esti—

(D

mated Y values computed at the mean of the time var—
iable indicate that the Limited Branching system has
actually operated with a slightly lesser quantity of
investment funds devoted to liquidity needs than has
the Unit Banking system. The estimated Y value for

the Limited Branching system is 0.014H in contrast to

that of the Unit Banking system of 0.2235.

to Equation (2), United States Government Securities

The Cross Section Regressions: Table Vl-Sa

 

The Limited Branching system was found to be not sig—
nificantly different from the Unit Banking system as
regards this component.

The Time Series Regressions: Tahle VT-Sb

 

The Limited Branching system was not found to be sig—
nificantly different from the Unit Banking system in
this regard.

The Estimated Y Values: Table Vl—8b

 

In all years but one (l9u9), the Limited Branching
system is indicated as devoting a lesser proportion
of its investment funds to the purchase of United

States Government Securities (0.3196) than did the

Unit Banking system (0.3357).

158

Y Variable Time Trends: Table VI_9

 

In this component, the Limited Branching system has
the lowest intercept of all three systems. The time
trend line is declining with a steepness just slightly
greater than that of the Unit Banking system. The
estimated Y value computed at the mean of the time
variable indicates that this system has actually oper-
ated with a proportion of its portfolio devoted to
United States Government Securities at the mid—point
between the other two systems. It, nevertheless, is
Operating more efficiently in respect to its local

market than is the Unit Banking system.

As to Equation (3), Other Securities

The Cross Section Regressions: Table VI—Sa

 

The Limited Branching system was not found to be sig—
nificantly different from the Unit Banking system as
regards this component.

The Time Series Regressions: Table Vljfih

 

The Limited Branching system was found to be signif—
icantly different from the Unit Banking system in this
regression over time.

Estimated Values of the Y Variable: Tabl VI—8c

\L

For thirteen of the twenty-one years, the Limited
Branching system reflected a higher estimated Y value

than did the Unit Banking system. The average

159

estimated Y value for the twenty-one year period for
the Lindixmi Branclmiu; systenivwh: 0.0791 in; comparwfll
with 0.0777 for the Unit Banking system. The dif—
ference was only 1.8 percent higher.

Y Variable Time Trends: Table VI1Q

 

Hemoving the determining influence of the K variables
and evaluating the proportion of the portfolio devoted
to "Other Securities" over time indicates that the
Limited Branching system has the highest intercept of
the three systems. However, the slope of the time
trend line is less steep than either of the three sys-
tems while all three are increasing. The estimated Y
values computed at the mean of the time variable again
places the Limited Branching system at the mid—point
between the two other systems. It is operating some-
what more efficiently as regards its local state econ—

omy than does the Unit Banking system.

As to Equation (4), Loans and Discounts

 

The Cross Section Regressions: Table Vl—Sa

 

No significant difference is observed between the
Limited Branching system and the Unit Banking system

in this component.

The Time Series Regressions: Table Vl~5b

 

The Limited Branching system is indicated as being
significantly different from the Unit Banking system

in this time series regression.

viva-

160

The Estimated Y Values: Table VI-Bd

The average of the estimated Y values computed at the

mean of the X variables for the Limited Branching sys«

tem was slightly higher (0.3767) than for the Unit

Banking system (0.3625) and proved to be higher for

nineteen of the twenty-one years of the study. How—

ever, the estimated Y values for the two systems are

very similar and cannot be deemed to be importantly

different.

The Y Value Time Trends: Table Vl-9

 

The Limited Branching system depicts an intercept

lying about mid—point between the other two systems.

However, the trend line is increasing less rapidly

than the other two systems. The estimated Y values

computed at the mean of the time variable is

indicated

to be very similar to that of the Unit Banking system.

Analysis of Signs and Relative Regression
Coefficients: Tables VI-E & VI—7

 

 

Because identification problems exist in the
sions, analysis of signs and relative magnitude of
sion coefficients is precluded. It was hoped that
analysis could have been performed to indicate the

:response of one system in contrast with another as

regres—

regres-

such

relative

a result

(sf policy changes which might affect the X variables.

Tfiiere are evidently variables which are not included in the

rwegressions which, had they been present, would have made

161

such analysis possible. It must be remembered that this
type of analysis was not within the stated intent and pur—
pose of the study so it is not believed that such intent
and purpose is violated by this difficulty. The determin—
ation of the additional variables necessary to the consid-
eration which would provide a more consistent pattern of
signs and levels of significance is beyond the limits of
this study as they were defined in the hypothesis.

Estimated Values of Y Variable (Tine Series):
Table VI-8e

 

 

Because of the wide differences in relative values
between the various systems as revealed by the Time Series
equations in comparison with the Cross Section equations,
little or no importance is given to the estimated values

indicated on this table.

Summary

In summary, it can be stated that the study, in view

of the hypothesis posed, indicates the following:

(1) The Statewide Branching system is significantly
different from the Unit Banking system in the
manner in which it responds to similar motiva-
tions as regards the following:

(a) The proportion of its portfolio devoted to
liquidity needs.
(b) The proportion of its portfolio devoted to

United States Government Securities.

l illlldl' till]! Ill.

162

(c) The prOportion of its portfolio devoted to
Loans and Discounts.

Therefore, as regards these specific components

of the banker's portfolio, the hypothesis of the

study is disproved and, therefore, is rejected.

(2) The Statewide Branching system is not indicated
to be significantly different from the Unit
Banking system as regards the proportion of its
portfolio devoted to Other Securities. Any dif-
ferences indicated in analytical techniques other
than the regressions were small in magnitude and
are, therefore, not considered to be si,nificant.
The hypothesis of the study, therefore, in this
component is proved and is accepted.

(3) The Limited Branching system is found to be very
similar to the Unit Banking system. While anal-
ysis indicated some differences, they were gen—
erally of a minor nature and are not deemed to
be significant. Therefore, the hypothesis of
the study is proved in this regard and is
accepted.

The direction of the differences, when noted, were
differences which would lead to the conclusion that the
Statewide Branching system is, indeed, more beneficial for
‘the generation of economic activity in the state in which

it is located than is the Unit Banking system. The study

 

163

revealed that the Statewide Branching system can and does
generally operate with a lesser proportion of its investment
portfolio devoted to satisfying liquidity needs and that it
devotes a significantly larger proportion of its investment
portfolio to Loans and Discounts--a type of investment
activity local in orientation and capable of making more
significant contributions to the generation of local econ—
omic activity than would the Unit Banking system or the
Limited Branching system.

The study is unique in its approach to the evaluation
of a banker's decision—making process and its conclusions
are significant in the often-debated question as to how much
and how far branching privileges should be extended in order
to benefit the economy of the state in which the market
structure is located.

It is felt that the study has substantiated the thesis

preposed by Professor Weston.

BIBLIOGRAPHY

1614

BIBLIOGRAPHY

Adams, Silas Walter. The Legalized Crime of Banking and a
Constitutional Remedy. Boston: Reader Publishing
Company, 1958.

 

Alhadeff, David A. Monopoly and Concentration in Banking.
Berkeley: University of California Press, 195A.

"Reconsideration on Restriction on Bank Entry."
Quarterly Journal of Economics (May, 1962).

Allen, Frederic Lewis. Only Yesterday. New York: Bantam
Press, 1931.

 

. Since Yesterday. New York: Bantam Press, 19A0.

 

Annual Report. Washington, D. C.: Board of Governors of
Federal Reserve System.

 

Annual Report. Washington, D. C.: Comptroller of the
Currency, Department of the Treasury.

 

Annual Report. Washington, D. C.: Federal Deposit Insur—
ance Corporation.

 

Assets, Liabilities, and Capital Accounts, Commercial and
Mutual Savings Banks. Washington, D. C.: Federal
Deposit Insurance Corporation.

 

"Bank Resources and Economic Growth." he Bulletin. C. J.
Devine Institute of Finance, Graduate School of
Business Administration, #195 (Old Series). New York:
New York University Press, 1956.

 

Banking Concentration and Small Business. Staff Report to
Select Committee on Small Business, House of Repre—
sentatives, 86th Congress, Second Session. December
23, 1960.

Ikinking Law Journal. Encyclopedia of Banking LRBS.
Hartford: LaMont—Cross, 1963.

I3attenberg, R. A., and Ward, J. H., Jr. Applied Multiple
Linear Regression. (PRL—TDR 63-6). Texas: Lackland
Air Force Base, 1963.

 

165

166

Bell, Frederick W., and Murphy, Neil B. Costs in Commercial
Bankine. Research Report #Al. Boston: Federal
Reserve Bank, 1968.

 

H

Benston, George J. ”Branch Banking and Economies of Scale.
Journal of Finance, XX, #2 (May, 1965).

 

"Economies of Scale and Marginal Costs in Banking

 

Operations." National Banking Review, II (June,
1965).
Bibliograph of Studies in Banking Markets. Banking Markets

 

Unit, Division of Research and Statistics, Board of
Governors, Federal Reserve System. August 15, 196A.

Burstein, M. L. Monev. Cambridge: Schenkman, 1963.

Byrdsall F. History of the Loco—Foco or Equal Rights Party.
New York: Clement & Co., 18A2.

 

"Changes in Banking Structure, 1953—62." Federal Reserve
Bulletin. Washington, D. C.: Board of Governors,
Federal Reserve System, 1963.

 

Chapman, John M., and Westerfield, Ray B. Branch Banking,
its Historical and Theoretical Position in America
and Abroad. New York: Harper, 19A2.

 

 

 

Cohen, Kallman J., and Reid, Samuel R. "Effects of Regula—
tion, Branching, and Mergers on Banking Structure and
Performance." Southern Economic Journal, XXXIV, #2
(October, 1967).

 

7

"Commercial Banks and Economic Growth." The Bulletin, C. U.
Devine Institute of Finance, Graduate School of
Business Administration, #12. New York: New York
University Press, 1960.

 

Commission on Money & Credit. Mpney & Credit, Their Influ~

ence on Jobs, Prices, and Growth. Englewood Cliffs:
Prentice-Hall, 1961.

 

 

"Commission on Money & Credit, Review of Report of the."
Hearings Before the Joint Economic Committee.
Congress of the United States, 87th Congress, First
Session. August 1A—18, 1961.

 

Committee on the Working of the Monetary System, Report of
the. London: Her Majesty's Stationery Office, 1959.

 

Conant, Charles A. A Modern History of Banks of Issue.
London: Putnam, 1927.

 

167

"Concentration of Banking in the United States." Staff
Report of Board of Governors. Federal Reserve System.
Submitted to Subcommittee on MonOpoly of Select Com-
mittee on Small Business, United States Senate.

September 10, 1952.

"Conflict in Federal and State Banking Laws." Washington,
D. C.: Committee on Banking and Currency, 88th Con—

gress, First Session, 1963.

I

Crosse, Howard D. "Bank Structure and Competition.‘
Journal of Finance, XX, #2 (May, 1965).

Dorfman, Joseph. "The Economic Philosophy of Thomas Jeffer-
son." Political Science Quarterly, XV (March, 19AO).

Draper, N., and Smith, H. ’Applied Regression Analysis.
New York: Wiley, 1967.

Eckhardt, H. M. P. "Branch Banking Among the State Banks.”
The Annals of the American Academy of Political and
Social Sciences, XXXVI,#3 (November 3, 19Al).

Edwards, Franklin R. Concentration and Competition in

 

 

Commercial Banking: A Statistical Study, Research
Report #26. Boston: Federal Reserve Bank of Boston,
1964.

Fischer, Gerald C. Banking Holding Companies. New York:
Columbia University Press, 1961.

Galbraith, John A. The Economics of Banking Operations.
Montreal: McGill University Press, 1963.

Galbraith, John Kenneth, Jr. The Great Crash. New York:
Houghton-Mifflin, 1961.

(haldfeld, Steven M. "Commercial Bank Behaviour and Economic
Activity." Econometrica. New Amsterdam: North
Holland Publishing Company, 1966.

(havernor General, Alaska Territory. Postwar Alaska.
Washington, D. C.: United States Department of the

Interior, Division of Territories and Island Pos-
sesions, 1949.

erunley, Lyle E. A Study of Scale Economies in Banking.
Kansas City: Federal Reserve Bank, 1962.

Grmuanbaum, Stuart I. ‘Banking Structure and Costs: A Sta—
tistical Study of the Cost—Output Relationship in
Commercial Banking. Ph.D. dissertation, Johns HOpkins

University, 196A.

168

. ”Competition and Efficiency in the Banking
Industry——Empirical Research and its Policy Impli—
cations.” Journal of Political Economy, Supplement
(August, 1967).

 

Gurley, John G.,and Shaw, Edward S. Money in a Theory of
Finance. Washington, D. C.: Brookings, 1960.

 

Hamilton, Alexander. "Opinion on the Constitutionality of
the Bank," February 23, 1791. Living Ideas in
America. Edited by Henry Steele Commager. New York:

Harper, 1951.

 

Hammond, Bray. Banking and Politics in America, from the
Revolution to the Civil War. Princeton, New Jersey:
Princeton University Press, 1957.

 

 

. "Jackson, Biddle, and the Bank of the United
States." Journal of Economic History (May, 19A7).

 

Hansen, Alvin H. Monetary Theory and Fiscal Policy. New
York: McGraw-Hill, 19A9.

 

Hester, Donald D., and Zoellner, John F. "The Relation
Between Bank Portfolios and Earnings: An Econometric
Analysis." Review of Economics and Statistics,
XLVIII (November, 1966).

 

Historical Statistics of the United States, 1789—19A5.
Washington, D. C.: United States Bureau of the
Census, 19A9.

 

Historical Statistics of the United S ates from Colonial

Times to 1957. Washington, D. C.: United States
Bureau of the Census, 1960.

 

 

Hogenson, Palmer T. The Economics of Group Banking. Wash—
ington, D. C.: Public Affairs Press, 1955.

 

Hoggson, Noble Foster. Banking Through the Ages. New York:
Dodd, Mead & Co., 1926.

 

iiorvitz, Paul M. Concentration and Competition in New
England Banking. Research Report #2. ioston:
Federal Reserve Bank, 1958.

 

 

. "Economies of Scale in Banking." Ppiyate
Financial Institutions. Commission on Money and
Credit. Englewood Cliffs: Prentice-Hall, 1963.

 

 

iinox, John J. History of Banking in the United States. New
York: Bradford Rhodes & Co., 1900.

 

169

Jefferson, Thomas. "Opinion on the Constitutionality of the
Bank," February 15, 1791. Living Ideas in America.
Edited by Henry Steele Commager. New York: Harper,
1951.

 

Lanzillotti, Robert F. Testimony of, in "Conflict Between
Federal and State Banking Laws.” Hgarings. Committee
on Banking and Currency, House of Representatives,
88th Congress, First Session. April 30—May 6, 1963.

. "The Growing Threat of Monopoly in America."
Speech delivered before Independent Lanking Associa-
tion, New Orleans, Louisiana, March 27, 1962.

 

Lawrence, Joseph Stage. Banking Concentration in the United
States. Bankers' Publishing Company, 1930.

Lawson, William John. History of Banking. London: Richard
Bentley, 185D.

 

McCulloch v. Maryland (A Wheaton 316).

Miller, Harry E. Banking Theory in the United States Before
1860. Cambridge: Harvard University Press, 1927.

Miller, John G. The Federalist Era, 1789-1801. N w York:
Harper, 1960.

 

Minsky, Hyman P., ed. California Banking in a Growing
Economy: 19U6—1975. Berkeley: Institute of Business
and Economic Research, 1965.

 

Morris, Richard B. "Alexander Hamilton After Two Cen—
turies." The Basic Ideas of Alexander Hamilton.
Edited by Richard B. Morris. New York: Pocket, 1957.

 

Motter, David C., and Carson, Deane. "Bank Entry and the
Public Interest: A Case Study." National Banking
Review. June, 196“.

 

"National Bank Act." Revised Statutes of the United Statei.

Washington, D. C.: United States Government Printing
Office, February 25, 1863.

 

Ostrolenk, Bernard. The Economics of Branch Banking. New
York: Harper, 1930.

 

Pesek, Boris P., and Saving, Thomas R. Money, Wealth, and
Economic Theory. New York: Macmillan, 1967.

 

 

. The Foundations of Money & Banking. New York:
Macmillan, 1963.

 

 

170

Phillips, Almarin. "Competition, Confusion, and Commercial
Banking. Journal of Finance, XIX, #1 (March, 1964).

 

Pierce, James L. "Commercial Bank Liquidity." Federal
Reserve Bulletin. August, 1966.

 

Prather, Charles L. Money and Banking. Chicano: Richard
D. Irwin, 1949.

 

Robbins, Sidney M., and Terleckyj, Nestor B. Money
Metropolis A Locational Study of Financial Activities
in the New York Region. Cambridge: Harvard Univer—

sity Press, 1960.

 

 

 

Robinson, Roland I. The Management of Bank Funds. New
York: McGraw-Hill, 1962.

 

Roosa, Robert V. Federal Reserve Operations in the Money
and Government Securities Markets. New York: Federal
Reserve Bank, 19561

 

 

Russell, William R. "Commercial Bank Portfolio Adjust-
ments." American Economic Review, Supplement (May,

196A.)

 

Schlesinger, Arthur M., Jr. The Age of Jackson. New York:
Little Brown, 1950.

 

Schweiger, Irvin, and McGee, John S. "Chicago Banking, The
Structure and Performance of Banks and Related
Financial Institutions in Chicago and Other Areas."
Journal of Business. Chicago: University of Chicago,
Graduate School of Business, 1961.

 

Shull, Bernard, and Horvitz, Paul M. "Branch Banking and
the Structure of Competition." National Banking
Review, 1, #3 (March, 196A).

 

. "The Impact of Branch Banking on Bank Perform—

ance." National Banking Review (December, 196A).

 

 

Southworth, Shirley Donald. Banking Facilities for Bankless
Towns. New York: American Economists Council for the
Study of Branch Banking, 19141.

 

£3tarnes, George T. Sixty Years of Branch Banking in Vir—
ginia., New York: Macmillan, 1931.

 

fStatistical Abstract of the United S ates. Washington,
D. C.: United States Department of Commerce.

 

171

Stolper, Gustav. The German Economy, 1870 to the Present.
New York: Harcourt, Brace, and World, 1967.

 

Suits, Dan B. "Use of Dummy Variables in Regression Equa—
tions." Journal of American Statistical Association,

 

52 (1957).

Survey of Current Business. Washington, D. C.: United
States Department of Commerce.

 

 

Trefftzs, Kenneth L. The Bankless Communities in Illinois.
Urbana: University Press, 1939.

Tussing, A. Dale. "Stimulating Bank Competition: Comment."
Journal of Finance (December, 1965).

 

United States v. Calumet National Bank of Hammond, Indiana,
and Mercantile National Bank of Hammond, Indiana.
Civil Action #3727 (DCND Ind 1963).

United States v. Continental Illinois National Bank & Trust
Company of Chicago, et a1. Civil Action #61e lAAl
(DCND Ill. 1961).

United States v. Crocker—Angelo National Bank, Citizen's
National Bank and Transamerica Corporation. Civil
Action $31808 (DCND Cal 1963).

United States v. National Bank of Lexington, 8“, S Ct

1033 (196M).

United States v. Philadephia National Bank, 201 F Supp 3A8
(Ed Pa 1962).

United States v. Philadelphia National Bank et 31: United
States Supreme Court, October term, 1962, #83.

United States v. Philadelphia National Bank, 37“ US 321
(1963).

Westerfield, Ray B. Historical Survey of Branch Bankinn in
the United States. New York: American Economists
Council for the Study of Branch Banking, 1939.

 

 

Weston, J. Fred. Economic Analysis of the Proposed Merger
of the Crocker-Angelo National Bank and the Citizen;:
National Bank. Affidavit submitted to Comptroller-of
the Currency, 196A.

 

 

 

Review of "California Banking in a Growina
Economy." Edited by Hyman P1 Minsky. National
Banking Review, 3, #4.

 

 

172

Whittlesey, Charles R., Freedman, Arthur M., and Herman,
Edward S. Money and Banking: Analysisuand Policy.
New York: Macmillan, 1963.

 

Willitt, Virgil. Selected Articles on Chain, Group, and

 

Branch Banking. New York: H. W. Wilson, 1930.

 

"'Tlll'liifliijfllliflilfl

3 030

     

1111111“