Full text of Monthly Income Payments in the United States, 1929-40

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

MONTHLY INCOME PAYMENTS
IN THE UNITED STATES
1929-40

U. S. DEPARTMENT OF COMMERCE
BUREAU OF FOREIGN AND DOMESTIC COMMERCE

U. S. DEPARTMENT OF COMMERCE

Harry L. Hopkins, Secretary
BUREAU OF FOREIGN AND DOMESTIC COMMERCE
James W. Young, Director

+

Economic Series No. 6

MONTHLY INCOME PAYMENTS
IN THE UNITED STATES
1929-40
BY

FREDERICK M. CONE
National Income Division

UNITED STATES
GOVERNMENT PRINTING OFFICE
WASHINGTON ; 1940

For sale by the Superintendent of Documents, Washington, D. C.

Price 10 cents

PUBLICATIONS OF THE NATIONAL INCOME DIVISION
National Income at Nearly 70 Billion Dollars in 1939. Reprint from Survey
of Current Business, June 1940. Presents summary data for the decade 1929
through 1939 on national income by industrial source and type of payment, and
compensation of employees by industrial groups. Includes revisions from
previous income publications. No charge.
Income in the United States, 1929-37. November 1938. Contains more de­
tailed information on national income for the years 1929 through 1937, including
estimates of total income and composition of income by industrial source, and
compensation and number of employees by industrial divisions. A chapter is
devoted to discussions of the income concepts and terminology and the sources
and methods of estimation. Price 10 cents.
National income in the United States, 1929 35. November 1936. This
bulletin includes a comprehensive discussion of the concepts and scope of the
official estimates of national income and describes in detail the methods of
estimation involved. Price 25 cents.
State Income Payments, 1929-38. Reprint from Survey of Current Business,
April 1940. Presents income payments by type of payment for each of the States
from 1929 through 1938, as well as per capita income payments by States. In­
cludes also a brief discussion of the trends indicated by the data. No charge.
Concentration and Composition of Individual Incomes, 1918-37, October 1940.
Prepared for the Temporary National Economic Committee. An analysis of the
changes in the shares of total individual income received by high-income recip­
ients and in the composition of these shares. Also contains data on the con­
centration of purchasing power and the effects of relief and veternas’ adjusted
service payments on income concentration. Price 20 cents.
Economic Conditions in the Osteopathic Profession. July 1939. Contains re­
sults of a survey of income in the osteopathic profession covering selected years
from 1929 through 1937. No charge.
Incomes of Chiropractors and Chiropodists. August 1939. Presents distri­
butions of income and average incomes for each of the professions for selected
years from 1929 through 1937. No charge.
Economic Conditions in the Dental Profession. September 1939. Contains re­
sults of a survey of income in the dental profession covering selected years from
1929 through 1937. Presents data relating income to type of practice, years of
practice, type of income received, and geographic location. Price 10 cents.
Indebtedness in the United States, 1929-39. Reprint from Survey of Current
Business, June 1940. Presents summary data for the decade 1929 through 1939
on Federal, State, and local public debt, and long-term and short-term corporate
and noncorporate private debt, compiled on a net-debt basis. This study differs
in concept and scope from earlier studies made by the Department of Commerce
which were confined to long-term debts. No charge.
N o t e . —These bulletins may be obtained on request from the Bureau of Foreign
and Domestic Commerce at Washington, or from any of its district offices, with
the exception of National Income in the United States, 1929-35, and the study
of income concentration. The latter reports may be obtained from the Superin­
tendent of Documents, Washington, D. C. (Postage stamps are not acceptable
in payment for any bulletins for which a charge is indicated.)

n

CONTENTS
Foreword______________________________________________________
Introduction-----------------------------------------------------------------Chapter I
Concepts and scope of the estimates__
..
__
National income and income payments.
.
Salaries and wages_____________ ___ ____ _ ____ ____________
Dividends and interest---------------- . _
__
Entrepreneurial income______
Net rents and r o y a l t i e s . ------Direct relief, gifts, and inheritances
. _______
Other labor income. _______ — ..... — ___ ____ ______________
Other forms of income_______________________________________
Chapter II
Interpretation and use of the estimates_____________________________
Income payments and consumer purchasing power ------------I ncome payments as an index of general economic activity________
Appendix.—Sources and methods_____________
TABLES
1. National income and income payments by years, 1929 38__
2. Income payments by months, 1929 40
CHARTS
1. Income payments and department-store sales_____________________
2. Income payments and commodity production. ... . . . . ----------------3. Indexes of salaries and wages by major industrial groups___________
4. Income payments and bank debits__
5. Methods of adjusting monthly data to annual averages------------ill

Page

iv
1
3
3
4
f>
5
6
(i
7
8

11
12
16
27
10
22
15
18
21
21
28

FOREWORD
This bulletin is another in a series of publications prepared by the
Department of Commerce on the subject of income in the United
States. The Department’s annual estimates of the national income,
which provide break-downs by industrial sources and by types of pay­
ment, have proved widely useful and have stimulated a rapidly grow­
ing interest in the measurement of our income. In addition, the
Department’s annual estimates of the flow of income to individuals
in each of the various States have been extensively used in the business
community as well as among research workers and Government
officials.
The annual figures, while valuable and essential for many purposes,
are not sufficiently current to provide a measure of short-term changes
in the economic activity of the Nation as evidenced by the flow of
income to individuals. To meet this need, the Department first
published estimates of monthly income payments early in 1938.
This series, extending back to January 1929, has been revised and
brought up to date. It is published each month in the Survey of
Current Business and provides a measure of month-to-month changes
in the aggregate income flow to individuals and a break-down by types
of income payments. This series is particularly useful as an indicator
of changes in general business activity and is helpful to business in
providing a measure approximating the purchasing power of
consumers.
This bulletin is published for the purpose of describing in detail the
sources and methods underlying the estimates and discussing the va­
rious uses and limitations which should guide those who use the statis­
tics. The estimates and the bulletin have been prepared in the
National Income Division of the Bureau of Foreign and Domestic
Commerce. This work is under the supervision of E. A. Tupper,
Chief Statistician of the Bureau of Foreign and Domestic Commerce,
and under the general direction of Robert R. Nathan, Chief of the
National Income Division. Frederick M. Cone has been primarily
responsible for the development of methods and techniques employed
in the preparation of the figures.
J ames W. Y oung , Director,
Bureau oj Foreign and Domestic Commerce.
A ugust 1940.
IV

MONTHLY INCOME PAYMENTS IN THE UNITED
STATES, 1929-40

INTRODUCTION
Early in 1938 the Department of Commerce published its first esti­
mates of monthly income in the United States. Since that time a
number of changes in the content of the series and in the methods of
estimation have been made. Further revisions of equal importance '
may be undertaken in the future as new source material becomes avail­
able and as the application of the series to the purposes it is intended
to serve, reveals its adequacies for those purposes. It is believed,
however, that the concept of income payments and tin1 sources and
methods involved in its measurement have become sufficiently crystal­
lized to render advisable a more comprehensive description of the
series at this time.
For those who have been using the monthly series since its original
publication 12 it may be, useful to indicate briefly the major revisions
which have subsequently been made, before proceeding to describe
the estimates as they now stand. The original series represented an
outgrowth of the Department’s annual estimates of the national
income and was prepared largely in response to a demand for a measure
of current changes in the, Nation’s income. Since it was not feasible
to construct a satisfactory index of current changes in the item of
business savings, the objective at that time was limited to a break­
down by months of what has been termed “ income paid out” in the
annual estimates. This is a measure of the total compensation disbursed
by business enterprises for services rendered by individuals. The sum
of income paid out and business savings comprises the national
income.3
It soon became apparent, however, that current changes in income
disbursed by business enterprises were likely to be misleading as an
indication of movements in the net value of goods and services pro­
duced, because of the marked quarterly changes in the residual item
of business savings. Also, it appeared that the series would probably
find its most important application as an indicator of changes in the
current purchasing power of consumers. Therefore, in correct ing the
monthly series to accord with the revised annual income estimates

1 Owing to the comprehensive nature of income aggregates, new information is constantly coming to hand
which makes advisable the revision of many individual components. Despite the frequent apj>earanec of
such additional source materials, revisions in the income aggregates have in the past, been of a moderate
character. Thus, since its original publication in 1934 the Department's estimate of the national income for
1929 has varied from 80.8 to 83.0 billion dollars—i. e., within a range of less than 3 percent. 'The estimated
trend in national income has been confined to much narrower limits. Thus, in all the various revisions
the year 1933 expressed as a percentage of 1929 has varied only from 51.5 to 52.1.
* “Monthly Income Payments in the United States,” February 1938 issue of the Survey of Current
Business.
3 For an extended discussion of the concepts and scope of the nat ional-income estimates, the reader is
referred to chapters 1 and 2 of National Income in the united States, 1929-35.

1

2

MONTHLY INCOME PAYMENTS, 1929-40

from 1929 to 1937, inclusive, it was considered advisable to depart
from the concept of income paid out and to construct a series that
would more closely represent the actual payment of income to individ­
uals. A brief description of the revised series was published in the
October 1938 issue of the Survey of Current Business.4 The esti­
mated volume of income payments was later raised somewhat for the
whole period as a result of revisions in the annual figures,5 hut the
concept of income payments as a measure of the flow of consumer
purchasing power is essentially that which was briefly described in the
October 1938 issue of the Survey of Current Business.
4 “Revised Estimates of Monthly Income Payments in the United States, 1929-38.”
* “National Income in 1938 at 64 Billion Dollars,” June 1939 issue of the Survey of Current Business.

C hapter I

CONCEPTS AND SCOPE OF THE ESTIMATES
In recent years national-income estimates have been applied to a
variety of uses, only two of which are noted here. The national
income proper continues to be generally accepted as the most compre­
hensive index of the product of the Nation’s economic activity and of
its economic welfare. More recently, however, the violent cyclical
movements of the past 10 years have concentrated attention upon the
purchasing-power aspect, of the national-income aggregate. The
national income is a measure of the net value of the Nation’s economic
output and at the same time a measure of the command over this
output. It is the demand or purchasing-power aspect of the national
income which has assumed increasing importance in recent years.
Its segments have considerable usefulness as measures of fluctuations
in consumer purchasing power, and as determinants of such important
variables as savings, investment, and consumption. In this con­
nection, two important changes in method arc demanded. In the
first place, it is immediately evident that the year, which is the unit
of time forced upon the national-income estimator by prevailing
accounting procedure, is too large a unit for the analysis of short-term
fluctuations or for the determination of current business and govern­
mental policy. Secondly, some of the value judgments that com­
plicate the work of the national-income estimator are not relevant to
the purposes noted above. In particular, the effect of a given amount
of personal income upon consumer purchasing as a whole must be
conceived of as being independent of whether or not this income
represents a reward for a service rendered.
In the analysis of cyclical fluctuations, interest ultimately attaches
to the flow of funds from industry and Government to the consumer
and his disposal of these funds, which in turn constitutes a demand for
consumable goods and services and for capital goods. For the most
part, however, consumers acquire a command over goods only in
return for definite services which they currently render in the pro­
duction of these or other goods. Therefore, the present monthly
series on income payments in the United States is, as already noted,
closely related to the Department’s estimates of the national income,
and for this reason and also because the latter figures are compara­
tively well known to the general public, it will be useful, in describing
the content of the series, to relate the treatment of its several items,
wherever possible, to the treatment of the same or comparable items
in the national income as estimated by the Department of Commerce.
NATIONAL INCOME AND INCOME PAYMENTS

In the annual income reports of the Bureau of Foreign and Domestic
Commerce, national income is defined as the net value of all com­
modities produced and all services rendered within the United States.
3

4

MONTHLY INCOME PAYMENTS, 1929-4Ô

The significance of the concept has been developed at some length in
the various publications6 of the Bureau on this subject, and it is not
necessary to enlarge upon the concept here, except as it incidentally
enters into the description of the monthly series.
In any given year the business enterprises producing the com­
modities and services that make up the national income may not
pay out to the individuals contributing- to their activities exactly
the value of the year’s net product. In some years business units
increase their net worth by retaining a share of their net product.
These undistributed earnings are termed positive business savings.
In other years the relative inelasticity of business costs with respect
to business income forces business units to sustain their payments
to individuals by drafts upon business capital and surplus. This
excess of income disbursed over the value of product is called negative
business savings.
The share of the national product distributed to individuals each
year includes the salaries, wages, dividends, interest, and rents and
royalties disbursed by business enterprises in return for the personal
services or capital which individuals have contributed to the process
of production, in addition to what individual entrepreneurs take
either in cash or in kind from their business for their own personal
expenditures. This aggregate has been defined in earlier studies as
“income paid out.” Until recently, the aggregate disbursements of
income to the agents of production (income paid out) might well have
been taken as representative of the flow of purchasing power to
individuals considered as ultimate consumers. The severe economic
distress of the past 10 years has, however, occasioned several important
changes in public fiscal policy, which have served to negate the identity
between the distributed portion of the national income and con­
sumer purchasing power and to make desirable a separate series
measuring the latter quantity. The character of these changes will
be evident in the discussion of the separate items included in income
payments.
SALARIES AND WAGES

Salaries and wages included in income payments are identical with
those included in the national-income estimates, save for one important
exception. Since 1936 national income includes, along with the com­
pensation received directly by the employee, also that portion of his
wage or salary which he contributes to the Social Security programs,
as well as the Social Security contribution of his employer to the
employee’s account. It is reasoned that such contributions represent
on the one hand a labor cost to the employer, and on the other a share
in the net product which accrues to the employee’s account though
not immediately available to him Such contributions are excluded
from income payments because they are definitely not at the em­
ployee’s immediate disposal. On the other hand, employee benefits
are included in income payments (cf. “Other labor income,” below).
Income in kind is inducted in salaries and wages in both series for
those industries in which this type of income is of importance. The
inclusion of this nonmonetary type of income in a series designed to
represent the flow of funds or purchasing power to consumers may

» Cf. especially: National Income, 1929-32; National Income in the United States, 1929-35; and Income in
the United States, 1929-37.

MONTHLY INCOME PAYMENTS, 1929-40

5

seem questionable at first sight, but payments in kind constitute an
effective command over consumer’s goods, though this command is
exercised by the employer rather than the employee. Payments in
kind constitute about 2 percent of total salaries and wages but are
particularly important in certain industries, notably agriculture,
domestic service, water transportation, and in the military and naval
services of the Federal Government.
DIVIDENDS AND INTEREST

In national-income estimates, institutional investors, such as
banks and insurance companies, are considered to be aggregates of
individuals, on the ground that dividends and interest received by
such institutions represent accruals to the credit of individual deposi­
tors or policyholders. Consistently with this procedure, the dis­
tributive shares in the national product do not include funds which
individuals actually receive, in the form of interest, from these insti­
tutional savings organizations. This treatment, not an entirely
satisfactory one, is necessitated by the consideration that it is at
present impossible to determine what proportion of withdrawals by
investors from such institutions represents the interest on invested
capital and what proportion represents the return of their principal.
This deficiency is particularly important in the case of income pay­
ments because the individual consumer cannot be considered to
exercise immediate command over the current investment income
accruing to his account. On the other hand, the individual can, and
frequently does, withdraw all or part of his capital, including the
accrued interest thereon, from the institution. Therefore, one pos­
sible solution for this problem might be the deduction from total
income payments as now measured of all interest income of such
institutions as well as the net of consumer deposits over withdrawals
from such institutions. Such a procedure would be tantamount to
isolating th(> demand for a certain type of goods, e. g., bank balances,
and calculating the residual purchasing power for all other goods anil
services. It seems likely that this is the direction that the statistical
analysis of cyclical fluctuations will ultimately take, but for the present
the inclusion of such capital transfers would involve a radical depar­
ture from the concept of income and would lead to an analysis of the
disposal as against the mere receipt of consumer income.
ENTREPRENEURIAL INCOME

Entrepreneurial income, as included in income payments, com­
prises the net income of farm operators and independent professional
men and the computed withdrawals of independent businessmen in
other lines of enterprise. Because the fields of agriculture and pro­
fessional service account for more than one-half of total entrepre­
neurial income, the aggregate is known by this more familiar term
rather than as entrepreneurial withdrawals. The withdrawal of an
entrepreneur is defined as that portion of his business receipts that
he withdraws from his enterprise for personal or nonbusiness uses.
It is obvious that this concept of an entrepreneurial withdrawal is
the proper one for inclusion in a measure of consumer purchasing
power, but there is great difficulty in determining the actual volume
of such withdrawals, particularly in agriculture and professional
230746°—40-----2

6

MONTHLY INCOME PAYMENTS, 1929-40

service, in which fields no distinction is here drawn between an entre­
preneur’s net income and his withdrawal for nonbusiness uses. In the
field of professional service, capital requirements are in general so
limited as to render unimportant the distinction between the entre­
preneur’s personal savings and his business savings. In agriculture,
it has been found impossible to make a satisfactory estimate of
entrepreneurial withdrawals. Furthermore, the limited information
now at hand indicates that farm purchasing varies substantially with
the net income of farm operators. Therefore, pending further studies
on this subject, agricultural net income has been included in income
payments.
NET RENTS AND ROYALTIES

Net rents and royalties are included in the national-income aggre­
gate as an independent type of income, but for lack of space are shown
in combination with entrepreneurial income in the monthly series.
They may be defined as gross rents and royalties less expenses, depre­
ciation, and depletion. Interest payments in this connection are
classified as expenses inasmuch as they appear elsewhere as a type of
income. In fact, net rents as thus computed represent to a consider­
able degree interest return on the equity of the owner or operator, and
on that account might well have been combined with dividends and
interest into an aggregate of property income. However, net rents
include an element of return for the efforts of the owner and operator,
and are, therefore, classified with entrepreneurial income, which is like­
wise a combination of interest income and labor income.
DIRECT RELIEF, GIFTS, AND INHERITANCES

Direct relief, gifts, and inheritances are all excluded from national
income estimates on the ground that, although they constitute in
many cases a substantial component of personal income, such receipts
are more in the nature of income transfers than rewards for services
rendered in the productive process, and therefore represent no addi­
tion to the income of society. Likewise, gifts and inheritances are
excluded from income payments because such additions to the income
of the recipient are considered to be simultaneously counterbalanced
by an equal deduction from the income (or wealth) of the giver or
from the wealth of the deceased. It is obvious, however, that such
transfers of income or wealth have a very marked influence upon the
distribution of individual income and consequently upon the disposi­
tion of available purchasing power as between different types of
consumer’s and capital goods, if not upon the volume of such pur­
chasing power.
Direct relief, as disbursed by governmental agencies, is not, how­
ever, excluded from income payments, since such relief is not imme­
diately offset by a like reduction in the income (or claim over income)
of the nonrelief portion of the population if the Government finances
relief by borrowing. In such a case the purchase of Government
obligations must be considered as entailing the same sort of decision
as any other act of individual investment, and, therefore, the inaugura­
tion of a direct-relief program does actually increase the volume of
funds at, the disposal of ultimate consumers. Of course, it is not in
general possible to say which functions of government are operated on
the basis of current tax receipts and which are operated on borrowed

MONTHLY INCOME PAYMENTS, 192 9-4 0

7

funds, but for the period during which direct relief has been of impor­
tance, there have been only isolated attempts to earmark tax receipts
for relief purposes, whereas the flotation of bonds for such purposes
has been a common occurrence throughout the period. The Federal
Government in particular has been financing a considerable portion
of its expenditures on borrowed funds during these years. For
these reasons all governmental expenditures on relief7 are included in
the monthly series. It should be noted, however, that the whole
problem of the effect of taxes upon the disposition of consumer income
is a complicated one, and certain aspects of this question will be noted
in another part of this bulletin.
The treatment outlined above refers to transfers of income among
individuals or from governments to individuals. In the case of con­
tributions to charity by business units, a somewhat different treatment
is required. Under prevailing legislation, business units are permitted
to enter such items as deductible costs, and since national-income
estimates lean heavily on corporation income-tax data, the reporting
of such contributions as a cost and their exclusion from the estimates of
national income in effect understates the computed national income
total by that amount. At the present time there is not sufficient infor­
mation available to permit a correction for this bias. Consequently,
such contributions are included in neither the national income nor the
monthly series on income payments. Such additions to personal
income by business units, while substantial in absolute amount,
represent but a fraction of the volume of direct relief or income trans­
fers between individuals.
OTHER LABOR INCOME

Other labor income comprises industrial pensions, industrial com­
pensation for injuries, and veterans’ benefits of various types, in
addition to social-insurance benefits under the terms of the Social
Security Act, the treatment of which has already been indicated in the
earlier discussion of salaries and wages. Income of this type differs
from direct relief, which is awarded on the basis of needs, in that it
represents a return for services rendered at some time or another; it
differs from salaries and wages in that the reward is separated in time
from the service which it represents and is, therefore, only indirectly
related to the current level of economic activity. This type of income
in reality comprises two distinct types, each of which requires separate
discussion.
1. Industrial pensions and compensation for injuries are included in
national income as well as in the monthly series. The ideal treatment
for this type of income would be that which is applied to the fiscal
operations of the Social Security Act, namely, the inclusion in the
national income of the employee’s full wage and the employer’s con­
tribution to such funds for the period during which the employer or
employee makes his contribution to the appropriate fund, and the
inclusion of employee benefits in income payments for the period during
which the employee receives them. Actually there is not sufficient
information to permit this procedure; in many cases the employer is
the only contributor, the plan is on a pay-as-you-go basis, and benefits

7 As in computing the national income, work-relief wages are included with other salaries and wages on
the ground that the value of works projects is commensurate with the cost of the projects, as with other
functions of government.

8

MONTHLY INCOME PAYMENTS, 19 29-40

are reported as a current expense. For these reasons the actual bene­
fits are included in both national income and income payments for the
period during which they are received.
It will be noted that no provision has been made in the construction
of the monthly series for individual annuity and accident insurance
policies, although the monetary consequences of such policies are very
much like those of the industry- or Government-sponsored plans that
are taken into account, namely, the increase in the demand for invest­
ment opportunities relative to consumption that occurs while the
insurance funds are being built up and the shift in demand while they
are being depleted. Private insurance and annuity policies should,
however, be considered as among the many goods that compete for the
consumer’s dollar, whereas he has no choice as to the disposal of a
portion of his income when the insurance plan is a matter of legislation
or of company policy.
2. Federal pensions and compensation to war veterans have, over
the period its a whole, averaged more than $400,000,000 per year and,
prior to the inauguration of social-insurance benefits in recent years,
constituted the largest single component of “Other labor income.”
In addition to its regular pensions to veterans, the Federal Govern­
ment since 1931 has disbursed nearly $3,000,000,000 to the veterans
of the World War in consideration of their services during that period.
Both types of veterans’ benefits are excluded from the Bureau’s
national-income estimates on the ground that the services for which
such payments are designed to compensate were rendered before the
period covered by the Bureau’s figures, but both are included in the
monthly series as constituting an important addition to consumer
purchasing power, especially in the years 1931 and 1936. Prior to
June 1936 payments to World War veterans took the form of cash
loans on the security of their adjusted service certificates; in that
month the Federal Government, in discharge of its obligations on
account of adjusted service legislation, remitted to the veterans more
than $2,000,000,000 worth of adjusted service bonds and some cash.
The larger portion of these bonds were liquidated during the year
1936, but the Treasury has been redeeming them in smaller amounts
during the past 3 years; the present rate is about $2,000,000 a month.
More properly, then, the payments to veterans represent in the first
instance loans on existing assets and in the second the gradual liquida­
tion of these assets rather than the actual payment of income as here
defined. Nevertheless, because these disbursements of the Federal
Government were in the nature of original receipts to the veterans and
because, owing to the large volume of the disbursements and their
wide distribution among 3,000,000 veterans, they represented an
important stimulus to consumption on two distinct occasions, they
have been incorporated into the monthly series. The treatment
illustrates the difficulty of distinguishing at times between capital
movements and income flow in the analysis of consumer purchasing.
OTHER FORMS OF INCOME

There are various other minor types of personal income which are
not included either in the Bureau’s national-income estimates or in
the monthly series of income payments.
Realized gains or losses from the sale of assets are not included in
national-income estimates because they are strictly price phenomena

MONTHLY INCOME PAYMENTS, 1929-40

9

and in no sense measure net value of product in a given period.
This type of income is likewise excluded from income payments but
on somewhat different grounds. It is assumed that the transfer of a
capital asset makes no contribution to mass purchasing power, since
the increase in funds available to the seller is simultaneously offset
by the immobilization of an equal volume of the purchaser’s current
income.8 Capital gains do, however, exert important influences upon
the distribution of current purchasing power and consequently
upon its disposal among different goods and commodities. In
particular, it is suspected that income arising from a realized gain on
capital transactions is subjected to quite different treatment in its
disposition than is income of a more regular nature. But this con­
sideration is only one of many psychological influences that indirectly
affect the direction of business activity and for this reason is not
involved in the present discussion.
Earnings from odd jobs or from illegal activities are excluded from
both series, the former because of insufficient data, the latter because
such activities are considered as being of no economic value within
the legal framework of the producing economy. The exclusion of
these two items from the income-payments series cannot be considered
as seriously impairing its usefulness. If such exclusion results in an
understatement of the dollar volume of consumer demand, this
understatement is offset by the exclusion of the same items from the
supply of goods and services (national income). In fact, for purposes
of comparison it is desirable that a consistent procedure be observed
for such items.
Imputed income from the ownership of durable consumers’ goods is
excluded from the income-payment series as representing no addition
to consumer purchasing power. This type of income is, however,
included in many estimates of the national income by nongovernmental
agencies, and it is likely that future studies by this Bureau may adopt
this procedure. In that event there will be a further important
difference between national income and income payments.
The accompanying table indicates the magnitude of many of the
items of which the treatment has been discussed above. Several
minor items have been grouped together in order to shorten the
table. These minor components include the effects of the Federal
civil-service retirement operations, State allowances to war veterans,
and some smaller items subjected to different treatment in constructing
the two aggregates.

8 This offset is not strictly true in a period of generally rising prices when capital appreciation is supported
by an increase in the volume of credit. For such a situation the volume of new credit is a more adequate
measure of the increase in consumer buying than the aggregate of capital gains as such.

Table 1.—National Income and Income Payments by Years, 1929-38
Item

Contributions to retirement systems for governmental
Additions:

1930

1931

82. 885
1,33»
81,54«

68.901
-5,033
73,934

54.310
-7. 490
61,800

131
60
421
168

140
94
445
191

149
158
525
907
211

160
326
548
147
227

82,064

74,524

63, 452

49, 319

1932

1933

1934

1935

1936

1937

50,304
-2. 407
52,711
3

55.833
-1.390
57, 223
7

64,792
-791
65, 583
299

167
580
431
54
246

180
828
353
26
271

192
1.099
385
19
282

205
672
399
1,427
268
1

46,830

54,006

58,809

67. 846

70,925 63. 459
-803 -2.284
71,728 65. 743
950
1, 119
329
295
122
108
215
223
837
1,008
409
398
128
57
269
268
1
396
96
35
3
10
71, 783 66,242

40,074 42,440
-8. 157 -3,246
48. 231 45,686

1938

1939
69.308
-294
1,196
319
112
234
1,067
422
34
279
429
109
15
70.096

1 National-income figures differ slightly from those published In the June 1940 Survey of Current Business. The national income totals for the years 1933 to 1939 have been sig­
nificantly revised owing to important revisions in estimated work-relief wages which were made subsequent to the publication of the national-income estimates for 1939. Several
additional changes of a minor character have also been effected for the years 1938 and 1939.
3 Exclusive of business savings in agriculture.

MONTHLY INCOME PAYMENTS, 1929-40

Other deductions:

1929

O1

C hapter

II

INTERPRETATION AND USE OF THE ESTIMATES
In the preceding chapter the content of the monthly series has been
described at some length. The mere enumeration of the various items
included in the series will have served to acquaint the reader with its
general nature and to indicate its possible uses. However, it seems
advisable to describe more fully the income-payments concept, even
at the risk of some repetition, and to define at some length the limita­
tions attaching to its use.
Although intended primarily as a measure of consumer purchasing
power, the monthly series is readily applicable to a number of other
uses. Among these, particular attention will be devoted to the use of
the series as an index of general business activity. For each of these
uses it may ultimately be desirable that a distinct measure of the
monthly fluctuations in income be computed and published. For the
present, however, it seems likely that the multiplication of income
series would merely result in greater confusion, and it seems
preferable to publish a single series, while noting briefly such ad­
justments as are desirable in each instance. Moreover, the break­
down of the present series is presented with these various appli­
cations in mind. Thus, the presentation of income payments by
type of payment is intended to be indicative in a very rough way
of the distribution of current income by size of income, which, as
well as the volume of income, is a very important factor in deter­
mining the direction of consumer demand. It is well known that
direct relief and unemployment benefits are subject to different
disposal than dividends and interest, which are for the most part
currently received by institutional investors or persons enjoying
a high level of income. However, this type of analysis is at present
limited by the wide variation in the average income from the im­
portant item of salaries and wages, even within each industry group.
On the other hand, the break-down of labor income by industrial
groups is designed to serve as an imperfect measure of the con­
tribution of the various industrial segments of the economy to cur­
rent economic activity. Similarly, the index of nonagricultural in­
come payments is presented primarily to provide a measure of general
economic activity that shall exclude the current contribution of
agriculture, and secondarily as a measure of the monthly purchasing
power of the nonfarm segments of the population.
The remainder of this chapter will be devoted to a discussion of the
two important applications of the series noted above, with special
attention to its use as an index of changes in consumer purchasing
power. As in the previous chapter, it will be convenient to start with
the better-known concept of national income.
In the estimation and analysis of the national income, it has been
found useful to distinguish between the business enterprises that con11

12

MONTHLY INCOME PAYMENTS, 1929-40

duct the Nation’s economic activity and the individuals whose services
make that activity possible. The concept of a business enterprise in
this connection includes not only business units in the narrow sense
but also such enterprises as governmental bodies, farms, private
hospitals, trade associations, and independent professional practices.
The year-to-year output of the Nation’s business enterprises is defined
as the national income. In return for their assistance in the productive
process, business enterprises disburse to individuals their share in the
monetary value of the year’s output. Individuals must, however, be
considered not only as the ultimate agents of production but also as
the ultimate end of production and their distributive shares in the
national product as the payments by means of which business enter­
prises enable individuals to consume the output of the economy.
Hence arises the notion of consumer purchasing power and its close
relationship to the national income.
As indicated previously, the two aggregates are not necessarily
equal and, in fact, are not likely ever to be equal over a limited period
of time. The reasons for this divergence are twofold. First, as noted
in chapter I, business units do not, over the short period, pay out to
individuals the exact value of the period’s output but retain the
ownership of a part of this output in the form of undistributed earnings
or under adverse conditions pay out more than the period’s net output,
notably by failing to meet depreciation and obsolescence charges.
Secondly, business enterprises not only make payments to individuals
who have contributed to the output of the period but also, in in­
creasing measure, to persons who have made such contributions in
the past or who have made no contributions at all. This lack of
identity between the value of current output and the aggregate flow
of income to consumers has been the subject of increasing attention
in business-cycle analysis. In particular, the question is asked:
Under what conditions is the current flow of income from business or
governmental units into the hands of individuals such as to enable
consumers to take of!' the market the final products of government
and industry without giving rise to those economic disturbances that
are commonly associated with the business cycle?9 Associated with
this question is a related problem of equal importance: how does the
“consumer disposal” of these funds affect the stability of economic
activity? Without entering into any of the theoretical aspects of
these related problems, the monthly series on income payments is
offered as a measure of the current flow of funds from business enter­
prises to individuals and constitutes at least a first step essential to an
analysis of the disposal of consumer income.
INCOME PAYMENTS ANI) CONSUMER PURCHASING POWER

In chapter I it was stated that the income-payments series was
intended primarily as a measure of changes in the current purchasing
power of consumers, but there was no attempt to define the term
“purchasing power.” In recent years the term has received consider­
able attention in economic literature, but so far no consistent definition
seems to have been evolved. On the whole, it does not seem profitable

9 It is not intended to imply that Consumer purchasing is the sole or even the most imjxirtant element in
the generation of cyclical changes in business activity. A similarly imj>ortant role must be assigned to
business purchasing, osi>ecially of raw materials and durable capital, but this type of demand is necessarily
excluded from this discussion.

MONTHLY INCOME PAYMENTS, 19 29-40

13

to speak of a community’s power to purchase in general, but rather
of its ability to purchase certain types of goods. Purchasing power
in the broad sense of the term is a function of wealth as well as of
income. Goods can be exchanged against goods, and the only
possible limit to the volume of such exchanges in a given period of
time is the volume of goods in existence and the velocity with which
such exchanges can be accomplished. Such a concept of purchasing
power seems to be as little susceptible to practical use as it is possible
of precise measurement.
It seems more profitable to restrict the term “consumer purchasing
power” to signify the ability of consumers to purchase the new goods
and services currently produced by business enterprises. This defini­
tion is more consistent with what is conceived to be the primary func­
tion of an economic society, namely, the production of new goods and
services, the process of exchanging assets assuming importance largely
to the degree to which it facilitates this production. Consumer pur­
chasing power in this sense is currently created only by the flow of
funds from business enterprises to individuals. That is to say, it
is impossible for individuals as such to increase their buying power in
the aggregate. It is possible for a given individual to increase his
own consumption of the current industrial output by the sale of what­
ever assets he may possess. But purchases and sales of existing assets
among individuals must cancel in the aggregate and, therefore, can
neither augment nor decrease the volume of consumer buying power.
Similarly the extension of new consumer credits results in no addition
to purchasing power insofar as the funds that sustain such credits
arise from the current savings of individuals.10
Even if it be conceded that the purchasing power of consumers,
as defined above, is restricted to the flow of funds from business
enterprises, it is none the less not valid to identify this aggregate
with “income payments,” which has been defined us the flow of
income from business enterprises. There are circumstances under
which the purchases and sales of assets may be said to exert an
influence on consumer buying power, namely, in those instances where
such transactions take place, not among individuals, but between indi­
viduals and business entities. Corporations can, for example by
the purchase of securities in the open market, increase the ability of
individuals to purchase the current industrial output to the extent
that these securities are purchased from individuals, and some author­
ities have stressed such operations as an important factor in cyclical
fluctuations. Similarly, business enterprises may absorb a certain
amount of consumer buying power bv the open-market sale of
securities. Therefore, the usefulness of the income-payment series,
as an index of consumer purchasing power, is limited by our present
inability to measure transfers of assets between individuals and busi­
ness units. In practice, the importance of this limitation is minimized
by the consideration that at the present time the series is likely to be
most useful in analyzing the demand for consumer’s goods in the narrow
sense and that the major portion of this demand arises from consumers
who are but little affected by considerations of t he type discussed above.

10 On the other hand, credits of this nature do havean important influence upon the directionof current
demand. Thecreation of installment credit, e. when financed by current savings, results in an increased
demand for consumer’s goods accompanied by a relative decline in the demand for capital goods. The
reverse of this shift in demand accompanies the liquidation of consumer credits.
230746° 40------------- 3

14

MONTHLY INCOME PAYMENTS, 1929-40

Moreover, it is in the power of business units in general and of
banks in particular to furnish a genuine stimulus to consumer buying
by the extension of new credits. This type of consumer borrowing
may take two forms: A retailer, for example, may elect to finance his
customer by leaving some of his profits in the form of accounts
receivable,11 or he may choose to borrow from a bank. In either
case, the business community must be considered to have supplied
consumers with buying power additional to that which represents a
return for services rendered. While changes in the volume of con­
sumer credit are normally small in relation to the average volume of
income payments, consumer credit is subject to wide fluctuations
and is, therefore, generally recognized as a factor of strategic import­
ance in the analysis of cyclical fluctuations. It is, however, subject
to independent measurement and will receive no further consideration
here.
Finally, some consideration should be given to the ability of the
monetary authority to make new money available to the public.
It may be profitable at this point to distinguish between consumer
purchasing power as a vague potential and consumer purchasing as
a realization of that potential. Income statistics as now presented
refer exclusively to the past; therefore, there seems to be little profit
in speculating as to how much new money the monetary authority
might have made available in the period under review and it seems
better to restrict our analysis to the actual volume of new money
created.1112
Similarly, when it is a question of estimating future purchasing
power, the ability of the monetary authority to create new money or of
the banks to extend new credits is of little consequence. Over the
period during which income payments can be forecast with any degree
of certainty, the volume of new money or of new credits that are likely
to be created can also be estimated with a fair degree of accuracy.
In fact, unforeseen changes in either of these items are likely to be
associated with simultaneous fluctuations in income. At any rate,
both of these forces are subject to independent measurement.
With the three qualifications noted above, the income-payments
series presented in this bulletin is believed to represent with a fair
degree of accuracy 13 the volume of funds currently made available
to individuals by business enterprises and, therefore, the current level
of consumer purchasing power as that concept has been defined above.
It is evident at once that it is impossible to work out a very close
correlation between the demand for investment goods and the flow

11 A rigorous analysis would classify this type of lending as merely involving a shift in demand (cf. lootnote
1»), since undistributed profits generally take the form of a demand for investment goods. However, since
business savings are excluded from this analysis, which is by daf'mition limited to consumer purchasing
power and excludes any considerat ion of demand on the part of business units, these credits should for con­
sistency be considered as an actual increase in the consumer demand for current output.
i* I n this respect the use of the term “power” is perhaps unfortunate: its use is justified, however, in con­
junction with the term “consumer,’since consumer buying power is limited by the actual volume of funds
made available by business enterprises. The volume of new money or new credits that might have been
but were not made available to consumers may be considered to be analogous to the idle labor and capital
which have been particularly characteristic of recent years. Such idle resources might have been brought
together to generate additional purchasing power, but, while their measurement is important as indicating
the expansive possibilities of the economic system, they may be disregarded in any analysis of its perform­
ance.
ls Inadequacies in the basic material have already been noted in Chapter I. Most important among
these is the want of statistical material that forces us to consider institutional investors, not as business
enterprises, but as aggregates of individuals. Minor déficiences arise, for example, from the exclusion of
the funds sj»ent by public institutions on such items as food and clothing, or from the omission of certain
business allowances which are charged to business expenses other than salaries and wages.

MONTHLY INCOME PAYMENTS, 1929-40

15

of current income. Even assuming that it were possible at present
to establish some relationship between the incomes and savings of
individuals, individual savings represent only a part of the total
demand for capital goods. Thus, in 1929, individual savings were
supplemented by an estimated $2,000,000,000 of business savings,
whereas in the years 1931-33 individual savings were offset by negative
business savings of approximately $20,000,000,000.
For this reason, the correlation of effective demand with income
flow must at present be confined to consumers’ goods as distinguished
from producers’ or capital goods. The sale of consumers’ goods at
retail affords the most fruitful subject for this type of analysis, partly
because such sales represent the one major type of consumer expendi­
ture for which current statistics are available, partly because retail
sales account for the major portion of consumer expenditures. Figure
1 illustrates the degree of correlation between the seasonally adjusted

Figure 1.—Income payments and department-store sales.

income index and the Federal Reserve Board’s adjusted index of
department-store sales, which may be taken as indicative of expendi­
tures of urban consumers on general merchandise other than foods
and those of an essentially nondurable nature. On the whole, the
relationship is striking for the period 1929-35, particularly when it is
considered that department-store sales represent but a minor portion
of all retail sales. However, during 1936 and the first half of 1937,
department stores failed to share fully in the sharp recovery in income
that characterized that period. Several factors may have resulted
in the divergence. Indications are that a large portion of the veterans’
bonus of 1936 may have gone into a more durable type of goods than
those purchasable in department stores, and it is unlikely that any
large portion of the extra dividends arising from the operation of the
undistributed-profits tax would flow immediately into departmentstore purchases. The sharply improved level of farm income in 1936
may have been another factor, since farm purchasing appears to only
a minor extent in department stores. Subsequent to the 1937 reces­

16

MONTHLY INCOME PAYMENTS, 1929-40

sion, income and department-store sales have followed generally
parallel trends but on different levels as compared with the 1929
averages. The lower level of department-store sales as compared
with income since 1937 may be explained in large part by the marked
growth in the sales of mail-order houses in recent years.
For purposes of analyzing consumer demand for the current output
of goods and services, it may sometimes seem desirable to deduct from
current consumer income those outlays over which the consumer may
be considered to exercise no choice, or at any rate only a limited
choice. Of primary importance among such outlays are, of course,
the taxes paid by individuals to Federal, State, and local governments,
since in the short run the individual consumer exercises no choice at
all over the disposal of tins portion of his income. On the other hand,
in preparing national-income estimates, it is useful to consider indi­
vidual taxes as a payment for the services customarily performed by
government, the desirability of these services being determined at
some period in the past by the democratic process. Whether or not
it is necessary or desirable to consider taxes as a species of demand
in national-income studies, in the case of income payments it is
obviously desirable to determine that portion of consumer income
which actually was or is being devoted to taxes in order to approxi­
mate more closely the residual income available for the purchase of
commodities currently coming into the market. Thus, in analyzing
the probable effect of an expected increase in industrial production
upon the consumer demand for the new goods arising from the pro­
ductive process, it would be desirable to be able to determine what
proportion of the increased consumer income could be expected to be
absorbed by increased taxation.
It is evident, however, that for such an analysis taxes differ only in
degree from various other types of fixed outlays or outlays which as­
sume a certain degree of priority in the individual consumer’s budgets.
Insurance premiums, for example, are not a fixed outlay for a given
individual, yet in the aggregate they fluctuate very moderately as
compared with other types of purchases. Other items that may for
some purposes be considered as fixed outlays are interest on individual
mortgages of a nonbusiness character and, over a shorter period, in­
stallments on consumer-goods purchased “on time.” 14
INCOME PAYMENTS AS AN INDEX OF GENERAL ECONOMIC ACTIVITY

It has been emphasized above that the income-payments series is
not intended to be indicative of changes in the output of the economic
community, but, inasmuch as by far the larger part of all individual
income received represents a payment for some service rendered in
the production of an economic good, the monthly series reflects in

'« It will be noted that in one instance the current series does make allowance for tax deductions from cur­
rent income i. e., for contributions to the Social Security programs. These contributions may be con­
sidered as mendy insurance premiums which, because the schedule of payments is prescribed by law as in
the case of any tax, can bo readily calculated. In future releases it may be desirable to adjust the income
series for tax outlays if studies of the tax payments of individuals make such an adjustment feasible. The
consideration of other types of fixed outlays appear to be somewhat further in the future, since primary data
are lacking at pnsent.

MONTHLY INCOME PAYMENTS, 1929-40

17

considerable degree 16 the current output of goods and services, and
represents the most comprehensive measure of general economic
activity now available. Tins is not to say that the income index is
thereby the most significant index of economic change. Perhaps
it is more useful to look upon this index as a background against
which to view the changes in the more sensitive indicators of business
activity. This relationship may be illustrated by a comparison of
income payments with some of the more restricted and therefore more
sensitive business indicators such as the index of industrial production.
During the past 3 years of pronounced business fluctuations, the
industrial production index of the Federal Reserve Board varied
much more than did the index of income payments. The production
index declined from 117 in August of 1937 to 76 in May 1938, a drop
of more than one-third as compared with a decline of less than oneeighth in income payments. Similarly, the production index rose
from 92 in May 1939 to 127 in December, a gain of about 40 percent,
whereas the income index increased less than 10 percent during the
same period. Of primary importance in the relative stability of the
income series is the inclusion in it of all the service and distributive
industries as well as the more volatile commodity-producing industries.
On the other hand, price changes enter into the income index and
tend to work for greater instability relative to physical output. The
net result of these influences is to make variations in the income
index vitally dependent on fluctuations in industrial output, but in
no very simple manner. Thus, during the 24 months following the
peak of 1929, a drop of one-third in the production index was associ­
ated with a decline of one-fourth in income payments. It is, there­
fore, indicated that whereas the drop in industrial production during
1937 and 1938 was as great as that during the period 1929-31, the
former was relatively restricted in its effects upon the economy as a
whole. Similarly, although the 40 percent rise in production during
the latter part of 1939 was associated with a rise of only 10 percent
in the income index, the higher levels of production, if sustained for
any length of time, might be expected to give rise to a greatly
increased volume of income payments.
The accompanying chart permits a more detailed analysis of the
relationship between income and industrial activity. For purposes
of comparison, it was desirable that the production series be adjusted
to represent the value rather than the volume of current output.
Therefore, the index of manufacturing production was multiplied by
the Bureau of Labor Statistics’ index of wholesale prices of finished
products, and the output of petroleum, anthracite, and bituminous

i* As noted in ch. I, the adjusted income index as now published contains several items, notably direct relief
and the soldier’s bonus, whose wide fluctuations in recent years are quite unrelated to business activity,
and which might easily be removed from the series. In fact, it. would be a simple matter to correct the
series for all item«* not related to business activity and to publish two adjusted series, one representing the
flow of income to consumers, the other the share of the national income currently distributed to the produc­
tive agents. It is felt, however, that the resulting gain inaccuracy would be more than offset by the possible
confusion caused by the publication of two series whose, month-to-month fluctuations would in most
periods be very similar. Therefore, the income-payments series alone is introduced in the following dis­
cussion. For measuring changes in economic activity some may prefer the adjusted index of salaries and
wages which does not reflect the secular increase in direct relief or the 1931 and 1936 payments to veterans.
This index, however, is also defective in several respects; e. g., agriculture is inadequately represented, since
agricultural wages represent a very small proportion of total income in agriculture and secondly, over the
period 1929-40 labor income has experienced a slight growth relative to total national income. On a monthto-month basis, however, the changes in the adjusted wage index are probably more representative than the
total income index of current changes in nonagricultural activity. This comparison i particularly true dur­
ing the period in which unemployment compensation benefits have begun to offset pay-roll changes in some
degree.

18

MONTHLY INCOME PAYMENTS, 1929-40

coal by appropriate price indexes. To afford further comparability
with the series on labor income in the commodity-producing indus­
tries, the value of construction contracts awarded (seasonally
corrected) was included in the composite production series and each
of the individual series weighted by the income produced in that
industry in 1929. All three of the composite series (production,
total income, and labor income in the commodity-producing in­
dustries—less agriculture) are shown in the form of 3-months moving
averages to eliminate minor variations and facilitate comparison.
As was to be expected, there is a close correlation between the
dollar value of industrial output and salaries and wages originating
in the commodity-producing industries. The income series is more
stable with respect to shorter fluctuations, but follows the dollar
value of production very closely on the long swings. Thus, by the

spring of 1933, income had reached a low point 64 percent below its
1929 maximum, whereas a month or two earlier the value of current
production had turned upward from a point 69 percent below its
predepression peak. In the summer of 1937 both indexes approxi­
mated 80 percent of their respective 1929 peaks. The relatively
greater stability of the income index may be explained in large part
by the heavy weight given to durable producer’s goods in the Reserve
Board’s index of manufacturing production.
The month-to-month fluctuations in the index of total income
payments are of a very mild character as compared with the series
discussed above. In general, minor fluctuations in industrial output
are not reflected in total income payments except as they take the
form of changes in the slope of the latter curve. Thus, the tempo'• With respect to fig. 2, above, it should be noted that this report was in proof form when the Reserve
Board completely revised its index of industrial production to provide a broader measure of current changes
in industrial output. The relationships noted above are, however, but little affected by the inclusion of
the new series in the commodity-value index. Prior to 1935 the latter index was virtually unchanged by
the substitution, although it showed slightly greater resistance to short-term fluctuations. Since 1935,
however, the new series shows a marked secular rise relative to the old series and also relative to the laborincome index and since 1937 has averaged considerably higher than either of these series on a 1929 base.

MONTHLY INCOME PAYMENTS, 1929-40

19

rary industrial recovery in the fall of 1932 resulted merely in slowing
up the decline in income payments which continued into the spring
of 1933. Similarly, the fluctuations in industrial output after the
speculative spurt in 1933 were reflected in a slower rate of growth
in income during 1934 as compared with the rapid increase during
1933 and that which began in June 1935. Reflecting the short-term
stability in income payments, the index has shown only five definite
reversals of trend in the 11-year period covered by the series—these
trend reversals being associated with the long cycle that reached its
peak in the fall of 1929 and its trough in the spring of 1933, and the
shorter cycles of the past 3 years, in particular the downward turn
in the late summer of 1937, the upward swing that began in the
middle of the following year, and the recent decline from the peak
of December 1939.
It will be noted that for the period 1929 to 1935, the value of
commodity production, exclusive of agriculture, shows a lead of
1 to 3 months over the index of labor income in the same general
industries. This relationship is particularly noticeable in 1929.
Although industrial production reached its peak about the end of
the second quarter of the year and declined sharply throughout the
third quarter, the income index advanced markedly throughout the
latter period and did not show a noticeable decline until the final
quarter of the year. The lead of production over income appears
also in the rise that began in the fall of 1932 and again in the spring
of 1933, and in every other instance in the following few years, but
in no case is the timing as clearly marked as in 1929. Since 1937
it is impossible to establish any definite tendency of this nature.
In respect to timing, the four trend reversals in the total index noted
coincide with trend reversals in the index for the commodity-producing
industries, although minor trend reversals in the latter series have no
counterpart in total income. This close correspondence is more or less
to be expected since the commodity-producing industries account for
about 30 percent of total income payments and represent by far the
most variable component of that total.
In general, then, it may be said that although the index of total
income reflects the general changes in industrial output in the past
10 years, the income index has, on the whole, shown much more
resistance to cyclical fluctuations than the former index, and shows
considerably more stability in the short run.17 In computing the
index from month to month, it has frequently been remarked that a
sharp rise in industrial production during a given month may be
accompanied by a very small rise in the income index. This must be
explained not only by the relative stability in certain components of
the dollar index but also by the fact that it takes a certain amount
of time for impulses generated in any given branch of the manu­
facturing industry to propagate themselves through the economic
system. Figure 3 illustrates the manner in which various industries
are affected by changes in the volume of current industrial output.

17 It is proper to point out that this short-run stability may reflect in part rather inadequate source
material and sometimes arbitrary methods of handling such material as is available. In fwticular, the
method of adjusting dividends and interest for seasonal variation is such as to yield a smooth curve free of
minor fluctuations; similarly, monthly withdrawals are estimated on the assumption that the entrepreneur
frequently draws upon his capital to maintain his scale of living and that withdrawals are, over the short
term, relatively independent of net income which may be supposed to fluctuate more or less with current
business activity.

20

MONTHLY INCOME PAYMENTS, 1929-40

More the pay rolls in the economic system are classified into three
groups of industries: Pay rolls in the commodity-producing industries
depend very closely even in month-to-month fluctuations on industrial
production; the distributive industries perform services for both pro­
ducer and consumer and, therefore, reflect both changes in industrial
activity and consumer expenditures; whereas salaries and wages
received in the service industries depend directly on consumer expendi­
tures and only in a general way on current industrial output. Govern­
ment pay rolls are not shown here since they can be assumed to bear
no relation to industrial production in the. short run. Of course, the
interdependence of the various industries is greater than that indicated,
since industrial output and consumer income are very closely related.
Therefore, month-to-month changes in current income reflect not
only current changes in industrial output but also the cumulative
effects of recent changes in the same aggregate.
ft is not to be expected that over longer periods of time the income
index will continue to reflect changed levels of industrial output,
owing to the varying rates of growth of different industries. Figure 3
indicates that, over the past 10 years, the service industries have
shown a secular growth relative to the distributive industries. No
such conclusion is possible with respect to the commodity-producing
industries, since it is impossible to determine how “normal” was the
situation existing in the first half of 1937. Collateral evidence on this
point, however, indicates a gradual decline in the relative importance
of the commodity-producing industries in the decade ended 1928 and
suggests that this relative decline in importance may well have
continued into the period covered by the monthly income-payments
series.
The accompanying chart (fig. 4) permits a comparison of the trend in
income payments with that in bank debits outside New York City over
the past 7 years. Both series are represented by 3-month moving
averages to facilitate graphical comparison. This comparison is of
particular interest because income payments have often been identified
with the gross volume of money transactions. Actually, income
payments represent only a small fraction of all transactions. Thus,
in 1939, income payments totaled 70 billion dollars whereas bank
debits in 141 cities outside of New York City totaled 218 billion
dollars. Debits of New York City banks accounted for an additional
150 billions. In addition to income payments, other important
factors in the debiting of accounts are interbusiness transfers and title
transfers among individuals, particularly stock-exchange transfers.
The importance of the latter item is greatly diminished, however, by
excluding the debits of New York City banks and by confining the
comparison to the period following the bank crisis of 1933. For this
period the two.series show a very high degree of correlation with
respect, to cyclical fluctuations. The income index is more stable in
the short run because of the relatively greater sensitivity of bank
di'bits to month-to-month fluctuations in general business activity.
For the years prior to 1933, the bank debits index exhibited markedly
wider cyclical fluctuations than the income index. In 1933 income
payments were 40 percent below 1929, whereas bank debits showed a
drop of 56 percent over the same period, reflecting partly the abnor­
mally large volume of stock transfers in the earlier years.

MONTHLY INCOME PAYMENTS, 1929-40

Figure 4.—Income payments and bank debits.

21

Table 2.—Income Payments, by Months, 1929 40
index of
total Total in­
income come pay­
payments ments Adjusted
Year and month adjusted
index

Salaries and wages

Millions of dollars

99. 5
99. 2
98.9
99. 0
99.0
99. 3
loo. 3
102. 2
101. 4
102. 0
99. 0
99. :>

6.937
6. 342
6. 553
6. 759
0,681
6.891
7. 180
6,739
6,988
7. 400
6,810
6. 784
82,004

99.5
99. 2
99. 3
99. 2
99.3
99. 5
100. 5
101. 8
101.4
101.5
99. 7
99. 1

4, 146
4. 176
4. 276
t. 341
4, 426
1,453
4. 355
i. 306
1. 494
1. 561
4,371
4. 304
52,299

1.641
1.691
1.748
1.801
1.861
1.885
1.868
1,910
1.917
1,824
1.784
1,684
21,717

1.117
1.093
1, 129
1. 140
1. 162
1. 165
1. 178
1.179
l, 184
l, 205
1, 166
1,184
13,902

970
970
978
978
978
977
981
984
980
995
986
997
11,780

418
419
421
422
425
426
328
323
407
437
435
439
4.900

98.2
97. 0
94. 9
94. 4
93. 3
91.7
90.0
88.6
88. 0
80.4
84.8
83.9

6,999
6, 211
6, 296
6,416
6. 309
6,418
6.408
5. 828
5,975
6,178
5,766
5.720
74. 524

97.5
90. 1
94. 6
93. 7
92. 6
91.4
90. 1
88.7
88.2
86.7
85.0
83.8

4.087
4.060
4.082
4.099
4,124
4, 102
3,888
3,811
3.886
3. 893
3, 734
3,660
47,426
“ ■

1,567
1,583
1,588
1,612
1,632
1.027
1.560
1,530
1,527
1,505
1,395
1,308
18. 434

1.116
1,082
I. 102
1. 100
1, 107
1.098
1,074
1,049
1,045
1,050
1,013
1,024
12,860

973
964
959
952
947
937
917
902
898
893
882
881
11,105

431
431
433
435
438
440
337
330
416
445
443
444
5,023

._ . _—

Total

. ... .

Adjusted
index
1929=100

5
5
5
5
5
4
5
5
5
5
5
6
60

77
70
77
77
78
78
78
78
78
80
79
79
935

1,347
795
898
1.038
857
1,007
1.306
738
856
1,084
901
1.021
11,851

1,362
1.290
1.297
1. 298
1.315
1.349
1, 436
1, 522
1,555
1,070
1,451
1,374
16;919

6,320
5. 807
6, 000
6, 189
6. 076
6. 236
6, 443
5, 939
6. 156
6, 432
6,078
6, 141
73,817

98.6
98. S
99. 1
99. 1
99. 3
99. 5
100. 5
101. 8
101. 6
101. 7
100. 1
99. S

7

80
79
81
81
81
75
87
83
83
85
86
88
989
' —- -

1,536
851
937
1.031
889
1,043
1. 242
742
772
968
803
898
11,715

1,289
1,214
1. 188
1. 191
1,208
1. 192
1. 184
1. 185
1.227
1. 223
1, 134
1.062
14,300
...----

6, 456
5. 730
5,823
5.915
5, 767
5,872
5,857
5. 277
5, 377
5,568
5,253
5, 293
68,188

98.7
97. 4
96. 3
95. 4
94. 3
93.0
92.0
90.5
89.8
88.5
87.0
86. 1

887

7
6
7
7
7
9
1
9
12
3
94
4
—1 ---------

.

— 1----- —

MONTHLY INCOME PAYMENTS, 1929-40

1929-100 ofMillions
dollars 1929-100
1929
January
February
March . . .
.Nlay... ..........
June...................
July
September.. .
October..
November.
Decomber.
Total
1930
January
February
March
April
M ay..................
June......... .........
July
August
September
October.
November
.
December
T o ta l..........

Nonagricultural
income

Social
EntreDivi­ preneuria
Direct Security
benefits dends
income
Com­ Distrib­ Service
and
other
and
other
and net
modity- utive in­ in­ Govern­ Work
relief and
Total producing
relief
labor in­ interest rents and
ment wages
dustries
royalties
dustries
come
industries

1931

82.7
81.9
85.4
85.1
80.0
77. S
70. G
74.7
72.9
71.7
70.8
G9.8

5,873
5,271
5,673
5,813
5.389
5, 431
5. 436
4.897
4,935
5,133
4,819
4, 782
63, 452

82.6
81.9
81.0
79.8
78.3
77. 1
76.3
74.8
73. 1
71.5
70.3
69.0

3, 480
3, 472
3,506
3, 494
3,486
3. 443
3,272
3. 190
3,211
3.202
3.090
3,019
39,865

1, 211
1. 229
1,245
1,255
1,258
1,233
1. 200
1,174
1, 136
1. 108
1,033
977
14,059

970
948
969
956
951
945
929
903
892
886
860
858
11,070

859
853
849
839
827
818
799
780
774
769
760
743
9,670

438
439
439
440
441
441
337
329
405
434
431
433
5,007

2
3
4
4
6
6
7
4
4
5
6
8
59

14
14
14
13
12
11
11
12
12
13
14
IS
158

88
92
395
442
182
124
121
110
107
110
106
114
1,991

1,283
733
793
909
754
914
1,087
080
706
886
724
801
10, 270

1,008
960
965
955
955
939
945
905
899
922
885
830
11,168

5,507
4,952
5, 332
5, 471
5.030
5,073
5,002
4, 564
4,597
4. 755
4. 476
4.505
59, 324

84.9
84.3
88.3
88. 1
82.7
80.7
79.7
78. 1
76.5
75. 1
73.8
72.9

68.3
60.8
64.9
62.7
61.0
58.8
57.3
57.0
56.8
56.7
56. 1
55.6

4,865
4.297
4. 295
4.307
4. 132
4. 162
4,070
3,693
3.828
4,048
3.804
3,818
49, 319

67.6
66.0
64. 2
61.9
60. 1
58. 1
56. 3
55.8
56.0
56.2
55.5
54.7

2,868
2,818
2, 792
2, 716
2. 677
2.591
2,392
2, 356
2.457
2,517
2, 448
2.403
31,030

892
887
866
832
S09
777
749
746
767
784
740
700
9,549

800
764
768
746
740
709
080
665
670
682
665
667
8.556

737
724
711
61«
683
660
635
622
622
626
618
612
7,943

428
431
433
435
435
436
319
313
388
414
408
410
4.850

11
12
14
10
10
9
9
10
10
11
12
14
132

17
21
25
24
25
26
23
26
27
30
37
45
326

110
103
102
102
101
97
115
119
103
100
98
98
1,248

1,084
602
646
765
643
790
888
531
555
706
553
aso
8, 393

786
753
730
700
686
658
652
661
686
695
673
642
8,322

4, 620
4,076
4.0-80
4,095
3, 907
3, 947
3, 844
3, 462
3, 567
3, 771
3, 547
3. 602
46, 518

71.3
69.6
68.0
65.7
63.9
61.9
60.4
59.9
59.6
59.6
58.7
58.2

55.6
54.6
53. 1
53.0
54. 4
56. 1
56. 6
58. 1
60.0
60.3
60. 6
62.4

4,044
3. 535
3. 513
3,611
3. 656
3,885
4. 016
3. 759
4. 036
4. 385
4, 096
4,294
46,830

54.4
53.4
51.2
51.0
51.9
53. 6
55.0
57.6
58. 7
59.2
59.7
62.8

2,319
2.285
2.230
2. 240
2,310
2.386
2. 336
2. 434
2, 565
2. 645
2. 621
2,751
29, 124

662
672
643
662
706
758
796
851
887
899
859
839
9,234

633
605
596
602
617
628
638
665
682
707
690
699
7, 762

602
589
570
569
572
579
580
597
609
623
625
629
7,144

401
395
391
378
380
378
277
273
342
371
370
372
4,328

21
24
30
29
35
43
45
48
45
45
79
212
656

48
50
57
52
49
47
42
42
41
44
53
55
580

98
96
97
95
93
109
77
77
77
75
74
74
1,042

938
501
527
613
532
625
798
452
483
723
516
643
7,351

041
603
602
611
672
718
763
7.54
870
898
830
771
8,733

3,839
3, 357
3,319
3, 395
3,373
3, 55(3
3, 643
3,414
3. 578
3, 897
3,691
3. 970
43,032

57.8
56.8
55.2
.54.7
55.3
56.8
57.3
59.4
60.3
60.8
61.6
64.0

MONTHLY INCOME PAYMENTS, 1929-40

January.-...............
February.. .. . ..
March
.........
April.
.............
May.
June.. .. __ ...
J u ly ____________
August.....................
September______
Octol>cr............. ......
November............
December.
Total. .. . ..
1932
January.
February
March............ .......
April.........................
May......................
June........................
July_____________
August ...............
September..............
October ... . . .
November_______
December...............
Total..........
1933
January
F eb ru ary
March
...........
April.........................
M a y .....................
June..........................
J uly ............................
August ..............
September ..............
October ...........
November................
December.................
Total...............

bO
cc

Table 2.—Income Payments, by Months, 1929-40—Continued

' ...............

Index of
total Total in­
income come pay­
payments ments Adjusted
index
Year and month adjusted

Salaries and wages

ill ions 1929-100
1929-100 ofMdollars
1934
64. 4
January. ______
64. 5
February
64. 8
March
______
64. 0
April
................
May ......................
65. 5
65. 6
June
................
July
.................
65.8
67. 0
August ..............
66. I
September ............
67.3
October ..............
67.2
November..........
67.4
December ...........
Total...............
1935
January
68.8
69. 6
February __
69. 6
March
.......
70.5
April
............
May . .....................
70. 6
June..........................
70.3
July........................
69. 9
71.9
August
72.8
September
74. 1
October
75.«
November
76. 1
December .
Total........... ~ .......

Social
Entre­ Nonagricultural
income
Direct Security
Divi- preneurial
l>enefits
income
and
other
dendsand
ottier
and net
Work
relief and
labor in­ interest rents and Total Adjusted
relief
royalties
index
wages
come

Millions of dollars

1929=100

4,685
4. 199
4. 274
4. 382
4. 244
4.471
4, 627
4,361
4. 585
4. 963
4, 475
4.740
54,006

64.2
64. 3
64. 6
63. 5
64.8
64. 7
64. 4
64.9
63. 5
64. 4
64.9
65.5

2,752
2,760
2,816
2,781
2,844
2. 649
2. 735
2. 750
2,781
2.886
2. 866
2,890
33. 710

830
893
944
977
1, 003
1.001
966
977
947
986
963
972
11,459

676
673
700
706
718
720
720
714
720
735
723
741
8,546

632
637
643
<>49
657
658
663
651
648
665
666
668
7, 827

367
375
377
378
387
393
311
310
376
405
407
405
4,491

247
182
152
71
79
77
85
98
90
95
107
104
1.387

56
58
69
<>9
68
64
62
68
67
76
81
90
828

74
72
75
76
85
82
82
82
82
84
84
83
961

1,003
547
547
705
451
639
857
494
629
825
463
777
7.937

800
762
767
751
796
837
891
967
1.026
1,092
981
900
10/570

4,338
3, 895
3.972
4.084
3.904
4,084
4. 189
3. 862
4.020
4. 332
3. 970
4. 333
48,983

65.0
65.3
66.0
65.4
66. 7
66. 5
66.6
67.0
65.8
66.8
67.4
68. 1

4,824
4, 495
4,676
4. 929
4. 615
4. 821
4. 774
4, 705
5, 127
5, 459
5,014
5.370
58,809

67.5
68. 2
68. 3
68. 7
69. 0
69. 1
69. 1
70.3
71.0
71.7
72.9
75.0
.........*

2,889
2. 928
2,977
3, 013
3. 031
3. 044
2. 936
2,981
3. 117
3,221
3,208
3,304
36, 649

973
1,021
1,044
1, 054
1, 053
1. 064
1. 046
1. 093
1, 132
1, 164
1, 136
1, 143
12,923

723
718
739
753
760
762
762
763
781
795
784
814
9, 154

674
<>80
<>85
689
691
695
694
693
702
713
718
727
8,361

402
403
407
415
418
423
333
335
416
451
450
456
4,909

117
106
102
102
109
100
101
97
86
98
120
164
1,302

98
94
102
102
97
89
88
91
89
94
84
71
1.099

83
84
85
86
86
87
87
86
90
89
89
88
1,040

879
520
623
810
476
674
717
.501
706
832
482
835
8,055

875
869
889
918
925
927
946
1,046
1, 125
1,223
1, 151
1,072
11,966

4,455
4,137
4. 294
4,514
4. 184
4. 381
4,318
4. 159
4,509
4, 745
4,381
4,837
52, 914

69.7
70.0
70. 1
70.5
70.8
70.8
70.9
71.9
72.6
73.4
74.2
75.6
■■

MONTHLY INCOME PAYMENTS, 1929-40

Com­ Distrib­ Service
modity- utive
Total producing
in­ in­ Govern­
ment
industries di istries dustries

bO

1937
January.................
February ................
March .. ...........
A pril........................
M ay............ .............
June..........................
J u ly ..........................
August
September . . . __ .
October
...........
November ____ .
December ............ .
Total..............
1938
January
February ................
M arch
April .....................
May ........................
June ........................
July
...........
August
.......
September ............
October
November
December ..........
Total...............

76.9
76.3
77.7
78.8
79.8
92.4
87.4
83.2
83.3
84.2
85.3
87.3

5,226
4.902
5. 188
5,335
5, 140
6,306
5,965
5,270
5, 842
6. 092
5. 625
6,955
67, 846

75.5
75.8
76.5
77.0
78. 1
78.6
79.3
80. 1
80.4
81.5
83.0
84.9

3,229
3,254
3, 327
3, 370
3, 430
3,462
3. 371
3,407
3.531
3.665
3, 670
3,733
41,449

1,097
1, 111
1,152
1, 186
1,232
1.254
1.255
1,293
1.307
1,361
1, 362
1,383
14,993

781
783
804
812
823
829
834
835
850
871
867
897
9,986

728
736
739
745
751
759
759
761
769
785
792
807
9,131

447
444
447
452
457
461
369
362
447
480
477
486
5, 329

176
180
185
175
167
159
154
156
158
168
172
160
2,010

62
62
61
56
51
50
49
50
53
55
58
65
672

88
88
88
87
87
887
500
164
135
122
112
114
2,472

876
568
733
812
523
799
875
472
852
895
507
1,809
9. 721

971
930
979
1,010
1,049
1, 108
1, 170
1, 177
1,271
1,355
1,278
1,234
13, 532

4,813
4, 532
4, 772
4, 884
4,646
5, 763
5, 357
4, 665
5, 148
5,320
4,946
6,349
61. 195

76.3
77.0
78.0
78.7
79.7
93.6
88. 1
83. 4
83.4
84.2
85.3
86.9

86.3
87. 4
89.4
89.0
88.9
89.3
89.2
89.0
87.3
86.6
84.9
83.7

5,810
5,417
5,944
6,015
5,699
6, 274
6, 133
5,863
6, 127
6,291
5,657
6,553
71,783

84.0
85.4
86.6
87.6
88.5
88.5
88. 5
89. 2
87.5
86.8
84.6
82.8

3,578
3.647
3, 764
3,822
3,884
3, 900
3, 772
3,800
3,849
3,904
3,737
3,640
45,297

1.309
1,365
1,430
1,483
1. 520
1,527
1,518
1,549
1, 518
1,521
1,384
1.259
17,383

848
850
890
804
912
919
918
929
929
942
912
924
10,867

794
80S
819
823
829
836
836
839
841
846
842
840
9,953

477
473
476
476
479
483
383
375
457
488
488
503
5,558

150
151
149
146
144
135
117
108
104
107
111
114
1,536

71
72
75
71
66
63
63
65
67
68
73
83
837

104
100
101
98
95
106
103
101
102
102
101
103
1, 216

901
488
788
866
.504
1, 015
932
622
785
852
480
1,561
9,794

1, 156
1,110
1,216
1,1.58
1, 150
1, 190
1,263
1, 275
1,324
1,365
1, 266
1,166
14.639

5,281
4. 938
5, 363
5, 488
5, 174
5, 703
5, 487
5, 216
5. 422
5, 535
5,000
6,002
64,609

86.0
87. 1
88. 1
88.9
89.3
89.6
89.5
89.5
87.8
87. 1
85.3
83.9

82.3
81.5
81.2
80. 1
79. 1
78.7
78.7
79.8
80.4
81.7
82. 1
83. 1

5,602
80.3
5, 104
79.9
5. 348
79.2
5,478
78.8
5. 168
78.4
5,543
78.0
5. 475
78.6
5, 183
80.5
5,674
81.3
82.0
5. 952
5,554
83. 1
6, 161
84.5
66,242 ................

3,420
3,407
3, 440
3, 446
3. 444
3. 443
3.340
3. 418
3. 570
3, 684
3,677
3,719
42,008

1, 145
1. 157
1. 160
1. 151
1. 156
1. 150
1, 155
1.211
1,255
1. 293
1,276
1.268
14.377

854
834
847
849
843
837
834
839
851
870
866
900
10, 224

816
800
794
791
780
778
772
111
787
799
810
824
9,528

487
489
496
500
504
510
404
403
489
524
526
533
5,865

118
127
143
155
161
168
175
188
188
198
199
194
2.014

89
90
91
85
81
80
80
80
80
80
83
89
1,008

103
120
146
139
140
146
143
152
145
139
131
129
1,633

878
4.56
600
760
458
814
814
434
712
760
473
1,099
8,258

1, 112
1,031
1,071
1,048
1,045
1,060
1,098
1,099
1, 167
1,289
1, 190
1, 125
13, 335

81.9
5, 106
4, 683
81.6
4, 888
81.4
5,029
80.5
4,710
80.0
5,069
79.7
4,956
79.8
4, 6S3
81. 5
5, 109
81.9
5,267
82.4
4, 987
83.3
5,679
84.4
60. 166 |................

MONTHLY INCOME PAYMENTS, 19 29-4 0

1936
January...................
February..................
March ....................
April ..................
May .....................
June..........................
July...........................
August
September................
October ..............
November ..............
December ..............

CtoT>

Table 2.—Income Payments, by Months, 1929 40—Continued
I ndcx of
total Total in­
income come pay­
payments ments Adjusted
Year and month adjusted
index

Salaries and wages
Com­ Distrib­ Service
modity- utive
Total producing
in­ in­ Govern­
ment
industries dustries dustries

1930
January
February
March
..........
April
May ....................
June .......... ......
July ................
August
September
October
November
December
Total........
1040
January ...........
February
M arch...................
April
May
__
June
July ___
August .......... .
September.
October
November
December
Total...............

N'onagricultural
Total

Millions of dollars

Adjusted
index
1929=100

83.4
83.7
84.6
83. 1
83.8
84. 1
83.6
85 2
86. 1
88.0
88. 5
90.0

5,720
5,208
5,771
5,674
5, 449
5,956
5, 736
5, 439
6,025
6, 259
5, 867.
6, 904
70. 096

84.3
84. 1
83.9
82.4
83. 1
84.3
83.8
84.8
85.0
86.9
87.7
88.9

3,585
3, 589
3,644
3,611
3,655
3, 723
3, 565
3,604
3, 738
3,911
3, 879
3,908
44,412

1,222
1,245
1,266
1, 245
1.272
1,321
1, 311
1, 359
1,398
1,472
1, 449
1,420
15,980

852
839
863
861
871
882
877
880
898
928
915
945
10,611

810
809
.808
810
816
826
817
816
827
846
850
862
9,897

518
518
524
522
528
536
419
417
504
541
537
547
6, 111

183
178
183
173
168
158
141
132
111
124
128
134
1,813

92
94
95
90
87
85
85
87
87
88
88
89
1,067

132
137
153
137
143
149
141
150
140
133
134
137
1,686

810
425
762
750
462
915
839
443
799
775
486
1,517
8,983

1,101
1,053
l' 117
1,086
1, 102
1,0*4
1, 106
1, 155
1,261
1, 352
1,278
1,253
13,948

5, 243
4,888
5, 295
5,214
4,962
5, 485
5,239
4,908
5,386
5, .>11
5, 239
6, 321
= =63,= =721=

00.3
89.7
88.4
88.2
88.6
88.7
89.3
90.3

6,093
5,604
5,987
5,965
5.689
6, 252
6,075
5.761

88.2
87.5
87.0
86.2
87.3
87.7
88.6
89.9

3,767
3, 742
3. 784
3, 784
3, 838
3, 864
3. 786
3,836

1,349
1,339
1, 352
1,356
1, 391
1,419
1,423
1, 189

902
882
900
900
908
915
920
925

847
845
845
845
854
,860
854
859

535
536
539
.540
548
550
440
445

134
140
148
143
137
120
118
118

95
95
94
92
89
86
85
85

148
151
155
152
166
166
167
164

840
447
820
799
472
1,021
886
469

1,243
1, 169
1. 134
1, 138
1, 124
1, 115
1, 182
1,207

5, 533
5,108
5,519
5,479
5, 211
5, 785
5, 534
5,210

S3 8
86 9
86 9
90.4
90. 2
89.6
89. 3
88.7
89.8
90. 7
91,1
92. 2
.................

MONTHLY INCOME PAYMENTS, 1929-40

1029-100 ofMillions
dollars 1929=» 100

Social
Entre­
Direct Security
Divi­ preneurs
benefits
income
other and other dends and and net
Work andrelief
relief
labor in­ interest rents and
wages
come
royalties

Appendix.—SOURCES AND METHODS

It should first be pointed out that for the most part the individual components
of the monthly series are based upon annual income estimates prepared in the
National Income Division. The sources and methods employed in constructing
the annual estimates have been discussed elsewhere,1and no attempt will lie made
to describe them here. Because the monthly estimates are corrected each year
within a short period after the completion of the annual income totals, the im­
portance of cumulative errors creeping into the monthly data is not very great.
On the whole, comparisons of the annual aggregate of monthly estimates with the
subsequent annual figures have so far yielded fairly satisfactory results. Thus,
for the year 1938, the following results were obtained:
Percentage Change in Income Payments by Type of Payment, 1938 to 1939
Item

Preliminary Revived
estimate,
estimate,
January September
1940 1
1940*
+6.3
+7.1
+3.2
+8.1

+8.7
+8.9
+4.«
+6.8

1 Based on the original monthly figures for 193».
* Revisions based on the annual income estimates for 1939.

It should Ire noted that the annual estimates of national income, particularly
for the years 1938 and 1939, are subject to further revision ns additional source
material becomes available. Though the absolute level of estimated income in
both years may change as the result of further information, experience has shown
that the trend from one year to another is likely to be but little affected by further
revisions. Even in the case of dividends for which substantial revisions may be
necessary on the publication of Statistics of Income for the year in question, the
monthly dividend estimates are likely to approximate the preliminary annual
estimate of dividends, since both are subject to the same limitation that they
depend too greatly upon the reported dividend disbursements of large corporations.
METHODS OF ADJUSTING MONTHLY INDEXES TO ANNUAL
ESTIMATES
Inasmuch as the monthly series is adjusted to annual estimates, special interest
attaches to the methods by which the adjustment is made. In only a few indus­
tries are the monthly pay-roll data sufficiently complete so that the annual esti­
mates represent merely a summation of the pay rolls for the individual months.
In many instances annual totals are available, whereas current estimates must be
based upon industry samples or upon series only indirectly related to income in a
given industry. In other instances periodical surveys or censuses must be used
to bring a monthly index into line with the actual trend in an industry.
In correcting the monthly income estimates to the annual totals, two general
methods have been used. One of the methods used is that adopted by the Bureau
of labor Statistics in adjusting its employment and pay-roll indexes to conform
with the trends of the industrial surveys of the Bureau of the Census. This is
essentially a straight-line adjustment between the annual averages shown in a
pair of census years. Underlying this method is the assumption that none of the
observed bias arose in the earlier census year, but that the causes producing this

1In detail in (he volume National Income in the United States 1929-35. and briefly in Income in the United
States, 1929-37.
27

M O N T H L Y INCOME PAYMENTS, 1929-40
28
bias began to operate in January of the following year and continued to operate at a
uniform rate until December of the second census year. A description of the
method is presented in Bulletin No. 610 of the United States Bureau of Labor
Statistics.
Because of its simplicity, the Bureau of Labor Statistics’ method was considered
satisfactory for periods of time longer than 1 year, but for adjusting the monthly
series to annual totals for each year it is believed to give too serious a distortion to
the month-to-month comparisons, since all of the correction would be made in the
second year of a 2-year comparison. This is especially true in the final months of
the year when an error is particularly important because all succeeding values of
the series must be distorted to the same extent.' Therefore, in making annual
adjustments, the method used was one based upon an unpublished paper by V.
Lewis Bassie of the U. S. Department of Commerce. This method involves the
assumption that the observed bias may have developed at any time between
January of the first census year and December of the second census year. The
most probable development of the bias is then calculated by constructing a
third-degree correction curve which is characterized by the following four limit­
ing conditions: (1) In the first year the average correction is zero; (2) in the

Figure o. Methods of adjusting monthly data to annual averages.

second year the average correction is such as will be necessary to bring the index
into line with the census results; (3) the correction shall be zero at the beginning of
the first year; and (4) the slope of the correction curve is zero at the end of the
second census year. Mr. Bassie’s mathematical formula, which was originally
derived in order to correct for biennial surveys, was then modified so as to be
applicable to those cases in which a survey (or at any rate information more
reliable than that available monthly) was available every year.
The accompanying chart indicates the difference between the results obtained
by the two methods described above. In constructing the chart it was assumed
that the necessary correction was one of 5 percent. A further difference should be
noted in the methods of correction. The straight line method is of an arithmetic
nature, the necessary correction for each month being added to the unadjusted
data. In the application of the corrections derived from the third-degree curve,
the corrections so obtained are added to or subtracted from 100 percent and the
resulting percentage multiplied by the monthly values of the unadjusted series.
Although this method will not reduce the total for the second year to the exact
survey total, the resulting error will be negligible. In those instances in which
the necessary correction is small and the seasonal fluctuations in the unadjusted
series of minor importance, some saving of labor is obtained, without a correspond­
ing reduction in accuracy, by applying an arithmetic correction derived from the
third-degree curve shown in the chart.

MONTHLY INCOME PAYMENTS, 1929-40

29

NATURE OF THE DOLLAR TOTALS AND ADJUSTED INDEXES
In the analysis of time series it has been found useful to distinguish four factors
which constitute the fluctuations in such a series. These factors arc, respectively,
the secular trend, and cyclical, seasonal, and random elements. The process of
adjustment for seasonal variation consists in eliminating the seasonal influence in
order to facilitate an analysis of the trend, cycle, and random elements.
In the case of the present series, neither the totals nor the adjusted indexes
can be considered to conform strictly to this usage. The character of the com­
ponent series will be evident from the methods of estimation outlined below, but it
may be useful here to indicate the general character of the unadjusted dollar totals
of income payments. In general the unadjusted series includes all four elements,
but in some instances random elements arc necessarily excluded by the method of
estimation. This would, in particular, be true wherever a seasonal movement is
superimposed upon a trend-cycle line constructed from the annual figures—which
was the procedure adopted in estimating labor income in several minor industrial
fields. In other instances, as in estimating net rents and royalties, it has not been
considered feasible even to impose seasonal fluctuations, and in such instances only
trend and cycle elements enter into the composite scries. For most of the fields
for which such procedures are necessary, however, as, for example, in estimating
rents and royalties or income in the service industries, random or seasonal ele­
ments are not believed to be of great importance. Such a generalization may
not, however, be expected to hold true of such critical periods as the earlv months
of 1933.
The adjusted index of total income payments contains even fewer random
elements than the unadjusted totals because of the special treatment accorded to
dividend and interest payments. Because the seasonal pattern, particularly for
dividends, is of a very unstable character, it is impossible to distinguish very
satisfactorily between seasonal and random elements, and, therefore, the seasonal
adjustment adopted is of such a character as to remove not only all seasonal
influences but also all random elements of a short-term character. In one other
respect the adjusted index does not confrom to general practice: namely, for relief
payments and other benefits it has not been considered advisable to attempt a
seasonal adjustment. For these items the movement appears to be highly irregu­
lar, and a sufficiently long period is not available to yield a satisfactory seasonal
pattern. Consequently, the seasonal elements undoubtedly present in relief
payments impart a slight seasonal influence to the adjusted index of total income
payments. On the other hand, relief benefits represent but a very small fraction
of all income disbursements, and, when properly weighted, their seasonal varia­
tions may be considered to be of negligible importance.
Some few remarks are in order here as to the purposes which it is believed that
the totals and the adjusted indexes will serve. In general it is believed that the
adjusted index will be the more useful at the present time because the wide sea­
sonal fluctuations in certain items, particularly dividends and interest, give a
rather misleading picture of changes in economic activity. Even in analyzing
consumer demand the unadjusted totals are of rather limited value because of the
present impossibility of correcting for the flow of funds into and out of savings
accounts and other deposits. Nevertheless, the unadjusted totals arc presented
for those who may wish to attempt such adjustments. However, as noted in the
body of this report, the series is at present most readily applicable to the analysis
of the demand for consumers’ goods in the narrow sense and for such expenditures
the seasonally adjusted index may be considered to give a rough adjustment.
Teachers, for the most part, actually receive income during 8 to 10 months of the
year, but their expenditures may be assumed to be spread out fairly evenly over
the 12 months of the year.
Because of the minor character of the short-term fluctuations in income pay­
ments and the possibility of changing seasonal influences in various industries,
too much importance should not be assigned to small month-to-montli changes in
the adjusted index. Except in periods of major fluctuations in industrial activity,
current comparisons are probably better limited to quarterly averages.
METHODS OF ADJUSTING FOR SEASONAL INFLUENCES
In adjusting the monthly income series for seasonal variation, each of the major
component items is adjusted independently. For salaries and wages, a seasonal
factor is computed for each of the major industries, and the seasonally adjusted

30

MONTHLY INCOME PAYMENTS, I929-4U

indexes thus obtained are weighted by the relative importance of the industrial
groups in 1929 in order to arrive at the seasonally adjusted index of employees’
compensation. In order to obtain the seasonally corrected index of total income
payments, the adjusted index of salaries and wages is in turn weighted by the
relative importance of this type of income to total income in 1929 and added
to the similarly weighted adjusted indexes for dividends and interest and for
entrepreneurial income. In the case of entrepreneurial income, the net income
of farm operators is the only component which need be corrected for seasonal
fluctuations, as the other components were assumed to show no seasonal fluctua­
tions (cf. under “Entrepreneurial income”).
As noted above, relief payments, benefits, and related items are not corrected
for seasonal variation. For those items (direct relief and other labor income)
which run back to 1929, the unadjusted indexes were weighted and added to the
cumulative total obtained above. For those items (work relief, agricultural
benefits, and adjusted service payments) which were not present in 1929, the same
effect is obtained by expressing them as a relative of the average of total income
payments in 1929 and adding to indexes for salaries and wages, entrepreneurial
income, and dividends and interest. Work-relief payments are included in the
adjusted index of salaries and wages as well as in total income.
Except in the case of dividends and interest, the ratio-to-moving-average is the
general method employed in correcting for seasonal influences. For all industrial
groups, it was considered desirable to work out two independent sets of factors,
one for the period 1929 33, the other for the period 1934 to date. The method of
correction adopted for dividends and interest yielded no usable seasonal factors
(cf. discussion under dividends).
SOURCES AND METHODS FOR SPECIFIC ITEMS
SALARIES AND WAGES

Estimates for this type of income are based primarily on the monthly pay-roll
indexes of the Bureau of Labor Statistics. For the most part, these indexes
have proved rather satisfactory in reflecting changes in pay rolls when checked
by the various industrial censuses. It should be noted that the Bureau of Labor
Statistics indexes are based on reports referring to typical pay periods within the
month, generally the pay period ending nearest the 15th of the month. There­
fore, they do not reflect total pay rolls for the month, but rather a rate of flow
of wages for a week in the early or middle part of the month.2 Therefore, those
indexes are not affected by the length of the month or by the number of Saturdays
or pay periods occurring within the month.
In the process of adjusting these indexes to annual wages, the total pay roll for
the year was allocated according to the rates of pay indicated for each month.
Where wages are paid only for time actually worked there is a possible source of
error in this practice. Since the weekly pay-roll periods in question may end
anywhere between the 12th and the 18th of the month, this period will in certain
years include such holidays as Labor Day, Armistice Day, and in certain States
Lincoln’s Birthday and the Election Day holiday. In such years the estimated
rate of pay for the month affected will be lower than the true rate, in some indus­
tries perhaps one-sixth lower. If these holidays were included in the pay-roll
period each year, the process of adjustment for seasonal variation would eliminate
this seasonal factor from the adjusted index. But since most of these holidays
may in some years fall within the pay period, and in other years outside it, there
remains a residuum of error even in the seasonally corrected index. Any errors
introduced by holidays, strikes, or other factors, are not, however, cumulative in
nature, but are automatically corrected for the succeeding month.
The industries for which the Bureau of Labor Statistics monthly pay-roll
indexes are currently used are the following:
Manufacturing, wages only.
Brokerage.
Mining, wages only.
Wholesale and retail trade.
Street railways.
Year-round hotels.
Telephone and telegraph.
Laundry and dry-cleaning establish­
Electric light, power, and gas.
ments.
Insurance.
* For some industries, notably minim- pay rolls are made up on a semimonthly basis so that the p a y ­
roll indexes cover pay rolls for the first 15 days of the month.

MONTHLY INCOME PAYMENTS, 1929-40

31

For the most part, these indexes are available back to 1929, but in some minor
instances it was necessary to resort to extrapolation for the early years. In the
aggregate the industries covered by the Bureau of Labor Statistics pay-roll
indexes account for approximately 47 percent of all salaries and wages paid in
1938.
In addition to the Bureau of Labor Statistics data, official reports of pay rolls
on a current basis are available for several other important industries, including
the executive, legislative, judicial, and military services of the Federal Govern­
ment, the majority of steam railways, the various forms of work relief, and for
those banks which are members of the Federal Reserve System. In the aggregate
these reports covered an additional 16 percent of all salaries and wages in 1938.
In addition, as noted below, information of this character will soon be available
for the pay rolls of State and local governments.
For another large group of industries, information of a less satisfactory character
is available. In recent years State departments of labor and industry have been
collecting statistics relating to pay rolls in an increasing number of industries.
For other industries, information relating to production or shipments can be
taken as indicative of current changes in employment and pay rolls in particular
lines of industry. Industries susceptible to this type of treatment account for a
further 17 percent of total wages and salaries.
There remains about 20 percent of all salaries and wages for which no current
information is available. This group includes, however, several important indus­
tries, such as education, which are subject to very small changes from year to
year, but contains in addition several large industries, mostly of the service
group, which experience definite cyclical fluctuations. The method used to
estimate month-to-month changes in this area is indicated below.
In the following discussion describing the particular method used to estimate
month-to-month changes in the salaries and wages for each industry, no reference
will be made to those industries for which the Bureau of Labor Statistics collects
current information, since the Bassie method, discusser! above, was used in each
instance.
A G R IC U L T U R A L W A G E S

The basic material on employment and wage rates in agriculture is taken from
pages 8 to 11 of the January 1939 issue of Crops and Markets. The product of
hired labor and the weighted average wage rate is adjusted to the annual totals.
The necessary adjustments are usually very small. Employment and wage-rate
figures for the first of each month are assumed to apply to that month. Since
wage-rate data are compiled only on a quarterly basis it is necessary to estimate
wage rates for the other 8 months. For this purpose indexes of seasonal variation
were computed for the months of January, April, July, and October. It was
observed that the seasonal fluctuations in wage rates roughly parallelled the
seasonal variations in employment. (Seasonal variation in agricultural employ­
ment from p. 17 of Trends in Employment in Agriculture 1909 1936 W. P. A.)
Seasonal indexes of wage rates for the other 8 months were, therefore, graphically
interpolated by comparison with the seasonal indexes of employment for those
months. Nonseasonal fluctuations in wage rates are interpolated by a straightline method. In making the current estimates, it is usually necessary to estimate
wage rates 1 or 2 months in advance of their compilation by the Bureau of Agri­
cultural Economics. For this purpose the main reliance is upon the seasonal
index, but in times of substantial nonseasonal fluctuations in agricultural employ­
ment a roughly determined cyclical pattern is applied.
M A N U F A C T U R IN G

S A L A R IE S

Several attempts have been made to construct a satisfactory indicator of changes
in this important source of income (salaries in manufacturing constituted nearly
8 percent of total salaries and wages in 1938). However, because of the dearth
of source material, the present series is projected largely on the basis of the rela­
tionships that are observed to hold between salaries and wages in this industry
over the past 10 years. In distributing the annual totals for the period 1929-38,
some help was obtained from an index of clerical salaries in Wisconsin factories
(recently discontinued). Some assistance was also rendered by a semiannual
index of clerical salaries in manufacturing industries compiled by the Bureau of
Labor Statistics for the years 1935-38, but as yet unpublished.
Both series suffer from the limitation that they apply exclusively to clerical
employees, who represent only about one-third of the total salaries in manufac­

32

MONTHLY INCOME PAYMENTS, 1929-40

turing. In addition, the New York State Department of Labor publishes an
annua] study of the earnings of office workers employed in New York State fac­
tories. This tabulation has the advantage that it covers supervisory and technical
as well as clerical employees, but its usefulness is limited by the fact that it refers
to only 1 month (October) of each year. Nevertheless, it does serve as a useful
check on the estimated trend in manufacturing establishments during the earlier
part of the years.
The Pennsylvania Department of Labor and Industry has in recent months
teen publishing estimates of employment and weekly pay rolls of salaried workers
in manufacturing establishments in that State. While this series is of too recent
origin to be of value for historical purposes, it is expected that it will in the future
be of considerable value in estimating current changes in this field.
At present it is assumed that there is no seasonal variation in salaries in the
manufacturing industry. An instance of the relative seasonal stability in salaries
as compared with wages for an extreme case is given by a study of the Ohio
Unemployment Compensation Commission which indicated that the employment
of office and inside sales workers in Ohio canning establishments varied front 85.3
to 130.7 of the annual average, while the employment of all employees varied
from 45.4 to 343.2 of the average. Inasmuch as the extremes in the seasonal
variations for total manufacturing wages are within 5 |>ercent of the year’s aver­
age, the seasonal variation in aggregate salaries may be assumed to be very slight.
Using all this information, it was concluded that, in general, salaries follow the
trend of wages after seasonal correction, with a somewhat smaller amplitude of
the cyclical swings. There is also a tendency for salaries to lag behind wages,
but the relationship is not sufficiently definite to be taken into account in deter­
mining current changes in manufacturing salaries. It appears that in recent years
salaries have become more responsive to current business trends and in this respect
parallel more closely the month-to-month changes in wages.
M IN IN G S A L A R IE S

As in manufacturing, monthly salaries in this industry are based largely on
changes in the seasonally corrected index of wages.
C O N S T R U C T IO N

The Bureau of Labor Statistics publishes a monthly pay-roll index, which,
however, is not sufficiently complete for the purposes of this series, since it covers
only building construction and excludes road building and other types of engi­
neering projects. In addition to the monthly index, the Bureau of Labor Sta­
tistics estimates monthly employment in all contract construction in connection
with its monthly employment aggregates. This estimate, together with the
Labor Department’s index of average pay in building construction, is used to
distribute annual pay rolls as estimated by the Department of Commerce, and
also to project these estimates on a current basis.
S T E A M R A IL W A Y S , P U L L M A N A N D E X P R E S S

The compensation of employees of class I railways are reported monthly to the
Interstate Commerce Commission. These monthly reports are adjusted to the
total of employees’ compensation for all railways as shown in the annual publica­
tion entitled. “Statistics of Railways.” The pay rolls of the railway express
system and the Pullman Co. have been reported monthly to the Interstate Com­
merce Commission since 1934. Prior to 1934 the monthly break-down for these
companies is based on revenue statistics. It should be noted that the railway
companies report to the Interstate Commerce Commission the wages paid out
during the calendar month. This is the only industry, therefore, in which a
correction should be made for the number of working days in the month. Actu­
ally, the refinement to be effected did not seem to justify the labor involved, and
no correction was attempted.
The Interstate Commerce Commission’s figures are not made available in time
to be included in the estimates for the latest month. As a rough expedient, the
adjusted index of railway employment is assumed to represent the cyclical change
in compensation. The change in the employment index is then corrected to give
effect, to the ordinary seasonal fluctuations in compensation. The corrected
percentage change is then applied to the compensation of railway, Pullman, and
express employees for the preceding month.

MONTHLY INCOME PAYMENTS, 1 9 2 9 - 4 0

33

M O T O R T R U C K IN G

Current information relating to pay rolls in the motor-trucking industry are
confined to three series: (1) an index of wages for identical reporting firms classi­
fied as “ Motor freight, docks and warehouses” by the Pennsylvania Department
of Labor and Industry; (2) a similar index for the “Teaming, Trucking, and
Handling Industries” in Massachusetts; (3) an index of commodity movements
by motor truck prepared by the American Trucking Association. The value of
the two pay-roll series is impaired by the fact that neither index covers the “motor
trucking” industry alone. The Pennsylvania Department of Labor and In­
dustry’s index shows very wide fluctuations over the past few years, whereas the
Massachusetts index has been very stable. The American Trucking Associa­
tion's index refers largely to interstate trucking and shows wider seasonal and
cyclical fluctuations than would be expected of a more comprehensive index. An
additional limitation upon this index is its availability only sines January 1937.
From that date, pay-roll changes in the motor trucking industry have been based
on the three indexes. The indexes are plotted on a single chart and a composite
index estimated graphically by a rough weighting of the changes in the several
indexes.
For the period 1932 to 1936 the two State indexes form the basis of the monthly
break-down. For the period 1929 to 1931, the month-to-month changes are
based on monthly fluctuations in miscellaneous and 1. c. 1. carlondings, because
in Massachusetts such loadings were found to correlate closely with the pay-roll
index for the years 1932 to 1937.
M OTOR B U SSES

The motor-bus industry is relatively limited in scope and magnitude, excluding
as it does motor busses operated by electric railways or subsidiaries or their
successor companies. . These latter are covered in the Bureau of Labor Statistics
pay-roll index for street railways. For estimating the month-to-month trend in
pay rolls of independent bus conipanies, some weight is given to the pay-roll index
prepared by the Massachusetts Department of Labor, covering 90 percent of the
industry in that State. Some weight is also given to the volume of passenger
traffic on steam railways with special allowance for those periods in recent years
when changes in passenger rates have presumably influenced the competitive
position of the railroads.
STO RAG E

The month-to-month changes in pay rolls in the storage industry are estimated
by applying an index of seasonal variation to a smooth curve connecting the annual
averages. The index of seasonal variation is constructed from monthly pay-roll
data in the 1935 Census. As the revenue from the storage of agricultural products
constitute a substantial part of the revenue of this industry, some weight is given
to crop movements in constructing the curve. For current months, the smooth
curve is simply carried forward by reference to similar periods in the past, with
special weight given to the movement of crops.
W A T E R T R A N S P O R T A T IO N

Pay-roll estimates in the field of water transportation are of necessity based on
statistics relating to the monthly movement of freight by water. The only useful
series available are: (1) Volume of freight moving through the Sault Sainte
Marie; this series is believed to be a fairly good indicator of freight movements
and to a lesser extent of wages on the Great Lakes. (2) Ship clearances of
American vessels in foreign trade, which give some indication of wages of personnel
on United States vessels engaged in foreign commerce. (3) Movement of freight
through the Panama Canal in United States vessels—largely composed of inter­
coastal traffic but, of course, including some traffic from the east coast to the
Orient or to the western coast of South America and some traffic between west
coast and Europe, which, to this extent, is a duplication with (2). (4) Ship
clearances in foreign trade (United States and other vessels), which should provide
a rough measure of the wages of stevedores in boarding such freight. (5) The
War Department publishes monthly figures on freight movements on the Ohio,
Allegheny, and Monongahela Rivers, but these series are not indicative of any­
thing except traffic in the Pittsburgh district and should be weighted accordingly.

34

MONTHLY INCOME PAYMENTS, 1929-40

The principal deficiency is that there is no current index of freight movements in
coastwise trade which is the most important of all types of water traffic. In order
to give some effect to the available indexes of current freight movements, a
weighted index was constructed from the five component series listed above for the
period 1929 to 1938. In 1929, the salaries and wages of the labor involved in the
handling and movement of the traffic reflected in this index is believed to have
represented about 60 percent of the total salaries and wages in the water-trans­
portation industry as a whole. The residual 40 percent roughly measures the
salaries and wages of persons employed in coastwise movement of freight, including
the stevedoring necessary to load and unload such freight. The month-to-month
changes in this volume of income were based on the movement of freight by rail.
The combined index is then adjusted to the annual totals for each year. In
general, the index works fairly well for the period 1929-35 but has a downward
bias in recent years. The divergence is probably to be explained by the growth
of coastwise waterway traffic relative to railway traffic over the later period.
The Research Division of the United States Maritime Commission is now
compiling information relating to employment and pay rolls in water-borne traffic.
Therefore, reliable indexes will soon be available for these industries on a current
basis, but for historical purposes it is still necessary to rely on traffic statistics.
B A N K IN G

Bureau of Labor Statistics indexes are available for the («jriod 1932-35. In
1936, the Bureau of Labor Statistics discontinued this series on the advice of the
Central Statistical Board to the effect that fluctuations in employment and
pay rolls were too small to justify the collection of current information. In
addition, the Federal Reserve Board compiles data on a semiannual basis for all
member banks. For the years 1932 to 1935, inclusive, the Bureau of Labor
Statistics series was adjusted to the annual estimates for all banks exclusive of
building and loan associations (included with miscellaneous in monthly estimates)
and brokerage (estimated separately). For the years 1929 to 1931, inclusive,
the Federal Reserve figures were used. The Reserve figures showed a consistent
seasonal rise during the latter half of the year, which, by comparison with pay-roll
estimates for the Pittsburgh banks, was concluded to represent bonuses and
payments for overtime during the month of December. Therefore, a smooth
curve was fitted to the Reserve data for the first half of the years 1928-32, inclu­
sive, to get the trend exclusive of such “bonuses.” The monthly movement in
salaries was calculated from this curve, and the computed deficiencies for the
latter half of each year were added to the calculated December figure. These
monthly data for member banks were then adjusted to the annual estimates for
all banks. For current information, it is necessary to rely upon the Reserve
Board’s semiannual reports and upon pay-roll indexes published by the Massa­
chusetts Department of Labor and Industries and by the University of Pitts­
burgh (this index covers banks in western Pennsylvania). None of these sources
in banking salaries have been very mild in the years following the liquidation
of the early 30’s.
BROKERAGE

Month-to-month changes in pay rolls in the brokerage business have been
published by the Bureau of Labor Statistics since January 1932. For the years
1929 to 1931, monthly pay rolls were estimated by the graphical comparison of
pay rolls for the period 1932-39, with the volume of stocks traded each month on
thé Now' York Stock Kxchange. The general relationship noted for these years
is projected for the earlier years.
IN S U R A N C E

The Bureau of Labor Statistics monthly pay-roll index is available since
January 1932. For the years 1932 to 1938 this index was, therefore, adjusted to
the annual estimates for these years. For the years 1929 to 1931, a smooth
curve was fitted to the annual data for 1929-32. The monthly series thus
obtained was linked to the January 1932 figure for the adjusted Bureau of Labor
Statistics index and then converted backwards to the annual estimates.

MONTHLY INCOME PAYMENTS, 19 29-40

35

R EA L ESTA TE

The monthly income of employees in real-estate enterprises is estimated on the
basis of a monthly index of real-estate activity published in the Real Estate
Analyst. This procedure is justified by the circumstance that pay rolls in this
industry represent largely commissions which depend primarily on the volume of
real-estate activity. At any rate, the monthly pay-roll estimates may be ex­
pected to predict, with a fair degree of accuracy, changes in the annual aggregate,
since this is also based largely on the series appearing in the Real Estate Analyst.
L A U N D R I E S A N D D Y E IN G A N D C L E A N I N G

E S T A B L IS H M E N T S

The Bureau of Labor Statistics’ indexes for these two industries exclude clerical
and technical employees: wherefore, salaried employees in these industries are
included with “miscellaneous” employees. The coverage is 27 percent for
dyeing and cleaning and 30 percent for laundries. For the years 1931 to date,
the monthly indexes are adjusted annually to our estimates. For the years
1929 to 1930, the estimates are based on the employment estimates of the Com­
mittee for Economic Security.
Y E A R -R O U N D H O T E L S

W'AGES AND SU BSISTEN CE

The Bureau of Labor Statistics’ index covers 50 percent of the industry, includ­
ing both salaried and wage personnel. The index is based upon cash payments
only. It has been corrected to the 1933 Census, as have the annual estimates,
but, owing to the inclusion of subsistence in the latter, the percentage change
from 1929 to 1933 is not the same. However, the discrepancy was so small that
the month-to-month fluctuation in the Bureau of Labor Statistics’ index was
simply applied to the more inclusive dollar figures.
G RA TU ITIES

In the annual income estimates gratuities are estimated as a percentage of the
hotel expenditures of guests. In computing the monthly income of employees
from this source, the Horwath & Horwatli figures covering room rentals and
hotel restaurant sales were weighted according to the relative importance of each
as shown by the Census of 1935, and the resulting composite index adjusted to
the annual estimates of gratuities received by hotel employees.
SEA SON AL H O TELS

The Censuses of 1933 and 1935 yielded an index of seasonal variation of employ­
ment in each of these years. It was assumed that the seasonal variations in pay
rolls followed the same pattern, and the seasonal factors for 1933 were applied
to the annual wages for 1932 and 1933, the 1935 seasonal to all other years on the
assumption that the 1933 seasonal was typical of the worst years of the depres­
sion, the 1935 seasonal of relatively better years. The cyclical changes in this
series are estimated currently on the basis of pay rolls of year-round hotels and
the trend in certain luxury expenditures: for example, Pullman traffic.
R E S T A U R A N T S A N D D R IN K IN G P L A C E S

PAY ROLLS

Several States are now collecting and publishing pay-roll indexes for restaurants.
Massachusetts has been publishing such figures since September 1931; Kansas
since August 1932; New York State, November 1932; Maryland, January 1934;
Illinois, July 1935; Wisconsin, August 1937, and Minnesota since 1938. These
State indexes were weighted according to their importance in the Censuses of
1933 and 1935 and a chain index was constructed to cover the period 1932 to
date. For most of 1932 the index reflects merely changes in the Massachusetts
index, but becomes more comprehensive and more reliable as new States have
been added. The 7 States now represented in the index accounted for more than
40 percent of the salaries and wages in restaurants in 1935, so that, assuming 50
percent coverage, within each of the reporting States, the index may be assumed
to be based on reports from 20 percent of the industry. None of the State indexes

36

MONTHLY INCOME PAYMENTS, 1929-40

is believed to cover drinking places, so these are estimated separately for the
period 1933 to 1935. With 1935 as a base the index is adjusted to the annual
dollar pay rolls back to 1932. Estimated pay rolls in drinking places (cf. below)
are added for the years 1933 to 1935. For the period 1929 to 1931 a smooth
curve was drawn through the annual estimates and seasonally adjusted figures
read from the curve. The seasonal factors calculated from the index for the
period 1933 to 1938 were applied to this curve. When drinking places had been
added to restaurants for 1935, the restaurant index was linked at December of
that year and adjusted to the annual data for restaurants and drinking places for
the years 1935 to 1938, on the assumption that after 1935 restaurant pay rolls
reflected also pay rolls in drinking places. The separate treatment of drinking
places between 1933 and 1935 is necessary because of the unusual changes following
the repeal of the prohibition amendment.
Since the “drinking places’’ industry grew very rapidly after the repeal of pro­
hibition in 1933, it was considered desirable for purposes of historical accuracy to
estimate pay rolls separately for the years 1933 to 1935. Monthly fluctuations
in employment are given in the Censuses of 1933 and 1935, and are employed in
estimating monthly labor income for those years. The census figures, however,
are of no value as indicating the cyclical or seasonal character of fluctuations in
employment, inasmuch as the dominant influence during these years was the suc­
cessive legalization of liquor in the various States. In each year there was a
marked rise in employment during the months April to July and a slow' rise there­
after, presumably the effect of legislative activity during the early months of the
year. For the year 1934, the same characteristic rise was assumed; for later
years it is assumed that this influence has diminished in importance and that any
further changes have been largely of a cyclical character. Hence the combination
of this item with restaurants after 1935.
G RA TU ITIES

In the annual estimates these are taken to represent 10 percent of restaurant
sales. Therefore it is only necessary to construct an index of restaurant sales.
For this purpose the Marketing Research Division’s index of restaurant sales
and the chain-restaurant sales carried in the Survey of Current Business until
recently (Childs and Waldorf still available though not published) were used.
A composite index was built up by giving chains and independents equal weight—
this only for the period beginning January 1937, because the Marketing Research
Division’s information is not available before that date. A seasonal w'as then
computed for this period and applied to data read from a smooth curve drawn
through annual data for restaurants only for the period 1929-35. For the years
1933 to 1935 it was necessary to add the sales of drinking places (cf. pay rolls).
The growth in pay rolls of drinking places over this period was assumed to be
similar to the growth in pay rolls. For the period 1935 to date the same pro­
cedure was adopted as in the case of pay rolls, but, since the index was not availa­
ble for 1936, recourse was had to a smooth curve drawn through annual data for
restaurants and drinking places. The seasonal factors computed above were
also applied before adjustment to the annual data.
INCOM E IN K IN D (MEALS)

Pay rolls are used as a trend and adjusted to the annual estimates of income in
kind. The adjustment is a rough one; the ratio of actual annual estimates to
annual figures as given by the pay-roll trend is calculated for each year, and this
ratio is assumed to vary continuously w-ithin each year. The ratio for each
month as determined by inspection is multiplied by the trend as calculated above.
D O M E S T IC S E R V I C E

As in the case of the annual estimates of the number of domestic servants, the
monthly fluctuations in domestic-service employment are those estimated by the
Committee for Economic Security. Monthly rates of pay are then calculated by
drawing a smooth freehand curve through the average monthly W'age for each
year. The product of these two series yields the monthly income of domestic
servants. In the original compilation, the estimates were carried forward on the
basis of employment alone, but the resulting annual figure for 1937 was so far
below the actual total that it was considered desirable to extrapolate wage rates

MONTHLY INCOME PAYMENTS, 1929-40

37

also, and this is now being done. Employment estimates are a composite of
estimates of employment in manufacturing, trade, etc.
P R IV A T E E D U C A T IO N

Estimates in the field of private education are largely a matter of determining
the monthly distribution of the annual salaries and estimating very roughly, on
the basis of the general movement of total income payments, the percentage
change of the current year over the preceding year. Year-to-year variations are
slight. The method of determining the monthly distribution is a variation of that
used for public education (q. v.).
O T H E R IN D U S T R IE S

Included in this catch-all group are the following items: building and loan
associations and miscellaneous industries of the finance group; the bulk of the
industries classified as service industries in the annual estimates, including the
employees of independent professional practitioners; and finally, the industries
grouped under “miscellaneous” in the annual estimates. These combined indus­
tries accounted for $6,145,000,000 in 1929, or 12 percent of the total compensa­
tion of employees in that year. Inasmuch as the larger part of these industries
are engaged in the production of services the demand for which does not vary
significantly from one month of the year to another, it was assumed that there
were no seasonal variations in employment and pay rolls in this group of industries.
By experiment, it was found that the pay-roll series for the distributive industries
gave a fairly good fit to the annual estimates of this miscellaneous group, and
therefore the seasonally adjusted index of labor income in the distributive group
was applied to the 1929 total for these industries. The adjustments neces­
sary to bring the annual totals thus obtained into line with the estimated annual
totals are indicated below: 1930, +0.6; 1931, +1.5; 1932, +2.8; 1933, +1.2;
1934, +0.5; 1935, -1.3; 1936, -0.2; 1937, +0.6; 1938, +0.5.
It is evident that these industries are subject to somewhat smaller cyclical
fluctuations than is the distributive group, and allowance must lie made for this
difference in periods of rapid changes in activity.
g o v er n m en t

:

g en er a l

c o n s id e r a t io n s

The total salaries and wages of governmental units as included in this series
will not be found to be identical with that presented in the annual income esti­
mates, for two reasons: (1) Retirement contributions of governmental em­
ployees have been deducted in the monthly series; and (2) certain types of com­
pensation representing a reward for part-time services have been classified in the
annual estimates with other labor income in order that salaries and wages may be
comparable with the number of employees. Since the number of employed per­
sons is not estimated in connection with the monthly series, such types of income
have for convenience been included with salaries and wages.
FED ERA L GOVERNM ENT

CIV IL EX EC U TIV E O TH ER THAN POSTAL SERV ICE

Payroll statistics are available only since November 1933. Prior to that date,
it is necessary to rely upon midmonth employment figures multiplied by monthly
estimates of average pay. These monthly estimates of average pay are con­
structed by drawing a smooth curve through the fiscal year averages with cor­
rection for the legislative changes in the rates of pay in July 1932 and April 1933.
The monthly pay-roll estimates were then adjusted to the fiscal-year totals, but
inasmuch as the necessary adjustments were small, an arbitrary smoothing method
was employed.
POSTAL SERV ICE

The same method was used except that it was necessary to make some provi­
sion for the large sums paid out to temporary employees during the Christinas
season, inasmuch as these employees would not be included in the midmonth
employment totals. As was found to be the case in recent years, these additional
payments were assumed to be closely related to the expansion in department
store sales during the month. The monthly estimates so obtained were adjusted

38

MONTHLY INCOME PAYMENTS, 1929-40

to the fiscal-year totals by an arbitrary smoothing method. The necessary
adjustments were small in every case.
M ILITARY, L EG ISLA TIV E, AND JU D IC IA L

Subsequent to November 1933, payments to employees in these branches of
the Federal Government are reported monthly to the Bureau of Labor Statistics.
Prior to that date a smooth curve was drawn through the fiscal-year averages.
A D JU S T M E N T S F O R R E T IR E M E N T S Y S T E M S

The annual congressional appropriations on account of the Civil Service Retire­
ment which are included in national income have been dropped from this series
Similarly, employee contributions to the retirement fund have been deducted
from the salaries of executive employees. Employee contributions to the fund
are available only on a fiscal-year basis, so that it is necessary to calculate such
payments as a proportion of salaries for each year and apply this proportion to
the individual months. On a current basis, deductions are estimated by applying
the average for the latest fiscal year available to salaries paid out during the
current month.
Similarly, the contributions of States and local governmental units to their
own retirement funds have been excluded from the monthly series, and the
employee contribution deducted from salaries, and wages. Since contributions
are available only on an annual basis, a procedure similar to that used for Federal
employees is adopted.
STATE AND LOCAL G O VERNM ENTS
PU BLIC EDUCATION

As there is no current measure of cyclical changes in teacher’s salaries, the chief
problem here was to allocate the annual figures on a monthly basis. For this
purpose, the source used was the March 1932 Research Bulletin of the National
Education Association. Table 11 in that bulletin gives, for cities with popula­
tion greater than 2,500, the number of months over which teachers’ salary pay­
ments extend by size of cities. The cities in each size group were then weighted
by their pay roll for the year 1931-32 (cf. p. 35 of Statistics of City School Systems
for 1931-32, U. S. Office of Education), and there was thus obtained the percentage
of annual teachers’ salaries that extended over 8, 9, 10, 11, and 12 months. Where
salaries were paid during 11 months, it was assumed that August was the month
during which no payments were made; for 10 months, July and August; for
9 months, July to September;3 for 8 months, June to September. On these
assumptions a monthly distribution of all teachers’ salaries was computed which
indicated a constant rate for October to May, with a reduced rate for June and
July and very low rates for August and September. When corrected for seasonal
variation, this series would show an increase or decrease in October of each year
and a constant rate over the following 11 months. This annual change is de­
termined on the basis of the rough estimates of informed workers in this field
The method assumes that salary changes are effected only at the beginning of
the school year While this may be the general practice, it was certainly not
maintained during the depression years, when many local governments were forced
to retrench whenever possible. On the other hand, such salary cuts often took
the form of a more or less voluntary return of part of the teachers’ salary and in
such cases the salaries reported by local governments might not necessarily be
reduced. Therefore, no attempt was made to account for changes in personnel
or salaries within the school year. In recent years the year-to-year changes in
the volume of teachers’ salaries have been small.
STATE AND LOCAL CONSTRUCTION

The monthly estimate of salaries and wages on public construction are based on
a confidential estimate of monthly employment by the Bureau of Labor Statistics.
To these employment estimates is applied the average compensation for this type
of work computed from employment and pay-roll statistics that the Bureau of

* When1teachers are paid on a monthly basis, at least some of these payments might be made at the end of
June. It was assumed, however, that even in such cases the payment was more reasonably allocated to
the following month during which, presumably, it was first available for expenditure.

MONTHLY INCOME PAYMENTS, 1929-40

39

Labor Statistics has been collecting from a small number of State and local
governments during recent years. As the average compensation showed little
variation over these years, the same average was used for all the years 1929-39.
O TH ER STATE AND LOCAL GOVERNM ENT PAY

By the method outlined above, annual estimates were obtained of the com­
pensation of public employees on construction projects, which constitute the most
variable of the functions performed by local governments. These annual estimates
are then subtracted from the annual estimates of total compensation of State and
local government employees, exclusive of public education, and a smooth curve
drawn through the annual averages. The result is a much smoother curve than
could have been drawn if construction were not deducted, and there can be less
hesitation in extrapolating the curve in the current period.
Since January 1940 the Bureau of the Census has been collecting quarterly
information on the employment and pay rolls of State governments and munici­
palities. It is hoped that this material will be very helpful in improving the
estimates of the current income of governmental employees. Similar material
for the years 1929 to 1939 will soon be made available by the Bureau of Labor
Statistics.
W O R K R E L IE F

This type of income since 1933 includes relief and nonrelief earnings on work
relief and administrative projects under the Federal Emergency Relief Adminis­
tration; earnings on the Civil Works program of the winter months of 1933 34;
and the income of enrolled personnel in the Civilian Conservation Corps (the
salaries of nonenrolled personnel being included for the most part in the pay rolls
of other governmental agencies) and the income of employees on Work Projects.
Monthly data have been supplied by the Work Projects Administration and,
inasmuch as these monthly figures are aggregated to give the totals appearing in
the national income estimates, no adjustment is necessary. The present series
also includes an estimate of work-relief wages for the years 1930 to 1932. This
type of income became of increasing importance during 1932, and therefore, in
order to prevent a sharp discontinuity in the work-relief and employees’ income
series in January 1933, work-relief wages have been roughly estimated for the
years 1930-32. The resulting estimates for work and direct relief are less reliable
than that for their aggregate, but any resulting bias is not carried into the totalincome-payments series. For the method used to arrive at this estimate, see
under “Direct relief.”
ENTREPRENEURIAL INCOME AND NET RENTS AND ROYALTIES
N E T IN C O M E O F F A R M O P E R A T O R S

This item represents the most important component of entrepreneurial income—
both because of its magnitude, averaging as it does one-third of total entrepren­
eurial income as here defined, and because of its variability. Thus, in 1932
agricultural net income was 70 percent lower than in 1929, whereas all other
entrepreneurial incomes showed in the aggregate a decline of only 30 percent for
the same period. The net income of farm operators has in recent years been
composed of two elements: (1) the net income arising from the production and sale
of farm products, and (2) Government benefit payments. The former item is
computed by deduction of estimated costs from gross income. The latter item
is carried in its entirety to net income.
For estimating agricultural net income on a monthly basis, the Agriculture
Department’s estimates of cash income from farm marketings is used. The
Department of Agriculture’s unadjusted dollar series is then adjusted to the
annual estimates (exclusive of Government benefits) of the net income of farm
operators, which include the value of farm products consumed on the farm. The
net-income figures are naturally more variable from a cyclical standpoint than
are the gross-income figures, cost items being relatively more stable. Conse­
quently, in projecting this series into the current period, account is taken of this
tendency in accordance with the observed relative movements in gross and net
income for similar points in the cyclical trends of the past 10 years. To the
monthly net-income figure as thus obtained is added the governmental benefits
received by farmers.

40

MONTHLY INCOME PAYMENTS, 1929-40

In correcting farm income for seasonal fluctuations, a single set of factors is
computed for the period 1929 37. Farm benefits are not adjusted for seasonal
influences. The seasonal factors thus computed are not entirely satisfactory,
and even after correction the agricultural-income figures show marked month-tomonth fluctuations which are communicated in reduced measure to the index of
total income payments.
The index of nonagricultural income, therefore, serves the important function
of abstracting these random fluctuations and of presenting the nonseasonal
changes in the more stable forms of income. In computing nonagricultural in­
come, the items deducted are: Agricultural net income (as defined above),
agricultural wages, and interest and net rents arising from farm property. It
should be noted that a substantial volume of these latter two types of income are
received by persons not living on farms, particularly those receiving mortgage
interest. On the other hand, farm incomes are supplemented by part-time earn­
ings in industry, by an unknown volume of nonfarm property income, and by a
small amount of direct relief, including emergency subsistence payments. Con­
sequently, the nonagricultural-income series represents income payments arising
from nonagricultural activities, rather than income received by the nonfarm
population.
The difference between total income payments and nonagricultural income
payments will not, however, be found to be identical with income produced as
incorporated in the annual income estimates. This divergence is accounted for
by the circumstance that agricultural net rents are in the annual aggregates
included in the finance group along with net rents and royalties on nonfarm
property.
O T H E R E N T R E P R E N E U R I A L IN C O M E

As noted in the body of this report, this item represents largely withdrawals
rather than income in the accepted sense—save in the important case of independ­
ent professional men, for whom, because of the small scale of capital involved, it
is impractical to distinguish net income from personal withdrawals even on an
annual basis. Information relative to the month-to-month fluctuations in this
type of income is entirely lacking on either a historical or a current basis. There­
fore it is necessary to assume that withdrawals are not subject to seasonal influ­
ences and, in projecting the estimates, to base such projections on the study of
past performance with special reference to the seasonally corrected indexes of
salaries and wages in the distributive and service industries. The resulting
series is relatively insensitive to fluctuations in economic activity. If this result
fails to accord with the common conception of the entrepreneur as the bearer of
risk, it is none the less reasonable to assume that, for entrepreneurs as for other
persons, expenditures can be expected to be substantially more stable than income.
And since the personal expenditures of entrepreneurs determine largely his busi­
ness withdrawals, these latter may also be expected to show a certain degree of
stability. At any rate, the monthly series has so far predicted with fair accuracy
the changes in entrepreneurial withdrawals as entered in the annual income
estimates.
N E T R E N T S A N D R O Y A L T IE S

Rents and royalties, as included in the national-income estimates as well as
in the present series, represent the net return from the ownership and operation
of rented property, after the deduction of various cost items. Therefore this
item is more properly classified with entrepreneurial income rather than with
property incomes (dividends and interest) in the narrower sense of that term.
An additional reason for combining this item with entrepreneurial income lies in
the paucity of data available for both items; in contrast it is felt that the trend in
dividends and interest can be determined with a fair degree of accuracy.
As in the case of “other entrepreneurial income” the treatment of this item is
limited to an attempt to determine trend and cyclical influences from the annual
data. In projecting the annual data into the current period some assistance is
rendered by the Bureau of Labor Statistics index of housing costs and by the
housing component of the National Industrial Conference Board’s index of living
costs. Both indexes must, how'ever, be considered of very limited value for this
purpose- first, because residential rents account for only 50 percent of all net
rents and royalties; secondly, because both indexes can at best be taken as re­
flecting gross income from residential rents, whereas what is desired is some meas­
ure of net income from such sources. Net rents and royalties showTmuch wider

MONTHLY INCOME PAYMENTS, 1929-40

41

fluctuations than other entrepreneurial income as included in the present series.
The principal reason for this is that rents and royalties include an element of
business savings. However, because of the small size of the rents item, as well
as the great difficulties involved in estimating withdrawals for those persons
receiving such incomes, no attempt lias been made to eliminate this element
from the monthly income estimates.
DIVIDENDS AND INTEREST

Dividends are estimated largely on the basis of the series of dividend payments
compiled by the New York Journal of Commerce, supplemented by special studies
for critical periods (as when the Journal of Commerce series was radically altered
in 1934). The Journal of Commerce series was selected because it comprises
actual payments to shareholders during the current month, rather than dividend
declarations as measured by the New York Times series, and is forthatreason more
comparable, in respect to timing, with the estimates of labor income.
Interest payments on corporate and governmental indebtedness are also esti­
mated to a large extent, especially in the earlier years, on the Journal of Com­
merce series. For the years 1929 to 1935, the Journal of Commerce sories is
adjusted to the annual estimates of long-term interest for corporate and govern­
mental bodies. Despite the relatively stable nature of the series, the adjustments
for these years are surprisingly large. Beginning with 1930, interest payments
by the Federal Government are taken directly from the Daily Treasury State­
ments, and the trend of interest payments on other corporate debt is determined
on the basis of special studies of issues, retirements, and refunding operations in
the security markets. Seasonal fluctuations are still based upon the Journal of
Commerce series. For all years, interest payments on farm and urban home
mortgages and on the indebtedness of unincorporated enterprises are assumed to
have no seasonal fluctuations and are, therefore, estimated by drawing a smooth
curve through the annual estimates and extrapolating this curve forward on the
basis of fragmentary information on changes in the volume of mortgages and
interest rates on such mortgages.
It is a matter of some difficulty to correct dividend and interest payments for
seasonal variation, and so far none of the orthodox methods lias been found to be
applicable. The principal difficulty lies in the fact that, although there is a
definite seasonal pattern for dividends, the pattern is subject to rapid changes
which seem to be cumulative in nature. An index of seasonal variation for
dividends has been calculated by the ratio-to-moving-average method for three
periods, and these indexes are compared below:
Indexes of Seasonal Variation for Dividends
Month

1929-31 1932-34 1936-37
161
89
97
97
93
107

197
98
93
97
78
101

100 July.
44
76
101
39
130

Month

1929-31 1932-34 19:16-37
131
76
78
96
88
88

136
73
66
108
76
78

108
44
111
41
306

As might be concluded from a comparison of the seasonal factors for 1929-31
and 1932-34, it would have been possible to correct dividends for seasonal varia­
tions over the years 1929-35 by computing a seasonal factor every year from the
variations for the latest 3 years, and such a method does give a fairly smooth
curve, which, however, still exhibits random fluctuations such as are not observable
in seasonally corrected employees’ compensation. Hut at any rate, this method
would not be applicable to the later years during which the undistributed-profits
tax was in force. As a matter of fact, dividends could hardly have been said to
have a true seasonal pattern during this period. This tax, as already remarked,
imposed a penalty upon corporations not paying dividends in the years in which
they were earned. Therefore, insofar as a full year’s operation proved to be more
profitable than had been expected in the earlier months of the year, there would
tend to be large dividend payments in the final month of the year. Insofar as
operations for the year as a whole did not come up to earlier expectations, there
might well be a tendency for dividends to fall off at the year end. Thus, in 1936

42

MONTHLY INCOME PAYMENTS, 19 29-40

when business activity experienced a sharp rise during the last half of the year,
December dividends accounted for nearly 30 percent of all dividends disbursed
during the year. On the other hand, the closing months of the year 1937 were
marked by a sharp drop in industrial activity, and December dividends amounted
to only 22 percent of all dividends. The year 1938 was one of greatly reduced
business activity, and, therefore, dividends in the final month represented only 17
percent of the year’s total. The sharp rise in corporate incomes in the last quarter
of 1939 was again reflected in the fact that December dividends accounted for
fully 24 percent of the year’s total.
There was a well-defined seasonal pattern for interest payments during the
early years of the period under review, but with the inauguration of the Federal
Government’s refunding program, and the smaller-scale refunding operations of
public-utility corporations, this seasonal pattern has fluctuated widely in recent
years. Owing, therefore, to the difficulty of correcting dividends and interest for
seasonal variations in recent years by the orthodox methods, it was decided to
determine the nonseasonal movements for the whole period by merely computing
a 13-month moving average centered on the seventh month and smoothing
through this average at critical points. The results obtained by this method
are satisfactory in most respects, but it is necessary to project this index for­
ward to cover the latest 5 months. This projection is based largely on a com­
parison of movements in this average with movements in the adjusted index of
employees’ compensation for similar periods in the past. To the extent that this
comparison proves faulty, it is necessary to revise the adjusted index of total
income payments from time to time. While this procedure may occasion some
inconvenience to those who use the series, it is to some extent justified by the fact
already indicated above that the actual payments of dividends (and in some
periods, interest) do not reflect current business operations, as do, for example,
production indexes and pay-roll statistics, but refer rather to the earnings of some
preceding period.
The construction of a smooth trend for dividends and interest, even by the
use of rather arbitrary methods, is believed to be justified by the uses to which
the adjusted index may be put. In its primary use as an index of consumer
purchasing power, it seems desirable to minimize the effects of month-to-month
changes in property incomes because (1) so much of this type of income accrues
to trust accounts of various kinds and so is not immediately available for con­
sumption, and (2) of the remainder, it is likely that the expenditure based upon
this income is extended over a period of several months rather than being con­
centrated in the months in which the income is received (cf. similar treatment in
case of entrepreneurial income). Secondly, with respect to the use of the index
as an indicator of general economic activity, month-to-month fluctuations in
dividends are misleading here, as they should be largely offset by changes in
business savings. Thus, the high level of December dividend payments must be
considered to be largely offset by negative business savings, particularly in years
of depressed activity. Similarly the low rate of dividend disbursements in sur­
rounding months is accompanied by positive business savings—that is, by an
accumulation of net income to be paid out quarterly or semiannually, according
to the usual dividend-paying practice.
DIRECT AND OTHER RELIEF

Included in this type of income are “general relief,” “payments to recipients of
special types of public assistance” (old-age assistance, aid to dependent children,
and aid to the blind), and “emergency subsistence payments to farmers.” The
earnings of relief persons employed on work-relief projects under the F. E. R. A.
have been deducted from the general relief estimates published in the Social
Security Bulletin and included in the work-relief estimates; cf. table 3, “Statistical
Summary of Emergency Relief Activities, January 1933 through December 1935.”
Subsequent to January 1933 the relief estimate are those of the Social Security
Board and the Works Progress Administration. Neither of these agencies has
released relief estimate for earlier years, and therefore these were constructed in
the following way.
Of primary importance were the estimate appearing in appendix B, p. 69, of
Trends in Different Types of Public and Private Relief in Urban Areas, 1929-1935,
Publication No. 237 of the Children’s Bureau of the Department of Labor. In
this appendix are given estimate of work relief, direct relief, and categorical
assistance by months for the period 1929 to 1935, for 120 urban areas. For each

MONTHLY INCOME PAYMENTS, 1929-40

43

of these three types of relief there was constructed an index for the years before
1933 by linking the Children’s Bureau estimates to the Works Progress Adminis­
tration estimates for the first 4 months of 1933. (Because of the unsatisfactory
nature of the estimates for the early months of 1933 this was considered a more
conservative procedure than that of linking the indexes at January 1933.) By
adding the monthly estimates for the three types of relief, a first estimate for all
relief from 1929 to 1932 was obtained, but it was still necessary to correct this
estimate for the varying trend of urban and rural relief. Tlie only estimate
available is one for 1932 (Trends in Relief Expenditures, 1910-1935, by Anne E.
Geddes, p. 44). From this source, public relief was estimated at $447,000,000 for
1932, whereas the estimate based solely on the information for 120 urban areas
was $531,000,000, indicating that the upward trend in rural relief from 1932 to
1933 was more marked than that indicated for urban relief by the Children’s
Bureau series. On the assumption of a similar differential trend for 1929 to 1931,
the annual estimates for this period were also revised downward and the monthly
estimates adjusted to these annual estimates by the Bassie method.
Work-relief payments were estimated roughly on the basis of the proportions of
work-relief expenditures to all other relief expenditures as estimated by the
Children’s Bureau for urban areas for the years prior to 1933. There is no esti­
mate of total work-relief wages comparable to the Works Progress Administration
estimate of total disbursements for 1932. Therefore the estimates of work-relief
wages and direct relief must be regarded as less reliable than their estimated total
for the years before 1933.
OTHER LABOR INCOME

The characteristic common to the income included here has been noted in the
text of this report-—in general, the disbursements included represent compensa­
tion for past services. The most important components, their relative importance,
and the method of estimation are noted separately below. In general they repre­
sent the results of official tabulations, or of arbitrary smoothing of annual data.
P E N S IO N S T O V E T E R A N S

Pensions to veterans constitute the largest single item for the period as a whole,
and for recent years have approximated 35 million dollars a month. The monthly
entries for the whole period are furnished by the United States Veterans’
Administration.
A D JU S T E D -S E R V IC E B E N E F IT S TO W O R L D W A R V E T E R A N S

This item is now of diminishing importance. Estimates for the years 1931 to
1937 were constructed on the basis of data supplied by the United States Veterans’
Administration and the United States Treasury Department. Current figures are
taken from the month-end statements of the Treasury Department.
Adjusted-service benefits constituted an important addition to income in 1931
and 1936, and for these years produce severe breaks in the otherwise fairly smooth
curve of total income payments. For some purposes if would seem advisable to
smooth over these discontinuities, by an arbitrary process if necessary. If the
income index is considered as an index of economic activity, such benefits obviously
have no place in the index; even where interest attaches to the primary or purchas­
ing power aspect of the index, a comparison with department-store sales, or other
available indexes of retail sales, indicates that the bonus payments of the summer
of 1936 had little effect on retail activity in those months. Indications are that
the expenditure of the $1,200,000,000 of benefits disbursed during June and July
1936 was distributed over a period of several months preceding and following their
actual disbursement. For those who may be interested in making some such

M ONTHLY

44

INCOME PAYMENTS, 19 29-40

allowances, or who for other reasons may wish to eliminate adjusted-service pay­
ments from the income series, they are shown below:
Adjusted Service Benefit Payments
[In millions of dollars]

Month

1931

February................................
March...................................
June.......................................
July.........................................
August...................................
September..... .......................
October ___ _________
November_____________
December..............................

5
307
352
92
39
25
19
17
17
13
21

1932
19
12
10
10
9
9
21
26
11
8
6
6

1933
6
5
7
6
5
5
4
4
3
3
3
3

1934
3
2
3
2
2
2
2
2
2
2
2
2

1935
2
2
2
2
2
2
2
1
1
1
1
1

1936
1
1
1
800
413
77
48
35
25
26

1937

1938

18
13
13
10
7
18
12
9
8
8
6
6

6
5
6
5
4
8
5
5
4
4
3
3

1939
3
2
3
3
2
5
3
3
3
3
2
2

O T H E R F E D E R A L D IS B U R S E M E N T S

Included here are all annuities to former civil-service employees as well as refunds
to persons withdrawing from the retirement system. Also, as regards retirement
benefits to military personnel, monthly data are from official sources. In the
aggregate these items accounted for roughly $85,000,000 of income in 1939.
S O C IA L S E C U R IT Y B E N E F IT S

Social Security benefits have been of increasing importance since 1937, when
they totaled only $40,000,000. Unemployment-insurance benefits now consti­
tute the most important single component of “other labor income,” aggregating
$435,000,000 in 1939. Railroad retirement benefits accounted foranother $100,000,000 in 1939. Disbursements under the old-age-insurance provisions of the Social
Security Act amounted to only $12,000,000 in 1939, but, consequent upon the
recent liberalization of the provisions of this act, such disbursements are expected
to grow in importance during 1940.
AH monthly data are from official sources.
w o r k m e n ’s

C O M P E N S A T IO N , P R IV A T E R E T IR E M E N T B E N E F IT S , E T C .

All other labor income includes workmen’s compensation, benefits .under retire­
ment systems sponsored by private industry or non-Federal governments, and a
small volume of Federal Government pensions for which no monthly data are
available. Because of the lack of monthly data, month-to-month changes are
estimated by graphical interpolation of the annual data. In the aggregate, these
items show little change from year to year.

o