Full text of Economic Review (Federal Reserve Bank of Kansas City) : March 1961

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

March

1961

NEBR .

COLO.

Interpretation of Size-Cost
Relationships in Banking . . . . . page
Wheat UtilizationTrends and Prospects
Current Statistics

. .

3

page 10

.

page 16

Sub cr-iption to th MONTHLY R •vrnw are available to the public without charge. Additional
copies of any issue may be obtained from the
Research D epartment, Federal Reserve Bank of
Kansas City, Kansas City 6, Missouri. Permission
is granted to reproduce any material in this
publication.

Jnterprelation o/

Size-Cost Relationships
article in th last issue of this Review
dealt with the relationship of size and
costs at member banks in the Tenth Federal
Reserve District. Based on data covering the
p riod 1956-59 for a sample of about 270 District members, it was shown that ratios of
cos ts to assets were lower for large banks th an
for small banks , aft 'r a]lowan · was mad
for th ' influ •11c' on cos ts of va ri ou · chara -t risti ·s of th ' banks, such ·1s th p r · ntag of
ass ts in th form of loans and th r Jativ
volum of tim d posi ts. A major shar of th
cost economies of large-scale operations was
traced to the wage and salary component of
total expenses.
The February article left unopened several
lines of investigation that are pursued in the
following pages. An important question has
to do with the effects on costs of bank characteristics which w re not brought into th e
discussion. A se ond question d als with the
possible sources of cost advantages at larg
banks. If cost economies arc located mainly
in the wage and salary component of total
expenses, do they result from lower wage and
salary payments per employee or from other
sources? By extending the analysis in these
directions, the present article seeks to refin e
further the interpretation of size-cost relationships in banking.
convenient point of departure for th e disussion to follow may be found in a summary of the m thod of approach us ed in the
pr vious article and of the r sults obtained.

A

A Brief Review

To measure differences in costs among
banks that are associated with size, it is necesMonthly Review • March 1961

•

In

Banking

sary to make allowances for variations in osts
that are due to other cost-determining influences. The statistical method of multiple r gression and correlation analysis is ideally
suited to this purpose, because it p ermits estimates of the separate influence of size and
oth r factors that affect bank cos ts.
Exp rim ntation jn licatcd th at djff r nc s
in ratios of total ·osts to total assets among
all sam pl b ank ov r th p riocl ] 956-59
ould b st b a ountcd for in terms of six
charac teristics of the banks: bank size, th ratio
of time to total deposits, the ratio of total
loans to total assets, the ratio of securities
other than U. S. Government issues to total
assets, the ratio of consumer loans to total
loans, and the percentage growth of assets between 1956 and 1959. The average relationship found b etween total costs ( as a per cent
of assets) and bank siz , after allowing for
effects on costs of th oth r fiv charact ristics,
is shown in th e top pan 1 of Chart 1. Th
bottom panel shows the comparable relation
b tween wages and salaries ( as a per cent
of assets) and size among the sample banks.
The fall in total cost ratios with increasing
bank size is associated mainly with economies
in wage and salary expenses. The size-cost
line in the bottom panel of the chart indicates
a decline of .33 p ercentage points in the ratio
of wages and salaries to total assets ( in per
c nt) for a h tenf Id in crease in size of bank.
If it w re possibl to make a curate adjustments for wage and salary payments attributable to trust departments of the banks, the
decline would be slightly larger. This is because a omewh at greater proportion of wages
and salaries is accounted for by trust activities

3

Size-Co t Relationships in Banking
Chart 1.
R lationship of Cost Ratios and Bank Size
Sample of Tenth District Member Banks, 1956-59
PER CENT ( r~;; : ~L

Cf:~:T~O)

3 .2
2 .8
2 .4
2 .0
1. 6 [
PER CENT ( w~~ESrtrN:L

s:i::r-:s)

2. 0

1.61
1. 2

.e
.4

10

.5

ASSETS IN MI LL I ONS O

NOTE :

The

top

panel

50
OOL

of th e chart

AR

100

( LOO SC AL )

Is ba sed on th e equation :

Xi - l.377 - .394 log X2 .0162X ;d .0221X 4 -.01 26X 5 I .0156X 6 --J
.0078X 7 , where X1 is the ratio of total costs to total assets, and X2

...... X7 are the six ch racteristics of the bank s in the order indi cated in the text. All ratios are expressed in percentage terms ; size is
measured by asset s in millions . The bottom panel of the chart is
based on the equation : X 1 = 1.108 - .327 log X2 +. 0087X 4 -.0115
Xs .0083X 6 +. oo30X1, where X 1 is the ratio of wages and salaries
to total assets , and X2 • . • . • X7 are defined as above . The
variable X3 , representing the ratio of time to total deposit:;, was
omitted in the second equation for lack of statistical significance .

at larger banks in the District.
Trust activities, however, apparently do not
account for a substantial share of wage and
salary expenses even at the larger District
banks. This is suggested by th fact that, on
th av rag , trust departm nts ac ount d for
less than 4 per cent of total earnings at District m m hers with more than $50 million in
asset during the period 1956-59, and for less
than 2 per cent of total earnings at sample
banks with $25-$50 million in assets. At still
smaller banks, the fraction is correspondingly
less. Inability to separate wage and salary
payments attributable to trust activities from
other wage and salary outlays thus does not
affect s riously the b havior of osts in r lation to bank siz among Distri t m mb rs.
Wage and Sal ry Payments Pe

Em

I

yoe

A question may b e raised as to whether the
declin e in wage and salary ratios with increasing size of bank may reflect differences in
4

average wage and salary payments per employee. This question may seem rather peculiar at first, since average salaries per officer
and per nonofficial employee are widely believed to be higher at large banks. It is possibl , nevertheless, that average payments per
employee - without regard to official status could be larg r at small banks, since they hav
comparatively high r numb rs of offi ers in
relation to nonofficial employees.
The data in Table 1, showing average payments per employee by bank size, were
gathered for a small r sample of the 270 member banks includ d in th study. Calculations
of av rag paym nts per mploye were has cl
on total annua l wage and sa lar clisl ms mcnts
divided by th ' numb r of mplo
s at y ar
nd. 1h figur s do not r pr s nt av rag
mannual wag s and salari s of full-tim
ployees, because the number of employees
included part-time workers.
Average annual payments per officer and
per nonofficial employee are seen to increase
sharply with rising bank size, but the ratio of
employees to officers is much higher for the
larger banks. The result is that average annual payments p er p rson for all employees
are virtually the sam for all bank size group .
The d clin in wage and salary ratios with
increasing bank siz shown in Chart 1 does
not, therefore, simply r fleet diff ring amounts
of compensation per mployee. Rather, the
cost advantages of large banks are reflected in
lower total numbers of employees per dollar
of assets and a higher proportion of nonofficial
Table 1.
Estimated Averag

yment

Per Employee,

S'z
Sample of Tenth District Member Banks, 19S9
Bank Size
(Asset s in
millions of
dollars)

0-10
10-25
25-100
Over 100

Officers

Other
Employees

All
Employees

Average Ratio
of Nonofficial
Employees to
Officers

$7,740
8,970
10,440
11,690

$2,580
3,120
3,290
3,550

$4,530
4,540
4,520
4,500

1.90
3.27
5.00
7.76

Average Annual Payment per Person

Size-Cost Relationships in Banking

employees. The number of total mployees
p r $1 million of assets averages out to just
over 2 for banks with more than $100 million
in asset and to about 3.3 for banks with less
than $10 million in ass ts.
Testin

f r

t lnfl

nc

Th n xt qu stion to inv stigat is th posibility that th ability of larger banks to
op rat with small r numb rs of employ s
per doJlar of ass ts may reflect charact ristics
of assets and liabiliti s among the sample
banks that h av not y t b n consider d . Th
fa t th at cliff r n
in o ts among alJ sampl
rC' r ·la t d to sizC' and flv oth r harh e, ks
" tcristi ·s manif stly cl s , ot m '.-1 11 th, t
th r w re no oth r fa tors aff ting osts of
th banks. It only impli s that fort hni al reaons, th ability to id ntify additional sour
of cost variation was limit d. 1 Sine this is
the case, it is possible that by con idering
additional characteristics of the sample banks,
a clue may be found as to the source of cost
advantages enjoyed by larger District members.
Among th additional characteristics of
ass ts and liabiliti s that might h lp to xplain
diff r n
in o ts among banks are thr
whi h merit d tail d att ntion b caus th y
ar rath r los ly asso iat d with bank siz .
( 1) D mand d -po it liabilities of larger Di trict memb r banks are compris d rath r
heavily of interbank balances. Interbank deposit liabilities are uncommon among District
1 The inability to identify other cost-determining influences results, to some extent, from the fact that
charact ristics of the individual banks are intercorrelated. on qu ntly, only a limited numb r of ind p nd nt variabl s an b in lucl d in th r gr ssion
analysis without n ount ring prob! ms of inlcrcorrelation. Th int rpr tation of size- ost r lationships is
affected importantly if a characteristic is omitt d b caus of a stron int rcorr lation with bank siz . The
di u sion in thi s ction d als with charact risti s
omitt d from the r gression analysis for thi reason.

Monthly Review • March 1961

banks with less than $25 million in assets, but
all larger sample banks have some correspondents among their deposit customers. If high
ratios of int rbank to total demand deposits
tended to b associated with low ratios of
costs to ass ts, this would help to account for
the cost advantag
njoy d by larg r banks.
( 2) Du mainly to differential res rv r quir ments b tw n r s rv city and country member , larg r District banks hold a gr at r fraction of their liquid assets ( cash and Government securiti s) in th form of cash. ( 3) Loan
portfolio of larg r banks in th District are
m r h av iJ w ight d with bu in ss loan ,
whil th ·on ntration , t ·mall r hanks i in
th · ar ·a of nonguarant' cl fc rm reel its. This
diff r n i, ass t stru tur woul l h Ip to xplain th cost advantag s of larg' banks if
busin ss loans g n rally ntail d lower administrativ costs than extensions of credit to
farmers.
Interbank Deposits

The influence on costs of differing relative
amounts of interbank deposits may be ascertained by concentrating attention on cost differences among larger banks in th District.
When this i don , th technical problem prohibiting in lusion of this chara t risti of deposit stru tur in th tati ti a] analysis for aJJ
sampl banks is not encount r d . Thus, for
District bank with over $25 million in assets,
it is possibl to obtain a measur of the decline in costs with increasing bank size both
before and after allowing for the influence on
costs of differing ratios of interbank to total
demand deposits.
The dotted line in Chart 2 shows the average relation hip between bank size and ratios
of wag s and alari s to total ass t before
r moving th ost influence of cliff ring ratios
of int rbank to total demand deposits among
the larger banks. The solid Jin in Chart 2
shows th av rage relationship between wage
and salar ratios and bank siz after a1lowance
for the cost effects associated with differing

s

Size-Cost Relationships in Banking
Chart 2.
Relationship of Wage and Salary Ratios
and Bank Size
Sample of Larger Tenth District Member Banks, 1956-59
PER CENT (

w~gE~o~~~ s~'s\\~~s)

1.2
I. I

1.0
.9

.8
.1

.6
0

100

200

300

400

ASSETS IN MI LLIONS OF DO LL AR S

NOTE : The dotted li ne Is based on the equation : X,
.517 .00090X
.0151X 4 .0176X 5 \ .0038X 6 I .00 28 X, , wh ere X, Is 1he
ratio of wa ges and sa lari es to total assets and X1, X4 . . . X,
are defined as In th e note to Chart 1. The so lid line Is based on
th e equation : X1
.559 - .00079X 1 - .0157X4
.0145Xs I .0036
X1> .0023X 7 - .0045X 8 , where X1 . . . . • . X, are defined as
above and X8 is the ratio of Interbank to total demand deposits
in per cent. The multiple correlation coefficient is .77 for the
first equation and .79 for the second.
In both equations, the variables X6 and X7 are not statistically
significant at the 5 per cent level, but are included to maintain
parallelism with the equations on which Chart 1 is based . Omitting
these two variables would have resulted in a higher absolute value
of the regression coefficient of the size variable in both equations ,
but by less than one standard error.
Size of bank is expressed in absolute terms, rather than in logs as
for the equations underlying Chart 1, because there is no clear
preference for a logarithmic measure either i n terms of goodness
or linearity of fit.

=

relative amounts of interbank demand balances.
Two implications may be drawn from the
fact that the solid line is the less steeply sloped .
First, high ratios of interbank deposits tend to
be associated with low cost ratios, other things
being equal, which accounts partly for the cost
advantages enjoyed by larger banks in the
District. But second, it is also evident that
there are other sources of cost advantage accompanying larger size, since the size-cost
line still is tilted downward after allowance
has been made for the relatively lower costs
associated with a high percentage of interbank
demand deposits.
Other Characteristics of A sset Structure

The distribution of liquid assets between
cash and Government securities, and differences in relative amounts of business and farm
6

loans extended by large and small banks, seem
to be much less important sources of cost
advantage to larger District members. The
influence on costs of these differences in asset
structure may be tested by examining banks
of approximately similar size, but with markedly different cost ratios, to see if cost differences among them are related to these
aspects of loan structure and liquid asset holdings. If these characteristics of assets are not
found to be significant in explaining cost differences between banks of similar size, it is
unlikely that they are important in explaining
differences in costs betw en banks in different
size groups.
To obtain the data for Tabl' 2, th sample
banks wer' divid ed into Rv size dass s.
Within ach size class, certain banks were
selected because they had unusually high
wage and salary expenses in relation to what
was expected on the basis of factors identified
in the statistical analysis described earlier as
being important cost determinants. Other
banks were selected because their wage and
salary ratios were much lower than expected. 2
Average ratios of business loans to total loans,
nonguaranteed farm loans to total loans, and
cash to total liquid asset holdings then were
comput d for each of the separate groups of
banks.
The reader should bear in mind, as the data
in Table 2 are examined, that the asset ratios
for individual banks within any given size class
show wide variation. Therefore, it is meaning2

Banks with extreme cost ratios were identified according to the value of the residuals ( actual minus
expected cost ratios) associated with the regression
equation th at underlies the bottom panel of Chart 1.
The cutoff points used to establish extreme residuals
were: ( 1) banks with $0-$5 million in assets-residuals exceeding 0.30 ( in absolute value) ; ( 2) banks
with $5-$10 million in assets-residuals exceeding 0.20;
( 3) banks with $10-$25 million in assets-residuals
exceeding 0.15; ( 4) banks with $25-$50 million in
assets-residuals exceeding 0.10; ( 5) banks with over
$50 million in assets-residuals exceeding 0.10.

Size-Cost Relationships in Banking
Table 2.
Cost Differences
Charact

nd Structural

i f

Sample of Tenth District Member Banks, 1956-59

Bank Size
(Assets in millions
of dollars)

0-5

High costs (12 banks)
Low costs (16 banks)

5-10

High costs (12 banks)
Low costs (8 banks)

10-25

Hi gh costs (9 banks)
Low costs (10 banks)

25-50

High costs (8 banks)
Low costs (7 banks)
Over 50
High costs (6 banks)
Low costs (5 banks)

Characteristics of the BanksAverage ratios of:
Nonguaranteed
Business Farm Loans
Cash to
to Total
Loans to
Liquid
Loans
Total Loans
Assets

38.7
41.2

9.0
11.2

49.5
41.4

37.2
36.0

23.9
21.7

25.6
26.9

40.6
36.6

28.0
27.3

21.6
20.8

40.0
50.5

38.2
39.6

3.7
12.8

46.6
50.8

45.5
46.3

6.0
3.5

ful to ask whether cost differences between the
banks in any size class are related to the
characteristics of their assets listed in the table.
For example, ratios of business to total loans
for high cost banks with assets of $0-$5 million
run from zero to 35 per cent. Among low
cost banks in that size class, the range is from
3 p r cent to 36 per cent. For banks with over
$50 mi1lion in assets, the range is from 35 per
cent to 61 per cent for high cost banks and
from 18 per cent to 70 per cent for the low
cost group. Averaging these ratios is done to
reveal any systematic association between the
costs of banks in each size group and the
character of their assets.
The first column of the table shows ratios of
cash to total liquid assets. Average ratios do
not differ greatly between high and low cost
banks in the $0-$5 million group, and for the
$5-$10 million group the difference is also
small. More importantly, the high ost banks
in the $5-$10 million group hold, on the average, a greater proportion of cash to total
liquid assets. Since costs are expected to fall
Monthly Review • March 1961

with an increasing proportion of cash assets,
the ratio of cash to total liquid assets does not
explain cost divergences for banks in this
group. Similar reasoning also applies to cost
differences of banks in the $10-$25 million
class. Only in the two largest size groups,
where the low cost banks hold the larger
relative amoun ts of cash, do differences in
ratios of cash to liquid assets help to explain
cost differences among the banks. The ownership of cash as a liquid asset, therefore, does
not appear to be a factor of fundamental importance in explaining why some banks' costs
ar cliff r nt from others, and it could not be
argu d onvincing]y that th greater rclati v
amount of ca ·h assets h Jd by larg r Di trict
m mb rs ac ·om ts for mor than a minor
share of th cost advantages they njoy.
Average ratios of business loans and nonguaranteed farm loans to total loans display
no systematic association with cost differences
among banks. High cost banks in some size
classes have the higher average ratio of business to total loans, and in others, the lower
ratio. The same is true of farm loan ratios.
Thus, while large banks extend a substantially
greater proportion of their loans to businesses
and a smaller proportion to farmers than do
small banks, this di.ff rence in the category of
borrower is not by itself sufficient to account
for the materially lower costs of larger District
members.
On the basis of this evidence, it seems reasonable to conclude that the cost advantages
enjoyed by larger banks in the District do not
stem simply from differences in the division
of assets between loans, securities, and cash,
from the type of loans made by major class of
borrower, or from differences in the class of
d positor. Rather, there are other sources
more intimat ly associated with size of bank
which must account for the ability of larger
banks to operate with lower numbers of total
employees per dollar of assets, and with a
higher percentage of nonofficial employees.
7

Size-Cost Relationships in Banking
The

ue t on of R lative Efficiency

It is possible that the cos t advantages of
larger banks stem mainly from their ability to
perform activities of all kinds with greater
efficiency. Certainly, a strong logical argument can be made for the view that increased
size in banking permits the organization of
banking functions along lin s that are likely
to add significantly to productivity. The gr ater use of mechanical and electronic equipment to facilitate the va t amount of bank
accounting tasks is a case in point, but
mechanization in the field of bankin g, mad
possible by recent innovations in· electroni
a o mting and computin quipmcnt, has y t
to r gist r its full ff ·ts at l istri ·t m mhcr
bank . How ost of Distri 't banks will l)l
inAu n d by in r as d mploym nt of u h
quipm nt- a d v lopm nt that is und r way
-remains to be seen.
A more important source of added efficiency
in the past probably has been the much higher
degree of functional specialization among
bank employees made possible by larger-scale
operations. No one observing the operations
of a relatively large and a relatively small bank
could fail to notice the far wider diversity of
tasks performed by an officer or mploy e of
the small bank. Incle d , at th v ry small st
Table 3.
Checking Account A t1v1t

at Kansas Banks,

1959
Size of Bank
(Deposits in millions
of dollars)
less than 1
1-2
2-3
3-5
5-7½
7½-10
10-20
Over 20

Average No.
Average Average No. ofTotal Items
Balance of Total Items
per$100
per
per Account
in Deposits
Account
per Month
per Month

$769
866
1,032
1,134
1,046
1,204
1,225
1,917

18.0
19.8
25.6
26.3
28.3
33.2
33.7
59.6

2.39
2.34
2.64
2.39
2.63
2.78
3.02
2.45

NOTE : Total Items include credit items to demand accounts, checks
" on us," local clearing items , and out-of-town remittance items .
SOURCE: 1960 Report, Bank Management Commission of the Kansas Bankers Association .

8

member banks conducting business in the
District- banks with assets of less than $1 million and a com pl em en t of only two or three
total employees-tasks performed by the senior
officer may include everything from comparatively routine maint nance of accounts to the
high st level manag ment d cisions. At th ,
Jarg r banks, th mor routine tasks ar taken
ov r by person Jess broadly skilled in th
field of banking- which acco unts for the relatively higher perc ntag of nonofficial employees at larger banks- permitting the official
staff to devot its attention increasingly to
cl cision-making fun tions. It would hardl y h e
surprising to find that lh c hi gher degree of
specialization among employers at larger
hanks was r fl ' ·kd i11 lowN bank ·osls.
Average Size of Loans and Investments

Th re is one important differ nee, how v r,
in the nature of larg and smalJ banks , hich
prohibits ascription of alJ of the cost advantages of large-scale operations to greater efficiency in the performance of identical tasks .
This is the fact that the cost of acquiring earning assets at small and large banks differs
fundamentally by reason of disparities in the
dollar amounts in which individu al earni ng
assets ar acquir d . This is cspe ·ia11y im portant with regard to Joans, incc administrative costs p r loan decline steeply with
incr asing size of loan , but it may also be tru
with respect to purchases of s curities. Differences in the average size of certain types
of loans-such as business loans-at large and
small banks are especially great. In a survey
of business loans at District member b anks
several years ago, banks with less than $10
million in assets reported business Joans with
an average size of Jess than $5,000. For banks
with more than $100 mi11ion in assets, th
av rag siz was about $65,000, or more than
t n times as Jarg .
A similar principle do snot n cessarily hold,
however, with resp ct to average size of deposits. For while the average dollar balance

Size-Cost Relationships in Banking

per account increases with bank size, so also
does the amount of activity per account. Data
published by the Kansas Bank Management
Commission show that, for banks in the state,
ch eking account activity- measured in terms
of numbers of items processed over a month
- p er dollar of demand deposits is not materially differ nt for banks with over $20 million in d eposits than for banks with less than
$1 million in d eposits . The relevant data are
reprodu ced in Table 3.
The cost ad vantages of larger banks in the
District, th refore, appear to result both from
hi ghe r cffi ·icncy in th' performanc of regular
hanking functions comparabl e to those performed al small hanks and from th e acquisition of c,nning assets in larger dollar amounts .
The opportunity to redu ·c cos ts in each of
th ese ways is c]osely tied to size of bank, and
the separate influ nee of each on costs cannot
b evaluated from available data.
A finding that large banks are able to operate more efficiently than small banks should
not be taken to imply that efforts to minimize
costs are pursued less aggressively the smaller
is the size of bank. For unless smaller banks
were uniformly less energe tic in their attempts

Monthly Review • March 1961

Table 4.
Measur

o

Relati e Cost Variation

Sample of Tenth District Member Banks, 19S6-S9

Bank Size
(Assets in
millions of dollars)

Measure of
Relative Cost
Variation

0-5
5-10

15.0
18.1

10-25
25-50
Over 50

15.7
18.6

11.4

NOTE , The measures of relative cost variation are determined
from the residuals (actual minus expected cost ratios) around the
regression equation that underlies the bottom panel of Chart l.
Residuals were averaged, without regard to sign, for each bank
size group and expressed as a percentage of the regression esti mate for the median bank size In each group .

to op rate with maximum effi cien ·y, ost ratios
at th small hanks would not only he hi ghcr,
on the average, hut they also would be expct ,c1 to display a signifi ·antly larger r ,Jative
variation from bank to h ank. Measures of
relative cost varia tion, shown in Tab]e 4, do
not reveal any systematic relationship between
variations in costs and bank size. Thus, the
comparatively greater efficiency of larger District members seems to stem, not from more
consistent efforts to hold down costs, but from
the opportunities presented by their larger
size to organize their activities in ways that
contribute to a larger output per emp]oyee.

9

Wheat UtilizationTrends and Prospects
HE USE OF WHEAT has tr nded downward
since 1945 while production has continu d
to increase. The cumulative effect of these
opposing trends has been a doubling of wheat
stocks. Year-end carryovers have increased
to th point that they exce ded production
for a ·h of the past 2 y ars and a suh tantial
in r as · is expect cl for this year. Most of
thcs ' slo ·ks ar' held by th
ov 'rnmc nt and
th
ost of maintaining th m has grown
rapidly.
Some stock of wheat is necessary as a reserve in case of crop failures or national
emergencies. Prior to the buildup of Government stocks, large quantities were held by
farmers, grain traders, millers, and speculators. Year-end carryovers from 1921 to 1937
averaged 186 million bushels, or about one
fourth the annual av rage production. A proportionate res rve in more recent y ars would
have amoun t cl to 250-300 million bushelsabou t half the average annual <lorn stic requirem nt. Private grain stor rs might maintain this amount if the Government ceased
storage operations, although its adequacy as
a reserve is debatable. The reserve probably
should be large enough to cover domestic
needs through 1 or 2 years of poor crops.
Also, it has been contended that reserves
should be sufficient to cover export commitments made under the International Wheat
Agreement and Public Law 480. Even under
th broader definition, the substantial carryovers of re ent years s em to be considerably
great r than are needed for reserves.
Th wheat surplus has continued to grow
despite various efforts to curtail it. This ar-

T

10

ticl explores the patterns and trends in wheat
utilization and seeks to project them through
the coming decade. Such information may be
helpful in evaluating the prospects for preventing further growth in the wheat surplus
through increasing utilization.
Domestic Utilization

Dorncsti · us' of wh 'al in ·r as d fairly
st 'adil y from 1910 to 1943, largely b caus of
incr ased use for feed, then declin d to 1957.
Sine 1957, wheat usage has risen slightly and
is currently about 13 per cent above the
1910 level. The primary use for wheat is for
human food. Seed and livestock feed have
each taken about one tenth of the total used
domestically in recent years and an insignificant amount is used for industrial purposes.
HUMAN FOOD. Wheat used for human food
has flu ctuated around 500 million bush ls
since 1909. The accompanying chart shows
that total use has been very stable, especially
since World War II. Consumption p r person decreased from 5.3 bushels in 1909 to
2.7 bushels in 1960, a decline of 49 per cent.
However, population shifts almost exactly
offset the effects of the dietary changes.
Wheat consumption appears to be related
to standards of living. In comparing countries
with different living standards, it has been
observed that ereal consumption tends to
incr ase with income in nations with low living standards. Also, there is a tend ncy to
substitute wheat for other c reals in th s
countries as income rises. At som point,
however, the use of wheat reaches a maxi-

Wheat Utilization - Trends and Prospects
Trends in Population and Wheat Processed
For Food
Continental United States
PER CENT OF 1910*

250

SEMl · LOG

r

SCALE

200

40
1910

'15

'20

' 25

'30

'35

'40

'45

'50

'55

'60

'65

'70

• In 1910, population was 92 .4 million, total wheat processe d for
food wa s 482 milli on bu shels, and per capita consumption was 5.2
bu shels .
NOTES: On a semi -logarithmic chart, equal slopes indicate equal
rates of change .
Projections of percentage measures of wheat processed and consumed from 1960 to 1970 are indicated at high, low, and median
levels.
SOURCE : U. S. Departments of Agriculture and Commerce.

mum and further increases in income are accompanied by shifts to noncereals, particularly livestock products. People demand a
more varied diet as their incomes increase.
The United States and other relatively affluent countries where wheat is the preferred
cereal have experienced declines in per capita
wheat onsumption as their p eople b ecame
relativ ly more wealthy.
Th low t point that wheat consumption
may reach is problematical. The Food Research Institut of Stanford University has estimated that wheat consumption might drop
as low as 2 bushels per person-not a great
deal below current U. S. consumption levels.
With present incomes, the Nation's consumers tend to buy the foods they prefer. To the
extent that wheat products are consumed
through preference, p er capita consumption
probably can b maintain d.
Efforts to increase domestic food consumption of wheat ar being made by various private and public agenci s. Research work is
being carried out in the development and
marketing of new products. For example, the
Monthly Review •

March 1961

U. S. Department of Agriculture is testing the
distribution of frozen bread. Efforts also are
being made through advertising to stimulate
the use of wheat as food. While it is possible
that these efforts may stem or even reverse
the declining trend in per capita consumption, there is no clear evidence yet that they
can do so.
In projecting per capita wheat consumption for the next decade, it seems difficult to
justify an estimate higher than the current
rate or lower than would be obtained by extending the trend line of the last half century. If thes assum d xtremes ar used, th
high 's timate for 1970 would be about 23/4.
bush ·ls and th low 'S timat about 2 ') ,. bushls. Alt 'rnaliv ly, assuming a slight forth r
de Jin and a I v ling off about 1965, an intermediate estimate of about 2½ bushels
would be reasonable. One of the recent population projections of the Bureau of the Census, assuming that the birth rate will remain
constant at the 1955-57 level throughout the
projection period, yields an estimate of 213.8
million people in 1970. Applying this population estimate to the per capita consumption
estimates indicates that wheat processed for
food in 1970 would amount to a high of 590
million, a low of 480 million, and a median
of 535 million bushels.
OTHER DOMESTIC USES. A large quantity of
wheat is used for feed, seed, industrial purposes, and shipment to U.S. territories, Alaska
and Hawaii. Feed use has fluctuated considerably during the past 30 years. It reached a
high of 511 million bushels in 1943, declined
to 39 million bushels in 1957, and has risen
slightly since that tim . It seems to have stabiliz d around 50 million bushels in the past
3 years. Fe d use eems unlikely to increase
much unless price supports are altered to
make wheat more competitive with other feed
gains . Alternatively, the use of wheat as feed
probably will not decrease much because

11

Wheat Utilization -

wheat is favored for poultry feed and there is
usually a substantial quantity which is unfit
for human consumption. Feed use in 1970
probably will be between 20 and 60 million
bushels, with 40 million bushels appearing
to be the most likely sum.
The amount of wheat used for seed has declined in recent years. Seeding rates have
dropped but, more important, acreage seeded
has declined. Yields per planted acre have
increased significantly and probably will continue to increase as cultural practices and
varieties are improved. Assuming a seeding
rat of about 1 bushel per acre, a yield of 25
bush els per planted acre, and production
equal to utilization , th e total utilizatio, proj ctions summariz ·d later in this artid • would
r quire a maximum of 60 million , a low of 40
million , and a median of 50 million bushels of
wheat for seed in 1970.
The remaining uses for wheat are minor.
Industrial use has not exceeded 1 million
bushels in any year since 1945 and there is
little reason to expect any increase. Shipments to Alaska, Hawaii, and the various territories- which have averaged about 4 million
bushels annually for the past 20 years-probably will change littl in the fores eabl
future.
Exports

Wheat is the world's leading food grain
and is produced and consumed in nearly
every country. Total production was 8.1
billion bushels in 1959, and it is estimated to
have been 8.4 billion bushels for the 1960
crop. The Soviet Union and the United States
are the largest producers-together accounting for more than a third of total world production. Furthermore, world wheat stocks are
comparatively large. The four principal exporting countries-the United States, Canada,
Argentina, and Australia - had 2.9 billion
bushels available for export and carryover on
October 1, 1960. This was about three times

12

as much as they exported in the 1959 marketing year. France, also, usually exports significant amounts and the Soviet Union has become a major exporter in recent years. International trade in wheat, exclusive of trade
within the Communist bloc, was 1.3 billion
bushels in the 1959-60 fiscal year.
The United States is the leading exporter
of wheat, accounting for about 38 per cent
of the world wheat trade during the past
decade. Many problems are involved in expanding wheat exports. The United States
has been the only major exporter with private
wh at trad e in recent y ars, although Arg ntin cl b ega n transf rring its wheat trade to
privat · hands in 1960. Most xporting ·ountri ·s and many importing ·ountrics hav state
trading monopolies. Price supports on wheat
are common throughout the world. Import
quotas and duties also are used frequently to
protect local products from international
competition. These circumstances have made
it difficult for U. S. exporters to compete in
world trade without subsidies.
COMMERCIAL EXPORTS. Shipments made
through regular trade channels which involve
foreign exchange credit are termed commercial exports. Substantially all comm rcial exports are made for dollars. Table 1 shows that
sales for dollars have been overshadowed by
noncommercial sales in recent years. Among_
the major difficulties involved in increasing
commercial exports are lack of foreign exchange in many food deficit countries, large
supplies of wheat in other exporting countries,
and trade restrictions in some importing countries. The United States alone can do little
to overcome most of these difficulties. Participation in th International Wh at Agreement represents a major effort to solve some
of the commercial export problems.
THE INTERNATIONAL WHEAT AGREEMENT.
The world wheat market has long been af-

Trends and Prospects
Table 1.

Commercial and Government

ea

orts - United States, 1948-59*

~hels in Thousands
Year beginning July
Item

1948

Total exports:

-

1949

1952

1~

126 5
of total 1 '

I

r

1953

1954

_

1955

1956

1957

1958

1959t

216,512 273,634 345,564 548,558 401,762 1 442,106 1 506,644

29,605
9

100,544 158,025 240,700 375,000 245,430 302,116 372,970
46
58
70
68
61
68
74

41,680 192,605 315,374 287,585
14
53
66
91

115,968 115,609 104,864 173,558 156,332 1139,990 133,674
54
42
30
32 I
39
32
26

256,790 172,968 159,341
86
47
34

I

~~~~~i~i

1951

1502,559 298,470 365,573 474,715 T 317,190

Government programs:
Quantity
376,011
Per cent of total
75
For dollars:

1950

_

Govern ment exports
by programs:
I
P. L. 480
Title I
Title II
I
Title Ill:
Barter
..... .
2,619 16,924
Donation s
... .
Marshall Plan
208, 503 137,945 138,856 l 37,163
Army Civilian Supply 1 167,508 118,845
31,493
5,254

I

Total

376,011

256,790 172,968 159.'.3!1

l

3,938

9,964

22,965 1 89,063
2,702
1,5171

29,605

23,802
15,991

94,300 200,500 178,035 1230,820 301 ,214
11,900 12,200 14,290 10,861 10,677

46,458

66 700

70,Bii

6n~~

963

I 20,154
~U~6 I gm ~~:6~~

86.900

9,501

Includ es flour as wheat equivalent.
t Prelim inary .
SOURCE: U. S. Department of Agriculture.

Monthly Review •

March 1961

iN:§

100,54~ 15~25 , 2~0,700 375,000 J 245,430 302,116 372,970

*

flicted by p eriodic gluts and deficits, accompanied by sharp price fluctuations. Efforts
to establish an International Wheat Agreement date back to 1932. However, the first
operational Agreement did not b ecome effectiv until 1949. The initial Agreement was
for a 4-year p eriod, and it has since b een
extended three times for 3-year periods. The
obje tives of the Agreement are to assure
markets for exporting countries and supplies
for importing countries at stable and mutually
agreeable prices. The Agreement stipulates
basic minimum and maximum prices in terms
of Canadian wheat in storage at a specified
point.
The initial Agreement provided fixed quotas
for the participating countries. However, the
importing countries were bound only at the
minimum price and the exporters only at the
maximum price. The most recent Agreement
specifies that each importing member country
is to purchase from the exporting members a
minimum percentage of its annual commer-

23,745

cial imports of wheat and flour. These percentages vary from 30 to 100 p er cent, the
weighted average b eing about 70 per cent.
Within the stipulated price range- the minimum up to the maximum - th e importing
members are committed to purchase their
specified percentages from th e exporting
m mbers. The exporting members are committed to sell as much as the importin g members want to buy when the world price is
within the Agreement range.
The 1949 Agreement was ratified by four
exporting countries and 40 importing countries. The 1959 Agreement was signed by
nine exporting countries and 30 importing
countries. The Agreement covered about 60
p er cent of the wheat and flour moving in
world trade from 1949 to 1952- ranging from
525 to 581 million bu shels. The quota dropped
sharply to 389 million bushels in 1953 when
the United Kingdom declined to join the
renewed pact. The 1956 renewal provided
for a total quota of only 295 million bushels.

13

Wheat Utilization -

The 1959 Agreement, which does not specify
a given quantity, would result in the trading
of about 470 million bushels on the basis of
1954 to 1957 trade figures.
ExPORT SUBSIDIES. Since the United States'
support prices are above world market prices,
American wheat can be moved into commercial export channels only by subsidy. The
U.S. obligation under the International Wheat
Agreement requires a substantial subsidy. In
recent years, all commercial exports have
been subsidized. Export subsidies_ averaged
62 cents a bush 1 durin g the period 1949 to
19.56. More r cntly, a paym nt-in-kind program has b ·en in stilut ,cJ.
nd ,,. thi s program, xport subs idi 'S ar, p aid in w h at
from ov rnm I t stocks rather than in cash.
This program has the dual objectiv of encouraging commercial exports and reducing
Government stocks.
PROSPECTS FOR COMMERCIAL EXPORTS. Commercial exports depend heavily upon Government policy. It seems appropriate to assume
continued participation in the International
Wh at Agreem nt and continued export subsidi s so long as dom estic prices are supported above world levels. How v r, it is also
appropriate to assume that the United States
will not risk serious international friction by
expanding commercial sales aggressively either
through higher subsidies or domestic support
policies which might result in "dumping"
operations. Presumably, the interests of other
wheat exporting countries will be respected
in future export policies.
Most commercial sales in recent years have
been quota sales under the International
Wh at Agreem nt. Table 1 shows that sales
for dollars have varied between 42 million
and 315 million bushels sine 1948. A USDA
projection in 1958 pr dieted commercial
wheat exports of 200-300 million bushels for
1960 and about the same for 1975. However,
14

in 1959, a new projection indicated exports
of 150 million bushels for 1970. Dollar sales
have averaged about 140 million bushels during the past 5 years. Under the assumptions
stated above, commercial sales in 1970 seem
likely to fall within a range of 100 million to
200 million bushels, with a median of 150
million bush ls.
No COMMERCIAL EXPORTS. Shipments under
Government programs that do not involve
foreign exchange credit or dollar claims are
termed noncommercial exports. These transa -tions includ sales for local curr nci s,
hart ,,., donations to for i 1 n gov rnmcnts, and
gifts to individuals throu gh priva t ' r Ii ,f organi za ti on . Lug ' quantiti s of wh , t w re
xported und r th U. S. Army's civilian supply program, which provided for wheat and
other foods to be given to people in occupied countries during and following World
War II. After the war, several relief and
rehabilitation programs were instituted, and
sizable quantities of U. S. wheat were distributed through UNRRA and various private
relief agencies. A billion bushels of wheat
have been exported under the massive Marshall Plan and related aid programs. An even
larger program has been developed und r the
Agricultural Trade Dev lopment and Assistance Act of 1954- Public Law 480.
PUBLIC LAw 480. This Act provides for a
multiple-purpose program involving the disposal of surplus agricultural commodities and
foreign aid for relief and economic · dev~lopment in underdeveloped countries. A recent
amendment to the Act provides for longrange contracts with recipient countries so
that they may integrate food shipments with
their economic development plans. Public
Law 480 has become a long-range program
and it might be continued for other reasons
even if it were no longer needed for surplus
disposal.

Trends and Prospects

Public Law 480 provides for several methods of distributing surplus commodities. The
largest volume moves under Title I which
authorizes sales for local currencies. Title II
provides for donations to foreign governments
to meet emergency relief needs. Title III
provides for the bartering of surplus commodities for strategic and other materials produced abroad and also for donations through
private charitable agencies such as CARE,
UNICEF, and church organizations. The
relative volumes of wheat moving under each
of these programs are shown in Table 1.
A portion of the proceeds from Titl I
sal es is ea rmarked for market development.
The [• orcign ;\ gri cu ltural Service, in ·oo pcration with various American farm organization s, has undertaken several projects to expand the demand for American agricultural
products abroad. These include trade fairs ,
demonstrations, technical aid to trade groups
such as millers and bakers, consumer education, and other similar efforts.
A substantial portion of the proceeds from
sales for local currencies is loaned to the
recipient countries for economic development
purposes. The process of capital formation in
underdeveloped economics tends to exert
strong pressures on food prices. Development projects increase local purchasing pow r
and may at th e same time reduce agricultural
production by shifting labor out of agriculture.
Food supplied under this program simultaneously reduces local inflationary pressures
and helps finance economic develop:r;nent projects. Thus, by helping underdeveloped
countries to build industries which will
strengthen their trade positions, the program
may help build permanent markets for American farm products.
PROSPECTS

FOR

NONCOMMERCIAL

EXPORTS.

Noncommercial exports depend upon Government policy even more than do commercial
exports. It seems likely that Public Law 480
Monthly Review •

March 1961

or some similar program will continue in the
foreseeable future . Even if the United States
manages to curtail its agricultural production,
the exigencies of the "cold war" and the needs
of the underd eveloped countries may prolong
the noncommercial exportation of food and
fiber indefinitely. The demand for expansion
of economic aid to low-income countries is
Jikely to remain strong. It may b e simpler to
finance such a program with surplus commodities than with international credit, particularly during periods in which balance of
payments problems prevail.
Noncommercial whea t exports declined from
.'376 million hush ' Is in J948 to 29 million
h11sliels in 1952, hut ret urn e d to th e 1948
leve l in 1956 and again in 1959. They h ave
averaged about 300 million bushels for the
las t 5 years and th ere are some indications of
further increase. A recent Title I agreement
with India calls for shipment of 587 million
bushels of wheat over the next 4 years. Additional long-term agreements with other countries are expected to follow. The Foreign
Agricultural Service recently estimated that
the underdeveloped countries will need to import 150 million bushels more wheat in 1970
than th e 300 million bushels th ey have been
importing in recent years. However, it seems
likely that some of this will b e supplied by
other countries, especially by the Soviet
Union . Also, the foreign exchange positions of
the underdeveloped countries may improve
to the point that these countries can increase
their commercial imports. If present Government policies continue, noncommercial exports seem likely to fall within a range of 300
to 450 million bushels annually for the foreseeable future, with a most likely estimate for
1970 of 375 million bushels.
Summary of Projections

The projections for wheat utilization in 1970
are summarized in Table 2. The median estimates indicate slight increases in domestic use

15

Wheat Utilization - Tre ds and Prospects
Table 2.

Projected Wheat Utilization for 1970
Million Bushels
High
Domestic Utilization:
Human food
Livestock feed
Seed
Industrial use
Shipments to territories
Total Domestic
Exports:
Commercial
Noncommercial
Total Exports
Total Utilization

Low

Median

590
60
60
1
4
715

480
20
40
1
4
545

535
40
50
1
4
630

200
450
650
1,365

100
300
400
945

150
375
525
1,1 55

and in exports as compared with recent years.
The ran ge in estima tes of total utilization from
945 million to 1,365 million bu h ls app ars
quit, la rg . How 'V r, it is easy to ·on · ·iv , of
c:ha11 g 'S in th ' int rnati nal situ ation or Gov' rt m •nl poli ·y w hi ·h woul l r suit in figur ' S
ith r abov " or b low thes limits. On the
other hand , a radical change in on factor
may be offset by an opposite change in anBANKING IN THE TENTH DISTRICT
Loans

Reserve
Coun try

City

Country

Member

Member

Member

Member

Banks

Banks

Banks

+5

+16

-1

Colorado

-3

- 1

+13

- 3

Kansas

- 3

+ 3

+25

+1

New Mexico *
Oklahoma *
Wyoming

- 5 + 10
- 2

+ 3

**

**

- 4

+8

**

**

* Tenth District portion only .
t Less than 0.5 per cent.

16

+2

+a

- 2

+1
+5
+5
+9

+ 3

+19

- 2

+ 3

- 3

+s

**

**

- 4 + 16

+3 +11

+1

**

+9

**

+1

+1

+1 +10

+2

- 1

Dec.
1960

Jan.
1960

Consumer Price Index

(1947-49 = 100)

127.4

127.5

125.4

Wholesale Price Index

(1947-49 = 100)

119.8

119.5

119.3

Prices Rec'd by Farmers

(1910- 14 = 100)

241

242

232 r

Prices Paid by Farmers

(1910- 14 = 100 )

301

298

299

TENTH DISTRICT BUSINESS INDICATORS
-

-

District
and Principal
Metropolitan
Areas

+6

+2

Jan.
1961

r Revised .

Jan . Dec. Jan . Dec. Jan.
1960 1960 1960 1960 1960

- 4

Nebraska

Banks

January 1961 Percentage Change From
Dec. Jan. Dec.
1960 1960 1960

Missouri *

Index

City

and
States

Tenth F. R. Dist.

PRICE INDEXES, UNITED STATES

Deposits

Reserve

District

other. Offsetting changes seem to be more
likely than additive changes in the same direction because the size of many of these
figures is determined largely by Government
policy. Thus, a substantial increase in domestic use or commercial exports might be accompanied by a reduction in noncommercial
exports.
Wh at carryov r stocks could continue to
increase if pr s nt conditions w re xtended.
1 he national wheat allotment cannot b e set
lower than 55 million acres under the pr sen t
law. This acreage could produce 1,375 million bushels if the yield averag d 25 bushels
per acr as previously ass umed. Taking the
proj t d wheat utilization 1 v ls in ·omparison , ·a rryov r stocks would rema in high at
h 'St and incrcas' rar idl y at worst. If th
rn 'dian utili zation 's timat a nd th 25-bushcl
yield o urred, only 46.2 million acr s of
wheat would b needed to balance production and utilization.

+5
+5
+4

Tenth F. R. Dist.

Value of
Check
Payments

Value of
Department
Store Sales

Percentage change-1961 from 1960

Jan .

Jan.

+12

+4p

Denver

+1 6

+ 7p

Wichita

+4

- 9p

Kansas City

+ 12

Omaha

+12

+a

Oklahoma City

+ 13

- 4p

+5

Tulsa

+10

-3

**No reserve cities in this state.

p Preliminary.

Op
+34