View PDF

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

Supplement
to the
Survey

of
Current
business




UNITED

STATES

DEPARTMENT

OFFICE

OF

OF BUSINESS

COMMERCE
ECONOMICS

. ' • .

-,f

I

,

••:

- >

"

*

'

.1

t

t".'*

-A... I

A
Supplement
to the
Survey
Current
business

^ *^* *

P R E P A R E D

B Y T H E N A T I O N A L
GEORGE
CHARLES

U .

S .

D E P A R T M E N T

F.

SCHWARTZ,

O F

SINCLAIR

O F F I C E

O F B U S I N E S S
M.

V

.

S

. G O V E R N M E N T




JOSEPH

P R I N T I N G

JASZI,

I N C O M E

D I V I S I O N

CHIEF

ASSISTANT

CHIEF

C O M M E R C E
WEEKS,

Secretary

E C O N O M I C S
ME EH AN,
O F F I C E

.

Director
W A S H I N G T O N :

1 9 5 4

Contents
PAGE

PART I.

National Income and Product, 1929-53: A Review

Basic Trends in the Economy
Growth of the Economy
Shifts in the Use of National Output
Changes in the Income Flow
Expansion of Personal Income
Fluctuations in Economic Activity
Gross National Product Patterns
Chronological Review: 1929-53
PART II.

The Conceptual Framework of National Income Statistics

Nature and Significance of the Income Data
Summary Construction of National Output Measures
Detailed Structure of the Accounts
Definitions of Concepts and Terms
PART III.

Sources and Methods of National Income Estimation

1

2
2
4
7
11
13
13
15
27

27
30
40
58
61

Introduction
61
Section:
1. Wages and Salaries
68
2. Contributions for Social Insurance and Other Labor Income. . 73
3. Income of Unincorporated Enterprises
76
4. Rental Income of Persons
86
5. Corporate Profits
92
6. Interest
97
7. Personal Consumption Expenditures for Commodities
103
8. Personal Consumption Expenditures for Services
117
9. New Construction
122
10. Producers' Durable Equipment
126
11. Change in Business Inventories
135
12. Net Foreign Investment
139
13. Government Receipts and Expenditures
143
14. Transfer Payments
149
15. Personal Saving
150
16. Capital Consumption Allowances
150
PART IV.

Gross National Product in Constant Dollars, 1929-53

153

Characteristics of Constant-Dollar Gross National Product
Statistical Sources and Methods

153
155

PART V .

Statistical Section

List of Annual, Quarterly, and Monthly Tables
INDEX




159

159
243
HI

Foreword
This volume carries forward the official work on United States national income initiated in 1932 in
response to Senate Resolution No. 220 of the 72d Congress. The information presented is designed to
meet in comprehensive fashion the needs of business and other users of national income data.
Since publication in 1934 of the first of a series of national income reports by the Department of
Commerce, steady progress has been achieved in extending the scope of the estimates, in improving
their quality, and in making them available promptly, as well as in sharpening the concepts. A principal contribution of the present report—which is closely similar in form to the 1951 NATIONAL INCOME
supplement so as to facilitate use by those familiar with that volume—is the presentation of estimates
incorporating data collected in the postwar industrial and population censuses.
In the preparation of these new estimates, opportunity was also taken to rework many of the income
and product series for the entire period back to 1929 in order to reflect additional data sources and
improvements in estimating techniques. A special feature is the presentation of constant-dollar gross
national product in 1947 prices instead of 1939 prices, as previously used.
The tables presented in this volume incorporate the results of the first comprehensive review of
sources and methods since the initial publication of the national income statistics in the form of an
economic accounting system in the 1947 NATIONAL INCOME supplement. While the changes that have
been introduced do not alter the overall picture of the United States economy afforded by the income
and product accounts, they improve the data in many detailed aspects.
The text material in the 1951 volume also has undergone review. This resulted principally in reworking the descriptions of statistical sources and methods to accord with the new estimates and bringing
up to date the summary of economic developments.
The statistical changes have been examined for the light they shed on the reliability of the estimating techniques. This analysis confirms the adequacy of these techniques to produce reliable preliminary measures of national output and its major components on the basis of incomplete information.
Revisions for some of the more detailed components, however, were substantial and underscore the
need which we have repeatedly stressed for further development of the primary data sources on which
the national income estimates are based.
The present report contains all the national income statistics of the Office of Business Economics
except the annual series on income by States and the distributions of family income by size classes.
With these exceptions it supersedes all previously ptublished figures, and the series contained in this
volume will be kept up to date in the monthly SURVEY OF CURRENT BUSINESS.
We take this opportunity to express our appreciation for the continued cooperation of the many
Government and private agencies that assist the Office of Business Economics in preparing these
economic guides. Acknowledgment is made in the accompanying statement which also recognizes the
work of the individual members of our staff.




Director, Office of Business Economics

Acknowledgments
The present edition of the National Income supplement was prepared under the joint direction of George Jaszi, Chief of the
National Income Division, and Charles F. Schwartz, the Assistant Chief.
In charge of major areas of the statistical work were Lawrence
Grose, Raymond Nassimbene, and Harlow D. Osborne.
Special acknowledgement is due to Edward O. Bassett
for his contribution to methodology in the complex task of
incoporating the data collected in the postwar industrial censuses
into the commodity flow estimates.
This volume builds upon the foundations laid by its predecessors—the 1951 and 1947 editions of the National Income supplement. Work on the 1951 edition was initiated by Milton
Gilbert, former Chief of the National Income Division, and carried
through under the direction of Mr. Jaszi and Mr. Schwartz.
The conceptual framework and statistical methodology underlying the estimates, which were explained in detail for the first
time in the 1951 volume, had been established initially in the
1947 report, under the direction of Mr. Gilbert. His principal
assistants in this major undertaking of fundamentally recasting
the official national income statistics were Edward F. Denison,
now Assistant Director of the Office of Business Economics, Mr.
Jaszi, and Mr. Schwartz.
Part I of this report, dealing with trends in national income
and product, was prepared initially by Carl P. Blackwell, formerly
of the National Income Division, and revised for the present
edition by George M. Cobren. The accompanying charts were
prepared under the direction of Edwin C. Warren, Chief Draftsman in the Printing Division of the Department of Commerce,
with the cooperation of Anna M. Guindon of the Office of Business
Economics.
Numerous staff members of the National Income Division participated in the initial writing of the various sections of the
technical notes in Part III and their revisions for the new edition.
Special credit is due to Harlow D. Osborne in connection with
the sections on income of unincorporated farm enterprises, rental
income of persons, corporate profits, new construction, net
foreign investment, and capital consumption allowances.
Acknowledgments to others are listed in the sequence in which
the technical notes appear in Part III. Wages and salaries—
Franklin M. Aaronson and Lawrence Grose; contributions for
social insurance and other labor income, and income of unincorporated enterprises—Lawrence Grose; interest—Elwyn T.
Bonnell; personal consumption expenditures for commodities—
Edward O. Bassett and Raymond Nassimbene; personal consumption expenditures for services—Carolyn G. Bernhard;
producers' durable equipment—Robert C. Wasson and Raymond




Nassimbene; change in business inventories—George M. Cobren;
government receipts and expenditures—Carl P. Blackwell;
transfer payments—Lawrence Grose; capital consumption allowances—Robert C. Wasson.
In addition, the Balance of Payments Division and the Business
Structure Division of the Office of Business Economics provided
materials relating to the descriptions of net foreign investment and
personal consumption expenditures for commodities, respectively.
In the initial preparation of the estimates of constant-dollar
gross national product described in Part IV, John W. Kendrick
was principal assistant to Mr. Jaszi. Major contributions were also
made by Edward O. Bassett, Carolyn G. Bernhard, Morris Cohen,
Joseph B. Epstein, and Millard L. Gallop.
The vast statistical work underlying the estimation of the
multiplicity of income and product series contained in this report
is the result of the cooperation of all the members of the National
Income Division and others in the Office of Business Economics,
and is founded on their effort and experience. However, in a
larger sense, the statistics rest upon the work of Government
statistical agencies as a whole and of private agencies as well.
These provide the basic source data and the considerable volume
of supplementary information needed to construct the national
income and product accounts. The statistical work of the Bureau
of the Census of the Department of Commerce, the Health, Education, and Welfare, Treasury, Agriculture, and Labor Departments, and the various regulatory commissions is of fundamental
importance in this regard.
Certain of the estimates themselves are prepared outside the
National Income Division: farm income by the Agricultural
Economics Division of the Department of Agriculture; direct
estimates of personal saving by the Securities and Exchange
Commission; new construction activity by the Building Materials
and Construction Division of the Department of Commerce, in
cooperation with the Bureau of Labor Statistics of the Department
of Labor; net foreign investment by the Balance of Payments
Division; and personal consumption expenditures for commodities
since 1940 by the Business Structure Division, except 1947 for
which benchmark estimates were prepared in the National
Income Division.
Progress in the national income field has been facilitated by the
Bureau of the Budget, not only by its direct support, but by its
continuing recognition and furtherance of the needs of the Office
of Business Economics' national income work in the development
of the Government's statistical program. Finally, it is recognized
that what has been achieved has been possible only by the support
and encouragement given by the Congress continuously since
these studies were initiated at the direction of the 72d Congress.

Government Purchases
$85 Billion

Investment
Personal Consumption

$50 Billion

$230 Billion

The 1953 Income-Product Coin




Indirect Business Taxes $30 Billion

Depreciation, etc,
$30 Billion

Employee

Corporate Earnings
$39 Billion

Compensation
$209Billion

Proprietors' and Rental
Income and Interest

$57 Billion

PART

I

National Income and
Product, 1929-53

AR eview
The national income statistics presented in this report cover a
quarter of a century of highly varied economic developments in
the United States.
In broad outline, this period embraces the following sequence of
events: The precipitous fall from the prosperity of 1929 into the
deep depression of the early nineteen-thirties; the subsequent
recovery, interrupted by the brief recession of 1938, but then
continuing through the remaining prewar years; the tremendous
performance of the economy in the prosecution of World War II;
reconversion and the postwar boom, with attendant inflationary
strains; the mild business recession of 1949; and the rapid recovery
of 1950, merging in the latter half of that year into the period of
hostilities in Korea.
The national income data provide a detailed statistical description of the way the economy has functioned under these widely
diverse conditions. They reveal important fluctuations and longterm changes in the volume, composition, and use of the Nation's
output, in the industrial structure through which it is produced,
and in the distribution of the resultant income.
The nature of national income statistics is fully described in
subsequent parts of this report. The following general summary
highlights their major features as a prelude to an analysis of the
functioning of the economy and of the path traversed in reaching
peak levels of income and production in 1953.

National product: The flow of goods and services
Total output is measured from two principal points of view:
As the summation of final products produced by the economy; and
as the summation of costs incurred in producing those products.
Both of these approaches yield comprehensive measures of economic activity, but the analytical breakdowns to which they most
readily lend themselves throw light on different aspects of the
economy.
The gross national product measures the Nation's output of
goods and services in terms of its market value. When expressed
in current prices, this series reflects the total dollar value of production; when expressed in constant dollars to eliminate the
influence of price changes, it provides an overall index of the



physical volume of goods and services produced by the economy.
In both current and constant prices, the gross national product
is broken down to show its disposition among broad groups of
users—consumers, business, government, and foreign countries.
The commodity and service composition of purchases by each of
these major groups is delineated in considerable detail for the
current dollar series and in summary fashion for the constant
dollar series.

National income: Earnings from production
Total output is also measured, in terms of the factor costs of
producing it, by the national income—the aggregate earnings of
labor and property which arise from current production. This
measure differs from the gross national product chiefly in that it
is computed after deduction of indirect business taxes and of
depreciation charges and other allowances for business consumption of durable capital goods.
The national income is broken down by distributive shares, by
industry of origin, and by legal form of organization. The first
of these breakdowns represents a classification of earnings according to the forms in which they accrue to residents of the Nation—
compensation of employees, profits of corporate and unincorporated enterprises, net interest, and rental income of persons.
The second indicates the use of economic resources and the contribution to total output by each of a number of industrial subdivisions, as measured by income originating in the respective
industries. The third shows an important special aspect of the
institutional structure of the economy—the portions of total economic activity (also measured by income originating) conducted
by various types of productive units, including corporations, sole
proprietorships and partnerships, other private business, government and government enterprises, and households and institutions.
In addition to a summary account showing the national income
and product, accounts are maintained for each of the major
sectors of the economy. These consist of current income and outlay
accounts for the business sector, for persons, for government, and
for the rest of the world in its transactions with the United States.

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS
A consolidated saving and investment account for the economy as
a whole is also provided. The system of accounts is designed to
furnish a description of the economic process in terms of the
expenditures and receipts of the various sectors, arranged to show
their interactions upon each other. The nature and classification
of the transactions recorded for each sector are governed to a
considerable extent, of course, by the central objective of measuring total output.

Personal income: Receipts of consumers
Particular interest centers upon the personal account, which
covers the activities of the consuming public. On the one hand,
it gives total personal income—the current income received by
persons from all sources, inclusive of transfers from the government and from business; and on the other, it indicates the disposition of personal income for consumption, taxes, and savings.
Personal income is a third major aggregate, generally coordinate
in significance for economic analysis with the national income and
the gross national product.
In general outline, then, this is the body of statistical information utilized in the the following review of economic developments since 1929. While this framework permits a comprehensive
study of such developments, it by no means covers all significant
elements in the operation of the economy. It does not, for example, provide direct information regarding the monetary and
credit system of the United States. In the discussion below, therefore, only incidental attention will be devoted to trends in that
field—not because they are lacking in importance, but because
they fall outside the scope of the statistics with which this review
is primarily concerned. Although a number of other relevant
factors are similarly excluded, the picture of the economy in
action which emerges from the national income accounts is a
highly significant one.

BASIC TRENDS IN THE ECONOMY
Despite the violence of the economic fluctuations which have
occurred, the outstanding feature of economic developments over
the past 25 years is the tremendous growth of the United States
economy.
The population increased by over 30 percent, from 122 million
in 1929 to 160 million in 1953, and the number of persons engaged in production rose in roughly similar proportion. This
larger work force was equipped with a greatly expanded volume
and improved quality of machinery and plant facilities, as well
as with such intangible assets as better education and advanced
scientific knowledge. Through utilization of these enlarged
human and material resources, the economy has been able to
produce a vastly increased flow of goods and services, including
a wide array of new products.

Growth of the Economy
The gross national product amounted to $365 billion in 1953,
as compared with $104 billion in 1929. This comparison reflects
the combined influence on the current market value of total
output of both greatly increased physical volume and much



Chiefly as a result of the inflation associated with World War II
and its aftermath, and the Korean conflict, the general level of
prices in 1953 was more than Iwo-thirds above that of 1929.
After allowance for this factor, the physical volume of the Nation's
output, as measured by the gross national product in constant
dollars, is found to have risen 105 percent over the period. In
terms of real output per capita, the increase amounted to 57
percent. See chart on page 5.
A simple and meaningful comparison of the long-term rate of
growth in national production is provided by the average annual
percentage increase in constant-dollar gross national product
from 1929 to 1953, which were both years of high utilization of
productive resources. According to this calculation, the rate of
expansion in the real volume of output has averaged over this
25-year period about 3 percent per year.
In part, this growth has reflected the gradual increase of the
Nation's manpower resources. The advance in production, however, has outstripped this increase by a wide margin, owing to
the achievement of sizable gains; in productivity per unit of
manpower utilized.

Large advance in productivity
Trends in productivity may best be examined in terms of gross
product, excluding that arising in general government because
the method by which the contribution of government to constantdollar national product is estimated makes no allowance for
changes in productivity.
With this exclusion, the real increase in output from 1929
to 1953 was almost 100 percent. During the same period the
number of persons producing this output—full-time equivalent
employees plus active proprietors—rose by 27 percent, or about 1
percent per year on the average. An annual rate of growth in
real product per person engaged averaging approximately 1%
percent is thus indicated.
Moreover, this rate of increase occurred during a span of years
in which average hours worked per week in the private economy
were reduced by about 10 percent. On a man-hour basis, accordingly, the rise in productivity has been considerably greater.
Real product per man-hour was well over half again as large
last year as in 1929, implying an average annual rate of increase
somewhat in excess of 2 percent.

Many factors influence productivity
It is important to recognize that productivity increases as
computed above, although expressed in terms of output per unit
of labor input, are attributable not only to labor, but jointly to
all of the factors influencing productivity. Foremost among these,
undoubtedly, have been the technological improvements and
increased amounts of capital equipment utilized by the Nation's
industries. Better organization and management of productive
operations have also contributed, as have advances in the education, training, and health of the population.
These types of influences affect directly the technical efficiency
of particular productive processes. In addition, the productivity
measures given above also register shifts within individual industries among products involving varying amounts of output per
unit of labor input and. furthermore shifr« nf wnrV»™ l™*,..,..™

National Output, Income, and Consumer Purchasing Power in 1953
Gross national product was 20% larger
than national income. It includes in addition
depreciation and indirect business taxes

Seven-eighths of the national income was
distributed as personal income

Personal income exceeded national income
distributed, as individuals also received
$19 billion in transfer payments

After taxes, $250 billion of personal income
was available for spending or saving

$30 Billion '

i"

--INDIRECT BUSINESS
-'TAXES

$30KBion

,

UNDISTRIBUTED ' v
EARNINGS
$38 fitl lion

TRANSFER PAYMENTS*
$19 Billion

I

PERSONAL TAXES
$36 Billion

NATIONAL ,
INCOME

DISTRIBUTED

DISTRIBUTED

PERSONAL.INCOME

EARNINGS

EARNINGS

AFTER TAXES

$305 Billion

$267 Billion

$267 Billion

$250 Billion

GROSS NATIONAL
PRODUCT
$365 Billion

NATIONAL
INCOME

$305 Billion

•INCLUDING $ 5 BILLION OF NET INTEREST PAIO BY GOVERNMENT




PERSONAL
INCOME
$286 Billion

DISPOSABLE
PERSONAL INCOME
$250 Billion

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS
An appreciable part of the productivity rise since 1929 can be
traced to a shift of the latter type. The proportion of the labor
force engaged in farming—where real product per man-hour is
substantially less than in the private nonfarm sector—has declined
markedly and almost continually. This shift of workers to nonfarm industries has in itself contributed about one-fifth of a percentage point to the average annual rate of growth in real product
per man-hour, quite apart from the improvement of productivity
in each of the sectors separately. Relative shifts of labor among
industries within the nonfarm sector have probably affected total
productivity in a similar fashion. While information is not available for precise calculation of the effects of these industrial shifts,
indirect evidence suggests that in the aggregate they may compare in importance with the farm-nonfarm movement.

Shifts in the Use of National Output
Along with the huge expansion of the gross national product
since 1929, there have been significant changes in its disposition
among major groups of users and in the composition of purchases
by each of these groups.
All major domestic purchaser groups have shared—though to
somewhat different degrees—in the increased volume of production. See chart on page 6. Net foreign investment, which
measures net purchases of United States output by the rest of the
world, is the only principal component of national product to
show a decline from 1929 to 1953.
The most notable change since 1929 in the use of the Nation's
output is a shift from private to government use. In terms of the
current dollar estimates of gross national product, government
purchases of goods and services, which absorbed 8 percent of the
gross national product in 1929, took 23% percent in 1953. Personal
consumption expenditures, on the other hand, dropped from 75%
percent of the total in 1929 to 63 percent last year. The proportion of the value of output going into investment was reduced
moderately from 1929 to 1953.
Percentage Distribution of Gross National Product
192<J

In current dollars:
Personal consumption expenditures
Qross private domestic investment
Net foreign investment
Government purchases of goods and services
Total
...
ID 1947 dollars:
Personal consumption expenditures
Gross private domestic investmentNet foreign investment...
Government purchases of goods and services

75.6
15.5
.7
8.1
-

1953

63.1
14.1
—.5
23.4

100.0 100.0
_

Total

71.9
__ 18.0
1.1
9.1

64.2
12.8
—.1
23.1

100.0

100.
0

In terms of the constant-dollar gross national product, the shift
to government use is similar. It is seen, however, to be at the
expense of both personal consumption and of investment, rather
than mainly of personal consumption, as indicated by the currentdollar figures. These value and volume relationships are summarized in the table above.

Consumption patterns reflect price shifts
Although the proportion of total national product going to consumers was smaller last year than in 1929, owing to the larger
share used for public purposes, the absolute volume of goods and
services purchased for personal consumption was, of course,



vastly increased. In terms of cons-ant (1947) dollars, the expansion amounted to 83 percent—a gain of two-fifths in real consumption per capita.
Reflecting also a 59 percent rise in average prices, the dollar
volume of total consumer outlays last year reached $230 billion,
as compared with $79 billion in 1929. The distribution of these
outlays by major objects of expenditure shifted markedly over the
two and a half decades. Nondurable goods absorbed an appreciably larger share of the consumer expenditure dollar, rising
from 47% percent in 1929 to 51% percent in 1953, while the proportion spent on services dropped from 40% percent to 35% percent last year. Durable goods purchases accounted for about the
same proportion (12-13 percent) in both years.
To a very considerable degree, these shifts reflect differential
movements of prices for major items in the respective expenditure
classes, rather than fundamental alterations of the consumption
pattern in real terms. In particular, much of the relative decline
in the importance of service outlays has stemmed from the
marked lag of rent and household utility charges behind the general upswing of consumer prices in the last decade, while most of
the increased relative importance of nondurable commodities is
traceable to the considerably above-average rise in prices for food
and clothing. The chart on page 7 shows rental outlays in
relation to total consumption.
Changes in real spending, however, have also occurred. Such
factors as the development and marketing of innumerable new
products, increased reliance upon private automobile transportation, and the expanding use of household appliances have induced
marked shifts in the pattern of spending for commodities. At the
same time, demands for some classes of services—of which the
employment of domestic servants is a conspicuous example—have
tended to decline or to lag behind the general advance.

Equipment share of investment higher
Gross private domestic investment last year amounted to $51%
billion, or 14 percent of total gross national product, as compared with about $16 billion, or 15% percent, in 1929. In real
terms, as shown in the accompanying table, the share of domestic
investment in the total was appreciably lower last year than in
1929, the difference in movement reflecting a sharper rise in the
prices of capital goods than of goods and services in general. It
may be noted that the constant-dollar data take only incomplete
account of quality change. Inasmuch as quality improvements
in fixed equipment probably exceeded those in other goods and
services comprising the national product, the indicated decline
in the proportion of real investment would be moderated if
adequate allowance for quality change could be made.
There was a marked shift from 1929 to 1953 in the general
composition of investment expenditures. New private construction put in place accounted for abo it 54 percent of the total in the
earlier period, but for only 50 pesxent last year, while business
purchases of durable equipment rose in relative importance from
36 to 47 percent. Net accumulations of business inventories
represented 10 percent of total domestic investment in 1929,
about three times the 1953 proportion.
The relative decline in construction activity was attributable
to the decrease in the proportion of expenditures for
nonresidential construction, consisting mainly of outlays for

The National Output
In Constant (1947) Dollars

•

Growth has averaged about
3 percent per year
400

300

I

200

GROSS NATIONAL PRODUCT

l

I I I I I I I I I I I I I'M I I I I 1 I I I I I [ I I I I I I I I I I I I I I I I I I I I I

1910

•

15

20

25

35

30

40

45

50

Since 1929
•




Real Output has
more than doubled

1929

•

Real output per capita
has increased by
almost three-fifths

1929

55

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Distribution of Increased Output— in 1947 Dollars
Of the $157 billion increase in the total volume of
National Output from 1929 to 1953 . . .

Consumers received almost three-fifths, o r . . .

Government one-third, o r . . .

and investment the remainder, o r . . .

$57 Billion

$11 Billion

business plant. Nonresidential construction expenditures were
reduced from 31 percent of gross private domestic investment in
1929 to 26 percent in 1953. In contrast, there was little difference
between the two terminal years in the percentage of outlays for
new nonfarm dwelling units. Thus, an outstanding feature of the
shift in investment was that business expenditures for fixed capital
facilities in 1953 were concentrated much more heavily upon
acquisition of new equipment, and proportionately less upon
plant expansion, than they were in 1929. See chart on page 8.
Over the 25-year period, these broad changes in the relative importance of construction and producers' durable equipment
were even greater in real terms than in current dollars, since
construction costs rose twice as much as equipment prices during
the period.

Shift in foreign transactions
Net foreign investment was a relatively minor component of
national product both in 1929 and in 1953. In large measure, the
shift from a positive foreign balance of $1 billion to a negative
balance of $2 billion reflects a change in the means by which
foreign countries financed their net acquisitions of United States
goods and services. In 1953, they obtained large quantities of
American exports by grants from the United States Government.
Such amounts are recorded in the national income statistics as



$89 Billion

Government, rather than foreign, purchases. Corresponding
exports in 1929, being then financed through regular commercial
channels, entered gross national product under the net foreign
investment heading. When allowance is made for this factor, the
net flow of United States output to other countries shows a
relative increase from 1929 to 1953.
The low ratio of net foreign investment to total production
should not, of course, be interpreted as an indication of the
importance of international trade to the domestic economy.
Actually, its importance is much greater than is suggested by such
a net concept. The net balance is; a composite of much larger
gross flows of United States output into export channels and of
goods and services produced abroad into domestic consumption,
capital formation, and government procurement. These gross
flows in both directions, however, were smaller in relation to
domestic economic activity last year than in 1929.

Growth of government purchases
Combined Federal, State, and local purchases of goods and
services rose from $8^ billion in 1929 to $85 billion in 1953.
As already noted, these purchases represented an increasing
proportion of total national output. See chart on page 11.
Over the two and a half decades, the entire increase in this
proportion was attributable to expanded Federal Government

NATIONAL INCOME,

activities. In the main, the expansion stemmed from the imposition upon the economy of a national defense burden much heavier
in recent years than in the prewar period. National security
purchases, which constituted less than 1 percent of gross national
product in 1929, represented more than 14 percent in 1953.
Part of the increase in the national security outlays of the
Federal Government may be traced to the large volume of foreign
aid—both military and economic—a type of activity which was
nonexistent in 1929. Federal purchases of goods and services for
all other purposes combined accounted for only a minor portion
of the 1929-53 expansion.

Changes in the Income Flow
Accompanying the expansion of the national output and the
shifts in its composition and use since 1929, there have been
marked changes in the size of the corresponding income flow, in
its industrial origin, and in the form of its distribution to residents
of the Nation.
The national income rose from $88 billion in 1929 to $305
billion last year—an increase of 250 percent. This rise, of course,
reflected not only the expansion of the physical volume of production but also the sharply higher prices prevailing in 1953.

Shifts in industrial pattern
Perhaps the most important of the changes in the income flow
since 1929 are those relating to its industrial origin. Such
changes are indicative of the way in which the allocation of
economic resources has been altered to meet the shifting character
of demand for the Nation's output.
Income originating in each industry measures the earnings of
the economic resources—both labor and property—utilized by it.
Accordingly, the breakdown of the national income by industry
of origin provides a measure of the net contribution of each
industrial segment of the economy to the total value, at factor
cost, of the net national output.
Private nonagricultural domestic industries accounted for
about the same proportion (83-84 percent) of the national income
in 1929 and 1953. The principal changes in the income originating in the remainder of the economy were in government,
which showed a substantial rise over the period, and in agriculture, which declined in relative importance.

195 4

EDITION

meet it, has centered upon commodities, including those required
for national defense. In general, fabrication and processing of a
progressively more complex character have also been involved.

Parallel grovrth in distribution
Also immediately affected by the relative increase in demand
for commodities were the wholesale and retail trade industries,
whose share of the private nonagricultural domestic total was 21
percent in 1953, as compared with 18 percent in 1929. This
expansion was closely allied with the growth of manufacturing
output, the bulk of which is distributed to ultimate buyers through
trade channels.
With the relative growth of manufacturing and trade, the
percentages contributed by all other private industrial divisions
declined except that for contract construction, which advanced
from 5 to 6 percent of the total, and communications and public
utilities which maintained the same proportion of the total in
both years. By far the greatest decline in relative position from
1929 to 1953 was registered in the finance, insurance, and real
estate division. Its share, which had matched that of wholesale
and retail trade in the earlier period, was down to about 10
percent last year.

Factors in decline of finance group
Two major factors contributed to this drop in the finance group.
Earnings in the real estate industry—especially on residential
property—were relatively depressed during most of the period,

Rent and Consumer Expenditures
The proportion of rent in total consumer
expenditures has risen since the war, but is
still lower than in the nineteen-thirfies
20 -

Large increase in manufacturing
Industrial shifts within the private nonagricultural domestic
sector are illustrated in the chart on page 12, which shows the
percentage increases since 1929 in national income originating
in several broad groups of industries. Since 1929 and 1953 were
both prosperous, full-employment years, this comparison is little
affected by cyclical influences. Corresponding data in somewhat
greater detail, covering also the intervening years, are presented
in the table at the end of Part I.
The most striking feature of the 1929-53 comparison is the
large increase in the relative importance of the manufacturing
industries. From 30 percent in 1929, their contribution to private
nonagricultural domestic income rose to 39 percent last year.
This rise is a direct reflection of the increasing degree to which



1930
MOTE:

NONFAHU

35

40

SPACE RENTS, INCLUDING IMPUTED

45

50

55

RENTS

both because the industry was little affected by the growth of
commodity output and because rents did not keep up with the

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Business Capital Investment
Business has spent relatively more on machinery and
equipment than on structures in the postwar period

i

1
I
I

PRIVATE
NONRESIDENT1AL
CONSTRUCTION

0 1 I l I I I I I i I l l I i I I I I I i I i I I I 1 I I i I I i i I I

1920

25

30

35

housing was held down by war and postwar rent controls. It
should be noted, however, that most of the relative decline in real
estate occurred long before the imposition of such controls.
Among the broad factors contributing to it was the necessarily
slow adjustment of the supply conditions emerging from the
building boom of the 1920's to the depressed housing demand of
the prewar decade, when there was a temporary slackening in the
rate of family formation and population growth. More recently
the industry has been operating in a more favorable economic
climate, and rents have risen relative to prices in general since the
gradual elimination of rent controls.
In banking and other financial industries, income originating
was much lower relative to the total than in 1929, owing mainly
to an approximate halving of average interest rates and to the
marked shift from external financing of business investment to
financing out of retained earnings. The large expansion in
public debt obligations held by banks by no means compensated
for the fact that the volume of private interest was sharply reduced
in relation to total economic activity.
Two other major industrial divisions experiencing fairly substantial declines in relative importance from 1929 to 1953 were
transportation and services. The share of the former in total
private nonagricultural domestic income fell from 9 percent to 6%
percent, entirely as a result of the much below-average expansion




40

45

50

I I
55

of the railroad industry. In the case of services, the decrease—
from 14 percent to 11 % percent—was centered in the private
household segment, where the relative decline of domestic service
was the most important factor.
The proportion of income originating in the remaining industry division, mining, was 2 percent last year, about % of a
percentage point below 1929.

Increase in Federal employment
Outside of the private nonagricultural domestic sector, a
significant change over the period under review was the sizable
increase in the contribution of government. As a percentage of
total national income, it rose from 6 in 1929 to \\% in 1953.
These percentages, it should be emphasized, represent only the
return to resources (in this case, labor) directly employed by
government and government enterprises—not the production of
other industries whose output is purchased by government.
The increase was almost entirely in the Federal Government
component, which expanded over the two and a half decades
from less than 2 percent to about 7 percent of the national income.
The compensation of military personnel accounted for more than
half of this growth, and much of the remainder was in civilian
payrolls associated with the expansion of the defense establishment.

NATIONAL INCOME, 195 4 EDITION

Income from agriculture
The share of the national income originating in agriculture,
forestry, and fisheries, which consists almost wholly of income
from farming, amounted to 5K percent last year. This was much
below the 9% percent contribution of the agricultural sector in
1929. Because of the erratic annual movements which characterize farm income, however, it is difficult to draw significant
long-run conclusions from a 2-year comparison of this type.
The decline in agriculture's percentage of the national income
from 1929 to 1953 was by no means commensurate with the
decrease over the same period in the proportion of the population
engaged in agricultural production. Accordingly, the net value
of output per person engaged in production rose by about 220
percent in agriculture, as against about 140 percent in the private
nonfarm sector.

Changes in distributive shares
Along with the shifts in the industrial origin of the national
income, there have been noteworthy alterations of its composition
in terms of distributive shares. Some of these alterations have
represented fundamental changes in the relative importance of
various forms of income as such, while others have merely reflected
the influence of industrial shifts in combination with the existing
differences among the respective industries as to prevalent forms
of organization and characteristic types of income arising therefrom. Both sorts of changes in the distributive-share pattern, of
course, are of considerable interest.
This breakdown of the national income is simply a classification
of total earnings, before deduction of direct taxes, according to
the forms in which they accrue—compensation of employees,
corporate and unincorporated business profits, rental income of
persons, and net interest. Such a classification, it should be
realized, does not reflect the relative distribution of total income
among various groups in the population, since many of these
have multiple sources of income. Nor do the distributive shares
indicate the relative remuneration of the various factors of
production in a theoretical sense; most of them include more
than one element of factor cost, and each of them represents only a
partial measure of the factor cost suggested by its caption.
Over these two and a half decades, there have been marked
increases in the relative importance of employee compensation
and of corporate profits and parallel declines in the other income shares. The percentage of the total going to proprietors
of nonfarm unincorporated enterprises, however, was only
moderately lower than in 1929.
Employee share of national income
Compensation of employees rose from $51 billion in 1929 to
$209 billion in 1953, or from 58 to 68}£ percent of total national
income. Much of this increase reflected developments which
occurred outside the ordinary business system. Income arising
outside business firms consists of a series of income flows which
originate in wholly unrelated activities. In most of these there
is only one type of income, and changes in the relative importance
of various income types for the nonbusiness segment combined
merely reflect changes in the relative importance of various types



The relative size of these flows has changed greatly since 1929.
The more important, and their movements relative to the national income from 1929 to 1953, are: (1) the compensation of
government employees, including military personnel, increased
sharply. (2) The compensation of domestic servants and employees
of nonprofit institutions declined relative to the national income.
(3) Income arising from individually-owned property was greatly
reduced relative to the national income. This income consists
almost entirely of rental income and interest, with compensation
of employees minor. Both rental income and interest originating
in this sector dropped sharply relative to the national income
total. (4) Interest originating in private households showed a
relative decline mainly because of the drop in interest paid on
loans from brokers. (5) Corporate profits originating abroad
(branch profits and dividends combined) increased, while interest
received from abroad dropped sharply.
The net result of these shifts was a very sharp rise in the proportion of all income originating outside ordinary business firms
which consisted of the compensation of employees. The offset
appeared almost entirely in interest and rental income.
Within the ordinary business sector, as well, important changes
have taken place in industrial structure and, largely for this
reason, in legal form of organization. Since the usual division
of income by type varies widely among industries and as between
corporate and noncorporate firms, these changes also alter the
share distribution of the total national income.
In the important corporate sector of the economy, accounting
for more than half of the national income, the ratio of employee
compensation to total income originating was substantially stable
at about three-fourths of the total in prosperous peacetime years
throughout the period from 1929 through 1951. In 1951 (the
last year for which final data are available) it amounted to 74
percent as compared with 74% percent in 1929. It then increased
to 76 percent in 1952 and 77% percent in 1953. Corporate profits
in 1952 were adversely affected by the steel strike and in 1953
by the business adjustment that started in the second half of
that year.
Some increase in the employee share is indicated by the data
to have occurred in unincorporated nonfarm enterprises from
1929 to 1953. Developments in the construction and service
industries were largely responsible for this rise. The change in
the labor percentage in these two industry divisions reflected
mainly a larger increase in the number of wage and salary workers
than in the number of self-employed, rather than divergent movements in average employee compensation and average net income
per entrepreneur. In agriculture the share of labor income has
shown no apparent trend since 1929. Although the 1953 percentage was above that in 1929, it had been lower as recently
as 1951.
An additional element stands out in the change in the employee
share of national income. This was the internal shift to a somewhat lower proportion of wages and salaries and a higher proportion in the form of supplements to wages and salaries. The
latter were an inconsequential element in 1929, consisting chiefly
of compensation for injuries. Their growth to significant proportions—3}i percent of national income in 1953—stems from
the creation and expansion of the various social insurance pro-

10

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

pension, health, and welfare funds. Employers' contributions to
these funds, both public and private, are viewed as supplementary
compensation of employees.
Shifts in proprietors' and rental income
The advance of entrepreneurial earnings from $14% billion in
1929 to $38% billion last year was less percentagewise than the
rise in national income, due piimarily to the relatively small rise
in the farm component. As a share of the national income, farm
proprietors' income declined from 7 percent in 1929 to 4 percent
in 1953. The erratic behavior of farm income makes it difficult,
however, to assess long-term trend from this type of comparison.
The farm income percentage was markedly above the 1929
figure as recently as 1948, and averaged close to 6 percent in the
ensuing three years. In view of the sizable decline in the number of farm proprietors since 1929, the lower percentage share of
1953 represented an improvement in the relative position of the
average farmer.
Nonfarm business and professional proprietors' income was 8%
percent of the total last year, about 1 % percentage points below
1929. A rise in the relative importance of the retail and wholesale trade component was more than offset largely by the fact
that entrepreneurial earnings in the service industries did not
maintain their relative standing.
One of the two distributive shares exhibiting a sharp proportionate decline over the past 25 years was rental income. The
$5% billion going to persons in this form in 1929 constituted 6
percent of national income, while last year's $10% billion
represented little more than half that much, percentagewise.
The principal reasons for the diminished importance of rental
income in relation to the total are those outlined above in connection with the real estate industry as a whole. It should be
remembered, however, that the rental income share—including
imputed net rent on owner-occupied nonfarm dwellings—consists only of net rents and royalties accruing to persons not primarily engaged in the real estate business. Other rents are merged
unidentifiably with noncorporate business earnings and with
corporate profits.

Influence of inventory profits and losses
The foregoing remarks are based upon measures of corporate
profits after inventory valuation adjustment. Profits before tax
as reported under prevalent inventory accounting practices,
which generally charge goods to cost of sales in terms of priorperiod inventory costs rather than current replacement costs,
showed a somewhat greater increase from 1929 to 1953. These
figures included moderate inventory losses in 1929, when—with
prices falling—book costs of goods sold exceeded replacement
costs, and included inventory profits last year, when—with prices;
rising moderately—the reverse was true.
Such inventory profits and losses—which become major elements in book profits in years of sharp price change—are eliminated, in order to secure an economically more meaningful
measure of income originating in current production, by application of the inventory valuation adjustment. In effect, this adjustment substitutes the current replacement cost of goods sold for
their book cost in the computation of profits.
Because of the very large increase in Federal corporate income
tax rates, the percentage of national income taken by such taxes,
in combination with similar State levies, more than trebled from
1929 to 1953. The share of profits after tax (including inventory
profits) has fallen from 9% percent in 1929 to 6 percent in 1953.
Most of this decline has occurred since 1950 as a result of the
increase in taxes during the Korean war period.
One of the salient trends in corporate financing during this
period—toward greater reliance upon internal funds—is reflected
in the divergence between the terminal years with respect to the
disposition of profits after tax. In 1929, 70 percent was paid out
as dividends and 30 percent retained, whereas in 1953 only 51
percent was distributed to stockholders and 49 percent was
retained. This disparity between the 2 years is reduced if inventory
profits and losses are excluded from the comparison. On this basis,
undistributed profits rose from 33 percent of profits after tax in
1929 to 46 percent last year. It may be noted that in all but one
of the postwar years (1952) the contrast with 1929 in the proportion of corporate earnings withheld and paid out was even greater
than in 1953.

Net interest
Combined profits and interest share
The corporate profits share of the national income—corporate
profits and inventory valuation adjustment—increased from $10
billion in 1929 to $38% billion last year—from 11% to 12% percent. This rise was a reflection of the greatly reduced burden of
corporate debt, and does not signify a commensurate expansion
of the property share of current income. In combination, profits
and interest originating in corporate business was fractionally
lower in 1953 than in 1929 in relation to total national income.
It was down also in relation to total income originating in corporations—the counterpart of the increase in the labor share
already discussed. In the main, however, the combined property
share of total income originating in corporate business has been
fairly uniform in prosperous peacetime years of the period under
review. This uniformity is brought out in the chart on page 14,
and contrasts with the extreme variability of the property income
share during the business cycle.



The remaining distributive share, net interest, fell from 7%
percent of national income in 1929, to only 3 percent last year.
The major factors underlying this decline are those cited above to
explain the diminished proportion of income originating in the
financial industries—namely, the virtual halving of average interest rates and the relatively small expansion of private debt
since 1929.
In addition, part of the decline i:> attributable to a statistical
peculiarity of the series. To offset the inclusion in business incomes
of government interest, which is viewed as a transfer in the
national income accounts, government interest received by
business is deducted from the interest component of national
income. Had the statistically more difficult procedure of deducting
it from business incomes been followed, the relative decline in net
interest from 1929 to 1953 would have been less, while the other
affected shares would have shown correspondingly smaller
increases.

NATIONAL INCOME, 1 9 5 4

1?

EDITION

GOVERNMENT PURCHASES
amounted to 23% of the Gross
National Product in 1953 as
contrasted with 8% in 1929...
BILLIONS OF DOLLARS

400
GOVERNMENT

with the bulk of the increase
in the government proportior
going for defense purposes

300
PRIVATE

PERCENT OF G N P

200

25

20

100

1929

OTHER FEDERAL
and STATE & LOCAL

1953

1929
Expansion of Personal Income
Personal income differs from national income by the exclusion
of those portions of income earned in current production which
are not paid out to persons, and by the inclusion of certain items
not arising in current production—chiefly transfer payments
and government interest.
In 1929, personal income totalled $86 billion; by 1953, it had
mounted $286 billion. On a per capita basis, the increase was
from about $700 to $1,800, or more than 150 percent.
Along with this advance in the total, there were significant
shifts in its composition. Most of these—the increased relative
importance of payrolls and the reduced proportions of interest,
291602"—54


1953

rental income, dividends, and farm proprietors' earnings—have
already been noted in the discussion of distributive shares.

Transfer payments increase in importance
In addition, there have been important changes, stemming
from the expanded role of government in the economy, in other
elements of personal income. Foremost among these developments
is the growth of transfer payments. From $1}£ billion, or less than
2 percent of personal income, in 1929, they rose to $14 billion,
or nearly 5 percent, in 1953.
Most of the rise was in the Federal Government component,
which in 1929 consisted chiefly of military pensions and related
items. By 1953, as a result of World War II and the Korean war,

12

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Growth in Private Nonagricultural
National Income Since 1929
Manufacturing, construction, and trade
have shown above-average growth
350 -

300 CO

Manufacturing

Construction
Trade

250

ALL INDUSTRIES

o
200 ©

£
u
c

I

150 -

100

Services
_ Transportation 8
Public Utilities
_
Mining

-'-"•-IT

Finance 8
Real Estate

50 -

1929
these payments were greatly enlarged, both absolutely and in
relation to personal income, and new classes of veterans' benefits
under the servicemen's readjustment acts were flowing in large
volume. Moreover, payments from Federal social insurance funds,
which in 1929 had been confined to civilian retirement and veterans' life insurance benefits, last year included not only increased
amounts under these headings, but also $4% billion of old-age
and survivors', railroad retirement, and unemployment insurance
benefits.
State and local government transfer payments, although overshadowed by those of the Federal Government, have also risen
markedly since 1929. State veterans' bonuses have contributed in
recent years, but most of the increase has been in special types of
public assistance for such groups as the blind, the aged, the
disabled, and dependent children. It may be noted that while




1953
these payments are made by State- and local governments, they
are financed in part by Federal grants-in-aid.

Growth of social insurance contributions
For the personal sector of the economy as a whole, the growth
of transfer payments has been partly offset by the concomitant
expansion of social insurance contributions. Confined in 1929
to a few public employee retirement systems and veterans' life
insurance funds, but since extended by the establishment and
development of the various Social Security programs, these now
take an appreciable portion of current personal earnings. To date,
contributions for social insurance have consistently exceeded
benefit payments from the funds—in most years by a sizable
margin. On balance, however, rhe combined effect of social

NATIONAL INCOME, 195 4 EDITION

insurance transactions and transfer payments from general
government funds has been a material net addition to currently
earned personal income.

13

to trace in a general way the course of economic development
during the period under review.

Gross National Product Patterns
Government interest increases with debt
Also of consequence in the expansion of personal income since
1929 has been the sharp increase of government interest payments
associated with the tremendous growth, mainly during World
War II, of the public debt. Owing to the inclusion of government
interest, which is treated in the national income accounts as a
transfer item, personal interest income declined much less in
relative importance from 1929 to 1953 than did the net interest
component of the national income.

Real income up substantially
Each of the elements of personal income is measured without
reference to the impact of direct personal taxes. Much of the increase in the total over the past two and a half decades, however,
has been absorbed by such taxes. Personal tax and nontax payments amounted to about $2% billion, or 3 percent of personal
income, in 1929. In 1953 they were 14 times as large, totalling
$36 billion and absorbing about 12)^ percent of personal income.
See chart on page 16. The relative increase was entirely in the
Federal Government component, and resulted almost wholly
from the broadened base and sharply higher rates of the individual
income tax. State and local personal taxes, although doubling
from 1929 to 1953, declined slightly as a percentage of personal
income.
After deduction of these taxes, there remained disposable personal income of $250 billion last year, as compared with $83 billion in 1929. Corresponding figures on a per capita basis were
approximately $1,570 and $680, respectively. With consumer
prices averaging 59 percent higher in 1953 than in 1929, the increase in real disposable income per capita was thus about 45
percent. See chart on page 19.
The proportion of disposable personal income spent for current
consumption last year was lower than in 1929—with 8 percent
going into personal saving as compared with 5 percent in the
earlier year. However, the true extent of the change may differ
somewhat from that indicated by the figures, because the saving
estimates are computed as residuals, and hence are sensitive even
to minor statistical imperfections in the measurement of disposable income and consumption expenditures.

FLUCTUATIONS IN ECONOMIC ACTIVITY
The substantial growth of the economy, as revealed by the foregoing summary comparison of national income and product data
for 1929 and 1953, was extremely irregular. Likewise, the associated changes in the economic structure did not occur in smooth
progression, but emerged from a series of fluctuations of unprecedented magnitude, including the great depression and the vast
expansion of World War II. In order to illuminate the processes
by which the economic scene has been transformed, it is desirable



Before this summary is given, some of the major factors in the
fluctuations of economic activity since 1929 will be highlighted
with the aid of two interrelated percentage distributions of the
gross national product—one by type of expenditure and one by
type of receipt—which are presented in tabular form at the end
of this summary.
The distribution by type of expenditure indicates the proportions of total output bought by each major sector of the economy:
by persons, by business (for fixed investment and inventory accumulation), by the rest of the world (net), and by Federal,
State, and local governments.
The percentage breakdown of gross national product by type
of receipt reflects broadly the corresponding distribution of currently-generated purchasing power, exclusive of borrowing transactions. It shows the proportions of the gross income flow received
by consumers as disposable personal income; by the business sector in the form of gross retained earnings, including capital consumption allowances; and by all levels of government in the form
of tax and nontax receipts net of amounts transferred (such as
interest and transfer payments) to other sectors.
Each of these breakdowns of gross national product is of interest
in itself, but they provide particularly valuable insight into the
functioning of the economy when studied in combination, with
the respective expenditure and receipt shares of each major sector
paired. Such an arrangement of the data is shown in the chart
on page 21.
In each panel of the chart are plotted the percentage receipt
and expenditure shares of one of the major sectors, the rest of
the world sector being included with the business sector for this
purpose. The shaded areas between the lines measure, respectively, the government surplus or deficit; personal saving or dissaving; and the excess or shortfall of gross investment in relation
to gross business saving—each expressed as a percentage of gross
national product. It will be noted that the three expenditure
percentages, represented by the solid lines, add to 100 in every
year, as do the three receipt percentages (dashed lines), except
for the statistical discrepancy between the estimates of national
income and national product.

Strategic role of investment
The middle panel strikingly displays the strategic role of investment expenditures in the business cycle. Clearly depicted are
their disproportionate collapse in the great depression and their
gradual rise in relative importance, briefly reversed in the 1938
recession, during recovery. Also illustrated is the severe cut in
private capital formation required during World War II and its
resurgence afterwards. In connection with the wartime figures,
however, it should be remembered that sizable installations of
plant and equipment undertaken directly by the Government are
reflected in the bottom panel rather than as business investment.
The uniformity of the proportions of output going into investment in prosperous peacetime years is noteworthy, as is the
contrast between these proportions, all in the neighborhood of 16

CORPORATE PROFITS SINCE 1929
Profits fluctuate more than other income originating in corporate business
175

150
CORPORATE PROFITS BEFORE TAX-

125
OTHER INCOME ORIGINATING IN CORPORATE BUSINESS
(COMPENSATION OF EMPLOYEES AND NET INTEREST)

100

o
z

75

AFTER INVENTORY VALUATION ADJUSTMENT

I

I

I

I

I

I

I

I

1950

25

I

I

1955

The proportional share of profits and interest is
relatively uniform in prosperous peacetime years

1945
CORPORATE LOSS

I

I

I

1930




I

I

I

I

I

1935

PROFITS AND NET INTEREST
AS PERCENT OF
••
INCOME ORIGINATING

AFTER INVENTORY

VALUATION ADJUSTMENT

I

0

20

40

60

80

_

100

120

140

INCOME ORIGINATING IN CORPORATE BUSINESS
(BILIION DOLLARS)

_L
160

180

NATIONAL INCOME,

percent, and the limitation of investment to 2 or 3 percent of the
gross national product at the bottom of the depression.
Gross business saving, while also showing a disproportionate
swing in the major business cycle, has been a great deal steadier
than investment, reflecting the relative stability of depreciation
allowances. In the expanding United States economy, gross
business saving has usually fallen short of gross investment.
Apart from the World War II period, this has been true in all
years except those of the great depression and 1938. The financing
of the excess of investment has required funds made available by
the other sectors, including those left on deposit with banks, as
well as those provided through purchase of stocks or bonds.

Cyclical stability of consumer spending
The percentage of output bought by consumers—charted in
the top panel—has exhibited a contracyclical tendency, as
attested by its rise from 1929 to 1932, in 1938, and in 1949. It
should be remembered, of course, that these were rising proporportions of a diminishing total output. Except in the mild recession of 1949, they represented declining physical volumes and
dollar values of consumption expenditures in absolute terms.
Conversely, falling percentages are observable in most years of
peacetime expansion in economic activity. These, however, have
been associated with increases in the absolute volume and value
of consumer spending.
Another outstanding feature of the top panel is the broad
picture it gives of the characteristics of consumer finances during
World War II. Disposable income was reduced by heavy taxation, but consumption expenditures were cut far more by the
diversion of productive resources to war use, in combination
with price controls and exhortations to save. The resultant
extraordinary volume of personal savings is clearly illustrated.
This phenomenon was not repeated during the Korean war
period, as the proportion of total output diverted to national
security use was much more moderate, and was financed largely
by taxation.

War generates large Government deficit
The counterpart to these wartime personal savings, as well as
to the concurrent excess of gross business saving over investment,
appears in the bottom panel of the chart, where the huge World
War II Government deficit stands out. Although government
purchases reached a peak of about 46 percent of total output,
government at no time claimed a commensurate proportion of
currently-generated purchasing power. Net government receipts
averaged 22 percent of gross national product at their relative
maximum in 1943 and 1944.
The difference between the share of output bought by the
Government and its share of receipts was financed by borrowing
on an unprecedented scale. The inflationary impact of this
deficit financing was restrained to a considerable extent during
the war period itself, despite scarcities of civilian goods and
services, by such factors as price control, rationing, and the
willingness of the public not only to buy Government bonds, but
also to accumulate other liquid assets. After the cessation of



195 4

EDITION

15

hostilities, however, the postponed effects of wartime deficit
financing contributed to the postwar inflation.
In the Korean war period, net government receipts again
mounted to almost 22 percent of the gross national product, but
government purchases of goods and services fell considerably
short of the World War II peak. Accordingly no comparable
deficits were generated in the later period.
It should be noted, especially in connection with the postwar
figures, that the relative rise in total tax and nontax receipts of
Federal, State, and local governments has been greater than that
of the net receipts item plotted in the chart. There has been a
material increase in the divergence between these two measures
because of the disportionate expansion of the volume of gross
receipts required to finance veterans' benefits, interest on the
public debt, and other transfers deducted in deriving the net
figures.

Basic expansion in role of government
A final feature of the three panels in combination is the shift
from private to public consumption, which has already been noted
in the foregoing discussion of long-run changes in the use of
current-dollar gross national product. The shift is evidenced by
comparison of the top and bottom panels, which show a distinct,
although irregular, uptrend in the government share, whether in
terms of purchases or of net receipts, together with a generally
offsetting long-term decline in the personal share. The latter
movement is reflected in disposable income, as well as in consumption expenditures. Primarily, of course, the rising percentage
share of government reflects the increasing responsibilities
assumed by the Federal Government with respect to national
defense.
Neither in gross investment nor in gross business saving has
there been any noticeable long-term change in relative importance, despite extreme cyclical and wartime variations.

Chronological Review: 1929-53
The year 1929 marked the end of an era of relatively full employment, business confidence, and general prosperity. Economic
activity had been advancing strongly, with only minor interruptions, for eight years.

Business decline: 1929-33
The 1929 downturn, the ultimate causes of which are still a
matter of controversy, was most clearly reflected in the collapse
of investment demand. Gross private domestic investment
dropped about one-third from 1929 to 1930, as new construction
and producers' purchases of durable equipment were cut sharply
and the accumulation of nonfarm business inventories ceased.
Foreign purchases also declined in 1930, although the drop was
not reflected in net foreign investment because of a matching
reduction in import demand.
With employment and incomes adversely affected by the sharp
reduction of investment, consumer purchases also decreased,
contributing to the general contraction and inducing still further
cuts in outlays for investment.

16

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Consumer purchases, however, held up much better in 1930
than investment demand. The aggregate income flow to individuals shrank less than production and the incomes generated by it,
as undistributed corporate profits absorbed a disproportionate
share of the over-all decrease in earned income. Also, consumers
tended to spend a higher proportion of current income or to
dissave in the attempt to preserve previous living standards.
Essentially the same pattern of cumulative decline persisted,
and in fact accelerated, during 1931 and 1932. By the latter year,
gross private domestic investment had fallen to the very low level
of less than $1 billion, as contrasted with $16 billion in 1929.
The further moderate decline of the gross national product in
1933 was in consumer purchases, where it reflected primarily
lower average prices rather than a further decrease in volume.

dollar volume, were considerably increased in relative importance
by the collapse of private demand.
More than half of the 1929-33 decline in the market value of the
national product stemmed from lower prices. As measured by the
gross national product in constant (1947) dollars, real output fell
by three-tenths.
Foremost among the factors underlying the shrinkage of real
output was the reduction of employment. At the depth of the
depression, the number of persons engaged in production was
almost one-fifth lower than in 1929, and unemployment was
almost 13 million—close to one-fourth of the Nation's labor force.
Moreover, average hours worked per week by those who remained
employed were considerably reduced.

Recovery: 1933-37
National product halved in value
Over the entire period of contraction from 1929 to 1933, the
gross national product dropped by nearly one-half, from $104
billion to $56 billion. At the bottom of the depression less than 3
percent of the Nation's output went into business investment, as
compared with 15% percent in 1929. Conversely, consumer purchases rose from three-fourths of the total in 1929 to five-sixths in
1933. Government purchases, although little changed in absolute




Some of the most serious deflationary forces underlying the
post-1929 collapse were by 1932 beginning to spend themselves.
Installations of new plant and equipment had virtually ceased in
most segments of the economy, and such gross fixed business
investment as did persist represented primarily the fulfillment of
minimum replacement needs. As replacements had been cut to
the bone for several years, the feasibility of further postponing
them was rapidly diminishing by the end of 1932. Business pur-

INCOME AND

TAXES

Personal and corporate taxes have taken an increased share of incomes
PERSONAL
INCOME
and TAXES

CORPORATE
INCOME
iind TAXES

CORPORATE INCOME
AFTER TAXES
$8.3 Billion

1929
$9.6 Billion

1953
$39.4 Billion

1953
$286.1 Billion

NATIONAL INCOME, 195 4 EDITION

chases of durable equipment, accordingly, fell no lower in 1933.
Private construction activity did continue downward, but the
drop was smaller than in any of the three preceding years.
Sizable inventory liquidation continued in 1933, but as it had
already carried working stocks close to a minimum even in relation to the low current volume of sales, the rate of liquidation
was considerably reduced. It had previously been possible for
businesses to meet the sagging volume of sales partly out of
relatively excessive existing inventories, with the consequence
that production—and hence total income—was reduced even
more than consolidated business sales. Now, however, this possibility was vanishing, and it became necessary to keep output at
least on a par with current demand. Here too, then, a weighty
deflationary force was exhausting itself.
With the marked retardation of income declines stemming
directly from reduced investment expenditures, the fall in consumer demand was measurably slowed in 1933. The stage was
thus finally set for recovery. It was evidenced in a few industries
as early as the fall of 1932, but appears to have dated generally
from the spring of 1933. Monthly personal income data show the
low point in March, after which there was a slow and uneven
rise during the remainder of the year.
With its decline arrested in 1933, fixed business investment
turned up moderately in 1934, when both construction and
equipment outlays began to expand again. Nonfarm inventory
liquidation ceased, and a general trend toward rebuilding of
stocks depleted during the depression set in. It was stimulated
not only by the emerging recovery of sales, but by the rise in
prices already under way during 1933.

Government supplements private recovery
In the meantime, the Federal Government had assumed an
active role in the economy, and was making strenuous efforts to
promote recovery. Along with the adoption of other measures, it
entered the market directly on an expanding scale, especially in
its work relief activities, and also provided substantial aid to
State and local governments.
With the increase of incomes generated by the pick-up of
business investment and the growth of government purchases,
personal consumption expenditures also rose in 1934. Their
expansion, in turn, fed the income stream and provided stimulus
for a further upsurge of investment. This was at first mainly confined to long-deferred replacement of capital facilities which had
deteriorated during the depression; but as profits reappeared and
business confidence in future prospects was gradually restored,
an increasing proportion went into wholly new plant and equipment, and inventories were expanded to meet the rising volume
of sales. Residential building, spurred in part by Federal aid to
homeowners, moved ahead once more, and total gross private
domestic investment advanced steadily from $\% billion in 1933
to $11K billion in 1937.
Consumer purchases also continued to rise. At $67 billion in
1937, they were 45 percent above the low mark of 1933. Although
their rate of increase was proportionately smaller than that of
domestic capital formation, they represented quantitatively the
largest element in the upward spiral of employment, production,
and incomes.
Apart from the newly expanded role of government, the whole



17

mechanism of the recovery was thus very similar to that of the
downswing, except that it operated in reverse and also more
slowly. Of the $35 billion increase in gross national product from
1933 to 1937, about 30 percent was in private domestic investment, raising it from 2% to 13 percent of the total. Consumer
outlays accounted for about 60 percent of the change—substantially less than their share of total output—and government
purchases, dropping slightly in relative importance, absorbed the
remaining 10 percent of the increment.

The recession of 1938
Incomplete as was the recovery of the economy by 1937, it
was interrupted by a downturn beginning in the latter part of
that year and extending through mid-1938. Although of brief
duration, this downturn was relatively severe. Within a few
months, unemployment rose sharply. Industrial production fell
by over one-fourth from August, 1937 to January, 1938, and
personal income dropped at a pace comparable to that prevailing
in 1931-32. The decline tapered off thereafter, however, and
production began to pick up again in the second half of 1938.
For the year as a whole, the decrease in gross national product
was about 6 percent.
The 1937-38 recession was much steeper in its initial descent
than the previous downswing, but it was of a less basic character.
Of the $5% billion decline in gross national product from 1937
to 1938, almost three-fifths was attributable to a shift from
accumulation to liquidation of business inventories. Inventory
shifts accounted for only 16 percent of the drop in output from
1929 to 1930.
Business plant and equipment expenditures contracted about
as sharply in 1938 as in 1930, but residential construction activity,
contrastingly, continued to rise, and consumption expenditures,
despite the drop in employment and personal income, declined
by only 4 percent, as compared with 10 percent in 1930. The
consuming public as a whole sustained its spending close to the
1937 rate by a $3 billion cut in personal saving. Moreover, the
moderate decline in consumer outlays which did occur was very
largely counterbalanced by increased government: buying and
net foreign investment.
Altogether, purchases of goods and services by final users of
the Nation's output declined by less than 3 percent in 1938, as
compared with 11 percent in 1930, and the major portion of the
swing in production was absorbed by the change in inventories.
Curtailment of production ceased as soon as the strength of final
demand became apparent. The drop in fixed business investment
proved to have been instigated by short-run considerations, rather
than by any fundamental lack of investment opportunities. The
basic underlying situation, in fact, was that large capital requirements accumulated during the depression still remained to be
fulfilled, and that many new investment opportunities stemming
from technological advances remained to be exploited.

Renewed recovery: 1938-41
Following the jar of the 1938 recession, the recovery was
renewed and continued steadily into 1941, when it was merged
with the first stages of military preparation for World War II.
All forms of business investment were expanding steadily

18

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

during this period. Purchases of producers' durable equipment
nearly doubled in dollar volume from 1938 to 1941, and private
construction activity also rose strongly. Inventory liquidation
ceased in 1939, when production was brought back in line with
current sales, and inventories were accumulated on a mounting
scale in the next two years.
By 1941, total gross private domestic investment was not far
from three times as large as in 1938. For the first time, it surpassed
the 1929 total, both in value and in physical volume. Net foreign
investment was also sizable in the three years following 1938,
being especially stimulated in 1940 and 1941 by foreign demand
for munitions and other supplies required for the Allied war effort.
Responding to the increased incomes generated by expanding
employment—and contributing, in turn, to the advance of profits,
business investment, employment, and incomes—personal consumption expenditures rose from $64% billion in 1938 to $82
billion in 1941. The relative rise was particularly marked—about
two-thirds—in outlays for durable goods. Higher prices figured in
the advancing rate of consumer spending, but the major portion
represented enlarged quantities of goods and services. The real
volume of personal consumption per capita increased 16 percent
from 1938 to 1941, and exceeded the 1929 figure from 1939 on.
Military requirements become dominant
Government purchases were approximately stable until the
latter part of 1940. After the fall of France, the national defense
program got under way on a rapidly expanding scale with
progressively greater influence upon economic conditions.
Throughout 1941 and thereafter it was the dominant factor in the
economy.
In this second stage of recovery from the depression, from 1938
to 1941, the dollar value of the Nation's output advanced by
almost 50 percent, to $126 billion. With a general price rise in the
neighborhood of 9 percent, the increase in the physical volume of
production was close to two-fifths.
The year 1941 was one of marked expansion, and on the average
did not represent full peacetime capacity. During much of the
year, there were idle manpower resources, as is attested by the
fact that the proportion of the labor force unemployed was
slightly higher than in 1930, the first year of the depression.
Nevertheless, the degree of recovery evidenced by 1941 was
impressive. In constant dollars, the gross national product stood
one-third higher than in 1929. This expansion of real output
was achieved with an increase of only 15 percent in the total
number of persons engaged in production. Moreover, average
hours worked per week were reduced considerably over the 12
years. In the private sector, where the increase in output was
about 30 percent, total man-hours utilized differed but little from
those in 1929.
That so large an increase in the volume of output was nonetheless accomplished was attributable to the rise in real product per
man-hour worked in the private economy. Although lagging in
the depression, productivity had advanced rapidly after 1934,
and by 1941 was near the level indicated by its long-term trend.
The larger output of 1941 was being distributed to major
economic groups in a somewhat different fashion from that of
1929. The most noteworthy change was in the share bought by
government. With substantial military preparations getting under



way, this share—including a slightly smaller portion for State
and local governments—amounted to roughly one-fifth of the
gross national product as compared with less than one-tenth in
1929. The proportion purchased by consumers, on the other
hand, was reduced from 76 to 65 percent, while the percentages
going into gross private domestic and net foreign investment
were little changed.

The war economy: 1942-45
Preparations for war began at & time when the economy was
operating at less than full capacity, with unemployed labor,
plant, and equipment, and an abundance of raw materials. At
first, because of the availability of these unused economic resources, war production could be superimposed upon the civilian
economy. It acted as a stimulant, and civilian production increased concurrently. Gross private domestic investment proceeded at a high rate, and consumer purchases—especially of
durable goods—were buoyant.
During most of 1941, the needs of the war program were thus
compatible with expanding civilian production. Moreover,
much of the capital equipment acquired during this period later
proved to be readily convertible to war production. Also, the
additions to the stock of capital, along with additions to the stock
of durable consumer goods, subsequently permitted the diversion
of more productive resources from civilian use than would otherwise have been possible except with sharper cuts in living standards.

Emergence of war-time problems
As the dimensions of the war effort expanded, however, serious
problems emerged. Although the rising volume of war production
generated a rapid expansion of incomes, it provided no goods and
services to satisfy the resultant growth of civilian demand. Instead, it impinged upon their availability as soon as the slack in
the economy had been taken up. Shortages of specific labor skills,
capital facilities, and raw materials began to be more and more
frequently encountered. After Pearl Harbor, it became obvious
that the war program would take proportions of output so huge
that they could not be provided by enlarged production alone,
and that civilian demand would have to be restricted.
During the period of transition to a full war economy, accordingly, a succession of measures was adopted with a view to ensuring maximum war production together with the optimum functioning of the civilian economy. Rates of taxation were steeply
increased, not only to help finance Government war expenditures, but also to restrict the amount of civilian purchasing power
available to bid for the limited volume of goods and services
remaining after military requirements had been met. Fiscal
measures were supplemented by the imposition of direct controls,
including priorities, inventory limitation orders, allocations, manpower regulations, price and wage controls, and rationing. In
addition, individuals were urged to restrict consumption voluntarily and to invest their surplus purchasing power in Government
bonds.
On the whole, the flexibility of the economy in the transition to
full-scale war production proved great and total production continued to rise rapidly despite conversion.

19

NATIONAL INCOME, 195 4 EDITION

War peak reached in 1944
In 1944, the peak year of war production, the dollar value of
the Nation's output was $211 billion—$85% billion higher than
in 1941. Because of the extreme changes in the nature and composition of output, it is difficult to measure how much of this
dollar increase reflected physical volume. According to the constant-dollar gross national product series, the expansion of real
output from 1941 to 1944 was more than one-third. Whatever
the precise figure, it is clear that there was a large rise in physical
production during World War II. The major underlying factors
were an extraordinary expansion of the labor force and employment, an increased stock of capital equipment, large-scale operations and technological progress in war production, and a better
utilization of labor and productive capacity in many civilian
industries.
The increase in physical output was by far the most important
source of war output. Next in importance were reductions in
gross fixed investment, consumer durables, and government nonwar purchases. According to available evidence, real consumption
of nondurable goods and of services was more than maintained in
total. While the wartime level of personal consumption is somewhat exaggerated in the constant dollar figures because of the
impossibility of measuring fully the actual price rise which

occurred during the war, it is evident that, in the aggregate, real
consumption was not curtailed.
Although personal consumption held up during the war, it was
much lower in relation to income than it would have been in
ordinary peacetime years of high economic activity. An unprecedented proportion of wartime incomes was absorbed by
taxes, and the saving rate was abnormally high. The latter
reflected a combination of restricted civilian supplies regulated by
price control and rationing and the response to patriotic appeals
for investment in war bonds.

War changes distribution of output
As a result of these wartime changes, the composition of gross
national product in 1944 differed drastically from that in 1941.
Government purchases amounted to 46 percent of the total in
1944. The comparable figure for 1941 was only 20 percent, including twice as large a proportion for civilian programs of
Federal, State, and local governments.
All other major uses of output were reduced far below their
usual relative importance, with the sharpest cuts being those in
domestic and net foreign investment and in consumer durables.
Private fixed capital formation fell to about 4 percent of the
national product as compared with 11 percent in 1941; and requirements were also met in part through a net drain upon both

Real Personal Income Per Capita in 1953

was 60 percent above 1929 before taxes




and 44 percent above 1929 after taxes
$1,532
$1,339

$956

1929

$927

1929

1953
IN 1947 DOLLARS

1953

20

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

business inventories and foreign sources of supply. The proportion
representing consumer outlays for durable goods was more than
halved as compared with that of 1941, and the relative share of
personal consumption expenditures for nondurables and services
also decreased—though by a much smaller margin.

Reconversion and postwar boom: 1945-48
At the end of World War II, the Nation faced a set of economic
problems which in some ways were the counterpart to those of the
original mobilization. It was widely recognized that the transition
from a situation in which roughly two-fifths of economic resources
were being employed in war production to one in which most of
the resources would again be devoted to civilian output could
only be accomplished in an orderly fashion by widespread
cooperation among all major groups in the economy. As in the
mobilization itself, the striking flexibility of the economy was
demonstrated.
A broad Government program designed to speed reconversion,
and to ease its impact upon the returning soldier and upon business, was enacted. With the quick resurgence of business, personal,
and foreign demands, as well as the vast Government programs
undertaken to aid in the rehabilitation of war devastated areas,
the immediate postwar economic decline was held to moderate
proportions.
Although war purchases of the Federal Government were cut
back with great speed—from an annual rate of $90 billion in the
second quarter of 1945 to $28 billion in the first quarter of 1946—
much of the slack was quickly taken up by the rapid expansion of
private spending. Total gross national product dropped 12 percent over these three quarters, or by $27 billion at annual rates;
thus more than half of the drop in war expenditures was offset.
Discharged servicemen and war plant workers were speedily
absorbed in civilian pursuits, and at no time did unemployment
rise appreciably above 2% million.
After the first quarter of 1946, the buoyancy of private demand
more than offset the moderate further declines in government
purchases. Strong inflationary pressures characterized the 194546 reconversion, even before controls were eliminated, and continued to dominate the economic scene for the next two years.
Influence of liquid saving and backlog demands
Several key factors underlay the strength of private demand.
During the war, both consumers and businesses had accumulated
an enormous volume of savings—much of it in liquid form. At
the same time, they had built up a backlog of urgent demands for
all types of civilian goods, and especially for durables.
As regards the purchasing power of consumers, it may be
further noted that in the brief contraction from mid-1945 to early
1946 the flow of disposable personal income was maintained.
Undistributed corporate earnings absorbed a large share of the
swing in total income arising from production, and the Government disbursed mustering-out payments and other veterans'
benefits in large volume and lowered the wartime tax rates. Not
only did the pent-up demand for durable consumer goods materialize as expected, but an insistent consumer demand for
nondurables and services also became an active and powerful force
in the economy.



Business speeds investment
At the same time, business plant and equipment investment
programs were pushed ahead fast in the immediate reconversion
period and continued to expand strongly thereafter. Inventories,
very low at the close of the war, had to be accumulated rapidly
to bring working stocks into line with the heavy volume of
business. The pace of residential building activity also accelerated
steadily.
Moreover, net foreign investment assumed a relative importance
far beyond its usual role. With financial support provided both
by wartime accumulations of gold and dollar balances and by a
large volume of United States Government loans, and under the
stimulus of world-wide shortages stemming from the impairment
of productive facilities abroad, net foreign purchases of American
output reached unprecedented proportions.
In combination, these heavy demands placed a severe strain
upon the productive capacity of the economy, which was reduced
considerably below the wartime peak. Despite low unemployment, the number of persons engaged in production was 7 million,
or 11 percent, lower in 1946 than in 1944. This was due to the
withdrawal from the labor force of sizable classes of individuals—
such as adolescents, housewives, and persons past the normal
retirement age—who are not ordinarily employed, but who had
been induced by special wartime circumstances to accept employment. In addition, average hours worked per week fell off as
overtime schedules were abandoned, and there appears to have
been some loss during the reconversion period in real output per
man-hour worked in private industries.
Rapid price rise until 1948
With the physical volume of production thus pressing against
capacity, much, if not most, of the pressure exerted by intensive
consumer, business, and foreign buying was reflected in price
movements. Prices were already advancing, though often in
covert fashion, in the early reconversion period. After the termination of wartime controls in the latter half of 1946, they spurted
up very sharply, and, except for a brief interlude of hesitation in
the spring of 1947, serious inflationary tendencies accompanied
the postwar boom until 1948.
In that year a better balance between supply and demand
emerged, and the price rise tapered off. This was brought about
partly through an appreciable expansion of real output and
partly through a diminution in the intensity of some of the demands, including those from abroad, from which the greatest
pressures had emanated.
There was a break in agricultural prices early in 1948. Although these recovered briefly, their downward slide, influenced
by the prospect of excellent domestic harvests and an improved
crop situation abroad, was resumed after midyear. Agricultural
prices are a substantial element in the total price picture, and
their decline was an important factor shaping the course of
economic developments during 1948.
More notable, however, was the increasing stability in consumer markets. The upsurge in personal consumption expenditures, stimulated by backlog demands and reinforced by large
holdings of liquid assets and a low volume of consumer debt
outstanding, had constituted one of the main foundations of the

Two Distributions of the Gross National Product




1. BY PURCHASES - SOLID LINES
2. BY RECEIPTS - DASHED LINES

L^ The sums of the receipts and expenditures are equal
0 But each part of the economy does not spend what it receives.
It saves a portion or draws on the savings of others
l) Shifting positions indicate the relative swings in
purchasing power and market demands
100
I
N
D
I

-

90

-

80

-

70

-

60

-

50

-

PERSONAL
OISSAVING

40

DISPOSABLE PERSONAL
INCOME

V
I
0

/

u
A
L
S

PERSONAL /

\

SAVING

\^

/

1

Siii

^ * "•"•'•<•

PERSONAL CONSUMPTION
EXPENDITURES
-

§
o

1
O
*o

8
-

B
U
S

I
N
S
S

-

GROSS
INVESTMENT
GROSS BUSINESS

§|k

\ _

SAVING

-%

20

EXCESS OF
INVESTMENT

\
V

30

S"~\

/ ^^^>-4
Y * ^ ^

i i i^ri

10

EXCESS OF , -

| | | 1 1AVI'N6I I V*f\

1 1 1 1 1 1 1 1 11

0

50
40

G
0
V
R
N
M
E
N

1

30

"o
c
o

a.
GOVERNMENT PURCHASES
OF GOODS AND SERVICES

20

I

O
"o

10
NET GOVERNMENT
RECEIPTS

1 1 1 1 1 11

1 I I I I. I I I

I I I I I I I I

1929 31 33 35 37 39 41 43 45 47 49

51 53 55

I

22

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Percentage Distributions of National Income by Distributive Shares, 1929-38
1929

Tetal national Income.

1930

1931

1932

1933

1934

1935

1936

1937

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

Compensation of employees
Wages and salaries
Supplements to wages and salaries.

58.2
57.4
.8

61.9
61.0

66.6
65.5
1.0

73.0
71.6
1.4

73.6
72.2
1.3

70.0
68.8
1.2

65.4
64.3
1.1

66.1
64.6
1.5

65.1
62.6
2.5

Income of unincorporated businesses.
Business and professional'._
Farm

16.8
10.0
0.8

15.2
9.8
5.5

14.6
9.3
5.3

12.5
8.0
4.5

13.9
7.9
6.1

14.3
9.3
5.0

18.2
9.4
8.8

16.1
10.1
6.1

17.2
9.6
7.6

Rental income of persons

6.2

Net interest

6.3

6.4

4.9

3.5

2.9

2.7

8.7
4.4
1.1
3.3
7.2
-4.0
4.3

2.7
-1.3

-4.6
-7.1
.9
-8.0
6.0
-14.0
2.5

-5.0
.4
1.3
-.9
5.1
-6.0
-5.3

2.2
3.5
1.5
2.0
5.3
-3.3
-1.3

5.1
5.5
1.7
3.8
5.0
-1.2
-.4

7.7
8.8
2.2
6.7
7.0
-.3
-1.1

8.4
8.5
2.0
6.4
6.4
.1

7.3

-

6.3

11.6
11.0
1.6
9.4
6.6
2.8
.5

Corporate profits and inventory valuation adjustmentCorporate profits before tax
Corporate profits tax liability
Corporate profits after tax
Dividends
_.
Undistributed profits.Inventory valuation adjustment

7.9

12.8

12.6

9.9

8.3

7.3

6.4

.8

-2.1
6.8
-9.0
4.0

1. Includes noncorporate inventory valuation adjustment.

Percentage Distribution of National Income by Sector of Origin, 1929-38
1929

National income

100.0

Government and government enterprises
Agriculture, forestry and fisheries
Rest of the world
Private nonagricultural industries

5.8
9.4

83.9

1930

1931

100.0
7.0
8.2
1.0
83.8

1932

1933

1934

1935

1936

1937

100.0

100.0

100.0

100.0

100.0

100.0

100.0

9.1
8.2
.9
81.8

12.1
7.8

13.3
9.2
.8

12.8
7.6

11.8
11.2

10.6

79.0

76.3

12.5
8.3
.6
78.7

9.8
.4
79.3

79.1

76.7

Percentage Distribution of Private Nonagricultural Income by Industrial Division, 1929-38
Private nonagricultural industries
Mining
Contract construction
Manufacturing
Wholesale and retail trade
Finance, Insurance and real estate
Transportation
Communications and public utilities
Services

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

2.8
6.2

2.6
5.0

2.0
4.5

2.0
3.1

2.1
2.5

3.0
2.8

2.8
3.0

3.0
3.9

3.3
3.6

28

29.7
18.1
17.2

_

28.7
19.3
16.7

25.4
19.9
17.7

21.4
18.9
20.1

24.6
17.8
18.7

28.2
20.8
14.6

30.5
21.1
13.6

31.7
20.7
12.9

33.1
20.9
12.4

28.4
22.6
14.5

100.0

3.8

9.0
3.9

8.8
4.4

8.9
5.4

95
.
6.8

9.9
6.6

8.8
6.7

8.5
6.2

8.4
4.9

7.9
4.7

7.7
5.1

14.0

_

14.5

16.1

18.2

18.1

16.1

15.3

14.6

14.1

15.0

Percentage Distributions of Gross National Product, 1929-38
1929

1930

1931

1932

1933

1934

1935

1936

1937

1938

By type of expenditure
Total

-

State and local

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

75.6
15.5
.7
8.1

77.9
11.3

80.4

84.3

16
.
.3

75.7
10.2

74.1
12.9

75.8

7.2
.3

-.1

.1

7.8
13
.

10.1

12.1

13.8

13.8

14.3

12.9

15.0

13
.
6.9

15
.
8.5

2.0

2.5

10.1

11.3

79.9
4.4
.7
15.0
4.6
10.4

77.6

.8

82.9
2.5
.3
14.4
3.6
10.7

4.0
9.7

5.8
8.5

5.0
7.9

6.2
8.8

100.0

Personal consumption expenditures
Gross private domestic investment
Net foreign investment
Government purchases of goods and services
Federal

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

79.6
11.0

81.6

83.7

83.2

80.0

78.2

77.1

96
.
-1.1

6.8
11
.
8.4
-.8
91
.

4.6
13
.

80.0
7.5
1.1
11.4
.2
11.2

80.4

.3
91
.
2.4

81.7
4.6
1.7
11.9
1.3

8.7
-.2

7.9
1.4

11.1

.5

100.0

By type of receipt
Total
Disposable personal income
Gross business saving'
Statistical discrepancy
Net government receipts
Federal'
State and local 3

67
.

1. Consists of undistributed corporate profits and corporate inventory valuation adjustment, capital consumption allowances, and excess of wage accruals over disbursements.
2. Consists of personal tax and nontax receipts, corporate profits tax accruals, Indirect




9.8
1.9
7.9

10.9
10.9

10.7

8.7
-.1

10.5

8.6

9.2

10.7

—. 3
13.5

13.2

1.6
9.1

4.8
8.7

3.9
9.3

.5

business tax and nontax accruals, contributions for social insurance, and current surplus of

f jvernment enterprises, less subsidies, transfer piyments, net interest paid, and grants to
tate and local governments

23

NATIONAL INCOME, 1 9 5 4 EDITION
Percentage Distributions of National Income by Distributive Shares,
1939

1940

1941

1942

1944

1943

1945

1946

1947

1948

1939-53

1949

1950

1951

1952

Line

1953

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

1

66.1
63.1
3.0

63.9
61.0
2.8

61.9
50.3
2.6

61.9
69.6
2.3

64.3
62.1
2.2

66.4
64.0
2.4

68.0
64.9
3.1

65.5
62.3
3.3

65.3
62.3
3.0

63.6
61.0
2.6

65.2
62.1
3.0

64.3
61.1
3.3

65.1
61.7
3.4

67.2
63.6
3.6

68.5
64.9
3.6

2
3
4

16.0
10.0
5.9

15.9
10.3
5.6

16.6
10.4
6.2

17.4
10.1

16.6
9.9
6.7

16.2
9.9
6.3

17.0
10.5
6.5

19.6
11.9
7.8

17.5
10.1
7.3

17.3
9.8

15.1
9.5
5.5

14.7
8.9
5.8

13.7
8.8
4.9

8.6

5
6

7.6

15.8
9.9
5.9

12.6

7.3

4.0

™

3.8

3.5

3.3

3.3

3.0

3.0

3.1

3.5

3.3

3.2

3.6

3.5

3.3

3.4

3.5

8

7.8
8.8
2.0
(S.8
5.2
1.6
-1.0

11.2
11.4
3.5
7.9
6.0
3.0
-.2

13.9
16.2
7.3
9.0
4.3
4.7
-2.4

14.3
15.2
8.3
6.9
3.1
3.8
-.9

14.0
14.4
8.3
6.2
2.6
3.5
-.5

12.6
12.8
7.1
5.7
2.6
3.1
-.2

10.2
10.5
5.9
4.6
2.6
2.0
-.3

9.6
12.6
5.1
7.5
3.2
4.3
-2.9

12.0
16.0
6.7
9.3
3.3
6.9
-3.0

13.8
14.8
5.6
9.1
3.3
6.9
-1.0

13.0
12.1
4.8
7.3
3.4
3.9
.9

14.6
16.7
7.4
9.2
3.8
6.4
-2.0

14.4
14.9
8.1
0.7
3.3
3.5
-.5

13.1
12.8
6.9
5.9
3.1
2.8

.3

12.6
12.9
6.9
6.0
3.1
2.9
-.3

9
10
11
12
13
14
15

6.3

5.5

4.3

3.1

2.1

1.8

1.8

1.7

1.9

2.0

2.4

2.5

2.4

2.6

2.8

16

Percentage Distribution of National Income by Sector of Origin,
1939

1940

1941

1942

1943

1944

1945

1946

1947

1948

1939-53

1949

1950

1952

1051

Line

1953

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

1

11.7
8.2
.4
79.7

10.7
7.7
.4
81.2

10.0
8.1
.3
81.5

11.9
9.0
.3
78.9

15.9
8.3
.2
75.6

18.5
7.9
.2
73.4

20.3
8.2
.2
71.3

12.6
9.7
.3
77.4

9.4
9.3
.4
80.9

8.9
9.4
.5
81.3

10.1
7.7
.5
81.7

9.8
7.2
.5
82.5

10.9
7.3
.6
81.2

11.8
6.4
.5
81.3

11.4
5.5
.5
82.6

2
S

6

Percentage Distribution of Private ^onagricultural Income by Industrial Division,

J
0

1939-53

100.0

100.0

lflO.O

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

2.7
4.0
30.9
21.6
13.7
8.1
4.9
14.3

2.8
3.9
33.7
21.6
12.4
7.6
4.6
13.4

2.7
4.9
38.7
20.2
10.7
7.4
3.9
11. i

2.4
6.0

2.1
4.2

2.2
3.1

2.1
3.3

2.1
4.7

2.6
5.3

2.9
5.7

2.5
5.9

2.5
5.8

2.5
6.0

2.2
6.1

2.2
6.0

41.8
18.7

45.2
18.5

44.9
19.2

40.2
21.7

34.9
24.7
10.4

36.8
23.4

37.0
23.1

7.4
3.4

9.6
7.2
3.2

35.5
22.9
10.7

37.5
21.9
10.4

39.0
21.3
10.0

38.0
21.5
10.3

38.6
20.8
10.5

10.2

11.3

12.4

11.9

11.4

12.0

11.5

11.0

11.3

11.5

Percentage Distributions of Gross National Product,

1939-53

1939

1940

1941

9.8
7.9
3.4

9.0
8.4
3.1
9.5

10.1

1942

1943

9.1
8.4
3.0

1944

9.9
8.2
3.3

1945

1946

1947

9.6
7.0
3.3

1948

6.8
3.7

1949

6.7
3.6

1950

6.6
3.7

1951

6.6
3.9

1952

6.4
4.0

7

8
9
10
11
12
13
14

Line

1953

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

1

71.4
13.1
1.5
14.0
6.1
7.9

65.1
14.4
g
19.7
13.4
6.2

56.4
6.2
—.1
37.5
32.7
4.8

52.2
2.9
—1.2
46.0
42.2
3.8

52.0
3.4
-1 0
45.7
42.1
3.6

57.0
4.9
.... «
38'8
35.0
3.8

70.1
13.0
2 2
14.8
10.0
4.8

71.0
12.8
3.8
12.3
6.8
6.6

69.0
16.0
8
14.2
8.2
6.0

70.2
12.6
.2
17.0
9.9
7.1

68.1
18.0
— 8
14.7
7.8
7.0

63.5
17.3
1
19.1
12.5
6.6

63.1
14.6

2

22.3
15.6
6.7

63.1
14.1
—.5
23.4
16.5
6.9

5
6
7

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

8

77.3
9.1
1.3
12.3
3.2
9.1

76.6
10.3
.8
13.3
4.7
8.6

73.9
9.1
.3
16.7

73.8
8.9
-.8
17.8
11.8
6.0

69.4
8.6
-.9
23.1
17.9
5.1

69.4
8.1
1.3
21.1
16.3
4.8

70.4
7.3
2 1
20!2
15.2
5.0

76.1
6.7
.4
16.8
11.0
6.7

72.8
8.6
.6
18.0
12.1
6.0

72.9
10.6
-.8
17.3
11.3
6.0

73.1
11.1
.0
15.7
9.0
6.8

72.3
10.0
.1
17.6
11.0
0.6

68.9
9.7
.4
21.0
14.6
6.J

68.4
9.9
.2
21.5
14.8

68.6
9.6

6.7

6.9

9
10
11
12
13
14

100.0
74.2
10.2
1 0
14.6
5.7
9.0

9.4
7.3

3. Consists of personal tax and nontax receipts, corporate profits tax accruals, indirect
business tax and nontax accruals, contributions (or social insurance, current surplus of




.3

21.5
14.6

3

government enterprises, and Federal grants-ln-aid, less transfer payments and net interest
paid.

24

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS
Selected Per Capita Income and Product Series in Current and Constant (1947) Dollars, 1929-38

Line

1929

1930

1931

1932

1933

1934

1935

1937

1938

Current dollars
Gross national product
National income
Personal income.
Disposable personal income
Personal consumption expenditures-

445
320
37fi
364

514
367
424
411
410

569
448
473
458
442

646
506
534
517
488

704
571
573
551
522

656
520
527
505
497

703
711

679
689

8S7
725
723

1,003
792
764

1,112
888

1,190
913

1,123
858
845

124,949

125,690

126,4S5

127,362

128,181

128,961

129,969

857
720
704
682
648

740
615
624
604
576

614
481
529
514
494

468
341
401
389
395

1,225
927

1,098
859
819

1,020
822
789

8M

121,881

123,188

124,149

Constant (1947) dollars
Gross national product
Disposable personal income'
Personal consumption expendituresAddendum: Population (thousands) >

1. This series was obtained by dividing current-dollar disposable personal income by the implicit deflator for personal consumption expenditures shown in table 41, Part V.

boom. As the more urgent backlog demands were satisfied, and
as the abnormally high spending rate of 1947 made inroads into
the net liquid asset positions of many consumers, the rising trend
of consumption flattened out in the latter part of 1948.
Closely allied with this tapering-off was the appearance of
substantial inventory accumulations, prevented in 1947 largely
by the intensity of consumer demand. These related developments, more than any others, slowed the price rise and brought
the inflationary spiral to an end during 1948. Late that year,
there came a general downturn in prices, and the postwar
economy entered a new phase.

Business readjustment and recovery: 1949—50
Businessmen adopted more cautious buying policies toward the
end of 1948, and the large inventory accumulations of that year
were sharply reduced in the first quarter of 1949. Substantial
inventory liquidation emerged in the next quarter, and the drain
upon stocks persisted during the remainder of the year. The
shift in the inventory position was reflected in a curtailment of
production mainly in the manufacturing industries, where the
bulk of all inventories held in the economy is produced.
By contrast, total final purchases—that is, elements of the
gross national product other than the change in inventoriesheld up extremely well during 1949. Consumer spending in the
first quarter dropped but slightly below its dollar volume at the
crest of the postwar boom, then climbed slowly upward again
during the remainder of the year. Residential building activity
decreased from a peak in the second quarter of 1948 but picked
up again in the spring and advanced strongly thereafter. And
government purchases, chiefly because of the expanding Federal
foreign aid and farm price support programs, more than offset
the moderate declines which occurred in business outlays for
plant and equipment. For the year as a whole, total final purchases actually exceeded those of 1948.
That the curtailment of employment and payrolls in the manufacturing sector had no greater impact upon consumption expenditures in 1949 was attributable in part to the payment of
sizable unemployment compensation benefits, and also in some
degree to the cushioning effects upon disposable personal income
of lower Federal income taxes as a result of the previous year's
Revenue Act. It may also be noted that dividends were sustained,
notwithstanding the sharp fall in profits. Perhaps more important



2. Continental United States including Armed Forces abroad, as of July 1.

than any of these factors, however, was the apparent willingness
of the consuming public as a whole to spend increasing proportions of current income to maintain living standards during the
recession.
It became apparent in the second half of 1949 that the curtailment of output had been excessive in relation to the existing
stable volume of business sales. Accordingly, production was
stepped back up, and the accumulation of inventories was resumed. Meanwhile, the recovery of residential construction had
grown into a sustained building boom, and consumer demand,
already strong, was being bolstered by large Government payments to veterans. These factors, moreover, were being reinforced by a renewed upturn in fixed business investment.
This widening resurgence of production generated increases in
employment and incomes, adding further impetus to consumer
purchasing. Before mid-1950, a business upswing of substantial
dimensions was under way and was carrying the economy close
to full-capacity operation.

Korean war period
It was upon this expansionary situation that the urgent demands resulting directly and indirectly from the outbreak of
hostilities in Korea were superimposed, as the Nation decided
upon a defense program encompassing a sharp step-up in direct
military potential as well as the establishment of a broad base of
productive facilities to permit quick economic mobilization in
case of full-scale war.
The ensuing rise in national output continued without interruption for three years, until the second quarter of 1953, when
gross national product reached a peak of $370 billion at seasonally
adjusted annual rates, one third above the rate of $276 billion in
the second quarter of 1950. About half of this rise represented
expansion in physical volume while the remainder reflected the
sharp advance in prices which occurred in large part during the
first nine months of the period.
In terms of current dollars, Federal Government purchases,
mainly for national security purposes, accounted for more than
two-fifths of the increase in national output. These purchases
tripled over the three-year period, and their share of total output rose from 8 to 17 percent. All other types of national expenditure combined rose by about one-fifth, with consumption, invest-

25

NATIONAL INCOME, 195 4 EDITION
Selected Per Capita Income and Product Series in Current and Constant (1947) Dollars,
1939

1940

1941

1942

1943

1944

1945

1946

1948

1947

1949

1939-53

1950

1951

1952

1953

Line

696
686
666
538
516

762
618
596
576
544

943
785
722
697
614

1,180
1,021
916
871
C65

1,408
1,246
1,107
977
735

1,527
1,320
1,197
1,060
794

1,526
1,295
1,224
1,075
870

1,480
1,270
1,259
1,126
1,037

1,611
1,368
1,322
1,173
1,145

1,755
1,512
1,424
1,279
1,211

1,725
1,449
1,386
1,261
1,211

1,879
1,582
1,497
1,359
1,279

2,126
1,795
1,654
1,465
1,350

2,204
1,853
1,727
1,509
1,391

2,286
1,911
1,792
1,567
1,441

1
2
3
4

1,202
926
888

1,299
981
927

1,486
1,113
981

1,658
1,245
960

1,820
1,277
961

1,938
1,312
982

1,880
1,282
1,038

1,654
1,247
1,149

1,611
1,173
1,145

1,663
1,211
1,146

1,619
1,203
1,165

1,745
1,280
1,205

1,833
1,290
1,189

1,874
1,306
1,205

1,921
1,339
1,282

6
7
8

131,028

132,122

133, 402

134,860

136, 739

138,397

139,928

141,389

144,126

146,631

149,188

151,683

154,360

157,022

159,629

9

ment, and State and local government outlays each registering
substantial gains.
In real terms the shifts in the use of output were more pronounced, with the Federal Government absorbing about twothirds of the increase in constant-dollar gross national product.
But on this basis also a margin was left which permitted an expansion in other purchases in spite of the increased demands of
national security. Total real consumption was about 9 percent
higher in mid-1953 than three years earlier, reflecting mainly a
larger volume of nondurable goods and services. On a per capita
basis the increase amounted to more than 3 percent. Real purchases of State and local governments showed a similar movement.
The rate of aggregate investment decreased somewhat in real
terms, although fixed investment was higher, due to a larger
volume of producers' durable equipment.

Inflationary upsurge
While the defense buildup was the dominant factor in the
economic situation throughout the three years following mid-1950,
its character and impact varied considerably during this period.
In the initial stage the increase in national security expenditures
was heavily concentrated in military payrolls, food, clothing, and
other related outlays which were capable of rapid expansion in a
comparatively short time. The strength of the Armed Forces increased by \% million persons in the first year of the defense
effort-—about four-fifths of the ultimate total increase—and
national security outlays mounted from 6 percent to 11 percent
of the gross national product.
At the same time there was a strong advance in investment by
businesses that were participating in the growing volume of
defense orders. This included investment in inventories and in
plant and equipment, the latter receiving a special stimulus from
the rapid amortization provisions that had been granted for
emergency facilities.
Although the demands stemming directly from the defense program contributed to expansionary developments, it was the waves
of anticipatory buying, set in motion by the expectation of commodity shortages and price increases, that determined the tone of
the business situation in this period. Consumer buying spurted in
the third quarter of 1950 and again in the first quarter of 1951,
following the further extension of the Korean conflict. Responding




5

to sharply rising sales, as well as to the same expectations that
motivated consumers, businessmen added greatly to their inventories, and fixed investment not related to the defense program
was also stimulated.
Although production continued to rise briskly, demand outstripped the supply of goods, and prices bounded upward.

Shift to defense production
By mid-1951 the situation had changed significantly. Output
was expanding at a rapid rate, and it became apparent that the
productive capacity of the economy had been underestimated.
Taxes had been raised, and the establishment of price and wage
controls gave some assurance to the public that inflation would be
kept in check. Other measures, including controls on credit and
on the flow of strategic materials, had been introduced to ensure
that the necessary resources would be channeled in an orderly
fashion into the defense program.
As a consequence of these developments, and the well-stocked
position of consumers after several months of extraordinarily
heavy buying, spending propensities eased markedly during 1951.
The rate of consumer spending out of disposable income, which
had risen to 96 percent during the buying surge in the first
quarter fell to about 91 percent in the succeeding quarters of the
year. Since the trend of income was strongly upward, the absolute dollar value of consumption in the final quarter of the year
exceeded the first-quarter rate in spite of this decline in the spending rate.
The easing in the urgency of consumer demand was paralleled
by a curtailment of business buying, which had been geared to
the same sort of anticipations and to the resultant abnormal
volume of sales. The shift entailed a very substantial inventory
readjustment—with a notable scaling down in the accumulation
of civilian goods.
In the setting of rapidly increasing physical production, of
Government measures designed to minimize the disruptive effects
of economic mobilization upon the economy, and of less urgent
private demand, it was possible to achieve a substantial expansion
in national defense expenditures and allied fixed capital outlays
concurrently with a high rate of civilian consumption and a
diminution of inflationary pressures.
The absolute increase in national security expenditures in the

26

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

second year of the defense buildup was nearly as large as in the
previous year although the percentage increase was smaller. It
was concentrated in hard goods, such as planes, tanks, and
electronic equipment. Deliveries of these items were reaching a
high volume, after design difficulties had been overcome and the
long lead-time involved in their production had elapsed. Defensecennected fixed investment also expanded rapidly in this period.
In contrast, the advance in total fixed investment was limited by
offsetting declines in investment not related to the defense effort.

General expansion
In the third year of the defense effort—from mid-1952 to mid1953—the Nation's capacity to produce continued to expand
substantially while the rate of growth of national security outlays
tapered off. This shift in the relation of supply and demand permitted the relaxation and subsequent removal of economic
controls.
Civilian demand proved strong, particularly in the durable
goods lines previously restricted by controls and material shortages.
Residential, institutional, and State and local construction,
business investment in plant and equipment other than in defenseconnected industries, and consumer and inventory demand for
automobiles and other durable goods became major factors supporting further economic growth. Together with the still rising
military demand they provided the basis for a balanced economic
expansion which, unlike that of the previous year, was shared by
most industries.
Agriculture was the principal sector of the economy that ran
counter to the generally favorable trend. As a consequence of
lower farm prices, income originating in agriculture was reduced
substantially during 1952 and 1953. The steady downward
pressure on prices reflected the unusually heavy output of farm
products—which reached successive highs in the 2 years—and
the appreciable decline in foreign sales. Domestic consumption
remained firm, however, and the full force of the other develop-




ments was not reflected in farm income, as large quantities of
the chief crops and dairy products were placed under loan to the
Commodity Credit Corporation, both in 1952 and 1953.
The restoration of civilian durable goods production was retarded by the steel strike which occurred in the second quarter of
1952 and affected end products mainly in the third. As a consequence final purchases and inventory accumulation of durables
were particularly heavy in the fourth quarter of 1952 and were
probably raised also in the opening half of 1953.

National output reduced
Total production reached a peak in the second quarter of 1953
and receded moderately in the latter half of the year. The major
change occurred in the net flow of goods into the inventories of
durable goods industries. These inventories had been replenished
and the immediate need for further accumulation had thus been
removed. In addition, sizable liquidations were made when it
became apparent that, relative to actual and prospective demand,
some of the earlier build-up had been excessive.
National security expenditures, entering a new phase, registered
moderate declines, mainly in the procurement of hard goods,
and consumer purchases of commodities, particularly durables,
also drifted downward. The effect of these reductions on total
final purchases was offset, however, by advances in other components of national expenditures, notably consumer expenditures
for services, State and local government purchases, and the agricultural price support outlays of the Federal Government. Hence
final purchases were maintained in the aggregate in the latter
half of 1953.
While the percentage decline in total national output was small,
it was heavily concentrated in durable goods and thus had a
disproportionate impact upon the durable goods manufacturing
industries where it led to substantial reductions in production,
employment, wages, and profits. The year 1954 opened with these
developments still in progress.

PART

II

The Conceptual Framework
of National Income Statistics
In the past two decades national income statistics have undergone a basic transformation. This is manifest in the considerably
broadened scope of the field.
The traditional purpose of national income research is to provide information on the outcome of economic activity through
comprehensive measures of the size, composition, and use of
national output. In the more recent period, the measurement of
national output has continued to be the basic aim. But, with the
growing realization that they can furnish a statistical picture of
the economic structure and process, national income statistics
have been used also to an increasing extent to facilitate an understanding of the factors which determine the outcome of economic
activity. Much more fully and systematically than in the past,
national income statistics have been designed with the dual objective of measuring the national output and placing it against
the background of the transactions which underlie its production
and distribution.
The national income statistics for the United States contained
in this report are constructed according to this broader plan.
They are therefore a comprehensive single source of integrated
information on the Nation's economic life.

NATURE AND SIGNIFICANCE OF THE
INCOME DATA
Because they reduce the voluminous detail of economic activity
to intelligible proportions, national income statistics have become
widely used as the factual background for economic analysis and
the preparation of economic programs. They provide the basic
statistical framework required for the study of long-term economic trends and of business fluctuations, and for the formulation
of business and government economic policies. Needless to say,
the statistics do not throw light on all aspects of the economy,
and often must be supplemented by other bodies of economic
information.
291692°—54

3




Two broad, practical uses of national income data may be
cited. These data are needed, in the first place, when the automatic working of the market mechanism cannot be fully relied
tipon. The mitigation of business cycles and economic mobilization for national defense are important instances in which an
understanding of the economic mechanism, such as is facilitated
by the use of national income statistics, is a prerequisite to intelligent action.
Secondly, even when active influencing of the broad course of
economic events is not the aim, it is desirable to have some knowledge of these events so that the best possible adjustments to them
can be made. For example, the businessman wants to gauge the
probable market for his output so as to obtain a more rational
basis for determining his policies; and the tax administrator must
estimate what governmental revenues are likely to be so that intelligent decisions can be made about matters relating to the expenditure and revenue policies of the various levels of government.
Whether for the purpose of exerting active influence on economic events or for passive adaptation to them, national income
data are the most important single statistical tool for orientation
in the economic world. They do not, of course, furnish direct
answers to the economic problems involving their use, but they
do provide the relevant, and often indispensable, statistical background for arriving at intelligent solutions. This statistical background consists of a quantitative description of the structure of the
economy over a period of years. The framework of this description is a national economic accounting system which summarizes
the transactions linking the economic units whose interplay determines the functioning of the economy.

Economic accounting system
The production and distribution of the Nation's output necessitate countless transactions of buying and selling, hiring labor,
investing capita], renting property, paying taxes, and other operations inherent in the functioning of the economic system. The
27

28

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

records of these transactions kept by the business, consumer, and
governmental- units participating in them obviously are highly
relevant for obtaining a statistical view of the economy because
they reflect the most concrete manifestations of the Nation's economic life. However, these innumerable records must be summarized into a limited number of significant categories if a comprehensible and useful description of the economic process is to
emerge. This is the basic task of the national economic accounting
system.
The plan of the accounting system underlying the United States
estimates is based upon a division of the economy into four major
sectors—business, consumers, government, and foreign. The economic behavior and motivation of these four sectors is quite different; to distinguish among them appears necessary for an understanding of the economy in terms of the interactions of its constituent parts.
In the construction of the economic accounting system, a national income and product account is first established. This account provides measures of total national output, which is the
sum of the outputs produced by the four sectors of the economy.
Next, accounts are set up for the sectors. In addition to showing
the portions of national output originating in each of them, they
are designed to depict the economic structure in terms of the interrelated transactions of the four major economic groups.
Specifically, four current accounts are shown, one each for
business, consumers, government, and the rest of the world. These
trace the transactions determining the current income of each of
the sectors, and what part of that income is used up and what
part is devoted to saving. The sector account for business is in
essence a consolidated profit and loss account for the business
system as a whole. For the other sectors, the accounts represent
current receipt and expenditure accounts, in conformance with
the nonprofit-making character of thdr transactions.
Most of the current transactions that appear in the account of
one sector are matched by corresponding entries in another.
However, this is not so with respect to the items of saving or investment. With these, the corresponding entry is found in the
capital or gross saving and investment account, which shows on a
consolidated basis the saving and investment for the economy as
a whole. This is the sixth account in the national economic accounting system.

Main advantage of accounting approach
The principal advantage of formulating and presenting national income statistics as a system of accounts has been intimated
in the preceding discussion. Such a system yields a set of interrelated tables which are a tremendous aid in revealing the structure of the economy and thereby contribute toward a better understanding of its functioning. Two aspects of the analytical value
of the accounting system may be considered.
It throws into clear relief the nature of accounting relations that
must always hold true among the component transactions summarized. The sense in which saving and investment are necessarily
equal is a prime example of such a relationship. The establishment
of an economic accounting system displaying this and other accounting relationships has been an aid to simplicity and clarity in
economic discussions.
Also, light has been thrown on the relative magnitudes of the



component flows of the economic process, and the study of the
functional relationships among them has been facilitated. In contrast to the accounting relationships, which are a matter of definition and must always hold, these functional relationships are
regularities that hold by and large as a matter of economic experience, but which can and do change in response to technological, institutional, and psychological changes. Measurement and
study of these relationships—such as those between consumption
and disposable income and between wages and profits—are
essential for an understanding of the working of the economy.
However, because they cannot be counted upon to hold without
fail, these relationships must be the object of continuing investigation.

Technical uses of economic accounting system
The establishment of a system of national economic accounts
has benefited also the producers of national income data. It has
aided their work in both its theoretical and statistical phases.
With respect to the former, it must first be recognized that national economic accounting has been of some aid in improving the
definition of national output. It is true that no genuinely new
criteria for solving definitional problems have been provided, and
also that many definitions of national output are compatible with
the principles on which the system is based. Yet, it has helped the
discussion in several ways.
A great deal of the discussion of definitions was obscured by the
failure to distinguish clearly between the income and product
measurements of output and by the lack of a clear grasp of the
relation between them. The development of national income statistics in an economic accounting framework has made for clarity.
Some of the larger issues involved in the definition of output were
brought into better focus, and a powerful tool was provided for
the consistent treatment of financial intermediaries, nonprofit
institutions, imputed income and product, and similar problem
areas in the formulation of national income concepts.
Economic accounting has contributed to problems of definition
also by depriving them of some of their importance. In substantial
part, these problems revolve around the question of whether or
not certain items—such as government interest, business taxes,
transfer payments, and subsidies—should be included in the aggregate measures of national output. Prior to the establishment
of the system of accounts, the decision to omit such moot items
from total output meant that their record, insofar as national income statistics were concerned, was lost. Since the items are germane to economic investigations in which national income data
are used, there often was reluctance to exclude them, even though
this may have been indicated from the standpoint of measuring
output. With the presentation of national income statistics in the
form of a complete statistical picture, the record of transactions
excluded from the principal aggregates is no longer lost. The problem of defining output can be faced squarely on its own merit.
Moreover, to the extent that fully satisfactory solutions to
definitional problems cannot be found, the economic accounting
system has made it easier to live with them. Many of the controversial items (irrespective of whether defined as part of the output
aggregates) are shown separately in the account tables, and alternative measures of output can be constructed depending on
particular needs and preferences.

NATIONAL INCOME, 1 9 5 4 EDITION

In two general ways, the accounting approach is an aid to the
statistical aspect of national income work. To begin with, it is of
considerable help in defining the task of statistical data collection.
Once the particular accounting framework providing the most
useful summary of the economic structure has been decided upon,
a comprehensive list of requirements for economic statistics
emerges rather automatically. A list obtained in this way provides
a useful guide for planning the collection of primary statistical
data so as to yield the information most relevant to economic
analysis.
The use of the accounting approach also facilitates the estimation of the various national income aggregates and their components from the available statistical material. It does this by making
clear that many items of information can be obtained from the
records of either the buyer or the seller, and hence affords flexibility in adapting estimating methods to available information.
In addition, this approach enables one to check every account
for internal consistency by comparing the debit and credit totals
as well as the relations among the various debit and credit entries.
It also enables one to derive as residuals components of the national economic accounts which cannot be estimated directly
from available data.

Improvements of the accounting system
In the derivation of the definitions and classifications used in
the national economic accounting system, an attempt is made to
set forth the distinctions that are most meaningful from the standpoint of economic analysis, taking account of the limitations imposed by the nature of the accounting data available for the four
major economic groups. Fortunately, there is a great deal of
parallelism between the requirements of economic analysis and
those of the accounting systems used by business and other economic units. Over wide areas no conflict arises from the fact that
economic questions have to be answered by reference to measures
constructed from such accounting data. On the contrary, a major
advantage of the system of national economic accounts is that it
summarizes the actual transactions of economic units as reflected
in their own accounting records.
However, some of the most difficult problems of national income estimation arise when the definitions underlying these
records do not yield the type of information demanded by economic analysis. National income work is continually concerned
with the modification of the basic accounting data in order to
improve their economic significance. Often these data can be
adjusted to meet the requirements of economic analysis, but when
the transition cannot be accomplished supplementary information must be introduced to complete the picture.
Comprehensive national economic accounting is a recent development the potentialities of which have not yet been fully
realized. The set of accounts presented in this report should not
be regarded as the definitive system. Apart from possible improvements in the formal design of the accounts, several elaborations
of the present system would be desirable.
For instance, only four major economic groups are distinguished, whereas, in view of their heterogeneity, further breakdowns would be useful for many types of analysis. Also, saving
and investment accounts for each of the four major sectors would
constitute an important supplement to the consolidated account



29

for the economy as a whole. The construction of balance sheet
accounts, showing the structure of the assets and liabilities of the
various sectors, likewise would expand the scope and usefulness
of the national economic accounting system.
It is to be emphasized, however, that further expansion of the
national accounts must be made with due regard to the flow of
statistical information (which would constitute a generally limiting factor) and to the danger of an overelaboration that might
add unduly to their complexity and not proportionately to their
value.
Coordinate in importance to further work on the conceptual
framework, articulation, and coverage of the economic accounting system is the improvement of its statistical reliability. For the
entries in the national economic accounts represent estimates
which are subject to error. The problem of statistical reliability
is discussed in Part III of this report.

The detailed statistics
In the preceding discussion the main emphasis was on the summary aspects of the economic accounting system underlying
United States national income statistics. However, sight should
not be lost of the wealth of statistical information that now exists
to elaborate and supplement various aspects of this accounting
system. For in many uses of the data it is specifically this information which is of primary interest and value.
Attention may be drawn first to the many statistical tables in
Part V of this report, of which the six national account tables
briefly described above are but highly condensed summaries.
These detailed tables present further information on the breakdowns of the income flow by type of income and legal form of
organization and of the product flow by type of product. Also
given are breakdowns of national income and its constituent
distributive shares by industry of origin.
Secondly, the conversion of gross national product and its components into constant dollars, which is presented in Part IV,
represents an important addition to the current dollar series in
terms of which the complete accounting system is stated.
Thirdly, the annual estimates of personal income by States, not
included in this report, may be regarded as the elaboration in a
regional dimension of the depiction of the economic structure.
Of a similar nature, as also involving further articulation of the
consumer sector, are the estimates of the size distribution of income prepared in the National Income Division.

Plan of the following discussion
In the following pages of this Part of the report, the conceptual
framework of the United States national income statistics is explained in greater detail. Since the measurement of output totals
is the prime objective of national income statistics and, moreover, is largely independent of the full-fledged economic accounting system depicting the economic structure, the derivation of
these totals is first explained, in a summary manner. Next, the
structure of the complete accounting system is developed. In the
course of this discussion, the more detailed aspects of the definitions of the output totals are also covered. A final section provides, for convenient summarization, a series of definitions to
which the national income and product aggregates and their
components conform.

30

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

SUMMARY CONSTRUCTION OF NATIONAL
OUTPUT MEASURES
In this section the basic notions underlying national income
and product are stated; the derivation of these measures in terms
of their conceptual content is explained; and the adequacy of
definition of the resulting aggregates is examined.

Basic Notions Underlying National Output
Measurement
Economic production
In the definition of a measure of national output, the first task
is to delimit economic production from the pursuit of other activities that resemble it in that they involve the use of human effort
and other resources and are useful. For instance, the production
of radio sets has its counterpart in the hobbies of the radio amateur, commercial shaves are akin to self-administered ones, and
the educational services of teachers often are supplemented by
those of parents. In spite of resemblances, a distinction must be
drawn between economic production and noneconomic pursuits.
For a measure of national output must, broadly speaking, be confined to the former; it cannot, in any systematic way, take account
of activities outside the economic sphere.
In the present report, the basic criterion used for distinguishing
an activity as economic production is whether it is reflected in the
sales and purchase transactions of the market economy. The exclusion of illegal transactions is a tradition-based convention
which is an exception to this general rule.

Product and income flows
A fundamental distinction relevant to the measurement of economic production so delimited is suggested by observation of the
operations of a typical business firm. On the one hand, such a
firm produces and sells a flow of product values. On the other
hand, it pays out (or retains) incomes that accrue in the course
of its operations. This double aspect of the activities of the single
business firm suggests that the measurement of national output
can be approached in a two-fold manner, either by summing
product values or by summing income flows. It will be seen that
the measure of national output in terms of product flows which is
obtained by pursuing this approach is the gross national product
and that the corresponding measure in terms of income flows is
the national income.

Final and intermediate products
In the measurement of national output via product flows, a
further distinction, between "final" and "intermediate" products,
must be made. A nonduplicative total is desired, one that is confined to the value of the final, or end, products of the economy and
excludes all others, labelled intermediate. To use a simple example, if the production process during a year involves the production of wheat, its milling into flour, and the baking of bread



which is sold to consumers, then the value of national output
should equal the full value of the bread and should not count
also the separate values of the wheat and flour which have been
used up in the course of producing it. This result is obtained by
counting only the value of the bread, as the end product, and
ignoring the other product values.
A distinction between final and intermediate products cannot
be drawn on the basis of the technical characteristics of the output involved. In the above example, for instance, flour is an intermediate product. If, however, the flour is sold not to bakeries,
but directly to housewives for home baking, it becomes the final
product of the economy, even though in a technical sense it is not
fully fabricated.
However, an effective criterion for distinguishing between final
and intermediate products can be established by reference to
business practices followed in the production of goods and services. There emerges a working definition of final product as a
purchase that is not resold, and of intermediate product as one
that is resold. A more technical, but sometimes more convenient,
phrasing of the same idea is that a final product is a purchase
that is not charged to current cost whereas an intermediate
product is one that is so charged. The phrase "during the accounting period" is sometimes appended to these formulations so as to
make them more exact.1

Imputations
In the measures of national output shown in this report, the
foregoing criteria are the basic tools for distinguishing economic
production from noneconomic pursuits and the part of economic
production which is final from that which is intermediate. However, modifications in the definitions are made in certain instances
to enhance the significance of the measurements.
The most important of these modifications concern the inclusion
in national output of the so-called "imputations," or items of production and income "in kind." For instance, food furnished to
employees would not become part of the national output if the
initial definition were rigidly followed. It would be an intermediate product, since it is an element of the current cost charges
of the employer furnishing the food. However, it seems desirable
to count it as part of national production, if only to secure uniformity of treatment with respect to employees who buy food out
of the correspondingly higher money wages given them. Other
imputations that are made in measuring national output are for
the value of food produced and consumed on farms, the rental
value of owner-occupied houses, and for nonmonetary income
and product flows arising in connection with financial intermediaries.

Charges against final product
In this report, the product measure of national output is derived by adding the values of final products and omitting intermediate products, as in the bread and flour example. It is termed
1. In order to simplify the discussion, changes in business inventories are not taken into
explicit account in this section. But it should be noted that, according to the above definitions, an increase in inventories is a positive component offinalproduct and a decrease in
inventories a negative component. (A detailed explanation of the treatment of inventories
in the national accounts is given later in this Part.)

NATIONAL INCOME, 195 4 EDITION

the gross national product. However, the same total could be obtained also by adding in the first instance the total product (final
and intermediate) of each producing unit and then deducting
for each the intermediate product it bought from other units.
For the economy as a whole, purchases and sales of intermediate
products would cancel, and what would remain would be the
value of final products.
The total value of final products can thus be broken down into
elements consisting of the total product of each producing unit
less its purchases of intermediate products. However, for each
producing unit the difference between the value of its product
and its intermediate purchases consists of the incomes that accrue
in the course of production (wages and salaries, interest, profits,
etc.) plus certain "nonincome" charges against the value of its
production, the most important of which are taxes (such as property, excise, and sales taxes) and depreciation charges for the
wear and tear and obsolescence of fixed capital.
Thus, since (1) the value of the final product of the economy
equals the sum of the total product of each producing
entity less its purchases of intermediate products, and (2)
for each producing entity, product less intermediate purchases
equals income plus nonincome charges against the value of production, it follows that (3) for the economy as a whole the total
value of final product equals the sum of incomes accruing in production plus nonincome charges against the value of production.

31

mation, and hence of net national product, cannot be calculated
since depreciation charges are not available on a basis of valuation comparable to that of the gross production of fixed capital.

Personal and disposable income
Another aggregate, personal income, measures the actual current
income receipts of persons from all sources. It differs from the
national income in that it excludes certain types of income which
accrue in production but are not received by persons (for instance,
the undistributed part of corporate profits) and, on the other
hand, includes certain types of income which do not arise in current productive activity but constitute personal receipts (such as
relief and unemployment benefits). Hence personal income, unlike the national product and national income aggregates, is not
a measure of national production. Since the bulk of personal
income is derived from production, however, it can in ordinary
circumstances be used as an indicator of productive activity and
has, in fact, gained prominence in this use as the only comprehensive income or product total available on a monthly basis.
Personal income net of taxes—disposable personal income—is
another useful aggregate, being the closest overall statistical approximation to consumer purchasing power derived from current
incomes.

National Income and Product Account
Factor incomes and other charges
A measure of national output in terms of total charges against
the value of gross national product, and therefore numerically
equivalent to it, can thus be obtained. However, this is not the
income flow measure generally used. The common measure,
national income, is derived by emphasizing the distinction, which
has already been emerging, between two types of charges against
the value of final product: factor costs and other costs.
Briefly, the sum of employee compensation, interest, and business incomes is considered to represent the remuneration of factors of production. In national income terminology, their aggregate measures the total factor cost incurred in producing the output of the Nation. By contrast, no factor incomes correspond to
the other charges against gross national product. Indirect business
taxes do not form the cost or income of any factor of production.
Depreciation and kindred charges reflect allowances for the consumption of fixed capital, and not its net income or return. Thus,
a further total is distinguished—the national income—which
represents the sum of factor incomes or factor costs. This is the
income measure of national output most widely used.
Two other magnitudes can be derived by recombining the
components so far discussed. Net national product may be obtained
as gross national product less depreciation and kindred allowances for the consumption of fixed capital. A corresponding
measure of total charges against net national product may also
be obtained by deducting these allowances from total charges
against gross national product. These two aggregates are in a
theoretical sense more clearly defined than the corresponding
measures of gross national output, since some duplication is involved by the inclusion in the latter of the production of fixed
capital which serves merely for replacement purposes. However,
as a practical matter, a fully satisfactory measure of net capital for


The foregoing basic notions underlying national output measurement are incorporated in the National Income and Product
Account shown in table I, which presents for 1950 the several
alternative measures of United States output.
The right side of this account shows the gross national product
as the sum of final product flows. The summary listing includes
personal consumption expenditures, gross private domestic investment (consisting of new construction, purchases of producers'
durable equipment, and the change in business inventories), net
foreign investment (reflecting, in general, net purchases by foreign nations), and government purchases of goods and services.
These components represent purchases of the four major sectors
into which the economy has been divided and conform to the
operational definition of final products as purchases not resold
during the accounting period. However, certain modifications
in this definition, especially with respect to imputed income and
product, have been made in deriving the national income and
national product measures.
On the left side of the account are listed the charges against the
value of gross national product. In principle, the sum of these
charges should numerically equal the value of gross national
product. However, because of statistical estimating errors, the
nature of which is discussed in Part III, this is not actually the
case. To secure balance, an item termed "statistical discrepancy"
is entered on the left side of the account.
Total charges are broken down into factor costs and other
charges. The former, consisting of employee compensation, net
interest, and the various types of business incomes, corporate and
noncorporate, are added to obtain the national income. The other
charges against the value of gross national product are arranged
so as to permit a further subtotal of charges against net national
product.

32

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Coverage of National Income and Product
Account
The economy covered by this account, and hence by the various income and product aggregates, is the continental United
States. Thus it does not coincide with the customs area of the
Nation since territories and possessions are excluded. Also, it is
important to note, the account measures the income and product
attributable to factors of production supplied by residents of the
country, rather than the income and product of factors physically
located in the country.
Not only individuals who contribute their labor and property
to the productive process, but nonprofit institutions and governmental bodies supplying capital resources are viewed as
residents supplying factors of production.2

Anatomy of the Output Totals
The foregoing discussion has attempted to give a general notion
of the major concepts underlying the measurement of national
output and of the nature and relation of the major alternative output measures. However, it has passed over much that is essential
to a precise understanding of the output concepts.
National income and national product are comprehensive
measures of total national output. They include not only business
production, but also production contributed by the nonbusiness
sectors of the economy—households and institutions, government,
and the rest of the world. Measurement of output in each of these
sectors is subject to special problems of its own. The principles
underlying the measurement of production in the various sectors
are next explained. The sector measurements based on these'
principles are then added to obtain the national income and
product totals shown in table I.

Measurement of business output
The bulk of national output originates in the business system,
and the framework adopted for the measurement of business output sets a pattern for the whole. In deriving national output as
the summation of the outputs originating in the several sectors
of the economy, it is therefore convenient to start with the business sector.
The Consolidated Business Income and Product Account, presented in table II, shows the portion of national output originating
in the business system. Business output is measured in terms of the
concepts underlying the major income and product aggregates
covered in table I and discussed in connection with it.
2. The meaning of the term "resident individual" is largely self-explanatory and agrees with
the definition adopted for balance-of-payments-purposes by the International Monetary
Fund. The classification of certain special categories may be noted explicitly. United
States Government employees, civilian and military, whose usual residence is in the United
States are considered United States residents even when stationed abroad. Residents of
foreign countries hired abroad by the United States Government are not considered United
States residents. This treatment applies symmetrically to employees of foreign governments
in the United States. In principle, foreign border workers working in the United States are
not regarded as United States residents. United States residents whose place of work is
across the border are so regarded. As explained in section 12 of Part III, however, the statistical information for implementing the definition of resident with respect to border workers
is lacking. Also, there are marginal oases in which it is difficult to decido whether an individual is to be classified as a resident or as a foreigner. In practice, these cases are quantitatively unimportant.



The right (or credit) side of the account shows the market value
of the consolidated production of the business system. On the left
side of the account appear the charges against this production.
The two column totals are equal in principle, for reasons which
have been stated in broad terms.
The nature of the equality can be understood in an alternative,
and perhaps more precise manner, if it is realized that the business
income and product account is similar to a profit and loss account
for the business system as a whole from which intrabusiness sales
and purchases on current account and intrabusiness flows of interest and dividends have been eliminated by a process of consolidation. The two sides of such an account must always balance
because profits are derived as the residual of sales and costs.
However, for statistical reasons equality is not achieved in practice, and in the estimates embodied in table II the "statistical
discrepancy" is entered on the debit side as a reconciliation item.
In addition to consolidation, certain other operations have been
performed to transform the column totals of the profit and loss
account into a measure of business output. Capital gains and
losses have been eliminated as not reflecting the value of current
production. Inventory change has been entered on the right side
of the account to convert sales into a measure of production, and
subsidies have been transferred to the left side as not being part
of the market value of the products shown on the right side.
Further, certain items have been netted, also with the aim of obtaining column totals that measure the contribution of the domestic business system to national output.
In the first place, imports, which in a consolidated statement of
domestic business would appear on the left side of the account,
have been transferred to the right side and netted against exports
(under consolidated net sales to abroad). Imports must be deducted from the credit side since they reflect foreign production
included in the value of the column totals prior to the deduction.
The fact that, in table II, they are netted against the export item
is merely a matter of convenience.
The second netting which is made in table II is in connection
with property income flows. In a consolidated account not further adjusted, receipts of interest and dividends from other sectors would appear on the credit side and would be reflected correspondingly in the total of the charges entered on the debit side.
Since receipts of interest and dividends from other sectors do not
represent output of the business system, it is necessary to remove
them from the credit side of the account and to net them against
elements of factor income in order to make the sum of factor incomes on the debit side of the account reflect the factor cost of
business output. In practice, interest received is netted against
interest paid under the heading "net interest," and dividends received are netted against dividends paid under the heading
"dividends." In these instances also, the particular matching
adopted is essentially a matter of convenience.
The components of business production listed on the credit side
of table II conform to the definition of final product specified in
connection with table I. It may also be noted that the entries are
very similar. Differences arise because table II is confined to
business output, whereas table I provides a summary of national
output which covers the nonbusiness sectors as well. Similar comments apply to the charges against the value of business product
listed on the debit side of the account. Factor income charges

NATIONAL INCOME, 1 9 5 4 EDITION

33

Table I.—National Income and Product Account, 1950

[Millions of dollars]
Compensation of employees:
Wages and salaries
Supplements

146, 526
7, 799

Income of unincorporated enterprises and inventory valuation adjustment
36, 140
Rental income of persons

8, 473

22

Personal consumption expenditures

194, 026

23

Gross private domestic investment

51, 219

24

Net foreign investment

25

Government purchases of goods and services

26

GROSS NATIONAL PRODUCT

— 2, 201
42, 023

Corporate profits and inventory valuation adjustment:
7
8
9
10
11
12

Corporate profits before tax:
Corporate profits tax liability
Corporate profits after tax:
Dividends
Undistributed profits
Inventory valuation adjustment

13 Net interest
14 National income

17, 829
9, 207
12, 934
—4, 864
5, 912
239, 956

23, 741
15 Indirect business tax and nontax liability
843
16 Business transfer payments
215
17 Statistical discrepancy
18 Less: Subsidies minus current surplus of government enterprises
204
19

Charges against net national product

264, 551

20

Capital consumption allowances

20, 516

21

CHARGES AGAINST GROSS NATIONAL PRODUCT.. 285, 067

differ from those in table I in magnitude only, because table I
includes nonbusiness items. Nonfactor cost charges are identical,
in magnitude as well as with respect to classification, because, as
will be seen, these charges are confined to the business system.

Treatment of taxes and subsidies
Some of the most important decisions that must be made in defining business output have to do with the classification of charges
against the value of business product between factor-cost and
nonfactor-cost charges, and involve in particular the treatment
of taxes and subsidies.
According to the definitions used in this report, employment
taxes under the social security and related programs are included
in the compensation of employees (the employers' share as a supplement to wages and salaries and the employees' share as a part
of them). This is done on the ground that these taxes are an element of the cost of hiring labor and accordingly should be included in a measure of total factor costs. A supplementary argument is that social insurance contributions reflect a benefit received by the employee in the wage bargain which should be
taken into account in calculating total employee compensation.
Contributions of self-employed persons under the recent extension
of the old-age and survivors insurance program are treated analogously; that is, the income of self-employed persons is measured
before deduction of these contributions.
Secondly, a distinction is made in this report between corporate
profits taxes, which are considered as part of factor cost—corporate profits are measured before deduction of these taxes—and
other (indirect) business taxes, which are considered nonfactor
charges—business incomes are calculated net of them. This distinction is based upon the following reasoning.



285, 067

Since national income is designed to measure output in terms
of the costs or incomes of the factors of production, it should
change only if either the volume of factor services or their unit
remuneration changes, and not because of a mere change in tax
rates. If it is assumed that corporation profits taxes are not shifted
and indirect business taxes are generally shifted forward, inclusion of corporate income taxes in national income and exclusion
of indirect business taxes from it are clearly indicated, since on
these assumptions mere changes in tax rates will not cause changes
in an income total so defined.
The classification of business taxes in this report is dictated by
the belief that the above assumptions about tax shifting are the
most realistic summary ones that can be made. It may be noted,
however, that the entire subject of tax shifting and incidence is a
rather controversial one and that definitive and final conclusions
are not available. Moreover, the operational definition of indirect
taxes is drawn in a rather summary fashion. All taxes that are
chargeable as business expense (other than employment taxes)
are classified as indirect business taxes. Thus certain taxes—business property taxes, for instance—are included in this group even
though a detailed study of their proper treatment from the standpoint of tax shifting and incidence might call for a different classification.
The classification of taxes in the national income represents an
instance in which an accounting distinction—whether a tax is
chargeable to expense or not—is adopted as a basis for classification and yields, broadly speaking, satisfactory results. In instances
in which it fails to do so and can be superseded by a superior
distinction—as in the case of employment taxes—modifications
are made. However, where the disadvantages of the accounting
distinction are less apparent and less easily remediable because of
conceptual or statistical difficulties, it is maintained.

34

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Subsidies, which are a type of business receipt, appear on the
debit side of the account (with a negative sign) because they are
not part of the market value of business output; the subsidized
products are included at their market values under business sales.
As a matter of long-standing convention, subsidies are regarded as
payments necessary to elicit factor services. Accordingly, they are
included in the sum of factor incomes (by considering them as a
gross receipt in the calculation of business profits) and are deducted in reconciling the factor income originating in the business
system with the market value of business output.

Other aspects of business output
The measurement of business output has, in addition to the
treatment of taxes and subsidies, many other aspects which must
be discussed—such as the precise delimitation of the business
sector; the classification of income shares and nonfactor charges;
the measurement of capital formation and consumption, including
inventory change; the imputations for wages and salaries in kind,
food produced and consumed on farms, and the rental value of
owner-occupied dwellings; the imputations made in the case of
financial institutions; and the special methods adopted in measuring the transactions of these institutions and of other organizations such as government enterprises.
Decisions that are made in these areas affect the definition of
national output and hence are relevant to the present discussion.
However, these decisions are regarded as less crucial to the
definition of national output than those concerning taxes and
subsidies that have been reviewed; also, the treatment of some of

them is rather complex. Therefore, their discussion is postponed
until later (to the detailed presentation of the business sector) so
as not to interrupt unduly the derivation of the national output
or to blur its outline by excessive detail.

Measurement of nonbusiness output
Production included in the measures of national output is not
confined to the business system, but occurs also in each of the
nonbusiness sectors of the economy: households and institutions,
government, and the rest of the world. In the household sector,
account is taken of certain services—such as domestic service—
which, although of a market character, are better thought of as
involving direct factor services rather than business production.
In the government sector, the services provided by the government to the community are accounted for. In the rest-of-the-world
sector, the production accruing to United States residents by
virtue of their net claims on foreigners is measured.
For each of these sectors, the measurement of output differs
basically from that employed for the business sector. The twofold
measurement of output in terms of product flows and factor costs
is not available for the nonbusiness sectors of the economy, and
factor cost must be used for both aspects of the value added by
them to total output. A single measure must be used in these
instances to depict both income and product originating because
there is no sales transaction involving the output produced as
distinguished from the purchase of the ingredient factors of production and supplies and materials. It may be noted parenthetically that the factor cost measurement of output in the non-

Toble II.—Consolidated Business Income and Product Account, 1950
[Millions of dollars]
Compensation of employees:
Wages and salaries:
Disbursements
Excess of accruals over disbursements
Supplements:
Employer contributions for social insurance
Other labor income
Income of unincorporated enterprises and inventory valuation adjustment
Rental income of persons
10
11
12
13
14
15
16

26

Consolidated net sales:

120, 565
0

27
28
29
30

3, 146
3, 440

31

Change in inventories

32

BUSINESS GROSS PRODUCT

To
To
To
To

persons
government
abroad
<
.
business on capital account

184, 340
17, 828
1, 302
. 43, 868
7, 351

36, 140
8, 473

Corporate profits and inventory valuation adjustment:
Corporate profits before tax:
Corporate profits tax liability
Corporate profits after tax:
Dividends
Undistributed profits
Inventory valuation adjustment

17 Net interest
18 Income originating

17, 829
8, 781
12, 360
— 4, 864
3, 708
209, 578

23, 741
19 Indirect business tax and nontax liability
20 Business transfer payments
843
21 Statistical discrepancy
215
22 Less: Subsidies minus current surplus of government enterprises
204
23

Charges against net product

234, 173

24

Capital consumption allowances

20, 516

25

CHARGES AGAINST BUSINESS GROSS PRODUCT. .. 254,689




254,689

NATIONAL INCOME,

35

195 4 EDITION

business sectors also conforms to the definition of final output
underlying national income accounting: the factor services purchased, in terms of which output is measured, are not resold.

Rest of the World
Finally, account must be taken of United States production
arising in the rest-of-the-world sector. As stated in connection
with the National Income and Product Account, the output of
Households and institutions
the United States economy is defined as that accruing to factors
Exhibit 1 presents the measurement of output originating in
of production supplied by residents of the United States, as disthe personal or consumer sector of the economy. Its measurement
tinguished from the alternative of defining national output in
is in terms of the direct factor services bought by households and
terms of the physical location of the factors of production.
Hence, to obtain a measure conforming to the definitions used
Exhibit 1.—Income and Product Originating in Households and Institutions,
in this report, there must be deducted from the components
1950
derived so far a measure of the output produced in the United
[Millions of dollars]
1. Compensation of employees:
2.
Wages and salaries paid
3.
Supplements paid:
4.
Employer contributions for social insurance
5.
Other labor income
6. Interest paid
7. Income originating and net and grvss product

Exhibit 3.—Income and Product Originating in the Rest of theWorld, 1950
6,312
__

18
53
1, 956
8,339

[Millions of dollars]
1.
2.
3.
4.

Wages and salaries (net)
Interest (net)
Dividends (net)
Branch profits (net)

_

5. Income originating and net and gross product

institutions, consisting of employee compensation and payments
of interest. The latter represent payments by households to nonpersonal lenders and payments by institutions. As in the case of
business production, discussion of more detailed points is postponed.
Government
The value added by government to total output is shown in
exhibit 2. As in the case of households and institutions, it is
measured by the value of factor services purchased, except that
in this instance only payments for labor are considered payments
for factor services. Interest paid by government is not counted.
Exhibit 2.—Income and Product Originating in Government, 1950
[Millions of dollars]
1. Compensation of employees:
2.
Wages and salaries
3.
Supplements:
4.
Employer contributions for social insurance
5.
Other labor income..
6. Income originating and net and gross product.

19,631
812
330
20, 773

The accounting for government production in general and
government interest in particular is, and probably will remain,
a controversial point in national output measurement. Basically,
this seems due to the fact that government activities are substantial and quite different in economic character from other
types of production. Accordingly, they are difficult to put on a
common denominator with other types of production in a summation designed to obtain a comprehensive measure of national
output.
The exclusion of government interest paid from factor incomes
stems, as a practical matter, from the fact that the bulk of government debt was created to finance wars and current expenditures.
In no commonsense use of the term can interest payments on
such debt be taken to represent currently produced goods and
services or the current use of economic resources. For example,
it seems sensible that a comparison of the prewar and postwar
volumes of production should not be distorted by the continuing
interest on the national debt that arose during the war. The
treatment of government interest is given further consideration
in the discussion of the government sector below.



-

-

-

18
248
426
574
1,206

States but accruing to foreign residents, and there must be added
a measure of the value of output not produced in the United
States but accruing to United States residents. The deduction
and addition are accomplished, on a net basis, by the separate
measurement of United States output originating in the rest
of the world, equal to the net international inflow of factor
incomes, as shown in exhibit 3.

National income and product, by sector of origin
In table la the contributions of each of the four major sectors
are summed to yield the totals of national output as shown in
table I, in a breakdown, however, which displays more clearly
their conceptual building blocks. The items in table la are crossreferenced to tables I and II and exhibits 1-3 so as to permit a
precise tracing of how national output is derived from the component output series of the four sectors.
Little need be said about table la since it represents only a
summary of the preceding discussion. However, certain of the
major points may be repeated for emphasis.
In table la, the measures of national output given in table I are
derived by summing the values added to total output by each of
the four major sectors of origin. For the nonbusiness sectors—
households and institutions, government, and the rest of the
world—contribution to total output is measured identically in the
product and factor income measures, since a distinction between
the two is not possible. For the business sector, however, the
product measure differs from the factor income measure. The
difference represents nonfactor charges against the value of business production, which are entered in table la and reconcile the
value of national income with that of the gross national product.

Anatomy of Product and Income Components
A tolerably precise derivation of the measures of total output
presented in table I has been obtained in table la. However, it
will be noted that the components of output as given in the second
table differ from the conventional classifications shown in table I.
Those of table I are drawn up, on the credit side of the account,
in terms of the purchases of product by the four sectors of the

36

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS
Table la.—National Income and Product Account, by Sector of Origin, 1950
[Millions of dollars]

Income originating in
Households and institutions (exhibit 1, item 7)
Government (exhibit 2, item 6)
Rest of the world (exhibit 3, item 5)

8, 339
20, 773
1, 266

Gross product originating in
Households and institutions (exhibit 1, item 7)
Government (exhibit 2, item 6)
Rest of the world (exhibit 3, item 5)

8, 339
20, 773
1, 266

Income originating in nonbusiness sectors

30, 378

Gross product originating

30, 378

Income originating in business (table II, item 18)

209, 578

National income (table I, item 14)

in nonbusiness sectors

Gross product originating in business (table II, item 32)

254, 689

GROSS NATIONAL PRODUCT (table I, item 26)

285, 067

239, 956

Other charges against net national product (table II, items 19, 20,
21, and 22)

24, 595

Charges against net national product (table I, item 19)

264, 551

Capital consumption allowances (table II, item 24)

20, 516

CHARGES AGAINST GROSS NATIONAL PRODUCT (table I,
item 21)
285, 067

economy; and, on the debit side, they do not separate the costs
incurred in production by each of the sectors. By contrast, the
classifications in both sides of table la refer to sector of production.
As the next step in explaining the content of national income
and national product, each of the components as given in table I
is derived from the measurements underlying table la. Product
and income items are taken up in turn.

Gross national product
Personal consumption expenditures
The personal consumption expenditures shown in table I are
derived as in exhibit 4. The bulk of them consists of purchases
Exhibit 4.—Derivation of Personal Consumption Expenditures, 1950

Government purchases of goods and services
The derivation of government purchases of goods and services
is indicated in exhibit 6. The procedure is similar to that employed for consumer expenditures. In addition to purchases from
business and purchases of direct factor services, which are entered
as the first and second items in exhibit 6, government makes purchases from abroad, which have not appeared so far in the accounting for national output. These purchases are entered as the
third item in exhibit 6 to complete the enumeration of government purchases as shown in table I.
Exhibit 6.—Derivation of Government Purchases of Goods and Services, 1950
[Millions of dollars]

[Millions of dollars]
Consolidated net business sales to persons (table II, item 27). 184,340
Gross product originating in households and institutions (exhibit 1. item 7)
-8,339
Net purchases of households and institutions from abroad-. 1,347
Personal consumption expenditures (table I, item 22)

account and the change in inventories, each as taken from the
business account (table II). The derivation is given in exhibit 5.

Consolidated net business sales to government (table II, item 28)
Gross product originating in government (exhibit 2, item 6)
Net purchases of government from abroad

17,828
20,773
3,422

Government purchases of goods and services (table I, item 25)

194,026

from the business system. Direct purchases of factor services by
households and institutions account for another part. These two
are entered as the first and second items in exhibit 4, with references to the previous tabulations in which they have been included. It will be apparent, upon reflection, that the sum of these
two items falls short of a complete enumeration of consumer expenditures by the amount of direct purchases from abroad
(mainly tourist expenditures). This item, which has not so far
appeared in the derivation of national output, is entered as the
third item in the exhibit.
Gross private domestic investment
Gross private domestic investment as shown in table I is derived
simply by combining consolidated net sales to business on capital
Exhibit 5.—Derivation of Gross Private Domestic Investment, 1950
[Millions of dollars]

42,023

Net foreign investment
With two exceptions, all of the elements underlying the construction of gross national product by sector of origin in table la
have now been used to derive components of gross national product as shown in table I. These two exceptions are "consolidated
net business sales to abroad" and "gross product originating in
the rest of the world." On the other hand, in the derivation of
personal consumption expenditures and government purchases
two items were added which have not appeared as components
of gross national product by sector of origin. These are "net purchases of households and institutions from abroad" and "net purchases of government from abroad." Hence, adding the first two
Exhibit 7.—Derivation of Net Foreign Investment, 1950
[Millions of dollars]

Consolidated net business sales to business on capital account (table II, item
30).
43,868
Change in business inventories (table II, item 31)
_
7,351

Consolidated net business sales to abroad (table II, item 29)
Gross product originating in rest of the world (exhibit 3, item 5)
Less: Net purchases of households and institutions from abroad (from exhibit 4).
Less: Net purchases of government from abroad (from exhibit 6)

Gross private domestic investment (table I, item 23).

Net foreign investment (table I, item 24)




51,219

1.302
1,266
1,347
3,422

-2,201

NATIONAL INCOME,

of these items to the components derived in exhibits 4, 5, and 6
and then subtracting the second two items will yield the correct
total of gross national product.
But the difference between these additions and subtractions is
numerically equal to "net foreign investment," the remaining
component of gross national product in table I which has not yet
been derived. This is so because the four items combined equal
the net receipts of the United States from abroad, from factor incomes as well as from the sale of goods and services. These net
receipts represent the balance of payments on current account,
and their financing implies a change in the net international
asset position of the United States, which is net foreign investment.3 The derivation of net foreign investment is shown in exhibit 7.
The derivation of gross national product as shown on the credit
side of table 1 is now complete. The next task is to derive the income flows shown in the breakdown on the debit side of that
table.

37

195 4 EDITION

Undistributed corporate profits
Undistributed corporate profits are obtained in exhibit 11.
Exhibit 11.—Derivation of Undistributed Corporate Profits, 1950
[Millions of dollars]
Business (table II, item 15)
Branch profits (rest of the world) (exhibit 3, item 4)

12,360
574

Undistributed profits (table I, item 11)

12,934

Corporate inventory valuation adjustment
This item is taken from table II.
Net interest
The calculation necessary to derive net interest, as shown in
table I, is given in exhibit 12.
Exhibit 12.—Derivation of Net Interest, 1950
[Millions of dollars]
Business (table II, item 1 7 ) . - .
Households and institutions (exhibit 1, item 6)
Eest of the world (exhibit 3, item 2)

National income
Wages and salaries
The derivation of wages and salaries is given in exhibit 8. It consists of a summation of payroll accruals in the four sectors of the
economy.
Exhibit 8.—Derivation of Wages and Salaries, 1950
[Millions of dollars]
Business (table II, item 2)
Household aad institutions (exhibit 1, item 2)
Government (exhibit 2, item 2).
Rest of the world (exhibit 3, item 1)

120,565
6,312
19,631
18

Wages and salaries (table I, item 2)

Net interest (table I, item 13)

3,708
1,956
318
1

5,912

Nonfactor cost charges
All the components of national income, as listed in table I, have
now been derived by allocating the entire national income as
constructed in table la. The nonfactor charges against the value
of gross national product shown in table I are all taken directly
from table II.

146,526

Supplements to wages and salaries
Supplements to wages and salaries are obtained in a similar
manner, as shown in exhibit 9.
Exhibit 9.—Derivation of Supplements, 1950
[Millions of dollars]
Business (table II, item 5)
Households and institutions (exhibit 1, item 3)
Government (exhibit 2, item 3)

6, 586
71
1,142

Supplements (table I, item 3)

7,799

Income of unincorporated enterprises and inventory
valuation adjustment, and rental income of
persons
These two items are taken directly from table II.
Corporate profits tax liability
This item is taken directly from table II.

Dividends
The calculation for dividends is shown in exhibit 10.
Exhibit 10.—Derivation of Dividends, 1950
[Millions of dollars]
Business (table II, item 14) — Rest of the world (exhibit 3, item 3)
Dividends (table I, item 10)

8,781
426
9,207

3. In the definitions adopted in this report, international cash gifts, which are also current
account items, are included with goods and services.




The Output Measures Examined
The measures of national output have now been derived with
sufficient precision to make possible their evaluation in terms of
the basic concepts that underlie them—of the national product
as an aggregate of final product values and the national income
as an aggregate of factor costs.4

The concept of final product
As noted earlier, measures of national output must be denned
essentially in terms of the market economy. They cannot encompass the broad range of nonmarket activities that may bear some
resemblance to economic production. In this report the market
economy is taken as the area over which sales and purchase
transactions occur. Once this criterion has been adopted, there
remains the necessity of distinguishing between final and intermediate production within the market economy.
The operational definition of final product underlying the
present national product measures rests on the obvious fact that
purchases not resold do not become elements in the value of other
goods and services, and that hence there is a presumption that
they should be counted in a comprehensive enumeration of the
4. For a more detailed discussion of these concepts see Milton Gilbert, George Jaszi, Edward F. Denison, Charles F. Schwartz, "Objectives of National Income Measurement,"
Renew of Economics and Statistics, August 1948.

38

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

value of the final output of the Nation's economy. It also takes
cognizance of the corollary fact that purchases which are resold
are used up in further production and included in the value of
other goods and services, and hence may be presumed to be intermediate products which should not be counted separately in an
unduplicated measure of production.
The practical consequence of this general definition is to enumerate capital formation and purchases by consumers and general government, and to exclude from final production the raw
materials used up by business in the course of further production.
Capital formation is clearly a part of final product to the extent
that it consists of items that are not used up but are added to
wealth. (Only the inclusion in gross national product of capital
formation for replacement purposes must be noted as a limitation
in this connection.) The inclusion of consumer and government
purchases and the exclusion of business purchases charged to
current account also are broadly reasonable.
Since the expenditures of individual consumers and nonprofit
institutions serving individuals are incurred largely to meet the
needs of individuals, they consist in the main of goods and services
that determine what is commonly regarded as the standard of
living. Government purchases consist essentially of goods and
services provided on behalf of the community as a whole, which
it has been found better to secure collectively rather than individually. They should likewise be included in a measure of the
total goods and services provided to satisfy the needs of the members of the community. In contrast, the bulk of business purchases
of goods and services consists of items that are raw materials in
the production process, rather than items that directly satisfy
human needs. Their separate count is accordingly not necessary
in enumerating the flow of final goods and services.
It is believed that this is a realistic description of the general
nature (of consumer, government, and business purchases and
that the criteria employed in United States national income
statistics for distinguishing between final and intermediate products are accordingly useful for segregating the major types of goods
and services provided to satisfy the needs of individuals.
It is evident, however, that, because of differences in institutional arrangements, certain anomalies may result from the
restriction of the national output measure to the market economy
and from strict application of the purchase-not-for resale convention. Thus, the dividing line between the final products
enumerated and similar nonmarket benefits excluded may not be
appropriate. For instance, literal implementation of the operational definition of final product would count the net value of
services rendered by rental housing but would exclude the
counterpart for owner-occupied housing. Moreover, within the
market economy the distinction between final and intermediate
product would sometimes be unsuitable. Food purchased by employees, for example, would be classified in national income and
product whereas food furnished to them by their employers
would not.
Peculiarities of this type can be dealt with to some extent by
appropriate modification of the definitions, but the potentialities
of this approach are limited. This is obvious with respect to the
extension of output measures beyond the market economy, but it
also holds true of modifications in the basic convention for distinguishing between final and intermediate production.



In the first place, it is not feasible from a purely physical standpoint to examine every purchase by consumers, government, and
business so as to determine which were simply means of facilitating
production, and hence intermediate, and which served an end
use, and hence were final products. As a practical matter, one
must generally deal with types of buyers and categories of goods
and services.
But more important, one must place basic reliance on a broad
convention because in most cases in point there is no alternative.
No precise line can be drawn between final and intermediate
products from mere observation of the nature of the products or
the uses to which they are put. It would be easy, for example, if
all consumer purchases were for goods like Sunday clothes and
holiday dinners, which are obvious elements of the good life, and
if all business purchases were raw materials for further processing,
which are obvious intermediate goods. Between these two extremes, however, there is a wide range of purchases for which
neither the motivation nor the use is so clear-cut and which must
be placed in one category or the other by somewhat arbitrary
rules.
For this reason any measure of total production must be somewhat conventional. For instance, it must overlook the fact that
the expenditures of individuals in their business capacity are
influenced by their standards as consumers, and that expenditures of consumers are influenced by their activities as producers.
It must overlook also the fact that the conditions under which
work is performed have an important bearing on the welfare of
individuals. These conditions are affected by business expenditures on goods and services that are classified as intermediate just
because there is no satisfactory way to take account of their
benefits in a quantitative measure of final output.

Adjustment by imputation
In the present estimates, adjustments have been made to take
account of institutional peculiarities to the extent of imputing
factor returns in the form of income in kind and entering corresponding inputed items in personal consumption expenditures.
Even in this direction cognizance has been taken, in the main,
only of sizable and unequivocal types of factor income in kind
which have come to be recognized through tradition as elements
of real income. It is apparent that other additions to the national
output could be made if the relevant information were available.
For example, income and consumption expenditures could be
imputed for recreational facilities provided by business, which
are not counted because they are charged to current cost by
business.
The limitation of imputations to cases clearly associated with
factor incomes serves to confine the field, but it is not a principle
of selection which could be firmly defended on theoretical grounds.
The services of the radio broadcasting and television industries
are an outstanding example of an item which is not listed in the
national product because it is financed by business via charges
that are made to current cost. Yet radio broadcasting and television are important forms of recreation, similar to legitimate
theaters and motion pictures for which explicit entries, representing admission fees, are made in consumer expenditures.
No imputation is made for radio broadcasting and television in
measuring national product. Formal neatness can be given to this

39

NATIONAL INCOME, 195 4 EDITION

omission by general reference to the limitation of imputations to
items that are associated with factor incomes, on the ground that
broadcasting services do not accrue to a distinct factor of production, but to the consuming public owning receiving sets at large.
However, it would seem preferable not to stress this point unduly
and to recognize the essentially arbitrary and tradition-based
nature of the decisions that must be made in this area.

practice to any detailed, selective functional classification designed to add up those particular items which may be considered
"final." However, when any sort of concrete, workable criterion
of intermediate product is applied, it becomes evident that the
present scheme of summary classification does not lead to significant distortion.

Need for market-type measures of output
Quasi-intermediate products
The process of reclassification of intermediate products could
be extended in the opposite direction by deducting from factor
income, and hence shifting from consumption expenditures to
business costs, business-type expenses incurred by individuals. As
an example, "miners' expenditures for explosives, lamps, and
smithing" (a small item of consumer expenditure in table 30,
Part V) are certainly the sort of cost ordinarily borne by business
rather than by the wage earner, and are unlike the vast bulk of
consumption goods. It is a peculiarity of the coal mining industry
that these materials customarily are paid for by the miner rather
than purchased by the enterprise and charged to current cost.
No attempt has been made to adjust the pattern of consumer
transactions along this line because there is no tradition of adjustment that provides an adequate standard of what is appropriate.
In any event, it would appear that the magnitude of such adjustments would be very limited, unless the concept of "production
expense" were stretched far into the broad region of mixed
motivations in which no useful and commonly accepted exclusions frc m final product can be made.
Similar considerations apply to government purchases as well
as to consumer purchases. It is possible to think of cases in which
the treatment of government purchases as final product would
not necessarily be the best procedure. For example, if certain
government purchases reflected clear-cut aid to business it might
be preferable to view them as "subsidies in kind" and, in accord
with the handling of subsidies, to eliminate them from government purchases and the national product. Such a treatment, it
should be recognized, would be somewhat artificial and statistically difficult, and would obscure the national economic accounts in their capacity as records of actual transactions, thus
rendering them less meaningful for many purposes. Also, the
feasibility of its application to government services used jointly
by business and individuals, like the maintenance of highways, is
highly questionable.
If government services consisted of the running of public recreation grounds on the one hand and of the provision of free raw
materials to business on the other, a classification of the latter as
subsidies in kind might be useful and important. This is not, however, the actual character of government operations. Clear-cut
types of direct subsidies in kind that are of any quantitative importance have not come to attention. Even if account is taken of
the more consequential cases of government services involving
the use of a public service jointly by individuals and by business,
the problem remains quantitatively small. It looms large only if
the concrete notion of aid to business is stretched to cover the
broad range of government services to the public which actually
reflect a complexity of causes and purposes and cannot appropriately be classified under any such narrow head.
Neither government expenditures nor consumer expenditures,



i n 1r»o-i«~ r\r

It is apparent that even if substantial departures from the present definition of final product were logically defensible and statistically feasible, they would not result in a measure of national
product that could serve as a substitute for the present one. This
measure is tied closely to the modern market-economy and is
obtained, broadly speaking, by summing actual transactions of
its major constituent economic groups. As such it is an important
element of the economic accounting system designed to facilitate
an understanding of the functioning of the economy in terms of
the interaction of these groups.
For this purpose, the definition of the consumer and government purchases components of national product is generally appropriate, as is the exclusion of intermediate production according
to the present definition of the term. Even if basic departures from,
these definitions could be justified on other grounds, the resulting
measures of national output would probably not be useful in the
study of business cycles, in problems of economic mobilization
and fiscal policy, in market research, and in many other types of
investigations in which national income statistics are increasinglyemployed.

The concept of factor cost
Underlying the definition of national income in terms of factor
cost is the general idea that the output of the Nation is the result
of the services rendered by agents of production who cooperate
in the production of that output. These agents of production are
the labor and capital, the entrepreneurial ability and natural
resources which are used in the production process. It is the
services of these agents or factors as valued in the market by their
earnings for which a quantification is sought in the national
income, to the extent permitted by the data available as statistical
raw material.
It is hardly necessary to stress the importance in studies of
resource allocation of such a measure of the services rendered by
productive agents. To give only a few examples, one may wish to
know the incomes of the various factors of production used in each
industry in order to compare the relative importance of industries, or to marshal information about the relative amounts of
factors of production that are available for allocation to various
uses, or to compare the relative importance of labor and property
factors in the outputs of various industries.
>
It must be recognized, of course, that the concept of factors of
production is not given precisely in economic theory but must, to
some extent, be formulated with reference to the problem at hand.
Furthermore, the use of factor returns for some problems of the
type indicated is limited by such facts as the temporary or permanent non-transferability of factors to other uses, and by monopoly
and imperfect competition. In addition, property income is only
tenuously related to the type of measure of the contribution of
nmnprtv

anH pntprnriQP lA/hirh miorlit

rl^air/^rfl fr\r

40

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

involving resource allocation. This is because it includes a
residual share, profits, which fluctuates widely over the business
cycle.
In spite of these limitations and difficulties, the idea of factor
costs has always been of fundamental importance in economic
analysis, and national income defined as an aggregate of factor
earnings is the only general measure by which the idea can be
quantified.
It is true that difficulties are encountered in the course of this
quantification. The assumptions about tax incidence that are
made in the classification of taxes as between factor income and
nonfactor cost, the somewhat conventional rationale that leads to
the calculation of national income gross of subsidies, and the
common sense consideration on which the exclusion of government interest is based are all open to question.
Yet it would appear that for the items that are large and of
strategic importance in the dynamism of the United States
economy, the assumptions made are sufficiently realistic to provide useful economic measurements. For instance, in spite of the
theoretical uncertainty which surrounds the incidence of corporate profits taxes, it appears in statistical investigation that
corporate profits before taxes are more invariant to mere changes
in tax rates than are profits after tax, and that the before-tax
series must be used in studying the economic regularities reflected
in the movement of the various income shares. Nor would any
realistic study of national output be advanced by the inclusion in
national income and product of interest paid by the government.
One important aspect of the factor-cost definition of national
income should be understood. The constituent income shares of
national income so defined cannot be construed as measures of
benefits accruing to the recipient groups. For instance, the definition of the income shares gross of direct taxes levied on them is
necessary in order to reflect the factor costs of current production,
but would not be appropriate for measuring benefits received.

DETAILED STRUCTURE OF THE ACCOUNTS
As noted at the outset, an important recent development in
national income research is the expansion of national income statistics from measures of the national output into an economic
accounting system giving a statistical picture of the economy.
The national output aggregates and their major components now
having been derived in broad outline, and presented in the National Income and Product Account in table I, the groundwork
has been laid for an explanation of how a comprehensive national
economic accounting system is constructed. Incident to this explanation, further light will be shed on the definition of national
output and its various components.
The plan is to derive a set of accounts which will summarize the
economic process in terms of the interrelated transactions of the
four major economic groups into which the economy can be
divided—business, persons (households and institutions), government, and the rest of the world. This will involve the construction of a current account for each of the constituent economic
groups or sectors and of a consolidated saving and investment
account for the economy as a whole.



The Business Sector
The business sector is defined broadly to include all organizations which produce goods and services for sale at a price intended
at least to approximate costs of production. In the main, it covers
all private enterprises organized for profit, both corporate and
noncorporate, including farm operators, independent professional
practitioners, and lessors of real property. Mutual financial institutions, cooperatives, and nonprofit organizations serving business are also included, as well as government enterprises. Owneroccupied houses and buildings used by nonprofit institutions
serving individuals are considered to be business establishments
selling their current services to their owners.
The business sector thus covers a wide variety of organizations,
and for some purposes it would be desirable to distinguish further
between corporations and unincorporated enterprises, financial
institutions and nonfinancial business, and so forth. Also, it would
be instructive to treat industries or industry groups as separate
sectors in order to reveal the flow of intermediate output among
them, and to show their complete sales, cost, and profit structures.
Such breakdowns of the business sector are not presented in
this report, although important elements of them are contained
in Part V. (See the tables on national income by legal form of
organization and by industry of origin.) To regard the business
system as an entity is sufficient for many purposes, and the statistical information for establishing further sectors within it either
is unavailable or could be assembled and utilized only at the
expense of disproportionate effort. However, a further development of national income statistics does lie along these lines and
would serve to integrate them with other studies of the industrial
structure, such as the interindustry relation studies.

The business account
The receipts of the business system and their disposition have
already been exhibited in table II, in connection with the derivation of national income and product via the summation of sector
incomes and products. This Consolidated Business Income and
Product Account serves as the current account for the business
sector in the present economic accounting system. Several essential features of this account—its basic affinity to a consolidated
profit and loss statement and the netting, transposition, and
classification of items necessary to obtain significant measures of
output originating in business—have already been explained. At
this point further characteristics will be considered.

Classification of business transactions
The right side of the business account consists of the consolidated net sales of the business system, adjusted for the change in
inventories in order to measure business output. Since this account
covers all types of enterprises that are included in the business
sector, the definition of "sales" is broad. For instance, fees for
professional services and gross rental receipts are included, although they are not always thought of as sales in the daily meaning of the term.
Sales are subdivided according to the four major purchaser
groups: consumers, government, business (on capital account)
and foreign nations. To a large extent the content of the items is

NATIONAL INCOME, 195 4 EDITION

adequately conveyed by their designations. Aspects needing further explanation will be taken up later, particularly in the discussions of capital formation (which concerns sales to business
on capital account and the change in inventories) and of imputations and the treatment of financial institutions (which affect the
definition of business sales to persons).
Types of factor income
The left side of the business account lists the charges against the
value of national product. These charges are classified into factor
costs or incomes and nonfactor charges. The former are listed in
five main categories—compensation of employees, income of unincorporated enterprises and inventory valuation adjustment,
rental income of persons, corporate profits and inventory valuation adjustment, and net interest.
The compensation of employees consists mainly of wages and
salaries but includes additional forms of compensation under the
heading of supplements to wages and salaries. Wages and salaries
include payments received in kind in addition to monetary remuneration. They are subdivided into "disbursements" and
"excess of accruals over disbursements" to take account of differences (due to retroactive wage adjustments) between amounts
charged to cost and actual disbursements to individuals.
Supplements to wages and salaries consist of employers' contributions for social insurance and of "other labor income." The
former item comprises employer taxes, or contributions, under
social security and kindred publicly administered schemes. The
corresponding employee contributions are included in wages and
salaries. Other labor income includes employer contributions to
private pension and related funds, compensation for injuries,
and certain minor items which are charged against the value of
business production and can conveniently be classified as factor
charges under the heading of supplements to wages and salaries.
Income of unincorporated enterprises and inventory valuation
adjustment, rental income of persons, and corporate profits and
inventory valuation adjustment cover the business incomes of the
private enterprises that are counted as part of the business sector
of the economy. Within the noncorporate part of the business
sector a distinction is drawn between unincorporated enterprise
and rental income. The former covers the earnings of sole proprietorships and partnerships (including farm and nonfarm businesses as well as independent professional practitioners) and of
producers' cooperatives; the latter consists of the net rentals of
individual landlords who are not primarily engaged in the real
estate business. The earnings of professional real estate operators
are classified under income of unincorporated enterprises. Both
the income of unincorporated enterprises and rental income include, in addition to monetary earnings, important items of income in kind.
The definition of monetary business earnings is in general accordance with Federal income tax regulations. Significant modifications are made, however, in the treatment of capital gains and
losses, inventory profits and losses, depletion charges, and receipts
of property income.
Business incomes in the national income and product accounts
are stated exclusive of capital gains and losses, because these do
not represent a return for the current use of economic resources.
The "inventory valuation adjustment" is designed to eliminate



41

from corporate profits and the income of unincorporated enterprises an element which is very similar to capital gain or loss.
The adjustment is often large and uncertain statistically, and
there is a great deal of interest in the corporate profit figure prior
to the adjustment. Hence the accounts are set up to give the
unadjusted figures and the adjustment separately, with the two
adding up to the proper total for purposes of national income
measurement. By contrast, ordinary capital gains and losses are
eliminated outright.
With respect to depletion charges, no deduction is made for
them in computing business net incomes. The value of new discoveries of natural resources is not counted as part of capital
formation or of profits, and consequently deduction of a capital
consumption charge for impairment of the stock of natural resources would be inappropriate.
Finally, all corporate receipts of dividends are netted out of
corporate profits (and dividends) to avoid double counting and
to arrive at income originating in the business system; and interest
and dividends received by the owners of unincorporated enterprises are considered to be received by them in their personal
capacity, rather than treated as an element of business income,
except in a few financial industries in which the earning of property incomes is an integral part of business operations.
In table II corporate profits before tax are broken down further
to show tax liability, dividends, and undistributed profits. A
similar breakdown is not presented for unincorporated business
and rental incomes. With few exceptions, there are no taxes levied
specifically against these types of income (business incomes are
merged with other types of income in determining individuals'
income tax liability). Moreover, in the noncorporate area a realistic distinction between distributed and undistributed business
income is difficult to establish in principle and to measure statistically.
Net interest measures the excess of interest payments of the
business system over its receipts. In addition to monetary interest
flows, it covers imputed interest arising in connection with the
operations of financial intermediaries.
Nonfactor charges listed in table II consist of indirect business
tax and nontax liability, business transfer payments, the statistical
discrepancy, subsidies minus the current surplus of government
enterprises, and capital consumption allowances.
The classification of indirect business tax and nontax liability
as a nonfactor charge already has been discussed in summarizing
the derivation of national output.
Business transfer payments
Business transfer payments represent transfers from business to
persons which are charges against business product for which no
return in the form of factor services is received. Major items included are corporate gifts to nonprofit institutions and allowances
for consumer bad debts. The nature of the latter item can be
understood as follows. Sales to consumers, on the credit side of
the account, are stated at their sales price and are not (except in
the case of professional fees) adjusted by an allowance for consumer defaults. However, the incomes of sellers and lenders, on
the debit side, have been calculated net of these defaults. Hence,
an accounting discrepancy between the two sides arises which
can best be resolved by regarding the value of consumer bad

42

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

debts as reflecting goods and services transferred from business
to consumers with no quid pro quo.
Subsidies are listed as a (negative) nonfactor charge against
the value of business output. They are not considered part of the
value of product, but are included as receipts in calculating
profits. The current surplus of government enterprises is an item
akin to business profits earned in the course of making the sales
listed on the credit side of the account. Hence it must be included
on the debit side to ensure balance. However, for reasons to be
noted later, it is classified as a nonfactor rather than a factor
charge.
The foregoing nonfactor charges reconcile the income originating in business with the market value of business net product.
To arrive at the total designated in table II as charges against
business gross product, capital consumption allowances must be
added.

Capital consumption allowances
Capital consumption allowances consist of depreciation proper,
capital outlays charged to current expense, and accidental damage
to fixed capital. The first of these items measures the wear and
tear and obsolescence of fixed capital and (with the exception of
agricultural depreciation, which is on a replacement cost basis)
is based on accounting practices used for tax purposes—largely
straight-line amortization of original cost to the current owner.
Capital outlays charged to current expense are an entry in lieu
of depreciation proper for items of durable capital listed on the
credit side of the account (on the basis of the durability definition
there adopted) but charged to current cost in accepted business
practice. It is apparent that the value of these items must be
entered on the debit side also in order to preserve the balance of
the accounts. In a stationary economy capital outlays charged to
current expense would, for business as a whole, approximate the
charges for depreciation which would have been made for these
items had they been capitalized instead of expensed. In a situation
in which net capital formation occurs, the entry will overstate
actual depreciation; when capital formation falls below replacement needs, it will fall short of an adequate capital consumption
allowance for the types of equipment involved.
Accidental damage to fixed business capital measures the value
of such capital destroyed by accidents. The accounting necessity
for an entry of this type stems from the fact that business profits
are net of such losses on the debit side while no corresponding
entry appears on the credit side. Its classification as a species of
capital consumption allowance is based on the practical fact that,
for the business system as a whole, the magnitude of the item is
steady and can be regarded as akin to regular depreciation. If
there were large fluctuations in these losses, a strong argument
could be made for treating these unusual deviations from the
average experience like capital gains and losses—that is, calculating profits and total income without deduction for them. This
treatment would prevent fluctuations in national income due to
the accidental destruction of fixed business property.
As has been noted, depletion charges are not deducted in calculating profits, since the value of the corresponding discoveries of
natural resources is not an element of capital formation or profits.
Similarly, these charges are not included with capital consumption allowances.



Interpretation of income share breakdown
The breakdown of the income shares given in table II reflects
to a large extent the actual institutional, legal, and financial arrangements in force at any particular time which determine the
form in which income accrues to individuals. An additional,
broader grouping sometimes found helpful consists of employee
compensation and net interest, which are contractual incomes,
and of other incomes, all of which are residual shares.
The recording of earnings in the forms in which they accrue
means, for example, that shifts in the legal form of organization
as between corporations and unincorporated enterprises, or
changes in the relative importance of internal and external business financing, will be reflected in the several income shares. In
many economic investigations which deal with the concrete arrangements of economic society, a breakdown of this type will be
appropriate. However, it will present handicaps in analyses in
which it is desired to abstract from such arrangements.
The attempt to use the income share breakdown to study the
ultimate factors which cooperate in production is a case in point.
It deserves special mention because it is directly suggested by the
definition of total income originating as the sum of factor incomes
or factor costs. In the light of this general definition, one might be
tempted to go further—to make identifications of particular income shares with the various factors of production envisaged in
economic theory.
Along these lines, it is possible to say that employee compensation consists of labor income, that unincorporated enterprise and
rental incomes are mixed returns to labor and other productive
factors, and that corporate profits and net interest are components which do not contain a labor return element, in any ordinary sense of the term. However, in view of difficulties attaching
to the factor of production notion and in view of the lack of
statistical information, one cannot go much further in the way of
identifying factors with the income shares, and it is important
to have in mind the limitations of the data for this general type
of use.
With respect to a segregation of the returns to the factor of production "labor," it should be noted that employee compensation
is heterogeneous in character. It includes the wages of the charwomen as well as the bonuses of the corporation executive. Moreover, it is not the only income share in which returns to labor are
reflected. In the income of unincorporated enterprises and the
rental income of persons the labor of the owner is an element,
although it cannot be quantified and segregated.
In connection with the classification of income shares other than
employee compensation, no identification can be made between
the rental income of persons and the rent concepts of economic
theory. Rental income is confined to the net rentals of individual
property holders (including imputed rentals on owner-occupied
nonfarm homes) whose main occupation is not the renting of
property. Rental income of professional real estate operators is
classified under the income of unincorporated enterprises; gross
rental receipts of corporations are merged with their other business receipts and (after deduction of costs) reflected in corporate
profits; both the imputed net rental value of farm homes and
agricultural net rents received by farmer landlords are included
in the income of unincorporated farm enterprises; and the return
on user-owned business real estate becomes a component either

NATIONAL INCOME, 195 4 EDITION

of the income of unincorporated enterprises or of corporate profits.
The dividing lines between profits and net interest also call for
comment. First, net interest represents the payments less the receipts of the business system. An increase in corporate interest
receipts from other sectors (most importantly from government)
is therefore reflected in a decrease of net monetary interest and an
offsetting increase in profits, and vice versa, even though no
change in the profit and interest flows that are an integral part of
business operations has occurred.
Also, the breakdown of property income between interest and
profits is affected by the manner in which interest flows are channelled through the economic system. For instance, if money is lent
by corporate or other professional lenders the interest paid on it
is counted in business receipts and reflected in profits. But if individual lenders are involved, the interest paid to them appears as
such in the business account.
Finally, of course, the breakdown between profits and interest
is influenced by the choice between external and internal financing. For all these reasons, it is more appropriate for some types of
economic analysis to combine the interest and profit shares than
to consider them in isolation.

Fixed investment
Fixed investment by business (business purchases on capital
account) includes new construction and durable equipment acquired by private business enterprises. New residential construction purchased by owner-occupants (as well as by business proper
for rental purposes) is included because home-ownership is treated
as a business in the national income accounts. Acquisitions of
fixed capital by nonprofit institutions serving individuals also are
included.
Fixed capital formation is defined, in the United States statistics, as including all newly produced durables (goods with an
average life exceeding 1 year) acquired by their ultimate business
users. Thus fixed capital formation is stated gross of capital consumption, and includes plant and equipment bought for replacement purposes.
From a theoretical standpoint, a net concept would be preferable. The definition of gross capital formation must needs be
somewhat arbitrary, because the size of the category will depend
on the particular definition of durability adopted. The shorter the
average life used in defining durability, the larger will be the
apparent volume of gross capital formation, although it should be
noted that the magnitude of this variation is very small for
alternative definitions of durability that one might consider in
practice. (See section 10, Part III.)
In addition, the considerations dictating elimination of intermediate production to achieve output measures without duplication also call for the statement of fixed capital formation on a
net basis, since, broadly viewed, capital outlays for replacement
purposes are really a species of intermediate product.
Measurement of fixed capital formation on a gross basis has
been advocated as being more appropriate for certain types of
analysis concerned mainly with the short-term availability of
resources. If fixed capital need not be replaced in the short run,
the total value of production available for alternative uses is
measured better, it is contended, by the gross than by the net
totals.
291602°—54
4



43

While this argument has some merit, there is no reason to
believe that gross fixed capital formation as measured in the
business account is appropriate for the purpose. It would give an
exaggerated view of production available for alternative purposes
in the short run if it included essential replacements that could
not be postponed without interfering with the operations of the
productive apparatus. On the other hand, it would fail to reveal
the true short-run production potential if there were ways of
utilizing the fixed capital more intensively than normal or of
postponing maintenance and repair, as distinguished from replacement. In short, the use envisaged calls for estimation of the
capital stock that is consumable in the short run, including the
stock of business inventories as well as consumer and government held tangible assets; it cannot be served adequately by a
fixed capital formation series defined on the gross basis used in the
national income and product accounts.
Measurement of fixed capital formation gross of replacement
may be preferable for certain purposes in business cycle and
market analysis. Replacement demand is an important factor in
aggregate demand and it is useful to have an explicit record
including it. However, in this as in the preceding case, the present
definition of gross capital formation is imperfectly adapted to the
purpose envisaged. For instance, it excludes maintenance and
repair outlays although these too have a significant bearing on
cyclical fluctuations in demand.
In any event, in designing an unduplicative measure of national
production these and related considerations must remain subsidiary. Defining fixed capital formation gross of capital consumption allowances must rather be justified largely on the
ground that it sidesteps the unsolved problems involved in the
definition of net capital formation.
The problem of what is meant by "keeping capital intact" is a
most controversial one in economic theory, basically because in
a dynamic economy the nature of capital equipment changes and
the notion of replacing worn out capital consequently loses its
simplicity. But even apart from theoretical difficulties, the statistical problem of estimating capital consumption in a manner
consistent with gross capital formation is quite formidable. The
bulk of the capital consumption allowances recorded in table II
is derived from financial accounting records and is on an original
cost basis. While from the standpoint of accounting consistency
these allowances are appropriate for inclusion on the debit side of
the business account—business profits are calculated as a residual
consistent with them—they do not measure capital consumption
on the current price basis which underlies the values shown for
fixed capital formation on the credit side, and hence cannot be
used to obtain a measure of net capital formation in current
prices.
The statistical information, on prices and on the age composition of the capital reflected in depreciation charges, that would be
necessary to convert these accounting charges into a price level
comparable to that of gross capital formation is deficient. Useful
approximations in this direction have been made (the pioneering
study in this field is Solomon Fabricant's Capital Consumption and
Adjustment, National Bureau of Economic Research, 1938). However, in the national income and product accounts no revaluation
of the accounting charges is attempted, and consequently a
measure of fixed net capital formation is not presented.

44

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

It will be noted that business profits, income originating in
business, and national income all incorporate the same depreciation charges which are considered inadequate as measures of
capital consumption for the purpose of arriving at fixed net capital formation, business net product, and net national product.
Net measures of income had long been established before the
problem of valuing depreciation charges was met in particularly
acute form in connection with fixed capital formation, and their
continued use reflects in part the accident of this historical
sequence.

Inventory change and the inventory
adjustment

valuation

The measurement of inventory change in business accounting
practice is subject to the same type of deficiency from the standpoint of national income and product as fixed capital depreciation. As a broad proposition, original cost instead of current replacement cost is used to value inventories consumed in the
process of production, and hence a measure of net change based
on business accounting records would be misleading for national
income purposes.
In this instance, however, an adjustment is made in the national
accounts to convert reported "book" value data to a current replacement cost valuation. The distortions that would result from
failure to make such an adjustment would be more disturbing
here. It is not possible to sidestep the issue by dealing with a gross
concept, as in the case of fixed capital formation; also, short-term
comparisons, for which national income data are frequently used,
are particularly affected by the methods of inventory valuation.
Moreover, the conceptual and statistical difficulties that stand
in the way of an adjustment, although formidable, are less overwhelming than in the case of depreciation. Use over a long period
of years is not involved, as with fixed capital formation, and consequently the conceptual problem of defining replacement which
stems from quality change and the emergence of new products
looms less large. In addition, information on prices and age composition is more readily available for inventory goods than for
capital goods.

Nature of inventory valuation adjustment
According to the prevalent methods of business accounting, the
book valuation of the physical volume of inventories used up in
production differs from current replacement cost in times of
changing prices.5 When prices are rising, book charges fall short
of current replacement cost; when prices are falling, they exceed it.
No deviation from a current price valuation occurs with goods
added to inventory during a given accounting period. These are
valued at prices current in that period.
The change in the book value of inventories represents additions
to inventories minus inventories used up. Hence it reflects not
only (1) the change in the physical volume of inventories valued
at current prices, but also (2) the excess of the replacement cost of
inventories used up in production over their book valuation.
The former element of book value change is appropriate
for inclusion as a component of national product, because it con5. The following discussion is concerned with the treatment of nonfarm inventories. The
estimates of farm inventory change are computed directly from data on physical stocks and
current prices. No problem of adjusting book value data is involved; that is, an "inventory
valuation adjustment" is not necessary.



forms to the principle of current price valuation applied to all the
other components. To include, however, the "inventory gain" or
"inventory loss" measured by the second element of book value
change would be misleading. In extreme cases the inventory
movement as indicated by the change in book values would differ
in direction from that of the actual volume of inventories. Therefore, the "change in inventories" line in business and national
product is derived by adjusting the reported book value change
in inventories to exclude the inventory gain or loss element.
For similar reasons, business profits as initially calculated on
the basis of business accounting methods of inventory valuation
are, for purpose of inclusion in national income, adjusted to
exclude inventory gain or loss. This is done in table II by adding
the "inventory valuation adjustment" to corporate and noncorporate business profits as estimated from "book" data reported
by business. When negative, the inventory valuation adjustment
measures the inventory gain, and when positive, the inventory
loss, which arises from the fact that inventories used up in production are not valued at current replacement costs. Its affinity
to capital gains and losses, which also are eliminated in calculating
national income and product, is readily evident.
The statistical methodology for estimating the inventory components of national income and product is explained in Part III
of this report. The following numerical examples may serve to set
forth more precisely the accounting principles involved.

FIFO method
Suppose that a firm had beginning inventories of 1,000 units
valued at $5 each, that it purchased during the accounting period
400 units valued at $3 each, and that it used up, in production
and sale, 300 units of inventories. According to the first-in, firstout (FIFO) method of inventory valuation, which charges inventories to cost of sales in the order of their acquisition, inventories
used up would be valued at $5 each, resulting in a total of $1,500.
The inventory change would be registered as minus $300, the
difference between $1,200 of acquisition and the $1,500 used up.
The book value change of minus $300 is the algebraic sum of a
physical volume change, in current prices, of plus $300 (100 units
at $3 each) and an inventory loss of $600, which measures the
difference between the book cost and the current replacement
cost of inventories used up ($2 on 300 units). Since purchases of
inventory goods are valued at current prices, the departure from
current valuation in the measure of inventory change reflects
entirely the manner of valuing inventories used up.
For national output measurement it would be misleading to
register an inventory decline of $300 when the volume of inventories has actually increased. Accordingly, on the credit side of
the business account, the change in business inventories is entered
at plus $300, equal to the physical change of 100 valued at $3
each, in conformance with the current-price valuation basis used
for the other components of the product and income flow. Correspondingly, on the debit side an inventory valuation adjustment
of $600 is added to business profits as based on the FIFO method
of inventory valuation. This adjustment corrects profits for the
difference between the book cost of goods sold ($1,500) and their
current replacement cost ($900). Essential to note is that the
adjustment is equivalent to the excess of the physical inventory
change in current prices ($300) over the book value inventory
change (minus $300). Needless to say, if current prices exceed

NATIONAL INCOME, 195 4 EDITION

book cost prices an inventory gain instead of an inventory loss
occurs, and the sign of the inventory valuation adjustment is
negative.
The above example is based upon the straight cost variant of
the FIFO method of inventory accounting. But a revaluation of
the inventory change reported by business is also necessary for
other business accounting methods in which the valuation of
inventories used up departs from a current replacement cost basis.
Consideration of two of these is pertinent: the "cost or market,
whichever is lower" practice of valuing year-end inventories,
which is frequently associated with the FIFO method, and the
last-in, first-out (LIFO) method of inventory accounting.
Under the cost-or-market practice, year-end inventories are
written down by businesses if market prices are below cost prices.
This practice generally necessitates a revaluation of book value
change for national income purposes. However, it should be recognized that the cost-or-market procedure is not the prime cause
leading to revaluation of book value changes in the national income and product accounts. It represents only a special case in
which revaluation is necessary because of a departure from the
current replacement cost basis of valuing inventories used up in
production.
For instance, if the cost-or-market procedure is followed in the
example given, the ending value of inventories will be reported
as $3,300 (1,100 units at $3 each, the lower market price). The
book value change of inventories will be minus $1,700, and an
inventory valuation adjustment of $2,000 will be needed to adjust
the change in book values to national income and product
definitions. However, even without exercise of the method an
inventory valuation adjustment (of 1600) is needed, as previously
shown.

LIFO method
The LIFO method of inventory accounting yields results most
akin to national income practice. As a general proposition, it
yields identical results when the physical volume of inventories
increases, but divergent results when the physical volume decreases. In the former case no revaluation of the book value
change is necessary, but in the latter an inventory valuation adjustment must be applied to inventories charged on a LIFO basis.
As long as the physical volume of inventories is increasing, inventories used up represent, according to the LIFO convention of
assuming that units acquired last are charged out first, current
acquisitions valued at current prices. There is no difference in this
case between the LIFO and national income methods of inventory valuation. This can be seen by applying the LIFO method
to the above numerical example. Inventories used up are valued
at $900, because they are assumed to represent the 300 units most
recently acquired at their current price of $3 each. The book value
change therefore amounts to plus $300 (purchases of $1,200
minus $900 of inventories used up), which is equal to the change
as measured for national income purposes.
However, when the physical volume of inventories decreases
"last-in" prices no longer represent current prices. Inventories
used up reflect past-period acquisitions valued at past-period
prices, which in general differ from current prices. To illustrate
this case numerically, it may be assumed that in the initial
example 500 units rather than 300 units are used up, so that the



45

physical volume of inventories decreases by 100 units. According
to LIFO practice, the inventory change would be minus $500.
This represents acquisitions of 400 units valued at $1,200, minus
400 units used up valued at $1,200 (corresponding to the acquisitions), and minus 100 units valued at $500, reflecting the cost
price of the units included in the initial stock. According to the
method adopted in this report, however, the inventory change
would be valued at minus $300, measuring the physical volume
change at current prices, and an inventory valuation adjustment
of $200 would be necessary. This adjustment would account for
the inventory loss which arises because the 100 units of inventories used up in excess of current acquisitions are valued at $5
each, $2 more than the current market price of $3.
The foregoing summary treatment of the problem of inventory
valuation should not create the impression that the subject is a
settled one. On the contrary, there is a great deal of discussion
among accountants and national income technicians both as to
the broad principles and detail involved.
It should also be mentioned that many simplifying assumptions
have been introduced into the summary in order to bring out
more clearly the basic nature of the problem. In their absence,
some of the generalizations made would have to be qualified,
although not changed in essence. In particular, the assumption
underlying the numerical examples, as well as some of the statements in the text, that the prices of inventory goods change discontinuously between accounting periods but remain constant
within them, has permitted the neglect of some complicating
factors which, although significant, are definitely of secondary
importance. This assumption should be noted specifically, because the fact of continuous price change during the year is quite
important from the standpoint of the statistical calculations
described in the section on Change in business inventories in
Part III.

Imputations
As noted earlier, the measures of national output presented in
this report cover not only output whose production and distribution give rise to explicit monetary transactions, but also certain
types of income and product flows which do not take monetary
form. It has also been pointed out that from a theoretical standpoint these imputations represent modifications of the operational
concept of final output, and that they are made to correct for
anomalies and other disturbing omissions that would otherwise
result. The imputations made are the result of concrete considerations and of the traditions of national output measurement.
They do not and cannot represent a logically clear-cut exhaustive
list, but merely a pragmatic selection among a wide variety of
possible imputations.
The general prodecure for allowing for nonmonetary income
and product flows in the national accounts is to imagine that the
flows in question take monetary form and to reconstruct the
accounts to reflect consistently these flows. The business account
is affected by four imputations: wages and salaries paid in kind,
the rental value of owner-occupied houses, food and fuel produced and consumed on farms, and nonmonetary income and
product flows arising in connection with financial intermediaries.
These will be discussed in turn.

46

A SUPPLEMENT TO THE SURVEY OF CURRENT

Wages and salaries paid in kind
An imputation is made for wages and salaries paid in kind in
the form of food and lodging in industries in which this type of
arrangement is of quantitative importance and is regarded as
involving a clear supplement to monetary wages and salaries.
Imputed items are valued at cost to the employer. Needless to say,
difficult and somewhat arbitrary decisions are involved in delimiting the area of this imputation and in establishing the proper
valuation.
In effect, the imputation takes the form of assuming that the
employer, instead of furnishing his employees with free food and
lodging, pays them corresponding amounts of wages, and that
the employees in turn use them to buy the items previously purchased by the employer. Wages and salaries (in income originating in business) and sales to persons (in business product) are
thus raised by corresponding amounts. In terms of the more technical implication of the procedure, intermediate purchases by
the employer are converted into factor costs (wages and salaries)
and final purchases (consumer expenditures).

Rental value of owner-occupied homes
The imputation for the rental value of owner-occupied homes
is made to provide comparable treatment between rented and
owner-occupied housing. It assumes that home ownership is a
business producing housing services which are sold to the homeowner in his capacity as tenant. These sales are estimated in terms
of the sum for which the particular type of home could be rented,
and the expenses of the home owners are deducted to obtain
imputed net rent. The imputed gross total becomes a part of sales
to persons, or consumer expenditures, and imputed net rent
becomes a part of the rental income of persons.
It may be wondered how the balance of the accounts can be
maintained if for imputed rents a gross item is entered on the
credit side and a net item on the debit side. The inconsistency is
only apparent. Adjustments corresponding to the expense items
which constitute the difference between imputed gross and net
rent are made simultaneously in several components of the gross
income and product flow, and secure balance.
One of the expense items, depreciation on owner-occupied
homes, is added to capital consumption allowances. In the
absence of imputation, it would not enter the purview of the
national accounts. Indirect business taxes are raised by the
amount of property taxes included in the expenses of owneroccupants. Otherwise, these taxes would be classified as personal
taxes (see the discussion of the personal account below). Mortgage
interest serves to raise the "net interest" item in the business
account. Without the imputation, it would be entered as interest
paid by the personal sector. Finally, all other expenses, such as
for supplies and materials necessary for the maintenance of
owner-occupied homes, are classified as intermediate business
purchases charged to current account. Without the imputation,
they would be counted as final products, as elements of business
sales to persons.

Food and fuel produced and consumed on farms
The imputation for food and fuel produced and consumed on
1 r frJlnwc
x




rental imrnirntinn Tn this

BUSINESS

instance the accounts are reconstructed to conform to a situation
in which the farmer sells the food and fuel to himself. An imputation for the full value of the food and fuel (at prices received by
farmers for this type of product) is made in business sales to
persons, and an imputed net profit on the production of this food
and fuel is included in the income of unincorporated farm enterprises. The apparent inconsistency of a gross imputation on the
product side and a net imputation on the income side is resolved
in a manner similar to that of the rental imputation.
Use of the rental and farm imputations, it should be noted,
avoids statistical difficulties that would occur if measurement
were restricted to monetary transactions. The imputation for
farm food and fuel, for instance, obviates the necessity, in arriving
at the income and product totals, of allocating expenses between
production for the market and for home consumption. Similarly,
it is unnecessary to allocate interest, maintenance, taxes, depreciation, and other housing expenses between owner-occupied and
tenant-occupied units. The statistical basis for making such allocations is tenuous. (Some of the detailed information shown in
Part V involves allocations of this type. However, within the
framework of the imputation procedure these do not affect the
income and product aggregates.) The estimation of consumption,
saving, and investment would be greatly complicated in the absence of a rental imputation because it would be necessary to
estimate the transfer of existing housing units between owneroccupied and tenant-occupied status. The amounts involved in
some periods are believed to be large but information required for
their determination is deficient.

Commercial banks and investment trusts
Imputed income and product flows arising in connection with
financial intermediaries involve some of the most complex constructions of national income and product measurement. Several
distinct types of procedures are involved, and will be discussed in
turn.
Exhibit 13.—Income and Product Account of a Commercial Bank,
Transactions Only

Monetary-

[Thousands of dollars]
Wages paid
Net interest paid
Interest paid on deposits
Less: interest received
Profit

50
—95
5
100
30

Service charge receipts
Less: current account purchases from
other firms

Income originating

—15

Product originating.

10

-IS

The ordinary methods of measuring value added to total output
in terms of income and product flows break down in the case of
commercial banking. This is because an element of the income
and product in this area does not take monetary form. An imputation is introduced to make it explicit and, as a consequence, a
much more satisfactory picture of value added emerges.
The problem is illustrated in exhibit 13 by means of an income
and product account for a commercial bank, drawn up in conformity with the principles of the Consolidated Business Income
and Product Account. Only a few essential transactions are
covered, in order to simplify the presentation.
On the credit side of this account, the value added to output by
the commercial bank is calculated in terms of monetary product
flnws hv HpHnrrinc from its sales rnnsistine' nf monetarv service-

NATIONAL INCOME,

charges, its current account purchases from other firms. (It will
be remembered that, for industry in general, this netting yields
the desired value of final production since sales and purchases of
intermediate products cancel for the economy as a whole.) In
terms of monetary income flows, value added is obtained on the
debit side by summing the distributive shares, with interest netted.
Since monetary service charges made by commercial banks are
low in relation to total costs incurred, income and product originating appears low—in the present example it is actually negative.
It is evident that the conventional method fails to give a proper
accounting of output originating in the commercial banking area.
Some income and product flows not taking monetary form must
occur, omission of which results in seriously incomplete measurement.
The product flows in question are identified as the services
rendered by banks without explicit charge to their depositors,
such as checking, bookkeeping, and investment services in connection with the handling of deposits. In lieu of monetary service
charges, banks finance the cost of these services by retaining part
of the property income earned by investing deposits instead of
paying it out to the depositors. This retained income is assumed
to represent the income flows not taking monetary form.
Exhibit 14.—Income and Product Account of a Commercial Bank, Monetary
and Imputed Transactions
[Thousands of dollars]
Wages paid
50
Net interest paid-.
0
Monetary interest paid on deposits
5
Imputed interest imid on deposits.. 95
Less: monetary interest received... 100
Profit
30

Service charge receipts
Monetary
Imputed
Less: current account purchases from
other firms

Income originating.

105
10
95

Product originating.

25

To obtain an adequate picture, the accounts are redrawn as
they would appear if this short-circuiting of income and service
flows had not occurred and, instead, commercial banks had (1)
paid out to depositors all property income earned on the investment of their deposits and (2) charged them fully for the cost of
the services rendered to them. An item for imputed interest paid
(equaling property income received minus interest paid on deposits) is entered on the debit side of the accounts. On the credit
side, an entry is made for imputed service charges (equalling
total operating expenses of banks, including profits, less monetary
service charges). It can be seen by reference to exhibit 13 that the
two must always balance: imputed service charges=wages paid
(50)+ current account purchases (25)+profits (30)—monetary
service charges (10)=imputed interest paid=interest received
(100) —interest paid on deposits (5) = 95.
This imputation is added, in exhibit 14, to the data shown in
exhibit 13.
The nonmonetary income and product flows having been made
explicit, a more adequate accounting of the value added by
commercial banks appears. Also, these banks are revealed in their
role of financial intermediaries. Interest is seen not to originate in
banking, but to be transferred by banking from the industries in
which it originates to the depositors to whom it accrues. (A minor
complication is introduced if, in addition to receipts of interest
income, receipts of dividends by commercial banks are taken into
account. See Part III, section on Interest.)



195 4

EDITION

47

Next, the imputed banking flows must be traced further through
the economy to determine their ultimate effect on the size and
structure of national income and product. This is done on the
basis of the ownership of the deposits in connection with which the
imputed flows arise. To the extent that these deposits are owned
by businesses, matching debit and credit entries are made in their
accounts—the debit for imputed service charges paid and credit
for imputed interest received. The balancing of the accounts is
not disturbed; for businesses affected, purchases of intermediate
products are increased by the amount of imputed service charges
and net interest paid is decreased by the amount of imputed
interest received.
Thus, to the extent that the deposits of commercial banks are
held by business owners, the imputation process does not change
the size of national product or income. Imputed service charges
cancel as intermediate products in the consolidation of the
business system, and so do the inter-industry imputed interest
flows. All that occurs is a redistribution in the industrial origin of
output, in the process of which the share of banking is increased
and that of other industries is reduced.
To the extent, however, that the ownership of bank deposits is
vested in persons, the results are different. Imputed service
charges made to persons constitute payment for a final product
and appear as a component of sales to persons under personal
consumption expenditures for "seivices rendered without explicit
charge by financial intermediaries, other than life insurance."
Imputed interest paid by banks to persons serves to increase net
interest by an identical amount. Both product and income originating are thus raised to reflect nonmonetary income and product
received by persons from banking. (In this discussion, it will be
noted, the accounting for imputed flows between banking and
government has been neglected. For this detail, see the section
on Interest, in Part III.)
The above description of the measurement of imputed flows
in banking is only a brief summary of a complex subject which
is still the subject of lively discussion among technicians in the
field. The procedure has been criticized in general as unduly
complex and, more specifically, as based on certain assumptions
of doubtful validity. Particular exception has been taken to the
assumption that all banking services not explicitly charged for
are rendered to depositors and that the borrowers of bank loans
are not involved, as well as to the assumption that these services
are distributed in proportion to the ownership of the volume of
deposits irrespective of turnover.
While these and other objections have some merit and it is
hoped that a simpler and more cogent solution may be found to
deal with the underlying problem of measuring the value added
to output by banking, it would appear that the present procedure,
all things considered, is the most satisfactory devised so far.
Whatever its particular limitations, it attempts to measure a real
element of income and product in the business economy and
permits a sensible solution to the problem of allocating income
by industries.
An imputation essentially similar to that for banking is made
in connection with investment trust type of financial institutions.
The precise mechanism of this imputation can be traced in the
light of the above discussion of banking on the basis of the additional detail provided in the section on Interest, in Part III.

48

A SUPPLEMENT

TO THE SURVEY OF CURRENT

Life insurance and mutual financial intermediaries
other than life insurance
The treatment of life insurance involves the second major type
of imputation which is made in connection with financial intermediaries. Imputations are introduced because the standard
national income and product classifications break down owing
to the combined saving and insurance functions performed by life
insurance. It is not possible to classify the explicit transactions
which occur between life insurance companies and their policy
holders into the conventional classifications of current receipts
versus capital transfers and of consumption and saving. Accordingly, in the income and product accounts imputed transactions
are substituted for the explicit transactions.
Specifically, claims and premiums are disregarded. Next, the
property income of life insurance companies which is withheld
to the account of policy holders is treated as if it were actually
disbursed in the current period. This item becomes imputed
interest in the net interest component of income. Finally, the
companies are regarded as explicitly charging policy holders for
their services, as measured by operating expenses. An imputation
equal to these expenses is entered in the business account under
sales to persons. It appears in personal consumption expenditures
as "Expense of handling life insurance."
That a balance between the income and product accounts is
secured if life insurance is treated in this manner can be seen
most simply by realizing that, as far as the totals are concerned,
life insurance companies have in effect been treated as individuals
rather than businesses. Claims and premiums have been cancelled
as though they constituted transfers among individuals; property
income received by these companies has been converted via the
interest imputation into property income received by policy
holders; and operating expenses incurred by the companies have
been converted by means of the service charge imputation into
final purchases made by policy holders. The balance of the Consolidated Business Income and Product Account thus reflects, in
essence, the balancing accounts of the business system other than
life insurance.
The effect of the treatment of life insurance on personal saving
may be anticipated at this stage. Since the property income and
operating expenses of life insurance are imputed to policy holders,
and receipts and payments of premiums and death claims are dis-

BUSINESS

regarded, a measure of personal saving results (in the personal
income and expenditure account described later) which consolidates the saving of life insurance companies with that of policy
holders.
An illustrative treatment involving mutual life insurance,
shown in exhibit 15, may serve to make this summary more
concrete.
The upper panel of the exhibit records the monetary transactions which occur in a simple economy involving life insurance
companies, other businesses, and persons. The lower panel reflects
the transactions that would be recorded in the national income
and product accounts. (Since nonbusiness production does not
occur in this example, a distinction between business output and
national output need not be made.) The lower panel differs from
the upper panel by excluding death claims and premiums and by
including imputed income and service transactions.
Gross national product (2200) is obtained from the lower panel
by adding sales (imputed) to persons by life insurance companies
(600), sales to persons by other business (1100), and business
capital formation (500). National income and personal income
(also 2200) are obtained by adding wages paid by life insurance
(200), wages paid by other business (1200), and (imputed) interest paid by life insurance (800). Personal saving (500) is obtained by deducting from personal income (2200) monetary and
imputed personal consumption expenditures (1700). It can be
seen that personal saving reflects the consolidated saving of persons and life insurance, as shown by the consolidated change in
their net asset positions (400 for life insurance and 100 for persons,
as indicated by the differences between the credit and debit totals
in the upper panel).
The treatment of stock life insurance companies is essentially
similar to that of mutual life insurance except that the operating
expenses of stock life insurance companies are measured to include
the companies' profits, which are correspondingly included in
total income.
Further detail on the specific items entering the calculation of
the property income flows (in the present example only interest
was allowed for), together with information on somewhat similar
imputations in connection with mutual financial intermediaries
other than life insurance, is given in Part III in the section on
Interest. This should be read in the general framework provided
above.

Exhibit 15.—Illustration of Treatment of Mutual Life Insurance
[Thousands ot dollars
Mutual Life Insurance
Debits

Other Business
Credits

Death claims .
Wages
Cost purchases

100
200
- 400

Wages
Cost purchases
Interest

200
- _ 400
_
*S00

Premiums....
Interest

700

Debits
300
800

Interest
Wages

1100
Interest
Sales.

1400
•Imputed.




800
- . . - «600
1400

Persons
Credits

800
1200

Sales
400
Sales
1100
Sales (business
capital account) . . . . 500

2000
Interest
Wages

800
1200

2000

Debits
Premiums
Purchases

300
1100

Purchases
Purchases

1100
*600

2000
Sales
Sales
Sales (business
capital account)

400
._ 1100

Credits

1400

500
2000

Death claims
Wages
Wages

1700

100
200
120©
150O

Wages
Wages
Interest

200
1200
*800
2200

49

NATIONAL INCOME, 195 4 EDITION

Government enterprises
In addition to financial institutions, other types of business organizations require special treatment in the national income and
product accounts. Government enterprises deserve explicit comment, not because they are important quantitatively in the
United States economic structure, but because they complicate
the accounts in a rather obtrusive way.
The distinction between government enterprises and general
government can be understood readily even though it cannot be
drawn with theoretical precision. Government enterprises are
those agencies of government whose operating costs are at least
to a substantial extent covered by the sale of goods and services,
in contrast to the general activities of government which are
financed mainly by tax revenues and debt creation. Government
enterprises, in other words, conduct operations essentially commercial in character even though they perform them under
governmental auspices. The Post Office and public power systems
are typical examples of government enterprises. On the other
hand, State universities and public parks, where the fees and
admissions collected cover only a nominal part of operating costs,
are part of general government activities.
Since a choice must be made, it is preferable to consolidate
government enterprises with business rather than with government. However, it seems desirable in handling these entities to
introduce certain departures from the standard procedures
adopted for ordinary business enterprises. (1) The profits of government enterprises are not included as part of factor cost in
income originating in business, but instead are treated as a nonfactor charge against the value of output (see "subsidies minus
current surplus of government enterprises" in table II). (2) The
capital formation of these enterprises (including both fixed capital
formation and inventory change) is classified in government
purchases rather than gross private domestic investment. (3) The
profits ("current surplus") of government enterprises are calculated without deduction either of net interest paid by them or
of depreciation. Therefore, depreciation charges of government
enterprises are not included in capital consumption allowances;
and net interest payments by government enterprises are not
counted as net interest payments by business. Since these modifications of the standard treatment of businesss enterprises do not
involve changes in the debit or credit totals, it can be seen that
the balance of the business account is not disturbed.
The effect of the treatment of government enterprises on the
government account (see table IV) may be anticipated at this
stage by noting that net interest paid by these enterprises is
combined with that paid by general government and that their
current surplus is treated as a receipt in the government account.
These steps, in conjunction with (2) above, serve to consolidate
the surplus (or deficit) of general government with that of
government enterprises. This is so because net interest paid plus
capital formation less current surplus of government enterprises
measures the net excess of their expenditures over their revenue.
Several considerations suggested the particular accounting for
government enterprises adopted in this report. With respect to
the profits ("current surplus") of government enterprises, it was
thought desirable to exclude them from factor charges because,
in a way too difficult to disentangle, they were recorded net of



losses which in effect reflected government subsidy operations
conducted through the medium of government enterprises,
mainly in World War II. The inclusion of government enterprise
losses due to subsidy operations would have offset the corresponding subsidies in total income originating, and would have run
counter to the general procedure of treating government subsidies as part of total factor cost.
The decision not to count net interest paid by government
enterprises as net interest paid by business (and, correspondingly,
to calculate the current surplus of government enterprises before
deduction of net interest paid by them) was closely related to the
decision not to treat their profits as part of factor cost. In general,
a meaningful total of factor cost with respect to property factors
can be obtained only if profits and interest are combined, and
the inclusion of net interest paid alone might have been misleading.
Next, government enterprise capital formation was combined
with government purchases rather than with private investment
because the dividing line between capital purchases by government enterprises and those by general government is quite
arbitrary. For instance, the construction of post offices is recorded
in the general budget of the United States rather than in the
accounts of the Post Office. Pending an exhaustive classification
of all government purchases of capital goods, it was thought
preferable to merge government enterprise capital formation
with government purchases.
Finally to be noted is that the government surplus or deficit
(consolidating both government enterprises and general government) which is obtained by this general procedure is the most
useful definition of government surplus or deficit for many types
of economic analysis.
The main aspect of the handling of government enterprises is.
their treatment as business-type organizations in order to avoid
the classification of their current expenses as final purchases.
Beyond this aspect, however, the treatment of government enterprises is in essence not more than a convenient means of disposing
of a type of operation that has not reached quantitative importance in the United States total income and product picture.
Were government enterprise operations to assume greater importance in the United States economy, it is entirely possible that
some modification of their treatment in the national income accounts would be called for.

The Personal Sector
The personal sector of the economy covers essentially the consuming public. It consists chiefly of individuals in their capacity
as income receivers and consumers, but it includes also nonprofit
institutions, private trust funds, and private pension, health, and
welfare funds.

Personal account

•

Unlike business transactions, which are summarized by a profit
and loss type of statement exhibiting the profit or loss realized in
the current period, personal transactions are summarized by a
statement of current receipts and expenditures. This difference
reflects, of course, the fundamentally dissimilar nature of the two
sectors of the economy.

50

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

The personal account, shown in table III, represents a consolidation of the accounts of all the persons who constitute the personal sector, just as the business account presented in table II
was derived by consolidating the accounts of all the firms included
in the business sector.
The personal account shows, in general, the transactions of
persons with the other sectors of the economy. Since nonprofit
institutions, private trust funds, and private pension and related
funds are regarded as part of the personal sector, income receipts
of these entities from other sectors are included in personal receipts and their purchases from other sectors are included in
personal expenditures.
Conversely, since the account is consolidated, most transactions
between these entities and individuals, as well as among individuals, are cancelled out in the process of consolidation. This
process of cancellation is not extended, however, to the transactions among persons that are regarded as purchases of the services
of factors of production—for instance, wages paid to domestic
servants and payments of wages and interest by nonprofit institutions. Instead, these transactions are reflected on both the receipt
and expenditure sides of the account, in order to preserve a record
of them which is needed in tracing the total flow of production in
the economy.

Classification of personal income
The classification of personal income on the right side of the
personal account in table III follows closely the classification of
the income items on the left side of the business account in table
II. In addition, however, to the incomes originating in business,
it also includes incomes received from general government and
from abroad, as well as incomes derived from production within
the personal sector.
The nature of the incomes derived by persons from the business
system has been covered in the discussion of the business account.

Separate entries for each of the items disbursed by the business
system can be found in the personal account except for interest
and dividend payments, which are included in the interest and
dividend entries but not available separately, for lack of statistical
information.
It will be noticed that only incomes currently received by persons
are included in the personal account. Thus, the wage and salary
component measures disbursements in the current period, and
differs from wages and salaries earned in the same period by the
"excess of wage accruals over disbursements" (see table II).
Similarly, only corporate profits distributed in the form of
dividends appear in the personal account. For unincorporated
enterprises, however, no useful distinction can be made between
distributed and undistributed income, and the entire amount is
transferred to the personal account. Finally to be noted is that
personal contributions for social insurance—including contributions by both employees and self-employed—are excluded from
personal income. Along with the contributions by employers on
behalf of their employees, they constitute receipts to government
rather than to individuals.
Income receipts from government consist of wages and salaries,
other labor income, interest, and transfer payments. With the
exception of interest payments, they are listed separately in
table III. Government interest payments to persons are included
in personal interest income.
The definition of each of these income receipts is similar to that
of the corresponding receipts derived from business and does not
require separate discussion. It should be kept in mind, however,
that although the formal definitions of the items are similar their
actual content may be very dissimilar owing to the different
nature of government operations. For instance, wages and
salaries received from government include military wages and
salaries, a type of payment which is not made by the business
sector. Similarly, government transfer payments include social
security benefits, relief, and various payments to former members

Table HI.—Personal Income and Expenditure Account,

1950

[Millions of dollars]
Personal consumption expenditures:

15

Wage and salary disbursements:

Purchases of direct services:

16
17
18
19

Business
Government
Households and institutions
Rest of the world

120, 565
19, 631
6, 312
18

20
21
22
23

Other labor income:
Business
Government
Households and institutions

3, 440
330
53

24

Income of unincorporated enterprises and inventory valuation adjustment

Compensation of employees:
Wages and salaries paid
Supplements paid:
Employer contributions for social insurance
Other labor income
Interest paid
Income originating in and net and gross product of households and institutions

6, 312
18
53
1, 956

36, 140

8, 339
25

9, 207

Personal interest income

10, 628

Government transfer payments

14, 304

29

Business transfer payments

30

Net purchases from abroad

Dividends

28

11

8, 473

27

Net purchases from business

Rental income of persons

26

10

Less: Personal contributions for social insurance

31

PERSONAL INCOME

184, 340
1, 347 j

12

Personal tax and nontax payments

20, 920

13

Personal saving

12, 104

14

PERSONAL OUTLAY AND SAVING




227, 050

843
2, 894
227, 050

NATIONAL INCOME, 195 4 EDITION

of the military establishment, all of which constitute payments
that are unique to government operations and have no genuine
counterpart in business.
Net wage and salary receipts from abroad appear explicitly as
receipts of the personal sector.6 Personal interest and dividend
receipts from abroad are included with other personal interest
and dividend receipts under personal interest income and
dividends.
Personal income derived from households and institutions
consists of income receipts of individuals for productive services
rendered within the personal sector of the economy. Incomes
included under this heading are those received for labor services
rendered directly to households, such as domestic service, and
the incomes received by employees of, and personal suppliers of
capital funds to, nonprofit institutions. (The labor income is shown
explicitly in table III, but the interest income is merged with
other personal interest receipts.) As has been noted, in order to
maintain a comprehensive record of total productive activity
these transactions are not cancelled in deriving the consolidated
personal account even though they occur within the personal
sector.

Relation of national income and personal income
The bulk of personal income is derived from production, and
personal income is therefore used widely as an indicator of
economic activity. However, it is not a measure of the value of
national output because it excludes certain incomes that accrue
in production but are not distributed to persons and includes
certain other income receipts that do not accrue in production.
The relation between national income, which is a measure of
output in terms of factor income flows, and personal income is
shown in exhibit 16.
Exhibit 16.—Relation of National Income and Personal
Income, 1950
[Millions of dollars]
National income
Less: Undistributed corporate profits
Corporate profits tax liability
Corporate inventory valuation adjustment
Contributions for social insurance..
Excess of wage accruals over disbursements
Plus: Net interest paid by government
Government transfer payments
Business transfer payments
Equals: Personal income

239,956
12,934
17,829
—4,864
6,870
0
4,716
14,304
843
227,050

In this exhibit, personal income is derived by deducting from
national income all incomes earned in current production but
not received by persons and by adding to it the incomes received
by persons but not earned in current production. The deductions
consist of all elements of the "corporate profits and inventory
valuation adjustment" component of national income except
dividends (undistributed corporate profits, corporate profits tax
liability, and corporate inventory valuation adjustment) and of
the parts of employee compensation and unincorporated enterprise income not regarded as distributed to individuals (contributions for social insurance and the excess of wage accruals over
disbursements). The additions consist of transfer payments from
government and business and of net interest paid by government.
The latter item represents the excess of the total interest payments
6. For certain statistical limitations of this item, see section 12 of Part III.




51

by government over its total interest receipts, and must be added
to national income because the net interest component of national
income falls short of the interest receipts of persons by that
amount. (For a detailed tracing of interest flows, reference is
again made to the Interest section of Part III).

Personal outlay and saving
The debit side of the personal account contains three general
categories: personal consumption expenditures, personal tax and
nontax payments, and personal saving.
Personal consumption expenditures consist chiefly of net purchases from business, corresponding to the credit entry for consolidated net sales to persons in the business account. Also included
are purchases made directly by persons from abroad (mainly
while traveling abroad, but including also international remittances) and purchases of direct factor services.
Purchases by persons of direct factor services measure production originating in the personal account. They are entered at
their full cost, which consists of the compensation of employees
and interest payments. As already mentioned, much of this expense is matched by a receipt entry in the personal account itself.
However, employer and employee contributions for social insurance appear as a receipt in the government account, while the
interest cost of production in the personal sector is composed of
payments to all sectors of the economy.
It will be noted that there is no entry for purchases of goods and
services from government. This is simply because all government
agencies which are conceived as selling their services for a market
price are defined as government enterprises and classified in the
business sector, so that consumer payments to them, such as for
postage stamps, appear as purchases from business.
Payments by persons to general government, consisting chiefly
iof direct personal taxes, are classified as personal tax and nontax
payments, the second general category on the left side of the
personal account. This entry does not include contributions for
social insurance, which are excluded in the computation of
personal income. It may be noted, however, that the total individual income tax, including the portion withheld at source, is
treated as though initially received by the personal sector. In
other words, personal income is measured before deduction of
this tax. The different treatment of income taxes and social
insurance contributions is somewhat arbitrary. Disposable personal income, of course, is measured net of both of them.
The amount remaining out of personal income after the purchase of goods and services and payments to government is
personal saving. It comprises the saving of individuals, including
owners of unincorporated enterprises, and the saving of the
organizations that are considered part of the personal sector,
namely private pension, health and welfare funds, private trust
funds, and nonprofit institutions serving individuals. It is the
algebraic sum of the saving and dissaving of these groups.
Seen from another aspect, personal saving measures the net
change in the asset position of persons as between the beginning
and end of the accounting period. Personal saving is made up of
the net increase in all the kinds of assets in which recipients of
personal income invest, offset by the net increase in all the kindsof liabilities which they incur.

52

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS
Exhibit 17.—Illustration of Treatment of Nonprofit Institutions Serving Individuals
[Thousands of dollars]
Nonprofit Institutions
Debits

Wages (t)_
Interest (2)
Interest (3)..
Purchases (4)
Cash relief (6)

Individuals
Credits

16
3
2
12
4

Gifts (6)....
Gifts (7)
Dues (8)__
Interest (9)._

36

Debits
11
6
..- 8
7
31

Gifts (6)
Dues (8)
Purchases (10)
Interest (11)
Wages (12)....

Credits
11
8
40
1
6
66

Included are not only the items commonly thought of in connection with personal saving, such as changes in cash and
deposits, in security holdings, and in personal indebtedness, but
also the net investment of noncorporate business in realty, equipment, and inventories. Personal saving also includes changes in
the reserves of life insurance companies and mutual financial
institutions, as explained above in the discussion of the treatment
of these entities in the national income and product accounts.
Assets are defined, of course, in the context of the conceptual
framework underlying national income measurement. Capital
gains and losses are not counted as changes in asset position, and
all consumer purchases of goods except residences are classified as
consumption rather than as investment.
A breakdown of personal saving by type of asset or liability is
provided in table 6, Part V of this report.

Imputations
As in the case of business output, the measurement of output
in the personal sector is not confined to monetary transactions,
but also takes into account imputed income and product flows.
The most important of the imputations is for the value of food
provided free to employees of households and nonprofit institutions. The imputation involves an increase in the wages and
salaries of employees, equal to the cost of the food to the employer,
on both the credit and debit sides of the personal account. The
reconstructed account depicts the situation which would prevail
if monetary wages were raised by the value of the imputation and
the corresponding food were purchased by the employees rather
than by their employers.
An imputation is also made for the value of free lodging furnished to clergymen, employees of nonprofit hospitals, and
certain quantitatively unimportant groups. None is made in the
case of domestic servants, because it is felt that, as a general
proposition, they do not regard the lodging furnished them as an
addition to income.

Nonprofit institutions
Nonprofit institutions include religious organizations, social
and athletic clubs, labor organizations, nonprofit schools and
hospitals, charitable and welfare organizations, and other nonprofit organizations furnishing services to individuals. It has
already been pointed out that they are consolidated with individuals in the personal account. While the principle of consolidation is clear, some fairly intricate manipulations of items are
involved. These may be illustrated by reference to exhibits 17 and
18, which are designed to show the treatment of the major
transactions involving nonprofit institutions. Certain simplifica


Corporations

Wages (1).
Wages (12)
Wages (13).Interest (2)
Cash relief (5).

Debits
15
6
45
3
4

Wages (13)
Gifts (7)...
Interest (9)

73

Credits
45
5
7

Interest (3)
Sales (4)
._
Sales (10)
Interest (11).

67

2
12
40
1
65

tions have been made to keep the example to manageable proportions. Investment by nonprofit institutions (counted as part
of private domestic investment) has been excluded, as have
government payments to nonprofit institutions (included in
government transfer payments).
Exhibit 17 shows a set of interrelated transactions of nonprofit
institutions, individuals, and corporate business. Exhibit 18 gives
the Personal Income and Expenditure Account corresponding to
these transactions. In essence, the latter account is obtained by
consolidating the transactions of nonprofit institutions and individuals. In this process, the tiansactions involving cash relief,
gifts, and dues cancel out. However, cancellation is not extended
to transactions between nonprofit institutions and individuals involving payments for factor services. The record of these transactions is preserved in the comprehensive accounting for national
output.
Exhibit 18.—Personal Income and Expenditure Account Based on Exhibit 17
[Thousands of dollars]
Personal consumption expenditures:
Purchases of direct services:
Compensation of employees:
Wages and salaries paid:
Wages (1)
15
Wages (12)
6
Interest paid:
Interest (2)
3
Interest (3)
. 2
Interest (11)
1
Income originating in and net and
gross product of households and
institutions
27
Net purchases from business:
Purchases (4)
12
Purchases (10)
40
Personal saving
2
PERSONAL OUTLAY AND
SAVING

81

Wage and salary disbursements:
Wages (1)-.
Wages (12)
Wages (13)
Personal interest income:
Interest (2)
Interest (9)
Business transfer payments:
Gifts (7)

PERSONAL INCOME

15
._ 6
45
3
7
5

81

Thus, employee compensation [wages (1)] and interest paid by
nonprofit institutions to individuals [interest (2)] appear on both
sides of the personal income and expenditure account and constitute components of "income originating and net and gross product
of households and institutions," which measures the value added
to national output by the personal sector. In addition, this measure includes the compensation of household employees [wages
(12)] and interest paid to the business sector by nonprofit institutions [interest (3)] and by households [interest (II)]. 7
7. The above summary covers the principal features of the national income accounting
treatment of nonprofit institutions. If allowance is made also for depreciation on plant and
equipment owned by nonprofit institutions, it can be seen that they are measured in personal
consumption expenditures by their current operating expenditures including depreciation
(but excluding domestic cash benefits paid). In the personal consumption expenditure estimates in table 30, Part V, the current operating expenditures of nonprofit institutions are in
some instances measured net of the institutions' receipts accounted for separately in consumer expenditures, such as receipts from the sale of food, housing, and recreational services.
Of further note, it is often not possible to measure current operating expenditures directly.
In such instances, indirect approximations—such as gross receipts less benefits paid—are
used, and the items in table 30 are labeled correspondingly.

NATIONAL INCOME, 195 4 EDITION

Private trust funds and private pension, health, and welfare
funds are also consolidated with individuals in the personal sector.
In these cases, however, the procedure is more straightforward
since the administrative expenses of these entities are small and,
in practice, can be neglected.

Further breakdowns of personal account
The personal account, like the business account, includes somewhat heterogeneous elements, and further breakdowns of it would
be useful. In particular, individuals might be distinguished from
the various types of quasi-individuals included in the personal
account; and, more important, individuals might be subdivided
into significant social groups, such as farmers, other businessmen,
independent professional practitioners, and wage earners, showing
separately the incomes of these groups and their disposition
among consumption, taxes, and saving. At the present stage, inadequacy of statistical materials limits the development of comprehensive measures of this sort.
The National Income Division does prepare, however, two sets
of estimates, not included in this report, that represent breakdowns
of the personal account which have wide analytical significance.
These are the series on personal income by States and by size of
income.

The Government Sector
The government sector includes Federal and State and local
general governments and the social insurance funds administered
by them. These funds comprise those set up under the Social Security and Railroad Retirement programs. State health insurance
funds, the retirement funds established for government employees, and military life insurance funds. The distinction drawn between general government and government enterprises, which
are included in the business sector, has already been described.

Government account
The transactions of government are summarized by a consolidated statement of receipts and expenditures, as presented in
table IV. In many ways this statement resembles the conventional
budgets of governmental bodies. However, there are several differences.
In the first place, the account shown in table IV is consolidated.
All levels of American government, the social insurance funds
administered by them, and the net expenditures of government
enterprises are covered. (However, separate breakdowns for the
Federal Government, State and local governments, and social
insurance funds are given in tables 8, 9, and 10 in Part V.)
Second, the account excludes receipts from the sale of, and
expenditures for the acquisition of, financial assets and secondhand fixed assets. Third, the timing of receipts and expenditures
differs from that of conventional budget statements, being synchronized with the timing of the corresponding expenditures and
receipts in the other accounts. Personal taxes are on a cash basis,
other taxes are on an accrual basis, and purchase entries reflect
time of acquisition rather than of payment. Finally, the classification of transactions differs from that of conventional budget statements, being adapted to the needs of national output measurement and general economic analysis.



53

Classification of receipts and expenditures
Most of the transactions contained in the government account
have already been discussed in connection with the business and
personal accounts.
The labor cost items, which appeared on a receipt basis in the
personal account, are on an accrual-cost basis in the government
account. The difference consists of employers' and employees'
contributions under retirement systems for government employees. In table IV the employer contribution is treated as a
simultaneous government expenditure and receipt. The employee
contribution appears as a deduction in the personal account
(under "personal contributions for social insurance") and as a
receipt in the government account.
Transactions with abroad, not yet discussed, appear explicitly
under "net purchases from abroad" and are included also in "net
interest paid." The former entry measures the excess of government purchases from foreigners over government sales to them
(cash gifts are treated like purchase and sale transactions in this
connection, but loans are excluded). Net interest paid is defined
as the excess of total interest payments by government to all sectors over total interest receipts from all sectors. Government
enterprises are covered in both payments and receipts.
Most of the salient features of the classification of government
receipts and expenditures have come to attention already, in
connection with the construction of measures of total national
output. However, certain points may be elaborated now that all
government transactions have been assembled.

The value of government output
Value added by government to national output, like value
added in all other nonbusiness sectors of the economy, is measured
by total factor cost incurred. In the case of the government,
factor cost is confined to the compensation of government employees. Interest payments are not counted. Two issues should be
considered as relevant to this treatment: first, the distinction between employee compensation and transfer payments; and,
second, the exclusion of monetary interest and the question of
substituting an imputed interest series for it.
Employee compensation versus transfer payments
A distinction is made between two types of government payments to individuals in their nonbusiness capacity—employee
compensation and transfer payments. The former is in return
for current productive services rendered; to the latter no such
services correspond. This distinction is a basic one, because it
segregates flows which are taken as measures of value added to
national output from flows which are not so regarded. Hence it is
important to see how it is made in practice.
Difficulties arise in the concrete interpretation of the term "productive service." For instance, in the classification of work relief
wages that were paid in the depression of the thirties, the question
arose as to whether they should be classified as labor returns or as
transfers. A more fundamental issue is raised by the national
income estimates of some foreign countries, in which military employee compensation has been excluded from factor income and

54

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS
Table IV.—Consolidated Government Receipts and Expenditures Account, 1950
[Millions of dollars]
16

Purchases of goods and services:
Purchases of direct services:
3
4
5
6

Personal tax and nontax receipts

20, 920

17 j Corporate profits tax accruals

7

Compensation of employees:
Wages and salaries
19, 631
Supplements:
Employer contributions for social insurance
812
Other labor income
330

8

Income originating and net and gross product

17, 829

18 j Indirect business tax and nontax accruals

23, 741

19 i Contributions for social insurance:
Personal contributions
Employer contributions:
Business
Government
20, 773
Households and institutions

9

Net purchases from business
Net purchases from abroad

2, 894
3,146
812
1&

17, 828

10

20
21
22
23
24

3, 422

11

Transfer payments

12

Net interest paid

13

Subsidies minus current surplus of government enterprises....

14

Surplus ( + ) or deficit (—) on income and product transactions
8, 113

15

GOVERNMENT EXPENDITURES AND SURPLUS

14, 304
4, 716
204

69, 360

treated as a transfer on the basis of some more ultimate, philosophical notion of productiveness.
In the United States estimates, the criterion for classifying an
item as employee compensation or as transfer hinges on the current performance of work. No attempt is made to probe into the
issue of whether the work is performed efficiently or whether, in
some more basic sense, it is "productive." The practical impossibility of drawing distinctions of this type has been covered inferentially in the evaluation of the basic notions underlying national
output measurement.
But even on the basis of the simple current-work-performed
criterion, a clear-cut distinction between wages and salaries and
transfer payments does not emerge in all cases. For it is not
always possible to say whether a specific payment is made for
the current work that is performed or for other reasons. Allowances for soldiers' dependents presented a classification problem
of this type. It was decided to classify them as employee compensation rather than transfer payments, although a case for the
opposite decision also could have been made. On the other hand,
retroactive terminal leave payments, bonuses, and other deferred
payments (such as under the "GI bill") to members of the Armed
Forces of World War II were considered transfers, as they were
disbursed at a date far removed from the time the military service
was performed. Counting these payments as compensation for
services would have necessitated allocating them over past years
on an accrual basis—a course which seemed artificial and would
have involved continuous revisions of the national income and
product estimates for the war period.

Government Interest
Government interest payments are not included in value added
to output by government because they are subject to fluctuations
which, it is believed, it would be artificial to regard as representing corresponding changes in the value of current production. It
may be added that business interest can be included in a measure



25

GOVERNMENT RECEIPTS

69, 360

of national output without such explicit consideration of its
behavior. Any fluctuations in it not reflecting productive activity
are offset by opposite changes in business profits. Hence, the
inclusion of business interest has no distortive effect on the output
measure and, in fact, is necessary to secure the correct total.
Thus, value added by government takes account only or the
services of the labor factor whereas the valuation of business output includes also returns to nonlabor factors. The question
accordingly arises as to whether an allowance should be made for
the services of government-owned property by the imputation of
a rate of return to it, somewhat analogous to the imputation of a
return to owner-occupied homes in the business sector of the
economy.
An imputation for government-owned property is not made in
the national income accounts for the United States because the
conceptual and statistical bases for making a realistic and useful
imputation are absent. The analogy to the housing imputation
does not hold. The bulk of this imputation is anchored to realistic
estimates of the gross rental value of owner-occupied houses
available from Census reports, based on comparisons of owneroccupied property with rental property of similar type. In the
case of the government no similar, market-based information to
establish the rental value of the vast bulk of government structures and equipment is, or can be, available. The rental value of
the highway system or of the Tennessee Valley Authority cannot
be estimated by reference to the rental value of property of
similar type.
In the absence of a realistic market evaluation of the rental
value of government property, its net return would have to be
derived by estimating the total value of government real capital
assets, segregating the part which is deemed to be in productive
use, and then applying to the latter a rate of return to reflect the
value added by the property. Clearly, each of these steps would be
highly speculative, and a measure of imputed return useful in
realistic analysis would not be likely to result.

NATIONAL

INCOME,

Decisions affecting valuation of business output
In the national product, output is valued at market prices—•
inclusive of indirect business taxes and exclusive of subsidies. The
manner in which these two items are denned therefore affects the
total value of national product.
Indirect business taxes versus personal taxes
Indirect business taxes are taxes (other than social insurance
contributions) that are chargeable to current cost by business
enterprises; and personal taxes are taxes paid by persons that are
not so chargeable.
This distinction leaves the treatment of retail sales and related
taxes somewhat ambiguous, since in some instances these taxes
are included in the sales price and charged to current expense,
and in others excluded and paid separately by the consumer. In
the latter case it would be possible to regard these taxes as personal
taxes and to list the corresponding purchases at values excluding
them. In this report the procedure of treating all these taxes as
indirect business taxes forming a part of market price has been
adopted, because it is thought to be the more meaningful from the
standpoint of studying market behavior.
Further emphasis on this type of study underlies a proposal to
depart from the accounting distinction between personal and
indirect business taxes used in this report. It has been suggested
that for analysis of consumer behavior all taxes that are closely
tied to consumer purchases should be treated as indirect business
taxes and included in personal consumption expenditures, regardless of whether they are chargeable to current cost by business.
For instance, automobile license and registration fees paid by
personal consumers would, according to this plan, be classified
in personal consumption expenditures and indirect business taxes,
rather than in personal taxes as at present. The logic would be
that payment of these taxes is a determinant in the choice of
consumers as between automotive and other types of expenditures.
While this suggestion has some merit, it would raise difficult
problems of classification. The influence of various types of taxes
on personal consumption is a matter of degree and does not
provide a clear-cut criterion of classification.
Subsidies versus purchases
Subsidies are monetary grants to business, and it is usually easy
to distinguish them from government purchases of goods and
services. However, in certain instances a subsidy element may be
included in the purchase price of an item in lieu of an explicit
subsidy. Cases of this type have involved payments by government enterprises (such as payments by the Post Office for air mail
contracts, which until recently have included an element of subsidy.) By virtue of the accounting for the current surplus and
deficit of government enterprises, as described earlier, indirect
subsidies of this type receive in effect the same treatment as
explicit subsidies. Payments of similar type on the part of general
government have not come to notice. To the extent that they
occur, they are regarded as purchases, and no attempt is made
to segregate the subsidy element.

Imputations
Imputations are made for wages and salaries paid in kind to
government employees. The most important of these imputations



1954

EDITION

55

is for food and standard clothing issued to members of the Armed
Forces. It may be noted that only standard, or personal, clothing
is included, not special clothing and equipment. Also, the rental
value of shelter provided is not allowed for. The principal line of
reasoning is that in many instances the provision of lodging to
servicemen does not reduce their cash housing expenditures and
hence is not a clear addition to their income.
In the recording of the imputations, the accounts are reconstructed to correspond to a situation in which the government paid out
to its employees additional wages equal to the cost (to the government) of the food and clothing provided and the employees themselves purchased these items from the business system. In the
government account (table IV), compensation of employees (and
income and net and gross product originating in government) is
raised by the value of the imputation. But total government purchases of goods and services are not changed because government
purchases from business are reduced by a corresponding amount.
In the business account (table II), there is a corresponding shift
from sales to government to sales to persons. Finally, in the
National Income and Product Account (table I), the imputation
raises personal consumption expenditures and gross national
product, on the one hand, and the compensation of employees,
national income and the sum of charges against the value of national product, on the other.
In view of the fact that an imputation is made for wages and
salaries received in kind, it may be wondered why transfer payments, and also subsidies, are confined to monetary transactions—
why specific goods and services given to individuals are not
included in transfer payments, and why those given to business
are not counted in subsidies. The basic reason for this apparent
inconsistency is that, whereas in the case of wages and salaries a
generally accepted procedure for imputation, partly pragmatic
and partly tradition-based, is available, this is not so with respect
to transfer payments and subsidies. The principal difficulties in
the way of establishing such a procedure have been set forth in the
discussion of the basic notions underlying national output
measurement.
By way of supplementary argument, it may be noted that the
introduction of transfer payments and subsidies in kind would
interfere with the function of the accounts as a record of actual
transactions. Of course, wage and salary and other imputations
that are made in the national accounts have a similar effect. In
these instances, however, it is felt that the resulting improvement
in the income and product components and totals outweighs the
disadvantages involved.
Two examples of the rather extreme complications of the
accounts that can arise from the introduction of imputations of
transfer payments and subsidies may be given.
Suppose, for instance, that services rendered by government
employees in the administration of relief programs were to be
regarded as transfer payments in kind. In the government
account (table IV), transfer payments would be increased and
compensation of government employees correspondingly reduced. In the personal account (table III), transfer payments
would be increased on the credit side and employee compensation (and hence purchases of goods and services and income and
net and gross product originating) on the debit side. In effect, the
government employees rendering the services classified as trans-

56

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

fer payments in kind would now be classified in the personal
sector, as household employees. Finally, in the National Income
and Product Account (table I), personal consumption expenditures would be raised, and government purchases lowered, by
the amount of the imputed transfer payments. It is apparent that
these reclassifications would greatly reduce the value of the
accounts as records of the income and product flows that actually
occur among the major sectors of the economy.
Or, assume that a subsidy in kind consisted of services rendered
by government employees. In this instance the reclassification
would involve labelling as subsidies in the government account
payments actually made for wages. In the business account wages
would have to be increased, to include wage payments actually
made by the government, and offset by a corresponding increase
in the deduction made for subsidies.

Social insurance funds
The consolidation of social insurance funds with the accounts of
general government is a straightforward procedure which need
not be detailed. However, certain differences between the treatment of social insurance funds and of privately administered
pension funds may be pointed out. These stem from the fact that
the social insurance funds are consolidated with government
while private pension funds are consolidated with individuals in
the personal sector.
With respect to the measurement of employee compensation
(and national income), the difference is merely one of classification between shares—employer contributions to governmentadministered funds are listed under "employer contributions for
social insurance," whereas employer contributions to private
pension funds are included in "other labor income." However,
the effect of the difference in treatment on personal income and
saving is more substantive. Employer contributions and property
income received by privately administered funds become elements
of personal income, and the saving of these funds part of personal
saving. Employee contributions into these funds, as well as benefit
payments made by them, are canceled as constituting transfers
within the personal sector. Employer and employee contributions
and property income received by social insurance funds, on the
other hand, enter government receipts; benefit payments made

by them are explicitly recognized (as transfer payments by the
government) as a component of personal income; and the saving
of social insurance funds is a component of the government surplus
rather than of personal saving.

Rest of the World Sector
The transactions of the rest of the world with the United States
are summarized in the rest-of-the-world account presented in
table V. The rest of the world covers foreign countries, territories
and possessions of the United States, international organizations,
and the United States monetary gold stock. The gold stock is included in this sector because net acquisitions of gold by the
monetary authorities from domestic sources are considered
foreign investment.
It may seem strange at first that the "rest of the world" is a
sector of the national economy, and, indeed, there would be no
need to consider it in the case of a closed economy which had no
dealings with foreign countries. In the real world, however,
trade and investment do cross international boundaries. Consequently, to complete the set of accounts it is necessary to include
one which summarizes the transactions of foreigners with the three
domestic sectors of the economy.
The rest-of-the-world account is a receipt and expenditure
account and, like the other accounts, consolidated. It bears a
close affinity to the balance-of-payment statement. It differs from
this statement mainly in arrangement, with respect to netting
and the classification scheme applied to the transactions involved.
The debit side of this account shows the net purchases of United
States goods and services by the rest of the world. It is divided
among net purchases of direct factor services and other purchases
(net) from business, government, and persons. Net purchases of
direct factor services from the United States by the rest of the
world, as shown by the net inflow of factor incomes to the United
States, measure United States national income and product
originating in the rest of the world.
Because of the inclusion of gifts in net purchases from United
States persons and government, net purchases of goods and
services by the rest of the world cover all its current transactions
with the United States. The excess of purchases over sales must
be financed by a change in the net international asset position (an

Table V.—Rest of theWorld Account, 1950

[Millions of dollars]
Net purchases of goods and services:

12

Net purchases of direct services:
Wages and salaries
Interest
Dividends
Branch profits

7

Income originating and net and gross product

8

Net purchases from business

9

Net purchases from government

NET DISINVESTMENT IN THE
UNITED STATES

- 2 , 201

— 3, 422

Net purchases from persons

—2, 201

13

2
3
4
5
6

Net disinvestment in the United States

— 1, 347

10
11

NET CURRENT PAYMENTS TO THE
UNITED STATES




18
248
426
574
7, 266

1, 302

- 2 , 201

NATIONAL INCOME, 195 4 EDITION

increase of United States claims on abroad or a decrease of foreign
claims on the United States). From the standpoint of the
rest of the world, this excess constitutes net foreign disinvestment
in the United States, as shown on the credit side of the account.

International gifts
Inclusion of cash gifts received by foreigners from the United
States in sales to United States persons and to the United States
Government, together with the corresponding inclusion of gifts
made by foreigners in purchases from the United States, secures a
treatment of these transactions which is appropriate for the
measurement of national product. Net gifts made by United
States persons and Government appear in personal consumption
expenditures and government purchases but are offset in the net
foreign investment component of national product.
Needless to say the treatment of gifts as purchases is a somewhat
unsatisfactory short-cut. It was adopted in order to simplify the
structure of the accounts at a time when the important role of
gifts in international transactions was not yet apparent. It has not
been discarded because the alternative, more elaborate treatment
also has serious shortcomings.
This alternative treatment consists of the establishment of a
separate category of international transfers in the income and
product accounts to cover the gifts now included in the purchases
and sales of goods and services. These international transfers
would affect, in addition to the rest-of-the-world account, the
personal and government accounts (tables III and IV). The
detail of the national income and product account (table I) also
would be affected. Personal consumption expenditures and government purchases of goods and services would be reduced by the
amount of net gifts made, and the entry under net foreign investment would be increased by a corresponding amount (with an
appropriate change in the designation of the term to indicate that
it would no longer reflect net foreign investment, but the balance
of transactions in actual goods and services).
The complication of the accounts which would be involved in
the establishment of a separate category of international transfer
payments, while considerable, would be warranted if it would
throw into clear relief the large international aid transactions involving the United States Government. However, this would not
be accomplished for the reason that international aid, in addition
to cash grants, involves also loans and aid in kind. The three forms
of aid often are almost indistinguishable from one another in their
economic aspects.
There is no clear-cut procedure available for distinguishing between international aid rendered in kind and government purchases for domestic purposes. For instance, in World War II it
would have been rather unrealistic to distinguish between lendlease, which presumably would have been classified as international aid in kind, and other government purchases for war purposes. Nor is there available an uncontroversial yardstick for
classifying loans into loans proper and those that in effect represent
international transfers. Since the relative magnitudes of the three
media for extending international aid of essentially similar nature
have been subject to considerable shifts, it would not have been
instructive to establish a separate category for cash gifts alone, the
only type of aid that could have been distinguished objectively.



57

It may be noted that conceptually similar problems arise with
respect to the domestic operations of government. Aid to individuals and business also involves monetary grants, goods and
services rendered in kind, and loans. In the case of these domestic
transactions, monetary aids (in the form of transfer payments and
subsidies) are recorded, but the two other types of aid are not
recognized as such. While from some standpoints this dividing
line is somewhat arbitrary, it is much more meaningful than a
corresponding line would be for international aid. In the domestic
sphere, shifts in the relative importance of the three media in
extending aid of essentially similar nature have not been present
in a comparable degree.

Treatment of gold
In essence, the treatment of gold production in national income
and national product is the same as that of any other commodity.
The distributive shares arising directly or indirectly in its production (together with nonfactor charges, such as indirect business
taxes and depreciation) are reflected on the debit side of the
national income and product account. The value of gold produced enters the credit side either as such or as an ingredient of
the value of some other final product, although the classification
in the product flow is less transparent than in the case of other
commodities. Domestic nonmonetary use of gold may be reflected
in any of the domestic components of national product—personal
consumption, domestic investment including inventory change,
and government purchases. To convert domestic nonmonetary
use into a measure of total domestic production, monetary use
and exports must be added and imports must be deducted.
These items—the change in the monetary gold stock and net
gold exports—which in combination measure net domestic
business sales of gold for export and monetary purposes, are
included in net purchases by the rest of the world from United
States business. In other words, the monetary gold stock is set up
as part of the rest of the world and its transactions with United
States business are treated as foreign transactions. Thus, changes
in the monetary gold stock not offset by gold exports and imports
come to be reflected in net foreign investment.
Silver is not regarded as an international monetary asset. It is
classified in the product flow exactly like any other commodity.

Gross Savins and Investment Account
The entries in the sector current accounts presented so far show
the current transactions of each of the four major economic
groups, yielding in each case a residual which represents a form
of saving. A logical extension of this system of sector current
accounts would be the establishment of corresponding sector
saving and investment accounts showing the disposition of these
savings in the form of net financial and real investment.
Sector saving and investment accounts of this type have not
been constructed on a comprehensive scale (a statement of this
type for persons is presented in table 6, Part V) and cannot yet
be made an integral part of the national economic accounting
system. A consolidated saving and investment account for the
Nation as a whole is presented instead, in table VI.

58

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

In the process of consolidation the financial investments involving transactions among the domestic segments of the economy
are canceled, and all that remains are matching flows of saving
and of domestic investment (fixed capital formation and inventory change) and foreign investment. The consolidated capital
account is obtained from the four sector current accounts by
assembling all items in these accounts that so far have been
entered only once because they constitute transactions not with
other current accounts but with the capital accounts of the same
or other sectors.
The content of the gross saving and investment account is
determined by the basic concepts and classifications underlying
national income accounting. Only business assets are included in
fixed capital formation and inventory change. Consumer- and
government-held tangible assets are not included in capital formation. It should also be recalled that fixed capital formation is
presented gross and that, therefore, capital consumption allowances appear as a component of (gross) saving.

DEFINITIONS OF CONCEPTS A N D TERMS
The following definitions of the national income and product
aggregates and their components are intended to give concise,
accurate descriptions of the coverage of the various series and,
at the same time, to call attention to the principal aspects of the
series which are not readily apparent from their titles. The definitions of the national aggregates should be considered in conjunction with the definitions of their components as the details of
the latter are not repeated in the former.

/. National Income and Product Aggregates
National Income is the aggregate earnings of labor and property
which arise from the current production of goods and services by
the Nation's economy. Thus, it measures the total factor costs of
the goods and services produced by the economy. The Nation's
economy in this context refers to the labor and property supplied

by residents of the Nation. Earnings are recorded in the forms ir
which they accrue to residents of the Nation, inclusive of taxes or
those earnings. As such, they consist of the compensation of employees, the profits of corporate and unincorporated enterprises
net interest, and the rental income flowing to persons.
Gross National Product or Expenditure is the market value o:
the output of goods and services produced by the Nation's economy, before deduction of depreciation charges and other allowances for business and institutional consumption of durable capita!
goods. Other business products used up by business in the accounting period are excluded. The Nation's economy in this context
refers to the labor and property supplied by residents of the Na
tion. Gross national product comprises the purchases of goods anc
services by consumers and government, gross private domestic
investment (including the change in business inventories), anc
net foreign investment.
Net National Product or Expenditure is the market value of the
net output of goods and services produced by the Nation':
economy. All business products used up by business in the ac
counting period are excluded. The Nation's economy in thi:
context refers to the labor and property supplied by residents o
the Nation. Net national product comprises the purchases of good
and services by consumers and government, net private domestii
investment (including the change in business inventories), anc
net foreign investment.
Personal Income is the current income received by persons fron
all sources, inclusive of transfers from government and busines
but exclusive of transfers among persons. Not only individual
(including owners of unincorporated enterprises), but nonprofi
institutions, private trust funds, and private pension, health, am
welfare funds are classified as "persons." Personal income is meas
ured on a before-tax basis, as the sum of wage and salary dLs
bursements, other labor income, proprietors' and rental income
interest and dividends, and transfer payments, minus persona
contributions for social insurance.
Disposable Income is the income remaining to persons after de
duction of personal tax and nontax payments to general govern
ment.

Table VI.—Gross Saving and Investment Account, 7950

[Millions of dollars]
Business purchases on capital account
Change in business inventories
Net disinvestment in the United States by rest of world

43, 868
7, 351
—2, 201

5
6
7
8
9
10
11
12

GROSS INVESTMENT



49, 018

Excess of wage accruals over disbursements
Undistributed corporate profits (domestic)
12, 3(
Corporate inventory valuation adjustment. .
—4, 8(
Capital consumption allowances by private business
20, 5
Foreign branch profits (net)
5
Personal saving
12, I1
Government surplus (-|-) or deficit (—) on income and product transactions
8, 1
Statistical discrepancy
2

13

GROSS SAVING AND STATISTICAL DISCREPANCY.

49, 0

NATIONAL INCOME, 195 4 EDITION

//. Components of National
Aggregates

Income and Product

A. National Income (as in table 1, Part V).
Compensation of Employees is the income accruing to persons
in an employee status as remuneration for their work. From the
employer's standpoint, it is the direct cost of employing labor. It
is the sum of wages and salaries and supplements to wages and salaries.

Wages and Salaries consists of the monetary remuneration of
employees commonly regarded as wages and salaries, inclusive of
executives' compensation, commissions, tips, and bonuses, and of
payments in kind which represent income to the recipients.
Supplements to Wages and Salaries is the monetary compensation of employees not commonly regarded as wages and
salaries. It consists of employer contributions for social insurance;
employer contributions to private pension, health, and welfare
funds; compensation for injuries; directors' fees; pay of the military reserve; and a few other minor items of labor income.
Income of Unincorporated Enterprises measures the monetary
earnings and income in kind of sole proprietorships, partnerships,
and producers' cooperatives from their current business operations—-other than the supplementary income of individuals
derived from renting property. As with corporate profits, capital
gains and losses are excluded and no deduction is made for
depletion.
Inventory Valuation Adjustment measures the excess of the
value of the change in the volume of nonfarm business inventories,
valued at average prices during the period, over the change in the
book value of nonfarm inventories. This adjustment is required
because corporate profits and income of unincorporated enterprises are taken inclusive of inventory profit or loss, as is customary
in business accounting, whereas only the value of the real change
in inventories is counted as current output in the national product.
No valuation adjustment is required for farm inventories because
farm income is measured exclusive of inventory profits.
Rental Income of Persons consists of the monetary earnings of
persons from the rental of real property, except those of persons
primarily engaged in the real estate business; the imputed net
rental returns to owner-occupants of nonfarm dwellings; and the
royalties received by persons from patents, copyrights, and rights
to natural resources.
Corporate Profits Before Tax is the earnings of corporations
organized for profit which accrue to residents of the Nation,
measured before Federal and State profit taxes, without deduction of depletion charges and exclusive of capital gains and losses.
Profits accruing to residents are measured by eliminating intercorporate dividends from profits of domestic corporations and by
adding the net receipts of dividends and branch profits from
abroad In other major respects, the definition of profits is in
accordance with Federal income tax regulations.
291692°—54

5




Corporate Profits Tax Liability comprises Federal and State
taxes levied on corporate earnings. Disbursements of tax refunds
are deducted from tax liability in the year in which the tax
liability was incurred.
Net Interest measures total interest (monetary and imputed,
private and government) accruing to United States persons and
governments minus total interest paid by United States governments. Government interest (Federal and State and local) is
deducted because it is not considered income arising in current
production. It is necessary not only to exclude the portion of it
paid directly to persons and governments, but also to deduct the
portion of it paid to business, because the latter is reflected in the
incomes paid out or retained by the business system. The imputed
interest component of net interest is measured in general as the
excess of property income received by financial intermediaries
from funds entrusted to them by persons over property income
actually returned in monetary form by these intermediaries to
persons. A portion of imputed interest is numerically equal to the
value of financial services received by persons without explicit
payment; the remainder represents property income withheld by
life insurance companies and mutual financial intermediaries on
the account of persons.
8. Gross National Product (as in table 2, Part V ) .

Personal Consumption Expenditures consists of the market
value of purchases of goods and services by individuals and nonprofit institutions and the value of food, clothing, housing, and
financial services received by them as income in kind. It includes
the rental value of owner-occupied houses but does not include
purchases of dwellings, which are classified as capital goods.
Gross Private Domestic Investment consists of acquisitions of
newly produced capital goods by private business and nonprofit
institutions and of the value of the change in the volume of
inventories held by business. It covers all private new dwellings,
including those acquired by owner-occupants.
Net Foreign Investment is the net change in international assets
and liabilities, including the monetary gold stock, arising out of
the current international flows of goods and services, factor incomes, and cash gifts and contributions. Thus it measures the
excess of (1) domestic output sold abroad over purchases of
foreign output, (2) production abroad credited to United Statesowned resources over production at home credited to foreignowned resources, and (3) cash gifts and contributions received
from abroad over cash gifts and contributions to foreigners. The
net transfer of cash gifts and contributions offsets corresponding
entries in personal consumption expenditures and government purchases of
goods and services.

Government Purchases of Goods and Services measures purchases of goods and services by government bodies, exclusive of
acquisitions of land and used depreciable assets and of current
outlays of government enterprises. It consists of general government expenditures for compensation of employees, purchases from
business (net of sales by government of consumption goods and
materials), net government purchases from abroad and inter-

60

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

private pension, health, welfare, and trust funds. Personal saving
national contributions, and the gross investment of government
enterprises. Therefore, government purchases of goods and servicesmay be in such forms as changes in cash and deposits, security
holdings, indebtedness, and reserves of life insurance companies
excludes transfer payments, government interest, and subsidies,
and mutual savings institutions, the net investment of unincorpoas well as loans and other financial transfers outside the scope of
rated enterprises, and the acquisition of real property net of
income and product transactions.
depreciation.
C. Personal Income and Disposition of Income (as in table 3,
Part V).
D. Reconciliation Items Between National Income and Gross
National Product (as in table 4, Part V).
Wage and Salary Disbursements is equal to wages and salaries,
except that retroactive wages are counted when paid rather than
Depreciation Charges represents the charges made by private
when earned.
business against receipts for the current consumption of durable
capital goods and comparable allowances for nonprofit instituOther Labor Income is the same as supplements to wages and tions. It includes depreciation charges against owner-occupied
salaries exclusive of employer contributions for social insurance houses. Depreciation reported by business is not adjusted for
in national income.
changes in the replacement value of capital goods, except for
farm enterprises.
Proprietors' and Rental Income is the sum of income of unincorporated enterprises and inventory valuation adjustment and rental income Accidental Damage to Fixed Capital measures the value of
of
persons as given in the components of national income.
the physical losses by fire, natural events, and other accidents to
fixed capital of private business, not covered by depreciation
Dividends measures cash dividend disbursements by corporacharges.
tions organized for profit to stockholders who are United States
persons.
Capital Outlays Charged to Current Expense represents new
construction and purchases of new durable capital goods included
Personal Interest Income measures total interest (monetary and
in gross private domestic investment that are charged as current
imputed, private and government) accruing to United States
expense by business rather than entered on capital account.
persons. The imputed interest component of personal interest income
is the same as in national income.
Indirect Business Tax and Nontax Liability consists of tax liabilities incurred by businesses, except corporate income taxes, and
Transfer Payments consists of monetary income receipts of
other general government revenues from business. It includes all
individuals from government and business (other than governsales taxes. It includes payments for such specific services as are
ment interest) for which no services are rendered currently, of
provided within the framework of general government activity.
government payments and corporate gifts to nonprofit instituIt excludes, however, purchases from government enterprises.
tions, and of individuals' bad debts to business.
Government receipts from the sale of surplus property are not
included in this item. Tax liabilities are net of refunds.
Personal Contributions for Social Insurance consists of payments by both employees and self-employed. Contributions of the
Subsidies Minus Current Surplus of Government Enterprises:
self-employed, which relate to old-age and survivors insurance,
were first made in 1952.
Subsidies are the monetary grants provided by government
to private business.
Personal Tax and Nontax Payments consists of the taxes levied
against individuals, their income, and their property that are not
Current surplus of government enterprises represents the exdeductible as expenses of business operations, and of other general
cess of sales receipts over current operating costs of government
government revenues from individuals in their personal capacity.
enterprises. In the calculation of the current surplus, no deducIt includes payments for such specific services as are provided
tion is made for charges to depreciation or other reserves and
within the framework of general government activity. It excludes,
interest is not counted in either receipts or costs.
however, purchases from government enterprises. Tax refunds
are deducted from payments as of the time of refund.
Subsidies and current surplus are shown as a single item
because of the difficulties involved in segregating subsidies paid
Personal Consumption Expenditures is the same as in gross
through Federal Government enterprises from other expendinational product.
tures of these enterprises.
Personal Saving is the excess of personal income over personal
consumption expenditures and personal tax and nontax payments. It consists of the current saving of individuals (including
owners of unincorporated businesses), nonprofit institutions, and




Statistical Discrepancy is the excess of the value of the estimated gross national product computed by the final products
method over its independently estimated value computed by
adding necessary conceptual adjustments to the national income.

PART

III

Sources and Methods of
National Income Estimation
INTRODUCTION
The statistical methodology underlying United States national income estimates is of interest to two broad groups—to users of the data and to producers
of similar and related estimates. The importance of the former group is obvious. The importance of serving the latter group is easily appreciated if it is
realized that it includes workers in the entire field of economic statistics, not
only in this country but all over the world.
Users of national income statistics are concerned with sources and methods
mainly from the standpoint of their bearing on the accuracy of the statistics.
The interest of this group in methodological detail is limited because much
of it is not germane to the question of the reliability of the estimates. Producers
of statistical data, on the other hand, want to learn about sources and methods
in order to derive assistance in their related tasks. They are likely to require
knowledge in greater detail.
In practice, however, this conflict of interests is not too serious. For, as will
become apparent, the user of the data cannot form a judgment of their reliability without a considerable acquaintance with statistical methodology. In
turn, the interest of the professional statistician in specific detail tends to be
limited, because much of it is not applicable to his situation. Hence, in writing
the statistical descriptions which follow, it seemed appropriate to plan a compromise that would be helpful to general users of the data as well as to technicians.

Salient Features of the Statistical Methodology

National product, which is a measure of total output in terms of product
flows, is obtained generally by adding component estimates of the purchases
of final products by major purchaser groups. Since measurement is restricted
to final product flows, the mutually canceling purchases and sales of intermediate products, which would be necessary to determine the industrial distribution of output, cannot be taken into account.
This lack is not felt to be a significant gap in the statistics, since the corresponding breakdown of national income serves most needs for an industrial
distribution of total output. (It may be noted that an indirect measure of
gross national product by industrial origin could be obtained by adding to
the national income originating in each industry other charges against gross
national product and deducting subsidies. However, most of these elements
of reconciliation have not been distributed on an industrial basis.)
Moreover, this lack of an industrial distribution of national product is
more than counterbalanced by the economic significance of the breakdowns
of the product flow by type of purchaser (and also by type of product) which
are yielded by the final product approach. In fact, these breakdowns provide
essential elements for extending the scope of national income and product
statistics beyond the mere measurement of output totals into a comprehensive
national economic accounting system providing a statistical picture of the
economy.
The components of national income and national product are the core of
this picture. The remaining statistical task consists of deriving supplementary
breakdowns and series necessary for a fully articulated picture of the economy,
mainly in terms of the interrelated transactions of its major constituent
groups—businesses, households and institutions, government, and foreign
nations.

Stages of statistical measurement

The reporting units

Since national income and product are measures of total national output,
it might appear that the most direct way to obtain these measures would be to
sum the values added to total output by each of the industrial sectors of the
economy. In terms of product flows, these values would, in general, be
measured as the total product of the industry minus its purchases of intermediate products from other industries. This difference equals the sum of
wages, interest, profits and other distributive shares accruing in the industry
plus certain additional charges against the value of its production. Seen from
this aspect, the summation of industrial values added would yield a measure
of output in terms of incomeflows.If national output were calculated according to this plan, its breakdowns by industry of origin would be the basic
statistical building blocks. In fact, however, the data on value added by each
industry are not available directly, and total output must be estimated by
other procedures.
National income, which is a measure of total output in terms of factor income flows, is estimated by summing estimates of the various distributive
shares. Data to estimate these shares are not always available on an industrial
basis, and only as the result of often complex supplementary calculations is
an estimate of national income by industrial origin obtained.

Only on a limited scale is reliable statistical information available from
individual consumers, because of the general inferiority of consumers' records
and other difficulties involved in the collection of data from them. Hence,
even over the broad areas in which national income and product flows reflect
transactions in which individual consumers are involved, the statistical information for making the estimates is usually derived from the other parties to
the transactions—chiefly businesses and government. These sources, in
general, are decidedly preferable. The character of the underlying records is
superior; the number of reporting units is smaller; and the reporting systems
necessary for the procurement of reliable information can be developed more
readily.




Major reporting systems
The entries into the national income and product tables are derived from a
multiplicity of statistical sources ranging over most of the essential phases of
the Nation's economic life.
A large body of statistical data is collected by the government, mainly with
the intent of providing information which is of general interest to broad user
61

62

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

groups. The various censuses—such as the Census of Manufactures, the Census
of Business, and the Census of Agriculture—are prime examples in this category. Of equal importance is the statistical information which becomes available as the byproduct of the administrative functions of the government.
The wage and salary data provided in connection with the old-age and survivors and unemployment insurance programs by the Health, Education,
and Welfare, and Labor Departments are an outstanding example. Another
is the information on the incomes of corporations and of proprietorships and
partnerships furnished as a byproduct of Federal income tax administration
by the Internal Revenue Service.
Government-produced statistics are the mainstay of national income and
product estimates, but they are supplemented by a wide variety of information obtained from private sources, such as trade associations, labor organizations, research organizations, private educational groups, and religious
and welfare organizations.
By any comparative standard the available information for estimating
national income in the United States must be judged abundant. Also, long
experience in the reporting and collection of economic data, the stringent
requirements of governmental administrative agencies, and emphasis on
sound statistical techniques make for a generally high degree of reliability.
The adequacy of the data has improved. Outstanding in this connection
has been the signal increase in the current reporting of economic information.
This development, closely connected with advances in the theory and practice of sampling techniques, has put the preliminary estimates of national
income and product, which are made pending the availability of final benchmark data, on a much firmer basis. And, indeed, it has made it possible for
these estimates to be made with little time lag and in considerable detail.

Estimating procedures
In spite of the abundance and general reliability of statistical information,
the task of the national income estimator is complicated because the basic
data are not collected in the framework of a coordinated statistical program
designed to fill the needs of national income measurement. Without significant exception, the reported information is not in a form in which it can be
entered directly into the national income and product accounts. It must be
processed further to fill gaps in coverage and to adjust for differences in
definition. This processing of the data involves procedures that are often
quite complex. Although the mathematical operations used are usually
simple, involved estimation may be necessary to combine and adjust a multiplicity of diverse sources to produce series that have the coverage and
definition required by the national income estimates.

Reliability of the Estimates
It is clear from this summary discussion of the derivation of national income estimates that they are subject to error. Hence it is important to evaluate the degree of their reliability.
A comprehensive statement of the degree of accuracy of a given estimate is
usually thought to involve the specification of a frequency distribution of
similar estimates around the universe value. In the field of national income
this ideal cannot be approximated. The many source materials and procedures
utilized are not of such a nature as to permit calculations of the probable
errors in the various income and product series.
The replacement of present estimating methods by the sampling approach
would not be a solution. Serious difficulties in applying sampling techniques
would be encountered with income and product components for which knowledge about the size and characteristics of the universe was lacking, or whose
composition was heterogeneous or subject to rapid change.
Moreover, the sampling errors that could be calculated would provide only
partial approximations to the errors in the final estimates. Faulty reporting,
willful misstatement, and negligent enumeration are all sources of error in
reported data (and hence in the estimates) which are outside the scope of
sampling-error measurement. Such sources of error might be checked upon
and allowed for; but in the sampling process, as well as in the varied other
methods now used in national income estimation, they generally are unknown
and hardly ever can be quantified. In practice, they are likely to be much
more important factors in the reliability of the national income estimates
than are sampling errors.



Thus, the reliability of the national income and product estimates cannot
be assessed with mathematical precision. Rather, the main approach must
be to make a detailed analysis of the statistical sources and methods underlying them and to use this as the basis for qualitative judgment. The general
aim must be to decide whether the reliability of the estimates is sufficiently
high to warrant the specific use intended, and, if this does not appear to be
the case, whether the plan of investigation can be simplified to take account
of the limitations of the estimates.
This task is admittedly difficult. It is complicated by the fact that throughout
the period since 1929 relatively few series of estimates are derived from the
same sources and methods, and hence have the same range of error in all parts
of the period. Many series are a time-period admixture of sources and methods
of widely varying type and quality.
Given this situation, it often will not be fruitful or possible to judge the
reliability of the estimates in an overall sense. The investigator not only will
have to take account of general features of the series involved, but often will
have to distinguish between benchmark and other-period estimates and make
such determinations as whether annual levels or year-to-year changes are
principally relevant and whether components are being used in isolation or
in relation to other components. Having determined what aspects of reliability are relevant in a particular instance, he can proceed to a study of the
national income data with these aspects in mind and obtain optimum results.

Factors affecting reliability
Consideration of four major factors should prove helpful in forming a
judgment about the reliability of estimates of the various components of
the income and product flow.
In the first place, one must consider whether the economic units (such
as businesses, governmental agencies, or individuals) are reporting on an
item which is represented by straightforward transactions of simple definition, or on an item which requires complex calculations on their part or is
somewhat vague in definition. In practice, the former case is likely to be
associated with the occurrence of monetary transactions.
The second factor to be considered is the quality of the records kept by
the economic units whose trarsactions are being measured. Lack of adequate
records leads to less reliable reporting or to an absence of reported data.
In either case, the reliability of the resulting estimates is impaired.
The third factor which should be given weight is the reporting system—
its character and the quantity of data it produces. The obvious distinction
here, as to the former, is between complete census-type coverage and
sampling. However, this distinction in itself does not throw much light on
the problem of reliability. While, other things being equal, complete enumerations are more reliable than samples—and, for that matter, large
samples are more reliable than small ones—the ceteris paribus qualification
in this instance deprives the statement of much of its practical sigrificarce.
So much depends on the quality of the censuses and of the samples—
including the skill and training of enumerators—that only a detailed investigation of all the relevant characteristics can yield well-founded conclusions regarding reliability. Needless to say, such investigations are difficult
undertakings and often may not prove conclusive. In particular, recent
advances in sampling techniques have considerably narrowed the area
over which a flat claim of superiority for the results of census-type reporting
can be made.
With respect to the quantity of information yielded by a reporting system,
it is first to be observed that large and frequent quantity does not necessarily, of course, make for reliable estimates. But smallness of quantity, even
of high quality, results in data gaps impairing the adequacy of an income
or product series.
The final point to be considered is to what extent the items that enter the
income and product accounts differ from those that are actually reported.
Such differences almost always imply that estimating procedures have been
introduced. This means an impairment of reliability of the final figures
which can be evaluated only by an examination of the procedures. In
general, a long and involved estimating chain can be taken as a sign of
statistical weakness, although this rule must be qiialified in the light of the
adequacy of the supplementary data introduced and of the cogency of the
procedures adopted. Simplicity of procedure, however, cannot be taken
as an evidence of absence of statistical weakness. It may only mean that

NATIONAL INCOME, 195 4 EDITION
reliable data for making necessary adjustments are not available, and that
summary, arbitrary assumptions have been used instead.

Application of factors to broad income and product
components
It may prove of interest and value to test some of the major components
of national income and product against these four criteria of reliability.
For brevity this is done in a very general way, and with frequent resort
to personal judgments of the type which have sometimes proved erroneous
in the past.
Considering first the components of national income, there can be no
doubt that wages and salaries rank highest in reliability. This conclusion
is based on the relative simplicity of the concept, the comprehensiveness
and high quality of the record-keeping and of the reporting system (both
to a large extent byproducts of the social security system), and the fact
that the adjustments to the reported totals that are necessary to bring them
into conformance with the requirements of the income and product accounts
are small and well-founded statistically. In this instance the statement
seems warranted that since 1939 the departure of the annual estimates
from their true value is probably very small. Any marked lowering in the
quality of the 1929—38 estimates is precluded by the fact that periodic
industrial-census results and the sample wage indexes compiled by the
Bureau of Labor Statistics are available to extrapolate the social security
based series.
The estimates of rental income of persons are on the other end of the
reliability scale. In this instance, a profit-type income is involved, the
definition of which to the reporting unit must always be complex and
somewhat vague. Both record-keeping and reporting systems are fragmentary and poor, and the estimating procedures which are necessary to convert
reported data into national income entries are unusually complex and
tenuous.
Estimates of the other income shares range between these two extremes on
the scale of reliability. Supplements to wages and salaries follow wages and
salaries closely. Large parts of them are as well-founded as the wage and
salary data, for reasons that are essentially similar. Supplements rank somewhat lower because their "other labor income" component includes certain
items which are statistically less well-founded, especially with respect to the
tentative estimates for the most recent years.
On the lower end of the scale the "income of unincorporated enterprises
and inventory valuation adjustment" may be considered as superior to
rental income. The problems of calculating entrepreneurial income confronting the reporting units are similar to those involved in the calculation
of rents—both constitute a type of profit income. Records and reporting
systems are, however, somewhat more satisfactory than in the case of rental
income; and the estimating procedures that are applied to the reported
data are somewhat more direct and incorporate better information. It
should be noted, however, that the entrepreneurial income estimates are
subject to significant shortcomings when compared with the other income
shares.
Broad generalizations of this type, it is recognized, are not apt to be of
much concrete help for any particular use of the entrepreneurial income data.
For one thing, the estimations of farm and nonfarm incomes, although
handicapped by certain common limitations, are fundamentally dissimilar.
There is no parallel in the nonfarm segment to the Department of Agriculture's systematic long-period study of farm income and development of
reporting sources. And the estimation of farm income has no counterpart
to the necessarily heavy reliance that is placed upon income-tax return information in deriving the net income of nonfarm unincorporated enterprises. Moreover, the data and procedures used to estimate farm income
since 1929 are characterized by substantial uniformity over the period,
whereas those underlying the nonfarm total vary widely in different subperiods.
Corporate profits before tax are a series whose probable deviation from
true universe values must be adjudged smaller than that of any of the other
distributive shares except employee compensation. The definition of profits
is not simple, and measurement at the level of the individual business firm
involves complicated computations that can be performed with varying



63

accounting criteria. Yet the quality of corporate records is surely good,
and the reporting system developed over a period of many years by the
Internal Revenue Service doubtless has gone far toward standardization
of reporting. This system has produced comprehensive annual data not
requiring unduly large estimating adjustments before inclusion in the
national income tables.
In judging the corporate profit series, two limiting aspects should not be
overlooked. First, the estimates for recent years are not based on Internal
Revenue Service tabulations, in which there is a 2-year lag, and are less
firmly grounded. Secondly, the addition of the "inventory valuation adjustment" to corporate profits before tax appreciably reduces the statistical
reliability of the profit series. This adjustment—designed to put inventories
charged to cost of sales on a uniform and current pricing basis, differing
from the diverse pricing practices followed in business accounting—is introduced because it is thought to constitute a significant improvement in
the economic meaningfulness of the statistics. But the adjustment is based
on information that is slender and procedures that are complex and subject
to error.
The remaining income share, net interest, is based in part on corporate
sources of data (obtained mainly from the Internal Revenue Service), but
its reliability is weakened by the inadequacy of information on interest
flows originating in certain major noncorporate areas, and by the general
lack of reliable data for the latest years.
Government purchases of goods and services are highest on the scale of
reliability among the components of gross national product. Were it not
for certain problems involving the timing of purchases which are important
when the level of government expenditures is changing sharply, the data
for the Federal Government would parallel the quality of the wage and
salary estimates, using the criteria of definitional clarity, quality of recordkeeping and reporting, quantity of available information, and the statistical
foundation of the estimating adjustments. However, the problem of timine,
together with the fact that the series on State and local government purchases
is less well-founded than the Federal series, reduces the reliability of the
combined government purchases series below that of wages and salaries.
The estimates of chanpe in business inventories probably rank lowest on
the product scale. The measurement of inventories presents substantial
problems of cost allocation and pricing- to reporting units, and for important
segments of the noncorporate economy records are not adequate and reporting is unsatisfactory, especially for recent years. More important,
however, is the fact that the change in business inventories represents the
difference between large and volatile annual totals, and hence is subject
to significant percentage errors. Also the reported inventory data require
substantial estimating adjustment. Involved in the measurement of nonfarm
inventory change is the inventory valuation adjustment, the inadequacy of
which was noted above in the comments on corporate profits.
Estimates of producers' purchases of durable equipment and of personal
consumption expenditures for commodities—both based largely on producers' records—follow government purchases in reliability on the product
side. The definitional problems confronting the ultimate reporting units do
not loom too large; record-keeping and reporting are relatively adequate;
estimating procedures applied to the reported basic data are fairly complex
and reduce reliability, particularly in the case of consumer commodities, but
on balance, the two series must be rated rather high. These generalizations,
it is important to add, refer to the methodology adopted for years for which
benchmark data are available. For estimates for other years, the evaluation
would be less favorable. As will be noted, however, in the sections covering
these series, actual experience to date with provisional estimates prepared
before the availability of benchmark data has been very satisfactory.
The series on personal consumption expenditures for services is based
largely, but by no means exclusively, on producers' sales records. The
concept of sales is simple, and comprehensive, census-type reporting systems
yield generally reliable data for items comprising the bulk of the total.
The estimation problem is simplified by the absence of intermediary enterprises between the producer and consumer and by the comparatively small
extent to which reported sales must be adjusted for the elimination of sales
to nonconsumers. But reliability is significantly lowered by the fact that
many of the comprehensive sources on which heavy reliance is placed become
available only rather infrequently. Of lesser effect is the inadequacy of
materials for many of the numerous smaller items forming part of the
services total.

64

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

The field of private construction is an extremely difficult one for statistical
estimation. The concept of "value of work performed" used in the new
private construction series is not a simple one on which to report, and
consequently little of the reported information is obtained on this basis.
Neither enumeration nor sampling is well suited for establishing universe
levels, and complex statistical methods are necessary to adjust for the
coverage gaps and timing deficiences of data secured through varied reporting systems.
The estimates of public construction are much more firmly based, mainly
because of the comprehensiveness and superiority of Federal agency records.
However, these estimates are not an independent component of gross national
product, but are used only in obtaining a breakdown of total government
purchases.
Estimates of individual series entering into the computation of net foreign
investment are based on a great deal of solid statistical information. But this
component of gross national product is derived as the difference of large
minuends and subtrahends. Consequently, substantial percentage errors are
likely, especially in years when the absolute magnitude of the component is
small.
Personal income is more reliable than national income. The major items
included in personal income but not in national income (government transfer payments and government interest) are reliable. The exclusions that are
made either do not affect reliability (such as the accurate deduction of social
insurance contributions) or else actually increase it (such as the deduction of
the corporate inventory valuation adjustment and all other components of
corporate profits except dividends).
With respect to the disposition of personal income, the major point relates
to saving. This item is the difference between large totals, and is therefore
subject to sizable percentage error. Supplementary information that should
be considered in interpreting the reliability of this component of the accounts
is discussed in the section of the methodological descriptions dealing with
personal saving.
Although a study of the methodology underlying the national income and
product estimates is the main basis for an evaluation of reliability, there are
two other types of evidence whose examination throws light on the subject.
The first is the record of the "statistical discrepancy;" the second is the record
of the revisions that are made of the estimates as originally published. These
two will be discussed in turn.

Significance of the statistical discrepancy
The "statistical discrepancy" measures the excess of the gross national
product as estimated by summing its component product flows over the gross
national product as estimated by summing components of the national income and all other charges against the total value of gross national product.
It arises because of errors in the component estimates, and hence is relevant
to the problem of reliability.
In the national income and product account the statistical discrepancy is
entered on the debit side, as an item reconciling national income with charges
against national product. This manner of entering the statistical discrepancy
is purely a matter of convenience. It permits the two most widely used aggregates—national income and gross national product—to be broken down into
component items which do not include the "statistical discrepancy." It does
not signify that the national income and the gross national product have been
correctly estimated, and that the error has been made in the estimation of one
or more of the items reconciling the two. Quite to the contrary, it is likely that
the aggregates are affected whenever a statistical discrepancy appears.
The statistical discrepancy appears also in the business income and product
account, and similar comments apply to the form in which the entry is made
there. It may be noted that in the accounts as shown the item always reflects
discrepancies between the estimates of business income and production. This
is so because business income and production are obtained statistically by
making consistent deductions for nonbusiness income and production from
the two sides of the national income and product account. Hence, the results
of all estimating inconsistencies, whatever their origin, are shifted into the
business account.
The adjustment for statistical discrepancies appears also in the gross saving and investment account. It signifies an error either in total saving and/or
total investment and in one or more of their components.
The statistical discrepancy is a measure of the difference in error between
the two estimates of the total gross national product. While its presence is



conclusive evidence that errors have been committed, a zero discrepancy
does not constitute proof to the contrary. Strictly speaking, the discrepancy
measures lack of consistency, and it does not register absolute errors which
compensate in the accounts. To the extent, however, that the sources and
methods of estimating the components of the credit and debit sides of the
national income and product account are independent—in the sense that
errors committed in estimating components on the one side do not involve
corresponding errors on the other—it is reasonable to give some weight to the
statistical discrepancy in evaluating the reliability of the totals. In these
circumstances, greater confidence can be attached to the value of the national
income and product totals if the size of the discrepancy is small than if it is
large.

Degree of independence of income and product estimates
It is important, therefore, to consider to what extent the two sides of the
income and product account are in fact independent as to statistical sources
and methods. Quantification is not possible, but certain relevant considerations can be presented. These are of a summary nature and should be supplemented by the detail contained in the following statistical notes.
It is not possible to classify the estimates of the various components of the
income and product flow into two neat groups, consisting of those that are
based on independent sources and methods, on the one hand, and of those
lacking such independence, on the other. In fact, the estimates range over a
wide scale.
Largely due to the utilization of social security data, the estimates of employee compensation are the outstanding example of a close approximation
to statistical independence on the income side of the accounts. Even this
statement has to be qualified, because certain components of these estimates
are entered identically on the product side (for instance, the sizable item for
domestic service).
A second example of a high degree of statistical independence is provided
by the estimates of government purchases, on the product side of the accounts. The degree of independence is probably somewhat smaller in this
instance, because estimates of government employee compensation, in the
national income, are based upon records and reporting systems related to
those upon which the estimates of government purchases are made. However, in spite of this qualification, the degree of independence is very large.
While, on the other end of the scale, there is considerable interdependence
of statistical methodologies, no major component of the income or product
flow can be said to lack independence completely. However, the opportunity
for consistent error is also wide. In particular, there is substantial interdependence between the business income components of the national income
and the estimates of several of the components of the product flow. This is so
because inventory and sales data are used in a related fashion.
The inventory valuation adjustment is perhaps the most clear-cut example
of such a relationship. It is made in the form of identical entries on the income
and product sides, consisting of an adjustment to the income of unincorporated enterprises and corporate profits, on the one hand, and to the change
in the book value of inventories, on the other, in the measurement of the
change in nonfarm business inventories.
Moreover, there is close interrelation between the estimates of unincorporated enterprise income and corporate profits and the change in the book
value of inventories. For benchmark-year estimates, errors in corporate inventories result in identical errors in corporate profits, since both estimates are
based on balancing corporate accounts submitted to the Internal Revenue
Service. For noncorporate business inventories, the offset is complete in the
case of farming, and it tends to hold for other unincorporated business to the
increasing extent that the estimates are based upon balanced accounting data
similar to those used for corporations.
Instances in which related sales or gross receipt series underlie the estimates
of both business profits and product flows are analytically similar, although
in statistical practice effective interdependence is usually reduced. Clear-cut
examples are afforded by the rent and professional service estimates. Errors
in gross dwelling rents and in the estimates of gross receipts of professional
practitioners from consumers affect both personal consumption expenditures
for services and the corresponding business incomes, although not necessarily
to the same extent. The same type of relationship can be found in other areas
of business income, unincorporated as well as corporate, although it may be
attenuated by the particular statistical procedures adopted, or harder to
trace because of a less explicit coordination of them.

NATIONAL INCOME, 195 4 EDITION
In the examples of interdependence hitherto given, all reference has been
to instances in which for accounting reasons errors in the estimates of the
product components must lead to offsetting errors in the income components.
It may be noted that complete interdependence is not involved. Independent
errors are still possible, for instance, when in deriving business profits current
expenses are wrongly reported or estimated.
So far no reference has been made to compensating errors which are a
matter of statistical probability rather than accounting necessity. For instance,
when wage and sales data are taken from identical sources, lack of complete
coverage of the basic reports may lead to similar errors in both items. However, this need not be the case. Both in census-type reports and in samplebased estimates, the error in the one component may differ from that of the
other. Prior to the introduction of the social security data, the United States
estimates were susceptible to this type of common error to an important
extent. Large segments of the estimates of wages and salaries and of the
product flow were derived from identical, industrial census, sources. Since
the utilization of social security data, there do not appear to be any important
areas of the income and productflowsin which this type of error is significant.
The statistical discrepancy measures the net residual of error which remains
after the best possible estimates of the various components of the income and
product flow have been made. If initial estimates of the components lead to a
sizable statistical discrepancy or to erratic movements in it, they are reexamined and an effort is made to trace the source of the discrepancy and to
eliminate it as far as possible. This reexamination of the initial estimates consists mainly of a critical comparison of the methodology of the component
estimates for error and inconsistency. This is an essential step of the estimating
procedure which cannot be taken by the individual estimators responsible
for the preparation of the component series, but must be reserved until initial
estimates of all the components have been prepared. While significant improvements can sometimes be made in this manner, a residual discrepancy
will remain.
The suggestion has been made that this residual discrepancy should be
eliminated, either by the exercise of further judgmental decisions of the type
used in reducing it from its initial size, or by the application of more formal
mathematical procedures that tend in the direction of greater objectivity.
Superficially, complete elimination of the statistical discrepancy would be
desirable, from the standpoint of convenience to the users of the data.
Basically, however, it would be harmful. A statistical discrepancy of substantial size or irregular movement reflects troublesome errors in the estimates.
If this is the situation, the users of the data should be aware of it so that they
can exercise due caution in the application of the estimates in economic
analysis.

Characteristics of the revisions
Recent-year estimates of national income and product are based on incomplete data and are revised as additional information becomes available. A few
of the components do not undergo significant revisions after the publication
of the initial estimates, but this is not the usual case. Fairly widespread revisions can be expected in the estimates for the two most recent years because
of the lag in the availability of Internal Revenue Service tabulations from
income tax returns, which serve as benchmarks for many of the component
estimates. Revisions extending further back reflect the incorporation of census
information obtained at intervals which can range up to 10 years. In certain,
much less frequent, instances, improved sources and/or methods may become
available which call for revisions over an even longer number of years.
Some inferences as to reliability can be drawn from the record of the
revisions which initial estimates of the income and product flows underwent
in subsequent years. Frequent and large revisions in the estimates are a positive evidence of lack of reliability. However, absence of sizable statistical
revisions can be taken as positive evidence of reliability only if the more
recent estimates incorporate additional information which is known to be
more reliable. In this case meaningful judgments as to the relative reliability
of recent-year estimates as compared with later benchmark estimates can be
made. But it is not possible to go further than this. Absence of revisions in
estimates of several significant components of the income and product flow
reflects only a lack of data accretion subsequent to the publication of the
initial estimates, rather than constituting a positive sign of reliability.
With reference to the national income and product aggregates, it may be
said that since their publication on a new basis in July 1947 the revisions in



65

the initial estimates have been very moderate—generally less than 1 percent.
Thus, for the years from 1942 through 1952—starting with the period in
which estimates presented in the 1947 supplement have undergone at least one
intermediate revision—the maximum difference between the present estimate
of national income and any previous estimate published in the 1947 and 1951
supplements and in July issues of the SURVEY OF CURRENT BUSINESS is less than
1 percent in 8 of the 11 years. The exceptions are 1943 (1.2 percent), 1947
(2.7 percent), and 1948 (2.1 percent). Over the same time span, the maximum difference between the present and any previous estimate of the gross
national product provided in the same reports is less than 1 percent in 6 years
and 1.1 to 1.6 percent in three years. It amounts to 2.6 percent in 1946 and
2.0 percent in 1948.
Revisions of the annual changes initially shown by the totals have been
much larger, of course. The worst record was with respect to the decrease
from 1945 to 1946. In this instance, the initial estimate was cut from $9.4
billion to $4.3 billion for the gross national product and from $4.6 billion to
$1.7 billion for the national income. The years immediately following the
war, it may be noted, saw drastic changes, giving rise to unusual estimating
problems, in both the composition of the national product and in the
legal-form and industrial structure of the national income.
Of particular interest from the standpoint of gauging statistical reliability
are the revisions incorporated in the present edition of the National Income
supplement. As will be evident from the statistical notes which follow this
Introduction, the major feature of these revisions is the incorporation into the
estimates of the results of the postwar censuses (the 1947 Census of Manufactures, the 1948 Census of Business, the 1950 Census of Population and
Housing, and the 1950 Census of Agriculture). Opportunity was taken at the
same time to introduce numerous other new data sources and improvements
in estimating techniques, some of them affecting series back to 1929.
As a consequence of the incorporation of the postwar census data, the new
estimates replace for more than one-half of the gross national product total
(but for a very much smaller proportion of the national income) preliminary
series which represented extrapolations ranging up to 11 years of benchmark
estimates based on prewar censuses. Evidence as to the reliability of the estimating procedures used by the National Income Division pending the
availability of census information is accordingly provided.
That this evidence is distinctly favorable can be seen from the fact that in
none of the years 1947 to 1950 did the estimates of gross national product
incorporating the census data differ by more than }{a of 1 percent from
the preliminary estimates which they replaced. It may be noted that a
favorable outcome of this general type had been expected with confidence.
This was because the estimates of the national income, which are dependent
on census data only in minor degree, had confirmed the gross national product
estimates in the interim period. The statistical advantages of estimating
income and productflowsin an interrelated accounting framework are clearly
apparent in this connection.
Needless to say, these favorable results should not be taken to indicate that
census information is inessential to the estimates incorporated in this report.
Census data provide indispensable benchmarks for these estimates, and the
above evidence relates only to the feasibility of extrapolating such censusbased benchmarks. Moreover, there is no assurance that the results of extensive extrapolation would be similarly accurate in future periods. Finally,
extrapolating errors in the components of the income and product flow were
larger than in the estimates of the aggregates, as will be noted in the subsequent sections dealing separately with these components. Frequent, comprehensive, and accurate census data are indispensable to the reliability of much
of the detailed income and product breakdown.
It is generally true that percentage revisions in the national income and
product totals are much smaller than those in some of their components.
To a substantial extent, this is due to the effect of offsetting errors. According
to past experience, such offsets may be counted upon to increase the reliability
of the estimates as their finest components are added to obtain broader
subtotals.
The national income and product series are published in somewhat
greater detail than is warranted by the statistical reliability of some of the
ultimate components. While it would be hazardous to attach precise significance to the level and movement of these components, offsetting errors
make it feasible to recombine them into reliable subtotals differing from the
published ones and better adapted to specific types of economic analysis.
It is in order to facilitate judicious recombinations of this type that some of

66

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

the detail (in particular recent-year detail on the industrial origin of some
of the income shares and on the product breakdown of some of the consumer
commodity and service flows) is published, and one should be aware that
the use that can be made of these series in isolation is limited.

Allowing for statistical error
While the foregoing survey may provide a sufficiently definitive basis
for the general conclusion that the estimated annual totals of gross national

product, national income, and personal income are subject to only a small
percentage of error, it clearly indicates that there is no easy way of providing
the users of national income data with measures of statistical reliability.
Relevant quantitative measures are not available; and, owing to the basic
nature of the data, the prospect of their ever being constructed on a comprehensive scale appears quite limited. A study of the statistical methodology
underlying the national income estimates, supplemented by analysis of the
statistical discrepancy and of the revisions, will remain the major avenue for

Exhibit 1.—Industrial Classification for the National Income

1

A. MANUFACTURING INDUSTRIES

Industrial content in terms
of the Standard Industrial
Classification 1942 edition
(basis for 1929-47 national
income data)

Industrial division or industry'

Manufacturing
Food and kindred products
Tobacco manufactures
Textile-mill products
Apparel and otherfinishedfabric products
Lumber and timber basic products
Furniture andfinishedlumber products
Paper and allied products
Printing, publishing, and allied industries
Chemicals and allied products
Products of petroleum and coal
Rubber products
Leather and leather products
Stone, clay, and glass products
Iron and steel and their products, Including ordnance
Nonferrous metals and their products
Machinery, except electrical
Electrical machinery
Transportation equipment, except automobiles
Automobiles and automobile equipment
Miscellaneous manufacturing

19 to 39.
20.
21.
22.
23.
24.
25.
._ 26.
27.
28.
29.
30.
31.
32.
19 and 33.
34.
35.
36.
37.
38.
39.

Industrial division or industry

Industrial content in terms
of the Standard Industrial
Classification 1945 edition
(basis for 1948-53 national
income data)

2

Manufacturing
Food and kindred products
Tobacco manufactures
Textile-mill products
Apparel and otherfinishedfabric products
Lumber and wood products, except furnitureFurniture and fixtures
Paper and allied products
Printing, publishing, and allied products
Chemicals and allied products
Products of petroleum and coal
Rubber products
Leather and leather products
Stone, clay, and glass products
Primary metal industries
Fabricated metal products, including ordnance.
Machinery, except electrical
Electrical machinery
Transportation equipment, except automobiles.
Automobiles and automobile equipment
Instruments
Miscellaneous manufacturing

19 to 39.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
19 and 34.
35.
36.
37 (exc. 371).
371.

38.
39.

B. NONMANUFACTURIXG INDUSTRIES

Industrial content in terms
of the Standard Industrial
Classification 1942 edition
(basis for 1929-53 national
income data)

Industrial division or industry2

Agriculture, forestry, and fisheries
Agricultural services, forestry, and fisheries

01 to 09.>
01 to 06.
07 to 09.3

Metal mining
Anthracite mining
.. _ ...
_
Bituminous and other soft-coal mining
Crude petroleum and natural gas
Nonmetallic mining and quarrying

10 to 14.
10.
11.
12.
13.
14.

.

Contract construction .
Wholesale and retail trade
Wholesale trade
Retail trade and automobile services
_.
Finance, insurance, and real estate
Security and commodity brokers, dealers and exchanges.._
Finance, n. e. c
insurance carriers
Insurance agents and combination offices
Real estate
..
...
Transportation.
_
_
Local and highway passenger transportation
Local railways and bus lines
_
Highway passenger transportation, n. e. c
Highway freight transporation and warehousing
Water transportation
._

16 and 17.
40 to 61 and 88.
40 to 47.
48 to 61 and 88.
62 to 70 (exc. 707).
62.
66.
63, 64, 65, and 67.
68.
69.*
70 (exc. 707).
72 to 80.
72.
73, 741, 742, 743, and 749.
73 and 741.
742, 743. and 749.
75 and 79.
76 (exc. 766).

1. Numbers refer to the code numbers in the Standard Industrial Classification Manual.
(Government Printing Office, 1942 and 1945 editions.)
2. All establishments operated by government agencies or corporations are classified in
the Government and government enterprises industrial division, regardless of their classification in the Standard Industrial Classification Code.
3. The National Income Division classification includes irrigation system operation in
"Local utilities and public services, n. e. c."
i. In National Income Division classification, includes insurance agents, brokers, and services, and establishments regularly engaged in any combination of real estate, insurance, loans,
or lega] activities when none of these activities alone constitutes the principal business of the
establishment.
6. Includes all Federal Government agencies and operations except those included in the
industry, "Federal—government enterprises."
6. The following list enumerates all Federal enterprises: Agricultural Marketing Act
Revolving Fund, Alaska Railroad, Army Post Exchanges, Banks for Cooperatives, Bonneville Power Administration, Boulder Canyon Project, Commodity Credit Corporation,
Defense Homes Corporation, Defense Plants Corporation, Defense Supplies Corpoi*ation.
Disaster Loan Corporation, Electric Home and Farm Authority, Emergency Crop and Feed
Loan Program, Export-Import Bank, Federal Crop Insurance Corporation, Federal Deposit



Industrial content in terms
of the Standard Industrial
Classification 1942 edition
(basis tor 1929-53 national
income data)

Industrial division or industry2

Transportation—Continued
Air transportation (common carriers)
Pipeline transportation
. . _.
Services allied to transportation
Communications and public utilities
..
Telephone, telegraph, and related servicesRadio broadcasting and television...
Utilities: electric and gas
.
. ..
.
Local utilities and public services, n. e. c
Services .
..
Hotels and other lodging places
_.
Personal services
_
Private households
...
Commercial and trade schools and employment agencies..
Business services, n. e. c
Miscellaneous repair services and hand trades
Motion pictures
.
.
Amusement and recreation, except motion pictures
Medical and other health services
Legal services
..
Engineering and other professional services, n. e. c
.Educational services, n. e. c
_
Nonprofit membership organizations, n. e. c.._
_
Government and government enterprises1
Federal—general government • 6
Federal—government enterprises T
State and local—general government 8
State and local—government enterprises
Rest of the world"

771.
78.
744, 766, 772, 773, and 80.
81 to 83.
81 (exc. 813).
813.
821, 822.
823, 83.3
84 to 96 (exc. 88), 707.
84.
85.
86.
874, 953, and 954.
87 (exc. 874), 942, and 707.
89.
90.
91.
92.
93.
941 and 949.
951 (exo. 953 and 954).
96.
97.

_.
..

Federal Housing Administration, Federal Intermediate Credit Banks, Federal Land Banks
(until July 1,1947), Federal National Mortgage Association, Federal Prison Industries, Inc.,
Federal Savings and Loan Insurance Corporation, Home Owners Loan Corporation, Inland
Waterways Corporation, Maritime Administration (operating activities), Metals Reserve
Company, Navy Ship Stores and Ship's Service Stores, Panama Canal Company, Panama
Canal Zone, Panama Railroad Company, Petroleum Reserves Corporation, Post Office,
Production Credit Corporations, Public Housing Administration, Reconstruction Finance
Corporation, Regional Agricultural Credit Corporations, RFC Mortgage Company, Rubber
Development Corporation, Rubber Reserve Company, Rural Electrification Administration,
Smaller War Plants Corporation, Tennessee Valley Authority, U. S. Commercial Company,
War Damage Corporation, War Shipping Administration (commercial operating and war
risk insurance activities).
7. Includes all State and local government agencies and operations except those included
in the industry, "State and local—government enterprises."
8. Includes State workmen's compensation funds, and business-type activities involving
significant amounts and accounted for as enterprises, mainly alcoholic beverage monopolies;
water, electric, gas, and transit systems; housing authorities; highway toll facilities; ports
and terminals; and airports.
9. Includes foreign countries, United States territories and possessions and international

NATIONAL INCOME, 195 4 EDITION

obtaining an evaluation of their reliability. If best use is to be made of national income statistics, their reliability will have to be evaluated concretely
on the basis of this evidence, from the standpoint of the specific economic
problem at hand.
While the task of evaluating statistical reliability confronting the user
of national income data is difficult, it should also prove rewarding. Analysis
of methodology and of relevant supplementary evidence will forestall many
misuses of the data. It will lead to more effective utilization of the data by
channeling them into uses warranted by their nature and degree of accuracy.
It will serve to make the informed user wary of many seemingly significant
conclusions that are drawn from small changes in the data which are
obviously well within their margin of error. Also, for any analysis an awareness of hitherto unknown limitations of particular national income series
may demonstrate the advisability of marshaling all other relevant information, within and outside the scope of national income. Clearly, if the evidence
is supportive and consistent, greater confidence can be attached to the
indicated conclusions than if the evidence is contradictory.
Finally to be noted is that the suggested approach to evaluation of reliability is methodic, even though presently permitting quantitative definiteness in only few instances. Large strides may be anticipated from the
integration into it of the results of further work and experience. For it must
be recalled that official national income work in the United States spans
merely two decades, and that the new and expanded series were established
only in 1947. The scope for analysis of methodology will be substantially
expanded over time with the continued improvement of source materials,
the opportunity for testing past sources and the procedures applied to them,
and a broadened basis for analyzing the record of revisions. The statistical
notes and related material presented, in this report represent only a start in
the indicated direction.

Aim and Plan of Statistical Descriptions
The following sections in this Part deal with the sources and methods used
in estimating the income and product flows. Insofar as feasible, the sections
are written according to a uniform plan.
An introductory part first discusses the general nature and reliability of
the series. Next follows a discussion of methodology, covering both base-year,
or "benchmark" year, estimates and their extrapolations, whenever such
a distinction is relevant. This discussion is intended largely as an evaluative
review but also contains considerable descriptive material, with the dual
objective of giving information about the principal methods used and affording an independent basis for judgment about reliability. Concluding remarks
are made on the characteristics of revisions in instances in which a separate
discussion of this subject appeared pertinent.
Only the annual estimates are covered by the sections on sources and
methods. No reference is made to the monthly or quarterly series, or to the
summary annual data derived from them which are published each February
in the Annual Review Number of the Survey of Current Business.
In general, more emphasis is placed on recent-period estimates than on
those for the period, say, 1929—39; and the discussion is aimed principally
at covering the totals of the various components, rather than their industry
or commodity breakdowns per se. Nevertheless, a considerable amount of
information on these breakdowns is introduced, as it often is relevant to an
evaluation of the broader categories and also is of substantial interest to
users of the estimates.
The various income and product components selected for discussion
cover all of those listed in the first four tables in the statistical section (Part
V) of this report. These are the summary tables on national income by distributive shares, gross national product or expenditure, personal income and
its disposition, and the relation of gross national product, national income,
and personal income. At the end of this introduction is provided a summary
of the stubs of these four tables, cross-referenced against the numbers of
the various sections on sources and methods.
In the following descriptions of sources and methods numerous "exhibits",
or supporting tables of data, are presented. Most of these exhibits refer to
1950. This year was chosen because the 1953 estimates (and the revised
figures for 1951 and 1952) were not completed until after the text had been
written and sent to press.
Although geared directly to the first four tables, the following sections on
methodology furnish partial or complete coverage of nearly all of the other



67

35 tables of annual estimates. Many of these tables relate to income flows by
industry. As already indicated, the discussions of the various distributive
shares give considerable attention to their industry breakdowns; but natiora
income by industrial origin, which is obtained statistically by aggregating
these individual-share breakdowns, is not separately discussed. It is convenient at this point, therefore, to give consideration to an important summary aspect of the industry data—the basis of industrial classification
underlying them.

Industrial classification of national income
The industrial distribution of national income is based primarily upon a
classification of establishments rather than of companies, or firms. Use of the
word "primarily" connotes a statistical exception (noted below), not one
of definition.
The establishment is the preferred unit since it yields an industrial classification much closer to an activity basis than does the use of the company. It
also largely prevents discontinuities due to mergers or other changes in the
structure of ownership.
Industrial classification by establishments, for example, places in bituminous coal mining a soft coal mine owned by a corporation engaged primarily in the production of iron and steel products, whereas classification by
companies places it in the iron and steel industry. The establishment classification, nevertheless, is quite different from an activity or product classification
since many establishments produce secondary products which fall within
industries other than those in which their major products are classified.
Force-account construction is an important special type of secondary product.
The establishment basis is used for the industrial classification of wages and
salaries, supplements to wages and salaries, income of unincorporated enterprises and inventory valuation adjustment, and interest paid by noncorporate
enterprises. But, because of statistical necessity, the company basis of industrial classification is used for corporate profits, the corporate inventory valuation adjustment, and interest paid and received by corporations.
The data for these items are all calculated from tabulations of corporation
income tax returns filed with the Internal Revenue Service. During the years
from 1934 to 1941 such returns were filed by every corporation, with certain
exceptions, and separately classified by industry. From 1929 through 1933,
and again in 1942 and subsequent years, affiliated corporations were permitted to file consolidated returns. More precisely, then, from 1934 to 1941 the
unit of classification for corporate profits and corporate interest is the individual corporation; from 1929 to 1933 and from 1942 on it is a corporate
unit consisting of either a single corporation or of affiliated corporations.
Because the bulk of total income originating is comprised of distributive
shares which are classified by establishments, and because of the probable
tendency for subsidiary activities of corporations operating in more than one
industry to be offsetting, it is unlikely that the industrial distribution of the
total national income is seriously distorted by the use of a company, rather
than an establishment, classification for corporate profits and corporate interest. This is a serious limitation, however, on the comparability of the
distributive share estimates for some industrial groups, and one which should
be considered carefully by those who use the data for particular industries.
The estimates of the national income and of the various distributive shares
by industry are based upon the Standard Industrial Classification Code,
which is published by the Office of Statistical Standards of the Bureau of the
Budget and recommended for use by all agencies classifying data industrially. Departures from that Code were dictated, for the most part, by statistical necessity.
For the years 1948 to date, the national income estimates are based upon
the 1945 edition of the Code for manufacturing and on the 1942 edition for
nonmanufacturing industries.1 The industry estimates for 1947 and prior
years, however, are based wholly upon the 1942 edition of the Code. This
difference introduces an element of noncomparability into the estimates for a
number of industries in the manufacturing division, since the 1945 edition
of the Code (now generally adopted by Federal statistical agencies) incorporated extensive changes for that division as compared with the 1942
1. Changes in classification for the nonmanufacturing industries were made in the 1949
edition of the Standard Industrial Classification Code. They are in general insignificant at the
level of detail shown in the national income classification and are not incorporated into the
present report. It may be noted that the State unemployment insurance wage data—the
principal statistical source for the national Income estimates—are not reported on the new
basis.

68

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

GUIDE TO SECTIONS ON METHODOLOGY
Income and product components

Section
Number

Income and product components

NATIONAL INCOME BY DISTRIBUTIVE SHARES

PERSONAL INCOME AND DISPOSITION OF INCOME—
Continued

Compensation of employees:
Wages and salaries
Supplements to wages and salaries

Transfer payments
Less:

Income of unincorporated enterprises and inventory valuation adjustment:
Income of unincorporated enterprises
Inventory valuation adjustment

3
11

5
11

Personal contributions for social insurance

Less:

Rental income of persons
Corporate profits and inventory valuation adjustment:
Corporate profits before tax
Inventory valuation adjustment

Section
Number

Personal tax and nontax payments

14
2
13

Equals: Disposable personal income
Less:
Personal consumption expenditures:
Durable and nondurable commodities
Services
Equals: Personal saving

7
8
15

Net interest

GROSS NATIONAL PRODUCT OR EXPENDITURE

RELATION OF GROSS NATIONAL PRODUCT, NATIONAL INCOME, AND PERSONAL INCOME

Personal consumption expenditures:
Durable and nondurable commodities .
Services

Gross national product
Less:

Gross private domestic investment:
New construction
Producers' durable equipment. .
Change in business inventories:
Nonfarm
Farm

9
10

Net foreign investment

12

Government purchases of goods and services

13

16

Equals: Net national product
Plus:

11
3

Capital consumption allowances

Subsidies minus current surplus of government
enterprises

Less:
Indirect business tax and nontax liability.
Business transfer payments
Statistical discrepancy

PERSONAL INCOME AND DISPOSITION OF INCOME

Less:
Undistributed corporate profits
Corporate profits tax liability
Corporate inventory valuation adjustment. . .
Contributions for social insurance
Excess of wage accruals over disbursements. . .

13
14

Equals: National income.

Wage and salary disbursements.
Other labor income

13

Proprietors' and rental income:
Income of unincorporated enterprises and inventory
valuation adjustment:
Income of unincorporated enterprises
Inventory valuation adjustment
Rental income of persons
Dividends
Personal interest income.

3
11
4
5
6

edition previously in force. Accordingly, in the industry tables in Part V,
estimates for the manufacturing industries which were basically affected by
changes in the Code cannot be shown for the entire period since 1929. For
other industries, estimates are shown for all years of the period in spite of
some noncomparability, but its approximate magnitude (for the year 1948)
is indicated by footnotes in all instances in which it is significant.
The exhibit on page 66 provides a comparison of the National Income
Classification with the Standard Industrial Classification Code.

1. WAGES AND SALARIES1

Plus:
Net interest paid by government.
Transfer payments

5
5
11
2
1

6
14

Equals: Personal income.

lag between preliminary and final estimates is short, and the largest revision
that has been required in recent years by the accession of later data has been
less then 1 percent of the total.
From the standpoint of sources and methods, the estimates may be divided
into those covered by the social security systems and those not covered. The
former include virtually the whole of industrial and commercial employment.
They account for almost 80 percent of total wages and salaries and almost
95 percent of private-industry wages and salaries.
The area of the economy not completely covered by the Social Security and
Railroad Retirement Acts, and therefore estimated independently, consisti
of government, agriculture, private households, and a few quantitatively
minor industries. The following tabulation shows a breakdown of wages and
salaries into the segments estimated from different sources.

Industries Covered by Social Security Programs

The annual estimates of total wages and salaries for the period since 1939
are extremely reliable. Over 90 percent of the total consists of reported payroll
information taken from accounting records of business and government. The

Total payrolls

1. This section covers also, at the end, "wage and salary disbursements," which is a
component of personal income.

The reporting system that has been developed under the Social Security
and Railroad Retirement Acts approaches the ideal as a source for income




69

NATIONAL INCOME, 195 4 EDITION

estimates. It has the advantages of comprehensive coverage, regularity of
reporting, and of being largely "self-policing," in that the wages reported by
employers—upon which the size of benefits partly depends—can be verified
by the employee. Because the reports from every firm list the employees and
their wages individually, unlike other enumerative surveys, the possibility of
omissions and accounting errors is minimal. Reported figures account for
practically the entire total of wages and salaries in covered industries; only
about 1 percent of the total must be filled in by estimation.
In industries covered by the old-age and survivors insurance program
(OASI), each employer with one or more employees files a quarterly list of
his employees and the taxable earnings paid to each employee—through
1950, the first $3,000 earned during the calendar year; thereafter, the first
$3,600. Taxable payrolls are compiled from these lists for each calendar
quarter and an estimate added for delinquent employers. The four-quarter
Exhibit 1.—Wages and Salaries, 1950
Item
Industries covered by social security programs
Industries not covered:
Federal Government
State and local governments
Farms.
__.
_
Private households
...
Nonprofit hospitals
Educational services, n. e. c. (part)
Nonprofit membership organizations, n. e. c. (part)
Insurance carriers (part)
_
._.
Agricultural services, forestry and fisheries (part) —
Federal Reserve banks...
_
Rest of the world
Tips (in all industries)
Total wages and salaries

Millions
of dollars

Percent

114,770

78.3

11,843
10,368
2,724
2,668
1,202
9S8
827
279
98
54
18
717

8.1
7.1
1.9
1.8
.8

146,526

100.0

which processes and tabulates these data. The Board has an accurate basis
for raising taxable wages to the total, by multiplying them by the ratio of
total wages reported by employers to the Interstate Commerce Commission
to the taxable wages of the same employers, which account for about 97 percent of taxable compensation.
Information necessary for computation of the "covered" wage and salary
total has become available 6 months after the period in which the wages were
earned. Thus in June of each year data have been available for the previous
calendar year.' Thefiguresaresubject to revisions of two types: (1) Correction
of errors in reported data and (2) substitution of actual data from delinquent
reports for the earlier estimates of delinquent wages. In the past, revisions of
either type have been negligible.
The relation of the reported to the estimated elements of wages and salaries
for covered industries is shown in Exhibit 2. This covers the 1950 estimates
prepared for publication in July 1954. The estimated portion was slightly
less than 1 percent of the total. In the first estimates for 1950, published in
July 1951, the estimated portion was somewhat higher—about 2 percent—
because of the necessary higher allowance for delinquent wages.
Exhibit 2.—Derivation of Wage and Salary Total for Covered Industries, 1950
[Millions of dollars]

Item

Reported
by
employers

Estimated

n

.6
.2
,1
.04
.01
.5

Industries under Social Security Act
Taxable wages
Nontaxable wages
Industries under Railroad Retirement Act
Taxable wages
Nontaxable wages
Total wages and salaries, covered industries.

sum of these totals represents a b o u t four-fifths of t h e total wages of these firms.
Nearly all of n o n t a x a b l e earnings in e m p l o y m e n t covered by the O A S I prog r a m is reported u n d e r the State u n e m p l o y m e n t insurance ( U I ) programs.
T h e State agencies obtain from employers covered by their p r o g r a m s regular
quarterly reports on b o t h total a n d t a x a b l e payrolls a n d s u m m a r i z e these
reports for t h e Bureau of E m p l o y m e n t Security of the D e p a r t m e n t of
L a b o r . T a x a b l e earnings u n d e r the O A S I p r o g r a m a n d n o n t a x a b l e earnings
reported to t h e State u n e m p l o y m e n t insurance agencies together h a v e represented in most years more t h a n 99 percent of total wages a n d salaries in
industries covered by the Social Security Act. 2
T h e two sources a r e not quite complete because some of the U I programs
e x e m p t firms with few employees (ranging at present from 1 to 7, according
to t h e individual State laws) a n d firms in business intermittently or for short
periods. As such firms a r e covered by the O A S I p r o g r a m , only their n o n t a x able wages m u s t be estimated. This is d o n e by a m e t h o d developed in the
Division of Research a n d Statistics of the D e p a r t m e n t of H e a l t h , Education
a n d Welfare. T h e taxable payroll of firms not covered by the State laws is
multiplied by a ratio of nontaxable to taxable earnings. T h e first of these
factors, t h e taxable payroll, is obtained by subtracting t a x a b l e wages paid
u n d e r t h e U I programs from taxable wages u n d e r the O A S I p r o g r a m . T h e
second factor, the ratio, is a p p r o x i m a t e d from U I d a t a a n d then adjusted, by
use of a 1943 O A S I study, to a p p l y to firms not u n d e r the U I p r o g r a m .
T h e overall wage a n d salary estimates for industries covered by the Social
Security Act a r e thus built u p as the s u m of (1) taxable earnings reported
u n d e r O A S I , (2) nontaxable earnings reported u n d e r U I , a n d (3) estimates,
based o n social security d a t a , of n o n t a x a b l e earnings in these industries not
reported u n d e r U I .
T o t a l wages paid u n d e r the R a i l r o a d R e t i r e m e n t Act are ascertained in
m u c h t h e same way. Wages taxable u n d e r this p r o g r a m — t h e first $3,600 for
each employee—are reported quarterly to t h e R a i l r o a d R e t i r e m e n t Board,
2. This percentage refers to the relation of the specified source materials in the 1939-50 period.
Since 1951, taxable earnings under the OASI program ($3,600) have no longer coincided with
those under the UI law ($3,000), and coverage of OASI has been expanded to include some
groups, chiefly agricultural and private household workers, not covered under the UI programs. The Bureau of Old-Age and Survivors Insurance, however, furnishes a special estimate ol what the OASI taxable wage total lot the year would have been in teims of the 1950
coverage provisions—thus permitting extension of the basic procedure being described,
which represents an integration of OASI taxable wages and UI nontaxable wages. This
special estimate is derived from sample data which serve, in efEect, to adjust reported taxable
wages for the year to a 1950 coverage basis.




Total

87,204
22,239

87,184
21,514

4,704
623

4,704

114,770

114,010

120
3 725

15
760

1. Estimated delinquency as of April 1954.
2. Nontaxable wages paid by employers not covered under State laws.

Industrial distribution of payrolls
The method used to derive an all-industry total of "covered" wages and
salaries cannot be followed satisfactorily for the separate industries, chiefly
because old-age and survivors insurance data have not until recently been
collected or tabulated on an establishment basis. As noted in the Introduction
to this Part, this is preferred to the company basis of classification.
The preparation of an industry breakdown of the covered payroll total
relies heavily on the Bureau of Employment Security's reports summarizing
wages and salaries under the UI programs, in which the establishment basis
of classification is used. This source and the Interstate Commerce Commission's Statistics of Railways reports together provide accurate, employer-reported data by industries for about 95 percent of total covered payrolls.
The missing part consists of the payroll of firms (1) covered by the OASI
program but not by the State laws and (2) covered by the Railroad Retirement Act but not reporting to the Interstate Commerce Commission. The
latter element, quantitatively very small and affecting only a few industries,
is estimated from data furnished by the Railroad Retirement Board. A satisfactory basis for estimating the former is furnished by special tabulations of
old-age and survivors' insurance data showing by industries the taxable payroll of the small firms not covered by the State laws. These are available for
the third quarter of 1940 and 1943 and the first quarter of each of the years
1945-49 and 1951.
For nearly all covered industries, the process of obtaining provisional estimates (before adjustment to the controlling total) involves simply the addition
of these OASI and Railroad Retirement Board data to the comprehensive
data reported under the UI programs and by the Interstate Commerce Commission. This general method was departed from only in those few instances
where more reliable data were available from other sources or where the
portion of employment in firms not covered by the unemployment insurance
data was so large as to suggest use of another procedure. The industries thus
3. At present, however, there is a time lag of one year in the availability of the special
estimate of total OASI taxable wages on the 1950 coverage basis. This is not a significant
limitation, as the taxable wage total for the most recant year can ba estimated within narrow
limits by means of the change in this aggregate shown by UI data.

70

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

receiving special treatment include "agricultural services," "forestry," "fisheries," "banking" (prior to 1943), "water transportation" (prior to 1947),
"medical and other health services," and "legal services." The data utilized
in making estimates for these industries were obtained from the population
and industry censuses, the Maritime Commission, governmental bi.nking
regulatory bodies, and special surveys of the professions (described in the
section on Income of unincorporated enterprises) conducted by the National
Income Division.
The summation of direct industry estimates from these general and special
sources yields a payroll aggregate falling short of the independent controlling
total in most years by only a fraction of 1 percent.4 A large part of this discrepancy can be traced to exclusions from the industry estimates of amounts
unclassified by industry in both the unemployment insurance data and the
special tabulations of OASI small firms, as well as to the omission of nontaxable wages from the latter. Adjustment to the controlling total is accomplished
by allocating the amount of the discrepancy among the industries in proportion to the estimates of wages not covered by the State unemployment insurance programs.
For the years 1929—38, before social security data were available, wages and
salaries for the "covered" segment of the economy were derived from diverse
sources, the most important being the periodic censuses of industry and business. These provided coverage of manufacturing,5 retail trade, wholesale
trade, most of the "covered" services, mining, construction, insurance, communications and public utilities, and parts of banking, highway transportation, and services allied to transportation. Reliable, comprehensive data were
available from Federal reports also for banking, railroads, pipeline transportation, and air transportation. For the industry groups for which censuses
furnished one or more benchmarks, the general procedure for estimating the
intercensal years was through interpolation or extrapolation by sample data
on payrolls and employment collected by the Bureau of Labor Statistics.
The series prepared for individual industries for the years 1929-39 were used
to extrapolate the 1939 estimates derived from social security data.

Industries Not Covered by the Social Security Programs
Federal Government

Statistics. Prior to 1934, when monthly payroll estimates were not available,
this conversion was made by monthly distributions of personnel.
Pay in kind includes the cost value of food consumed by the Armed Forces
and standard issues of personal clothing. This is estimated by the various
services from their cost records, since no suitable accounting data for the two
items are available. The procedure involves multiplication of the number of
personnel receiving the food or clothing by the estimated cost per person.
Necessary revisions of the calendar-year estimates of military wages are
usually small but on occasion have been as large as 5 percent.
Wages and salaries in Federal work relief projects, covering the period
1933—43, are compilations of the Department of Health, Education, and
Welfare from records of the various agencies which administered the
projects.

State and local government
Estimates of public education payrolls—classified in general government—
have been based since 1946 on information reported by the Bureau of the
Census. In the preparation of these estimates, three series of payroll data
have been utilized, covering (1) the school years 1945-46, 1948-49, and
1949-50, (2) 1 month of each quarter for the period 1946-49, and (3) each
month since early 1951. These data were derived from samples covering a
very large proportion of the universe (in general, 80 percent or more) and,
as indicated in Census Bureau reports, having a small degree of sampling
variability.
The school-year figures were converted to calendar-year totals for 1946,
1949, and 1950 by use of patterns shown by the monthly data noted in (3)
above. Estimates for the calendar years 1947 and 1948 were obtained by interpolation. For this purpose an annual series was prepared for 1946—49 which
represented for each year a combination of reported data for 4 months—see (2)
above—and estimates for the other months based on the later monthly payroll
distributions. For the period beginning with 1951, annual totals are the sum
of the monthly amounts reported by the Census Bureau. Totals derived from
this monthly series, it may be noted, check closely with the annual data collected by the Census Bureau in its comprehensive 1952 survey and published
i n Summary of Government Finances in 1952.

For the period 1929—39, the annual estimates of public education payrolls
are those prepared by the State, county, and municipal survey conducted by
Civilian payrolls of the Federal Government—executive, legislative, and
the Department of Labor with Works Projects Administration funds and
judicial—are reported monthly to the Civil Service Commission and the
published by the Department in Employment and Payrolls in State and Local
Department of Labor from records of the individual agencies. The data
Governments, 1929-39. This survey collected data from all States and a comprebecome available with about a 2-month lag; subsequent revisions due to late
hensive sample of local governmental units.
reporting are negligible. Two small items are estimated by the National InEstimates of public education payrolls for the years 1940-45 represent
come Division and added to the reported payroll total for the continental
interpolations of the 1939 Labor Department estimate and the 1946 CensusUnited States: (1) the pay of employees stationed abroad who are citizens of
based estimate by means of a series constructed from detailed payroll data
the continental United States, and (2) the pay of civilian employees of Army
given in the Biennial Survey of Education of the Office of Education. The 1939
post exchanges and Navy ship stores and ships' service stores.
and 1946 values in this series, it may be noted, agreed closely with the totals
Prior to September 1933, monthly payroll data for the civil executive service
used for those years.
were not available. The Bureau of Labor Statistics derived the 1929—33
annual estimates largely on the basis of Budget of the United States Government For the years 1940-50, total nonschool payrolls (except work relief)—general government and government enterprises combined—were estimated by
figures and detailed employment data collected by the Civil Service Comthe Census Bureau for 1 month in each quarter from a mail sample of governmission.
ment units. Months not sampled were filled in by the National Income DiviFederal Government civilian payrolls have been divided between general
sion by straight-line interpolation. Estimates for years since 1951 represent
government and government enterprises on the basis of these sources, supplethe sum of sample-based monthly figures reported by the Census Bureau. As
mented on occasion by direct reports from certain of the individual agencies
in the case of the public education series, the Census Bureau's nonschool paysuch as the Post Office Department, the largest of the enterprises. Little estiroll data have been obtained from samples accounting for a very large part
mation has been necessary.
of the estimated total and subject to small sampling variability.
Military wages are estimated as the sum of cash pay and allowances and of
The Census Bureau generally has provided separate data for 1 month in the
pay in kind. Information on cash wages is secured separately on a fiscal-year
year on government enterprise payrolls. These are interpolated for interbasis from the five armed services. These accounting (budgetary) records
vening months by total nonschool payrolls. Subtracting enterprise payroll
must be adjusted in some instances to eliminate nonwage items. The adjustfrom total nonschool payroll yields the general government portion of the
ments likewise are based on detailed data furnished by the services. The data
latter.
are generally adequate for the purpose, even though not developed for it. The
The annual estimates of nonschool (except work relief) payrolls for the
fiscal-year figures are then converted to a calendar-year basis by means of
period 1929—39 are those prepared by the State, county, and municipal
monthly estimates supplied by the armed services to the Bureau of Labor
survey noted above. Separate data were provided for general government
4. In 19.50, for example, the total of the direct industry estimates was lower than the $114.7
and government enterprises.
billion aggregate by only $500 million.
State and local government payrolls include also a small amount of income
5. For a description of the methodology used for this large industry, see the article on
in kind, representing the value of food and lodging received by government
wages and salaries by Edward F. Denison in the June 1945 Survey of Current Businsss, pp.
hospital employees and of food received by State prison employees. This has
23-24.



NATIONAL INCOME, 1954 EDITION
been estimated from data contained in two Census Bureau publications:
Patients in Mental Institutions and Prisoners in State and Federal Prisons and Reformatories.

State and local work relief wages, covering the period 1929—42, were derived largely from reports of the Federal Emergency Relief Administration,
supplemented by compilations of the Departments of Labor and Health,
Education and Welfare. For lack of complete data, the estimates for 1929—32
(totaling only $50 million for the period) are subject to sizable percentage

Farms
Farm wages, including both cash payments and the cost of board, lodging,
and other perquisites furnished to hired workers, are estimated by the Agricultural Economics Division of the Department of Agriculture. The cash
part, comprising more than four-fifths of the total, is taken as reported in the
Census of Agriculture. The census figures are extended to other years by
sample data on employment obtained monthly and wage rates quarterly from
a mail questionnaire of between 15,000 and 20,000 farmers.
When the results of the 1950 Census of Agriculture became available, it
was necessary to correct the 1949 estimates of cash wages derived from the
sample by only 3 percent. Larger corrections had been required to adjust to
the censuses covering 1929, 1939, and 1944.
The basic data for estimating wages in kind were obtained from two early
census enumerations and several sample surveys. Sample data on employment, wage rates, and the prices of perquisite items are used for extrapolation.

Private households

71

tary hospitals, as obtained from social security records, from independent
estimates for all privately controlled hospitals. The following description
relates to the latter.
Virtually complete data on the cash payroll of privately controlled hospitals
have been collected annually by the American Hospital Association since
1944. Only slight estimation is required to secure the aggregate, of which
nonprofit hospitals account for about four-fifths. The value of maintenance
furnished employees is added by the National Income Division. The value of
maintenance per employee receiving maintenance was derived from the 1935
Census of Hospitals and has been extrapolated to 1944 and later years by an
index constructed by weighting the Consumer Price Index by one and the
food component of this index by four. The proportion of employees receiving
maintenance was estimated from the 1935 census, and, in the absence of later
information, this proportion has been held constant. Maintenance accounts
for about one-fifth of total wages and salaries.
Prior to 1944 the only comprehensive payroll data were those provided by
the 1935 Census of Hospitals. The 1935 census total (cash and maintenance)
was extended back to 1929 mainly by data collected in a special survey of
hospitals made by the National Income Division in 1935.
Estimates through 1941 were prepared by extrapolating the 1935 payroll
by estimates constructed from such information as the number of hospital
beds and the average salary per bed, per capita expenditures for hospital inpatient service, and the average daily census of patients. Unemployment insurance data (covering largely proprietary hospitals) were used to interpolate
between the 1941 and the 1944 payroll estimates.
Since the hospital estimates for the period 1936—43 are based on diverse and
partial sources, it may be noted that the original estimate for 1944 differed
by only 2 percent from the final estimate subsequently derived from the basic
data provided by the American Hospital Association, adjusted to include
maintenance.

Estimates of the total cash pay of domestic servants are obtained as the
product of employment and average annual earnings. For the period beginning with 1940, the employment totals used represent monthly averages of
data collected by the Census Bureau in its continuing sample study of the
labor force reported in Current Population Survey. Prior to 1940, it was necessary
to base the movement of employment on indirect data (employment in inNonprofit membership organizations, n. e. c. (part)
dustrial and commercial pursuits), and the resulting series was adjusted to
an estimate of the number of domestic servants derived from 1930 Census of
The bulk of wages and salaries of nonprofit membership organizations not
Population data.
covered under the social security laws is comprised of the payroll of religious
organizations. This payroll is estimated separately for payments in cash and
For the years 1939 and 1949, estimates of average cash earnings of domestic
in kind.
servants were computed from data in the 1940 and 1950 Census of Population
showing the frequency distribution of such persons by detailed size-of-earnThe basic estimate of cash payrolls was derived from the 1936 Census of
ings classes. Estimates for other years since 1939 have been obtained by interReligious Bodies, reported totals being raised to allow for churches not repolating and extrapolating the 1939 and 1949 census-based averages by
porting. This figure has been extrapolated by multiplying estimated annual
means of the domestic service component of the Consumer Price Index of the
employment by an index of average salaries computed from data supplied the
Bureau of Labor Statistics. To obtain average earnings figures for the period
National Income Division by various denominations, with the movement
1929-38, the 1939 estimate was extrapolated by an index of the average
between 1940 and 1950 adjusted on the basis of earnings information reported
earnings of domestics prepared by the National Bureau of Economic Refor clergymen in the Census of Population. The employment estimates have
search from data collected from employment agencies by the National Income
been derived from data obtained from the following sources: the 1930, 1940,
Division and the National Bureau.
and 1950 Census of Population; the Official Catholic Directory; a special survey
The value of board furnished domestic service employees is based on a
of Catholic dioceses made by the National Income Division, covering 1929—
standard budget for 1935—36 derived from the National Resources Planning
35; and correspondence with other denominations.
Board study on Family Expenditures in the United States, 1935—36, moved by
The rental value of parsonages was assumed to be 10 percent of their value,
changes in food prices and multiplied by the proportion of employees reas reported in the 1926 and 1936 Censuses of Religious Bodies. The rent comceiving board in 1936, as estimated from a survey of employment agencies
ponent of the Bureau of Labor Statistics' Consumer Price Index has been
made by the National Income Division.
used for interpolation and extrapolation. The value of board received by
It should be noted that the level of private household employment indicated
Catholic clergymen—the other item of pay in kind—was estimated for 1934
for 1950 by the Census Bureau's Current Population Survey—as incorporated infrom data collected in the special survey by the National Income Division.
the present estimates—is about one-third higher than that shown by the
The extrapolation to other years has been made by the product of number of
decennial Census of Population. The Census Bureau has found that data perclergymen and the food component of the Consumer Price Index.
taining to employment of women 25 years old and over and in industries
Estimates of payrolls for other nonprofit organizations not covered by
where part-time and occasional work is common tend to be reported more
social security are based on the 1935 Census of Nonprofit Organizations,
accurately in its current survey than in the decennial census. It is believed
Office Buildings, and Miscellaneous. To obtain estimates for other years of
that for areas where labor force classification is difficult—including domestic
service—the Current Population Survey yields more reliable results by reason of the 1929-39 period, the 1935 census data were extrapolated largely on the
basis of the domestic current expenditures of private social welfare and relief
the greater skill and experience of its permanent enumerative staff.
agencies (as included in table 30, pt. V). For the period beginning with 1940,
the National Red Cross payroll has been obtained directly; and the remainder
Nonprofit hospitals
represents an extrapolation—with adjustment to the employment level indicated by the 1950 Census of Population—chiefly on the basis of payrolls in
Direct estimates of nonprofit hospital payrolls have not been prepared
"covered" nonprofit membership organizations.
separately. Instead, these are derived by subtracting the payroll of proprie


72

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Educational services, n. e. c. (part)

enterprises as estimated from tabulations of Federal income tax returns to
the Internal Revenue Service. For years other than 1939, 1945, and 1947,
Estimates of wages and salaries for private educational services not covered
fishery payrolls were estimated, in the main, through interpolation and extraby the social security laws are prepared separately for parochial schools, other
polation by the value of the catch series.
elementary and secondary schools, higher education, and a miscellaneous
Benchmark estimates of the payroll and employment of agricultural and
category of institutions and agencies. The general procedure is one of piecing
similar service establishments were prepared for 1939 from old-age and surtogether information on average earnings and employment (the latter advivors insurance data. In that year, although not subsequently, the OASI
justed to conform with results of the 1940 and 1950 Census of Population).
program provided complete coverage of this industry. Payroll estimates for
The Office of Education has been the principal source of data, including not
the period after 1939 were obtained by extrapolating the 1939 benchmark
only the Biennial Survey of Education and other published reports, but also nu- chiefly on the basis of data reported under the UI programs, which—with admerous special tabulations. Other sources include the National Catholic Weljustment for small firms not covered by them but covered by OASI—have
fare Conference, the National Education Association, State unemployment
accounted for about two-thirds of estimated total payroll in the industry.
insurance tabulations, and a special survey of Catholic dioceses made by the
Prior to 1939, payrolls were estimated as the product of employment and
National Income Division covering the period 1929-35.
average employee earnings. Employment was derived by extrapolating the
1939 figure by the Department of Agriculture's index of farm production.
The average earnings series for that period represents an extension of the 1939
Life insurance carriers (part)
average by the National Income Division series for miscellaneous repair
Until the 1951 amendments to the Social Security Act, insurance solicitors
services and hand trades, as adjusted for 1935-39 movement on the basis
on a commission basis were excluded from coverage of that act.6 The basic
of data reported for a group of agricultural service industries in the 1935 and
estimates of the number of such persons were derived from the 1935 Census of
1939 Census of Service Establishments.
Insurance and from the Institute of Life Insurance's Fact Book for 1940 and
1945—50. The 1935 census also presented data permitting the computation of
the average earnings of solicitors for that year. Such earnings could be estiRest of the world
mated for 1951 through a comparison of the average earnings of insurance
This series is described in the section on net foreign investment.
employees covered by OASI (including solicitors on a commission basis)
with the earnings of those covered by UI (not including solicitors). The 1935
and 1951 average earnings figures were interpolated and extrapolated by
Tips
average annual earnings of "covered" employees in the insurance industry.
Multiplication of employment by average earnings yielded the payroll series
Tips are treated in the category "not covered" since it is believed that the
used.
extent of actual coverage under the social security laws is small. The social
security regulations state that tips are considered wages only if the employee
renders to the employer an accounting of the tips. To the extent that tips are
Federal Reserve Banks
covered, however, they offset any exclusions of income in kind, which is
treated as being completely reported in the social security payroll data.
Wages and salaries are obtained from the annual reports of the Board of
Tips were estimated at $0.7 billion in 1950, and are included in the payroll
Governors of the Federal Reserve System.
estimates for retail eating and drinking places, railroads, taxlcabs, hotels,
personal services, and athletic and social clubs (classified in amusements and
Agricultural services, forestry, and fisheries (part)
recreation, n.e.c).
Tips in eating and drinking places were estimated for 1939 from wage
Only parts of the agricultural service, forestry, and fishery industries are
studies in restaurant occupations made by State labor departments of Illinois,
covered by the social security laws. These are deducted from estimates of total
New York, Ohio, and Rhode Island. The 1939 estimate has been extrapolated
payrolls (described below) to obtain the noncovered portions as shown in
to other years by the Office of Business Economics series on sales of eating and
Exhibit 1.
drinking places.
Forestry payrolls (and employment) are estimated separately for (1) gum
The estimates for the railroads (except dining and buffet cars) were preturpentine and gum rosin production and (2) other forestry. Logging, it may
pared on the basis of a survey for 1929 and 1933 conducted by the Brotherbe noted, is classified in manufacturing rather than forestry.
hood of Sleeping Car Porters for the National Income Division, and extraBase-period estimates for the first segment, which accounts for most of the
polated to other years by the number of berth and seat passengers in sleeping
industry, are derived from Census of Manufactures data. Extrapolation to
and parlor cars (reported annually in Statistics of Railways) and estimated
other years has been based mainly on a total labor cost series, computed from
changes in the percent of passengers tipping and the average tip. Tips in
Department of Agriculture data. This is derived by multiplying the physical
dining and buffet cars are estimated at 12 percent of dining and buffet car
volume of production of turpentine and rosin from gum by estimates of labor
sales (reported annually in Statistics of Railways).
cost per unit of naval stores.
Tips in the taxicab industry are calculated at 15 percent of operating
Employment in other forestry is estimated as a residual between total emrevenues, on the basis of several studies of the industry in the midthirties.
ployment for the industry as shown by the Census of Population and employThe studies for New York and Illinois used in estimating tips in eating and
ment in gum turpentine and gum rosin production. To obtain payroll, the
drinking places also provided the basis of the estimates of tips to hotel and
average earnings of turpentine and rosin workers are assigned to this residual
club food service employees in 1939. The 1939 figures have been extrapolated
group of employees.
to other years largely on the basis of estimated hotel and club receipts from
Wages and salaries offisherieswere derived for 1939 from employment and
meals and beverages. The allowance made for tips of nonfood employees of
earnings data reported in the 1940 Census of Population. The census data
hotels—while based in part on studies of the New York State Labor Departcould be used directly, with only minor adjustment. The 1950 Census of Popment—is, of necessity, essentially arbitrary.
ulation provided another benchmark figure on employment, but the earnings
Tips in personal services are estimated at 5-8 percent of the receipts of
data reported there (for 1949) covered the self-employed as well as employees.
barber shops, beauty parlors, and baths and masseurs. The higher figure has
Payrolls for 1949, therefore, were estimated in conjunction with the income
been used for recent years.
of unincorporated business, with 1950 census data furnishing a control total
for both. Addition of corporate profits to this total yielded a combined aggregate of payroll and business income which was extrapolated to 1945 and 1947
by means of data on the value of the catch furnished by the Department of
Wage and Salary Disbursements
the Interior. Payrolls for 1945 and 1947 were derived by subtracting from
this, factor-income total the income of corporate and noncorporate business
"Wage and salary disbursements," a component of personal income, is
equal to wages and salaries plus the excess of wage disbursements over wage
6. In previous issues ol the National Income supplement, such solicitors were classified
accruals.
as proprietors, and their earnings were included to "income of unincorporated enterprises".



NATIONAL INCOME,

Because of retroactive wage and salary payments, it is necessary to adjust
wages and salaries from an accrual basis to a payment basis for inclusion in
the personal income estimates. The adjusted series is termed "wage and salary
disbursements." The adjustment item, the excess of wage disbursements over
wage accruals, reflects the difference in timing, as between receipt and earning, of retroactive wages.
The following example indicates the procedure of moving from wages and
salaries to wage and salary disbursements. In 1946 a retroactive wage payment of $30 million was made by the Western Union Telegraph Co., under
order of the National War Labor Board. The award applied to work performed in the years 1943, 1944, and 1945, in the amounts of $2, $14, $14
million, respectively. Wages and salaries, which reflect earnings on an accrual
basis, in this specific case are $2 million in 1943, $14 million in both 1944 and
1945, and 0 in 1946. The adjustment item, the excess of wage disbursements
over accruals is $—2 million in 1943, $—14 million in 1944 and 1945, and
$30 million in 1946. Consequently wage and salary disbursements are 0 in
each year 1943-45, and $30 million in 1946.
The adjustment of wages and salaries to wage and salary disbursements is
not intended to correct discrepancies between the two arising from all retroactive wage payments, but only for the more significant of these actions.
While in the case of the Western Union award noted above, it was possible
to secure direct information on these payments, more often such information
has to be approximated from news accounts or other less formal sources.

2. CONTRIBUTIONS FOR SOCIAL INSURANCE and
OTHER LABOR INCOME
This section describes the derivation of (1) Employer contributions for social
insurance, (2) Personal contributions for social insurance, a n d (3) Other labor income.

Of these series, the first and third together comprise Supplements to wages
and salaries (shown by type in table 34, Part V), a component of the national
income. The incomes on which personal contributions for social insurance
are paid are measured in the national income before deduction of these contributions. Both employer and personal contributions, however, are excluded
from personal income (the latter handled as an explicit deduction item in the
computation of the personal income total).
The two series on contributions for social insurance comprise items for
which highly reliable data are obtained, with virtually no time lag, almost
exclusively from the accounting records of the agencies administering the
programs. Employer contributions for social insurance currently account for
one-half of total supplements to wages and salaries, although from 1937 until
recent years the proportion was much higher. Estimates of Other labor income
are not so reliable. They are based to a lesser extent on comprehensive accounting data, and these become available with a lag of two to three years.
Considerable estimation is involved for the most recent years.

Contributions for Social Insurance
Social Security Programs
Regular contribution reports filed by employers with the administering
agencies or with the United States Treasury are the source of data on contributions made by employers under the old-age and survivors insurance
program, the State unemployment insurance and cash sickness compensation programs, the railroad retirement and unemployment insurance programs, and the Federal unemployment tax. The reported contributions
data are lagged (usually one-quarter year) to time them with the accrual of
the wages and salaries on which they were levied, rather than with the receipts of the Government funds.
Like covered payrolls, discussed in the notes on Wages and salaries, these
contributions data are a byproduct of operations under the social security
laws. They are subject to revision only to the extent that wage reports are in
error and that estimates made for delinquent reports are corrected on the



195 4

EDITION

73

The industry distributions of employer contributions under these several
programs—which are required for the industrial breakdown of the national
income—either are available from tabulations of employer reports or can be
estimated satisfactorily from taxable wage data.
Data by industry on employer contributions under the State unemployment
insurance programs have been obtained for years since 1940 from the Bureau
of Employment Security of the Department of Labor. They are summations of employer reports. For the years 1936—39, when reported data
were not available on an industry basis, estimates were prepared by extrapolating the 1940 industry figures back to 1936 on the basis of the National
Income Division's estimates of wages and salaries, and then adjusting the
results proportionately to the all-industry total for each year.
Beginning with 1951, the industrial breakdown of employer contributions
under the OASI program has been obtained by multiplying the contribution
rate by data on total taxable wages as reported by the Bureau of Old-Age
and Survivors Insurance of the Department of Health, Education and
Welfare. For the years 1941—50, the industrial distribution of these contributions was derived as the product of (1) the contribution rate and (2) reported
taxable wages under the State unemployment insurance programs plus estimated taxable wages of firms covered by old-age and survivors insurance
but not covered by the State programs chiefly because of the varying size-offirm exclusion provisions of the latter.1 The old-age and survivors insurance
data for small firms were estimated from periodic special tabulations by the
Bureau of Employment Security of the Department of Labor. In the absence
of taxable payroll data by industry prior to 1941, the estimates in each industry for that year were extrapolated to the 1937-40 period by total wages
and salaries in the industry, with appropriate adjustment to the annual allindustry aggregate.
The Federal unemployment tax has been allocated by industry on the
basis of taxable wages reported under the State unemployment insurance
laws.
The all-industry totals for employer contributions under the railroad retirement and unemployment insurance programs have been allocated among
the few industries in which there is coverage on the basis of total payrolls, as
derived in preparing the estimates of wages and salaries.
Cash sickness contributions by employers, confined to a few States, have
been distributed by industry in the same proportion as unemployment insurance contributions in those particular States.
Of the various programs of social insurance listed above, employees contribute to all except railroad unemployment insurance.2 As in the case of
employer contributions, data on contributions by employees are obtained
from the administering agency or the United States Treasury, and are lagged
so as to make the timing coincide with the payroll disbursements on which
the contributions are based.
In addition to employee contributions under these programs and under
the Government retirement and life insurance systems covered below, the
series on Personal contributions for social insurance includes contributions for old-

age and survivors insurance by the self-employed. (See table 35, Part V.)
They were first paid in 1952, on 1951 incomes, under amendments extending
coverage of the OASI system as of January 1, 1951. Data on such contributions, which are paid annually by self-employed individuals with their returns on Federal income taxes, are furnished by the Bureau of Old-Age and
Survivors Insurance.

Government systems
Payments made by the Federal Government and by its employees to its
civilian employee retirement systems are obtained from records of the
Treasury Department or of the agencies responsible for the administration
of the systems. Data on the Government5s contributions are reported on a
1. The available tabulations of total taxable payrolls under the OASI program were not
used lor this period in obtaining an industry distribution because they were based in large
part on a classification of firms, rather than establishments, and therefore were not comparable to the estimates of wages and salaries. Use of unemployment insurance wage data
for tnis purpose was made possible by the fact that the $3,000 taxable wage provision under
the State laws coincided with the OASI provision (which was changed to $3,600 effective
January 1,1951).
2. Under the State unemployment insurance laws, however, employees have contributed

74

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

fiscal year basis, and the figures for adjoining fiscal years are averaged to obtain calendar year estimates. The Government's and individuals' contributions to the Government and National Service Life Insurance Funds are
compiled monthly from reports of the Veterans' Administration.
Estimates of national income by industry require a separation of Federal
Government contributions as between general government and government
enterprises. This poses no problem with respect to the numerous (and quantitatively minor) funds other than the civil service fund, since the functions of
agencies with separate funds are clearly defined. But since the Federal Government makes one lump-sum payment each year to the civil service retirement fund, it is necessary to make a statistical allocation of the data for this
fund between general government and government enterprises.
For ye?rs other than 1941—45, this allocation has been accomplished on the
basis of Treasury data on employee contributions to the retirement fund
classified by individual Federal agencies. To obtain estimates for the 1941—45
period, the 1940 estimate of the Government's contribution for enterprise
employees was extrapolated by the relative change in government-enterprise
employee contributions (adjusted for the July 1, 1942, rate change), and the
Government's contribution for general government employees was obtained
as a residual.
This procedure was adopted because, while employee contributions were
swelled by the expansion of payrolls, the Government's contribution was not
increased proportionately. Most of the war service employees, concentrated
in general governmental agencies, were expected to withdraw from the Federal service before attaining eligibility for retirement benefits, taking refunds
of their contributions instead of adding to the long-run liabilities of the fund
for benefit payments.
Estimates of employer and employee contributions to retirement systems
for State and local government employees have been furnished for years beginning with 1936 by the Division of Research and Statistics of the Department of Health, Education, and Welfare. Base year data for these estimates
were developed through a joint study by that Division and the Bureau of the
Census. Contributions for the fiscal year 1940-41 were estimated on the basis
of questionnaires on the operations of retirement systems sent to all State and
large local governmental units and to a sample of the smaller units. These
base year estimates have been extended to other years by data compiled by
the Census Bureau in its annual financial statistics series and in special reports, and from published annual reports for some of the larger retirement
systems. The fiscal year data have been adjusted to a calendar year basis by
averaging successive fiscal years. The estimates for 1929—35—definitely of
lesser reliability—were prepared by extrapolating the 1936 figure provided
by the Health, Education and Welfare Department by contributions data
obtained from a sample questionnaire survey of State and local government
units conducted by the National Income Division.
State and local government employer contributions have been allocated
between general government and government enterprises in proportion to the
breakdown of employee contributions for 1939 estimated from data furnished
by the Census Bureau. The enterprise portion is small, the 1939 estimate
amounting to only S3 million.

followed for later years, but greater use was made of data from the State
commissions.
The Division also estimates the relatively small amounts of court-awarded
benefits received by injured railroad and maritime workers, since these are
not covered by the compilations noted above. The estimates for railroad
workers are based on Interstate Commerce Commission data on payments
by railroads for injuries to persons (employees and passengers) and on the
numbers of employees and passengers killed or injured, as published annually
in Statistics of Railways. The estimates for maritime workers were derived for
1938 from data contained in the Labor Department Bulletin (No. 869) on
Workmen's Compensation and the Protection of Seamen, and extrapolated to other

years by maritime employment.
The industrial breakdown of the 1929—38 totals of workmen's compensation for injuries was prepared in several parts. Estimates were first made
separately for the following industries: Federal Government; railroads; maritime employment in water transportation; telephone, telegraph, and related
services; all of the mining industries except crude petroleum and natural gas;
and stevedoring, a component of services allied to transportation.
The sources of the data for railroads and maritime employment already
have been noted. Data for the Federal Government and for stevedoring were
obtained from the United States Employees' Compensation Commission.
Payments in telephone, telegraph, and related services were estimated by
raising the amounts of accident disability benefits and death benefits reported annually by the American Telephone & Telegraph Co. by the ratio
of total payroll in the entire industry to the payroH of that company. The
industry components of mining were estimated from tabulations of data reported by the accident compensation commissions in the principal mining
States.
For all other industries, total payments in 37 States for the period 1929-38
as a whole were tabulated from the report on Workmen's Compensation Experience Compiled in 7947 of the National Council on Compensation Insurance.3
To these figures were added data for three additional States compiled from
State reports. The 10-year aggregates for each industry were then distributed
by years in proportion to employment, and the resulting estimates for each
year were adjusted proportionately to the totals for all industries not independently estimated.
In the preparation of estimates by industry for years since 1939, the
Division's employment series has been used to extrapolate estimated 1938
payments in each industry except railroads, Federal Government, and
maritime employment (which have continued to be estimated separately by
the procedures indicated above). The results have then been adjusted to the
all-industry aggregate less data for the separately estimated industries.

Employer contributions under private pension and related
plans

Contributions in this category relate to the following types of private programs: (1) Pensions, (2) health and welfare programs, and (3) group insurance protection.
(1) The series on employer contributions to private pension plans covers
transactions under several different financial arrangements, including conOther Labor Income
tributions to self-administered plans (involving separate funds administered
by employers either directly or through a bank or other agent), the purchase
of group annuities under plans administered by life insurance companies
Compensation for injuries
(where usually there is no separate pension fund with segregated assets),
Estimates of benefits paid to workers (and their dependents or survivors)
direct payments by employers without the establishment of a fund (of minor
insured under State and Federal accident compensation laws, either on a
magnitude), and contributions to employees' profit-sharing trusts.
compulsory or voluntary basis, have been prepared annually since 1939 by
The totals for the period 1929-38 were built up as the sum of several estithe Division of Research and Statistics of the Department of Health, Edumated parts.4 Data for railroads were obtained annually from Statistics of
cation, and Welfare. These estimates are based on data for private insurance
Railways. Estimates for higher education were derived from contributions
companies from the Spectator Co.'s annual Insurance Yearbook, for State data furnished by the Teachers' Insurance and Annuity Association. Data
insurance funds from reports of the funds, and for self-insurers from informaon employer contributions by churches, the Young Men's Christian Assocition furnished by the State accident compensation commissions. Reports of
ation, and the Young Women's Christian Association were obtained annuthe United States Employees' Compensation Commission also are utilized.
ally for the period 1933—38 from the Church Pension Conference Report and
Data from these sources become available with about a year's lag, and a preliminary estimate is made by the Department of Health, Education, and
3. Data are given in tiais source for an extremely detailed industrial classification. The
time and labor in combining and converting them to the National Income Division classiWelfare from direct information for a few States and industrial injury data
fication prohibited the compilation ol separate data for each year.
compiled by the Department of Labor. The estimates for years prior to 1939
4. Contributions under profit-sharing plans, evidently ol inconsequential amount in this
were prepared by the National Income Division by methods paralleling those
period, were not included in the 1929-38 estimates for lack of data.




NATIONAL INCOME, 195 4 EDITION

75

Wage Developments. Studies by the Department of Health, Education, and Welextrapolated back to 1929 by church pension payments. The estimates for all
fare and the August 1946 Supplement to the periodical Labor and Nation also
other industries (in the aggregate) were derived for the years 1932—38 from
have been utilized.
direct and collateral information contained in the study by Murray Larimer
on Trends in Industrial Pensions. For the period 1929—31, with data lacking, Data on employer contributions to the principal plans have been obtained
they were made on the assumption that contributions dropped annually
for the most pprt directly from the administering organization. Thus, data
from 1929 to 1932 by one-half of the average annual drop from 1932 to 1934.
have been secured from the Amalgamated Clothing Workers of America,
the International Ladies Garment Workers Union, the United Mine WorkTotals for the period 1939—51 were derived as the sum of estimates for
ers of America, and the American Telephone & Telegraph Co.
employer contributions for pensions by religious organizations and instituFor other plans, estimates have been prepared by multiplying the number
tions of higher education (both obtained from the same sources as noted for
of employees covered by a plan (usually as reported in the general sources
the 1933-38 estimates), Federal Reserve banks (from the banks' annual recited above, or obtained directly from the union involved) by average earnports), and employer contributions in all other industries combined. The last
ings of employees in the industry to obtain estimated wages of covered workcategory was available for 1945—51 (as of July 1954) in the Internal Revenue
Service Statistics of Income publications, the reported data on corporate con- ers. These have been multiplied by estimated contribution rates (the modal
rate given in the above sources) to obtain the estimated amount of employer
tributions requiring only minor adjustments for inclusion in the series. It was
contributions.
estimated for the years 1943 and 1944 from special tabulations of pension
plans submitted to the Internal Revenue Service for approval for income-tax
The method of estimation, it will be noted, furnished the industrial disdeduction purposes under the Revenue Act of 1942. The tabulations were
tribution of employer contributions to health and welfare plans.
not designed primarily for statistical purposes, and the data on number of
(3) Employer contributions for group insurance are based upon studies
plans in operation and average annual contribution per plan derived from
for the years 1948 and 1951 made by the Department of Health, Education
them were not sufficiently accurate in time-period reference to yield more
and Welfare and for 1929 from information presented in Industrial Group
than rough orders of magnitude for these 2 years.
Insurance, 1929, National Industrial Conference Board. The 1948 and 1951
studies utilized reports by life insurance associations, the United States
Estimates for 1941 and 1942 were obtained by proportional interpolation
Chamber of Commerce surveys of accident and health insurance, Blue Cross
between the 1937 and 1943 values, using the number of plans in operation
and Blue Shield reports, and other sources. The 1929, 1948, and 1951 estias an interpolating index. Information on the number of plans in operation
mates were prepared for five categories of group insurance, and each has
was obtained from the Latimer study for 1937 and from the Internal Revenue
been interpolated and extrapolated by the National Income Division by data
Service for 1941, 1942, and 1943. Employer contributions in 1939 and 1940
on total premium receipts of the organizations furnishing the particular type
represent straight-line interpolation between the 1938 and 1941 estimates.
The corporate data reported in Statistics of Income—accounting for nearly of insurance coverage. The totals have been distributed by industry according to the relative distribution of employer contributions to private pension
all of the employer contributions total—do not become available until the
funds, after eliminating from the pension distribution the data for industries,
third year after the period to which they refer. In the interim, the current
such as coal mining and telephone and telegraph, which are known to have
estimates are prepared very largely by extrapolation on the basis of paywell established welfare funds (included in (2) above) providing group inrolls and of pension plan information from several sources. These include
the Bureau of Labor Statistics' monthly Current Wage Developments, which re- surance benefits.
ports coverage and contributions data relating to the more important new
plans and revisions of plans already in effect; data from insurance associaOther components
tions on premium payments under plans administered by life insurance
companies; and special tabulations of pension plan contributions reported
The remaining components of other labor income include pay of military
by corporations to the Securities and Exchange Commission. Extrapolations
reservists, Government payments to enemy prisoners of war, merchant marine
prepared from these materials have significant limitations, and the current
war-risk life and injury claims, directors' fees, jury and witness fees, compenestimates are subject to appreciable revision.
sation of prison inmates, and marriage fees to justices of the peace. These
Direct information on the industry distribution of employer contributions
items in 1950 amounted to only 11 percent of other labor income. Over the
to private pension funds is at present confined mainly to the comprehensive
entire period of the estimates only three—military reserve pay, payments to
Internal Revenue corporate tabulations for 1945-51 and to available data
enemy prisoners of war (covering the period 1943—46), and directors' fees—
permitting the preparation of separate estimates for several industries. These
have attained any appreciable magnitude. The other items have always been
include religious organizations; educational services, n. e. c. (taken as equal
negligible in amount.
to higher education); Federal Reserve banks; railroads; and telephone, teleData on the pay of military reservists have been obtained from the armed
graph, and related services (compiled from annual reports of the major
services or from the annual Budget of the United States Government on a fiscal
companies).
year basis, and averaged to obtain calendar year estimates. This type of supIn addition, an Internal Revenue Service tabulation of employer contribuplement increased sharply in the postwar period.
tions by industry covering the 5,116 plans approved through December 31,
Fiscal-year data on total Federal payments to enemy prisoners of war were
1944 was used as the basis of a distribution (apart from the separately estiobtained from the Department of Defense. These were converted to a calenmated industries) for 1944. For all other years of the period since 1929, the
dar year basis by use of data from that Department on the number of prisoners
1944 and 1951 estimates have been extrapolated, industry by industry, by
of war in the United States as an interpolating index. The estimated amounts
wages and salaries, with the resulting distribution adjusted proportionately
paid to prisoners of war working for civilian contractors were deducted from
to the all-industry aggregate exclusive of the separately estimated group of
the total, since such payments are included under wages and salaries, chiefly
industries.
in farming.
Employer pension contributions have increased greatly, both in absolute
The estimates of directors' fees are crude. They are prepared in two parts.
and relative terms, since World War II. For this period, except for the preFor the finance, insurance, and real estate industry, which accounts for about
liminary estimates for the two most recent years, it is to be noted that the
half of the total, the estimates are made by multiplying the ratio of directors'
pension series is based almost entirely, both in aggregate and by industry,
fees to compensation of officers in Federal Reserve member banks by the total
on comprehensive corporate tabulations of the Internal Revenue Service.
compensation of corporate officers reported by the Internal Revenue Service
(2) Contributions by employers under health and welfare programs have
in Statistics of Income. For other industries, a similar procedure is followed,
been derived by assembling data for the larger plans known to be in existence
but direct information on the ratio of directors' fees to compensation of
and preparing estimates for the remaining smaller plans.
officers was limited to the years 1929 and 1932.
Listings of existing plans, together with statements of their coverage and
Within each of these two broad groups—finance, insurance, and real
provisions, have been obtained mainly from publications of the Bureau ol
estate and all other industries—directors' fees are distributed by industry in
Labor Statistics, especially B. L. S. Bulletin 841 and the September 1946
proportion to compensation of corporate officers as reported by the Internal
Supplement to this Bulletin, Bulletin 900, and the monthly reports on Current Revenue Service.
291602°—54

6




76

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

3. INCOME OF UNINCORPORATED ENTERPRISES
Apart from farm income, which has been studied systematically by the
Department of Agriculture for many years, no comprehensive body of data
covering any appreciable time interval exists for the income of unincorporated
enterprises. Estimation in this field has generally required the laborious
piecing together and adjustment of various types of data from numerous
sources, some only inferentially connected with noncorporate business
income. In the light of the experience gained in estimation and the changing
nature of source materials, the National Income Division has periodically
reviewed, and materially revised, its estimating procedures in an effort to
develop more refined estimates. No such review, however, has as yet produced really satisfactory results, for refinement of estimating techniques,
unfortunately, is not an adequate substitute for reliable source materials.
Still, the latest such revision, prepared for this 1954 edition of NATIONAL
INCOME, can be said to mark a significant step forward in the work. The
incorporation into the estimates of data on noncorporate business receipts and
number of proprietors from the postwar industrial and population censuses
was in itself of value. But reference is had mainly to the availability of information from the Internal Revenue Service which afforded for the first time a
sound basis for adjusting for understatement the net income data which
previously had been provided from tabulations of unaudited tax returns
submitted by individual proprietors and by partnerships. With the benefit of
this information, the estimates of total nonfarm unincorporated business
income for the later period may be adjudged reasonably reliable as to general
level.
Exhibit 1.—Income of Unincorporated Enterprises, 1949
Millions o.
dollars

Industry
Farms

Percent

12,718

Business
Retail trade and automobile services
Contract construction
Services (except professional service)
Wholesale trade,
Manufacturing
_
Finance, insurance, and real estate
Transportation
Mining
_
Agricultural services, forestry, and fisheries
Communications and public utilities
Total

--

_.

._

11.3

2,386
987
227
219

7.1
2.9
.7
.7

17,144

Medical and other health services
Legal services
Engineering and other professional services, n. e. c
Accountants
___
___

37.8

3,819

Professional services

50.9

8,012
2,541
2,200
1,269

23.8
7.5
6.5
3.8

1,238
775
602
24 i
235
29

3.7
2.3

35,631

loa.o

General summary of sources and methods
The statistical approach and methods adopted in national income and
product estimation are primarily a function of the character of available data.
The force of this general proposition is clearly illustrated by the estimates of
the income of unincorporated enterprises. Three broad segments of the
estimates may be differentiated with respect to source materials and methods
used: Income of professional practitioners, business income (the nonfarm
total except for professional service income), and farm income. (See Exhibit
1.)
(1) Estimates of the net income of professional practitioners—amounting to
11 percent of the noncorporate total in 1949—are prepared very largely by
multiplying the number of persons engaged in independent practice in each
profession by their average net income. This method, adopted at an early
stage oi the official national income work, takes advantage of the basic data
on the number of practitioners from enumerations by the decennial Census of
Population (and of the records of professional associations and other sources
permitting extension of the census data to other years). The dearth of
requisite information on income, however, led the National Income Division



to undertake the collection of data on the average net income of independent practitioners in the various professions. Questionnaire surveys, first
made in 1933 to cover the years 1929-32, have been conducted at periodic
intervals. The results of three postwar surveys were published in the August
1949 Survey of Current Business for lawyers, the January 1950 Survey for dentists,
and the July 1951 Survey for physicians.
(2) Information for estimating the "business" segment of the income of
unincorporated enterprises—one-half of the total in 1949—has been generally
fragmentary. Comprehensive data are lacking except for 1945 and 1947, for
which years the Internal Revenue Service provided tabulations of the incomes of sole proprietorships and partnerships filing income tax returns. The
proportion of the total number of firms filing was very large because of
comparatively low income exemptions and the high levels of business activity.
For individual industries comprising this segment, the initial basic step in
procedure has been to establish a measure of the universe in terms of either
gross receipts or number of active proprietors. Gross receipts have then been
multiplied by a profit ratio (ratio of net income to receipts); the number of
proprietors, by an estimate of their average net income. For both methods,
the years 1939, 1945, and 1947 represent benchmarks, developed almost
entirely from Internal Revenue and Census materials.
Comprehensive data on receipts of sole proprietorships and partnerships
were reported for most of the larger industries for 1939 and 1947 or 1948 in
the industrial censuses and for all industries for 1945 and 1947 in the tabulations of income tax returns. Some of the censuses required some allowance
for undercoverage, and the Internal Revenue data had to be adjusted for
enterprises not filing returns and for differences in industrial classification
from that used by the National Income Division. Noncorporate business
receipts were also reported in a few of the 1929 censuses; but for all years
other than 1929, 1939, 1945, and 1947 or 1948 receipts almost always have
had to be estimated indirectly, generally by interpolation and extrapolation
by available data on total receipts (corporate and noncorporate combined)
or corporate receipts. The accuracy with which this could be done has varied
widely among industries.
Estimates by industry of number of active proprietors also have been
derived in large part from the industrial censuses, which for the period prior
to 1939 furnished information for this item more frequently than for noncorporate receipts. The other principal source has been the decennial Census
of Population. The 1940 and 1950 censuses, with their cross-tabulations of
the employed labor force by class of worker and detailed industry, were
more useful for this purpose than the 1930 census. Estimates of number of
proprietors by industry for noncensal years have been obtained by interpolation and extrapolation techniques utilizing, for the later period, chiefly the
business population series of the Office of Business Economics.
For all three benchmark years 1939, 1945, and 1947, the Internal Revenue
Service tabulations for sole proprietorships and partnerships filing returns
were employed in the estimation of profit ratios and average income per
proprietor by industry. It was necessary to adjust the data to represent the
universe through estimation of the profit ratios or average net income of
enterprises not included in the tabulations. Because of the difference in tax
return coverage, this adjustment was quite important for the 1939 estimates
but comparatively minor for the 1945 and 1947 estimates.
For extension of the base-year profit ratios or average income figures to
other years, Internal Revenue Service tabulations for sole proprietorships for
1941, 1943, and 1949 furnished the most directly relevant data. They required, however, difficult adjustments for differences in scope and other
aspects of noncomparability. To adjust, industry by industry, for firms not
filing returns in 1941 and 1943—as well as in benchmark year 1939—extensive use was made of the comprehensive IRS tabulations for 1945. Nonfilers were viewed as firms in the lower income and sales classes, and the
relationships found to obtain between the reporting (larger) and non-reporting (smaller) firms in the 1945 distribution were used to expand the markedly
less complete IRS tabulations for 1939, 1941, and 1943.
In the interpolation and extrapolation of the IRS-based estimates of profit
ratios and average income per proprietor, heavy reliance has been placed
on corporate profits data.1 These can be obtained annually from Internal
1. Where it appeared important to improve comparability with the noncorporate data,
compensation of corporate officers and sometimes net corporate monetary interest have been
added to corporate profits (before tax) prior to calculation of the ratio oi corporate profits to
sales or average income per corporation.

NATIONAL INCOME, 19 5 4 EDITION

Revenue tabulations except for the two or three most recent years, when
sample information must be employed. For the period after 1939, data for
smaller-sized corporations have been utilized. Prior to that, when smallcorporate data were lacking, the movement of noncorporate profit ratios or
average income was based to a great extent on data for all corporations.
While the source materials for estimating the income of unincorporated
nonfarai businesses have been generally unsatisfactory, a change in this

situation is in process. This steins mainly from the broadened coverage of the
Federal income tax and the plan of the Internal Revenue Service to mine
the comprehensive source of statistical data afforded by the returns. As
noted, the Service already has provided detailed industry tabulations of
proprietorship and partnership returns for 1945 and 1947 and of proprietorship returns alone for 1949. Proprietorship tabulations for 1951 are now in
process.
It is clear that prospect for further improvement in the basic data
situation for the entrepreneurial income segment of the national income
Exhibit 2.—Income of Unincorporated Enterprises, by Industry, 1929, 1939, must rest in the availability of comprehensive, current tax-return tabula1945, and 1949 l
tions. For, despite limitations, tax returns under a well-administered law
are the best source of annual information which it is feasible to secure on
[Millions of dollars]
the earnings of unincorporated enterprises. In the absence of such informa1945
1929
1939
1949
tion, the necessary resort to indirect data as the basis of estimation cannot
be very satisfactory, particularly on an industry basis.
14,017 111, 776 130, 941 33, 681
All industries, total
The reliability of any income estimates for noncorporate business is con6,033 4,378 11,972 12, 953
Agriculture, forestry, and fisheries
ditioned by the fact that this group characteristically includes a large
5,968 4,317 11, $21 12, 718
Farms
number of relatively small enterprises, which follow a diversity of account65
lil
148
235
Agricultural services, forestry, and fisheries.
ing practices and frequently maintain only rudimentary records. But now
54
243
108
61
Mining
2
3
5
5
Metal mining
that the income tax and the expanded social security system are covering a
2
3
4
1
Anthracite mining
very high percentage of business proprietors, it can be anticipated that
6
28
39
7
Bituminous and other soft coal mining34
63
169
42 I
Crude petroleum and natural gas
record-keeping and accuracy of reporting will improve with the passage of
10
11
6 !
26
Nonmetallic mining and quarrying
time. Not only will the individual firm require systematic records for
1,127
2, 541
646
Contract construction..
these purposes, but the growing audit program of the Internal Revenue
565
1,238
400 2, 361
Manufacturing
Service will tend both directly and indirectly to promote standardization
64
S87
215
88
Food and kindred products
and accuracy of reporting.
3
7
4
1
Tobacco manufactures
119
23
40
14
Textile-mill products
In this connection, it should be noted that results of the Internal Revenue
132
603
73
Apparel and otherfinishedfabric products.
185
85 II
38
30
Lumber and timber basic products
Service 1949 audit study have been incorporated into the estimates for the
93 / M73
18
17
Furniture andfinishedlumber products
30
5
Paper and allied products.
"business" segment of income of unincorporated enterprises. This was an
14
intensive audit, based on a representative sample, giving primary em165
130
101
52
Printing, publishing, and allied industries
04
31
Chemicals and allied products
15
31
phasis to business income reported by individuals. It was carried on by
5
7
Products of petroleum and coal
1
1
16
3
experienced field investigators through direct contact with taxpayers and
1
0
Rubber products
75
28
15
Leather and leather products
9 I
examination of their records.
42
47
23
Stone, clay, and glass products
12
137
26
Iron and steel and their products, including ordnance..
19
The IRS 1949 audit study permitted an adjustment to be made, by
123
24
»201
Nonferrous metals and their products
17
147
23
Miscellaneous manufacturing
18
industry, of the net income reported for tax purposes by individual proprietors. For lack of data, the same relative adjustment was applied to the
213
31
97
22
Machinery, except electrical
-31
3
10
3
Electrical machinery
income reported by partnerships. Analysis of collateral audit information
40
2
4
1
Transportation equipment, except automobiles.
14
indicated that, insofar as size-of-firm differences (as measured by "adjusted
2
14
2
Automobiles and automobile equipment
gross income") alone are determining, a lesser adjustment might have been
2,867 3,193 9,676
9,281
Wholesale and retail trade
382
1,269
478 1,739
Wholesale trade
in order. However, an estimate of the magnitude of difference involved
8,012
2,485 2,715 7, »37
Retail trade and automobile services-.
amounted to only a small percentage of total business income, and would
775
762
346
757
Finance, insurance, and real estate
tend to offset any extent to which the audit study could not uncover all
2
1
3
2
Banking
34
392
65
171
errors on the part of taxpayers resulting in understatement of income.
Security and commodity brokers, dealers and exchanges20
17
27
49
Finance, n. e. c.
162
226
144
340
(3) The Department of Agriculture furnishes the estimates of farm proInsurance agents and combination offices
186
119
330
350
prietors' income included in the national income. The estimation of
Real estate
220
249
426
602
Transportation
farm income is a principal part of the over-all statistical services rendered
0
0
0
0
Railroads
by the Agricultural Marketing Service. From the quinquennial Census of
0
0
0
0
Local railways and buslines
41
115
119
27
Highway passenger transportation, n. e. c
Agriculture, the Department's Crop and Livestock Reporting services,
289
173
452
Highway freight transportation and warehousing.
215
field surveys, and many other sources, it has developed both income and
1
4
10
Water transportation
14
balance sheet statistics for the agricultural industry.
0
0
Air transportation (common carriers) _
0
0
0
0
0
Pipe-line transportation
0
The aggregate net income of farmers is derived as the difference between
12
Services allied to transportation
5
17
3
!
gross income (calculated in detail by type of product) and production
Communications and public utilities
23
29
7 i
expenditures (estimated separately for about 40 different types of expense).
Telephone, telegraph, and related services-.
6
2i
6
Radio broadcasting and television
1
4
2
Such a complete development of income data by means of a synthetic
0
0
Utilities: electric and gas
0
4
19
Local utilities and public services, n. e. C—.
15
income and expense statement is unique with agriculture.
2,980 2,496 4, 530
6,019
Because the individual industry figures on noncorporate business income
Services
123
91
252
192
are subject to unusual estimating difficulties and are not so reliable as the
Hotels and other lodging places..
493
396
758
923
9
6
13
25
industry breakdowns for most other types of national income, only industrial
Personal services
345
168
172
451
Commercial and trade schools and employment agencies.
division totals are presented in table 17 of Part V. However, to show the
449
192 !
157
563
Business services, n. e. c
Miscellaneous repair services and hand trades
general composition of the division totals and the factors underlying their
112
50
28
73
Motion pictures..
157
48
52
130
Amusement and recreation, except motion pictures
movement over time, Exhibit 2 gives individual industry estimates of the
1,145
903 1, 527
2,386
Medical and other health services
income of unincorporated enterprises for selected years of a two-decade
763
571
553
987
Legal services
103
119
104
227
Engineering and other professional services, n. e. c
span. The exhibit does not extend beyond 1949, it may be noted, since
51
62
34
Educational services, n. e. c
62
that is the latest year for which noncorporate business income data by
industry are presently available from the Internal Revenue Service.
1 "Income of unincorporated enterprises" measures the net income of sole proprietorships
and partnerships, except that the series for wholesale trade includes estimated patronage
The following description of methodology is divided into the three areas
refunds and stock dividends paid by farmers' cooperatives (shown separately in table 12).
2 Because of changes in industrial classification, as discussed in the Introduction to Part III,
noted in the foregoing introductory remarks: Professional services, business,
basic data are not available for making 1949 estimates for the individual industry groups.
and farm.
Approximate eomparibility is possible, however, for the combined-group totals.




78

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

cians, for which the percentage was as high as 45. Because of larger-sizec?
samples, the sampling error in the later surveys (covering the period since
1937)—aside from the question of possible bias in the response—is probably
The professional services cover the following industry groups for which
smaller than in the earlier surveys. This consideration is especially important
estimates are published in table 17 of Part V of this report: Medical and
with respect to the legal profession, in which the dispersion of income, and*
other health services, legal services, accountants (included in business
consequently the sampling error, is larger than in the curative professions.
services, n. e. c ) , and engineering and other professional services, n. e. c.
The generally low rate of response to most of the questionnaire surveys
For the medical group, aggregate net income is the sum of series for physiraises doubt concerning the representativeness of the sample data on average
cians and surgeons, dentists, osteopathic physicians, chiropractors, chiropoincome. Various tests, however, point to the probability of no marked bias
dists, private-duty trained nurses, veterinarians, and miscellaneous curative
on this score. It has been possible in many cases to check the representativeand healing services. For the engineering and "other" group, income is
ness oi the sample with respect to known characteristics of the profession
estimated separately for engineering and architectural service and for other
sampled, such as State of practice, size of city, and age of practitioner. In
professional services, n. e. c.
addition, study of average incomes in the various professions by age classes,,
Of these various professions, by far the largest are physicians and surgeons,
years-in-practice classes, city-size classes, and regions—and, in cases where
lawyers, and dentists. These three groups accounted for four-fifths of prothe sample was large enough, by cross-classifications of these characteristics
fessional service income in 1949.
—has been reassuring. The income patterns revealed were both uniform
and plausible.
Further substantiating evidence is afforded by the result of the specific
Medical and legal services
check on response bias made in the 1950 physicians' income survey. As
For the medical and legal service industries the income of proprietors
explained in the report in the July 1951 Survey of Current Business, the first
measures the earnings of all professional practitioners from independent
and second follow-ups of the first response modified its findings only slightly
practice. It is derived by multiplying (1) the number of professional perand indicated the absence of appreciable bias. Mainly because the initial
sons in independent practice (full or part time) by (2) average net income
response rate was so high, however, this result is not conclusively applicable
figures determined from questionnaire surveys of the professions.
to the earlier surveys.
(1) For some or all of the years 1930, 1940, and 1950, data from the
The questionnaire surveys for physicians, dentists, and lawyers have prodecennial Census of Population permitted the derivation of the total numvided average income data for almost all years of the period since 1929. For
ber of self-employed persons for the three major groups—physicians, denthese three professions, average income estimates for years not covered by
tists, and lawyers—and also for veterinarians and miscellaneous curative
National Income Division surveys have been made in various ways. These
and healing services. For the other medical and health service groups, the
include interpolations by the results of questionnaire surveys by other
universe numbers of persons engaged in independent practice have been
agencies (for physicians, by Medical Economics magazine; and for dentists, by
estimated in a generally satisfactory manner from such sources as reports
the American Dental Association), by reference to the movement of income
of the professional associations, mailing lists of the Fisher-Stevens Service,
in one of the other two professions, or by using disposable personal income to
Inc. (a mail-service company), and, for 1929, the report of the Committee
interpolate gross receipts and industry (professional) payroll data to interpoon the Costs of Medical Care. This Committee's report also furnished an
late expenses. A recent-period procedure by the Division has been to make
estimate of the total number of physicians in independent practice in 1929.
short surveys (based on a mailing of only 3,000-5,000 questionnaires) in order
For interpolation and extrapolation, the American Medical Directory of to obtain interim information for extrapolation purposes. Reports on such surveys, covering lawyers, dentists, and physicians, were given in the July 1950
the American Medical Association has provided data for most years since
and July 1952 issues of the Survey of Current Business.
1929 on the total number of physicians (including those not in practice),
and for some years on the number engaged in private practice. The American
For the remaining group of smaller medical professions, surveys by the
Dental Association has furnished information on the number of dentists
National Income Division provided average income data for most years of the
for years since 1941, with estimates for 1929 and 1931—39 being obtained
earlier period. Later estimates represent extrapolations from survey benchby straight-line interpolation and extrapolation of the 1930 and 1940 censusmarks for the year 1941 in the case of private-duty nurses and veterinarians,
based figures. Straight-line interpolation and extrapolation has also been
and for 1937 for osteopaths, chiropractors, and chiropodists. Because of evinecessary to fill gaps in the data for the other medical professions. This
dence of some past period correspondence, the average net income of physiis not, however, a source of any appreciable error in the medical services
cians has been generally used as an extrapolating series for these groups.
total, as these professions form only a small part of it.
Little direct information has been available on average incomes in the
For the number of lawyers, interpolations have been based mainly on
miscellaneous curative and healing services.
data obtained from the Martindale-Hubbell Law directory and from compilations of the number of attorneys listed in a selected group of city direcOther professional services
tories.
Estimates of the two remaining professional groups—engineering and other
In the estimates for physicians, dentists, lawyers, and veterinarians,
professional services, n. e. c. and accountants—are of lesser validity because
adjustments have been made for the changing number of professional
of the lack of directly relevant information for most years.
practitioners in the armed forces after 1940, on the basis of data obtained
With respect to the former, the component series on engineering and
from the Procurement and Assignment Service of the National Roster of
architectural service has been derived from combinations of numerous
Scientific and Specialized Personnel and from the Department of Defense.
sources, including income-tax return data, a survey of consulting engineers,
(2) Most of the income surveys yielding the data on average net income
by the National Income Division covering the 1929-32 period, and an index,
used in the estimates for the medical and legal services industries have been
of the value of all engineering construction contracts awarded as reported by
conducted by the National Income Division, usually in cooperation with
Engineering Mews Record. The other, much smaller, component, other prothe professional association in the field. In most cases, separate computafessional services, n. e. c , has been estimated from tax return data and use
tions of average entrepreneurial income were made for nonsalaried pracof the total net income in legal services as an interpolating and extrapolating:
titioners (those deriving all their net income from independent practice)
series.
and for part-salaried practitioners (those deriving income both from independent practice and from salaried work).
The net income of independently employed accountants was obtained for
1929-36 by multiplying their estimated number by average income data
All of the periodic questionnaire surveys have relied upon a voluntary
collected in National Income Division surveys. The total net income of
response to mailed questionnaires. The respondents were not identified.
accountants for the period beginning with 1945 was estimated on the basisQuestionnaires were mailed either to all persons in the profession or to a
of tax return data. For intervening years, it was estimated largely from the
representative sample of the profession. Information was requested in each
linear regression between the net income of accountants and that of lawyers
case for a series of years. Usable replies have been received, on the average,
in independent practice.
from about 15 percent of the mailing, apart from the 1950 survey of physi-

income of Independent Professional Practitioners




NATIONAL INCOME, 195 4 EDITION

Income of Business Enterprises
The aggregate net income of noncorporate "business" enterprises—the
nonfarm total other than professional service practitioners—represents the
summation of separate estimates for about 65 industry subgroups. Many of
these, however, are quite small. Three important industries—retail trade,
wholesale trade, and construction—largely determine the accuracy of the
business total. They accounted for 69 percent of it in 1949.
The following discussion of this segment is divided into the several timeperiods characterized by general similarity of source materials and procedural problems. It describes the preparation of the estimates of noncorporate business income prior to correction for audit. To the extent based
on tax-returns data, the estimates for each of the periods reviewed were
adjusted, industry by industry, by reference to results of the Internal Revenue
Service 1949 audit study discussed above. The necessity of using a single
adjustment factor for all years was somewhat unfortunate, but nonetheless
represented a substantial improvement over the situation when no systematic
basis was available for any period to allow for the understatement of income
reported in compilations of unaudited tax returns. As described in the 1951
NATIONAL INCOME supplement, the allowances that could be made for this
factor were informal and indirect.

Benchmark estimates, 1945 and 1947
The tabulations from tax returns provided by the Internal Revenue
Service for 1945 and 1947 represented such high coverage of all proprietors
in the business segment as to constitute, after certain necessary adjustments,
benchmark materials.
For both 1945 and 1947, the Internal Revenue tabulations showed receipts,
expense, and net income information by industry, separately for sole proprietorships (from the business schedule of the individual income tax returns)
and for partnerships (from the mandatory informational returns). The information for sole proprietors in 1945 and partnerships in 1945 and 1947 was
classified by receipts size-class of firm. The tabulations were based on complete coverage of the larger sized enterprises and on estimates developed
through scientific sampling for the smaller enterprises. The definition of net
income employed was very similar to that used in national income measurement. Only a few adjustments had to be made, on the basis of reported
data, to secure uniformity. These included the addition of depletion charges
for proprietorships and partnerships and the elimination of capital gains
and losses and property income received from partnership receipts, which
were reported inclusive of these items.
The Internal Revenue tabulations, it was evident, differed in some respects
as to industrial classification from that used in the national income estimates.
Also, the tabulations were incomplete to the extent of not covering firms
which did not file tax returns. Adjustments for both of these deficiencies were
accomplished through comparison of the universe number of proprietors or
gross receipts in each industry, as estimated by the National Income Division,
with the data reported by the Internal Revenue Service.
The estimates of number of proprietors in 1945 and 1947—included in the
Division's annual series on number of persons engaged in production (table
28, Part V)—were prepared by interpolating base-year estimates derived
from the industrial and population censuses by an index of the number of
noncorporate firms, adjusted for differences in the estimated number of
active partners per partnership on the basis of Internal Revenue compilations.
The index of noncorporate firms was constructed, industry by industry, by
subtracting the number of active corporations in each year from the quarterly
average number of operating firms estimated by the Office of Business
Economics as part of its business population series. The sources and methods
underlying this series were described in an article in the January 1954
Survey of Current Business. For some industries, the 1945—47 movement shown
by this index was modified on the basis of Internal Revenue Service tabulations of the number of sole proprietors and partners.
In the estimation for 1945 and 1947 of noncorporate receipts—the universe
"control" adopted in the important retail trade sector and in some of the
service industries—benchmarks were developed from the 1948 Census of
Business. The reported data were adjusted upward to allow for exclusion of
concerns which, although in operation during 1948, had gone out of business
before enumerators' visit in 1949, or for some other reason were not enum


79

erated. These adjustments, which reflected also any differences in industrial
classification, were made largely by comparing the Census of Business data
on the number of employees or proprietors with similar data from the social
security records or the Census of Population and Housing (the latter data
obtained by interpolations of 1940 and 1950 Census enumerations).
To derive estimates for 1945 and 1947 (as well as for other years of the
period since 1939), these 1948 receipts figures and similar benchmarks for
1939 were interpolated and extrapolated by available receipts data. In the
case of retail trade, the index of noncorporate receipts utilized for the purpose
was obtained by deducting corporate sales (as reported annually in the
Internal Revenue Service's Statistics of Income) from total retail sales as estimated by the Office of Business Economics through 1951 and by the Census
Bureau thereafter (described briefly in the section on Personal consumption
expenditures for commodities). For personal service industries, the censusbased estimates were extended to other years by utilizing as indexes the series
on consumption expenditures for personal services (described in the service
expenditure notes in Section 8). For both retail trade and personal services,
the interpolations and extrapolations were carried out in considerable typeof-store or industrial detail in order to obtain the advantages of proper
weighting.
In most industries, the National Income Division estimate of number of
proprietors or gross receipts for 1945 and 1947 somewhat exceeded the
Internal Revenue figure based on reporting for tax purposes; and the profit
ratio or average income shown for one of the smallest receipts classes was
assigned the nonreported receipts or proprietors in order to derive net income
not reported on the tax returns. This, together with an adjustment for unreported net income disclosed by audit (based on the 1949 study), was
added to the amount reported in the tax-return tabulations to secure the
estimate of total net income for the industry.
In some industries, however, the Division's estimate of number of proprietors was less than that compiled from income tax returns. The differences
were regarded as a matter of industrial classification.2 In wholesale trade, for
example, the excess of proprietors filing income tax returns was adjudged to
belong in the Division's retail trade classification. The average income per
proprietor in retail trade was assigned to this group of proprietors, and their
total income was added to the amount reported on income tax returns to
account for part of the income in retail trade not reported to the Internal
Revenue Service. In manufacturing, where the Internal Revenue total number of proprietors greatly exceeded the National Income Division estimate
the receipts and net income of proprietors in the under $10,000 or under
$15,000 receipts classes (varying among the 20 industries) were shifted from
manufacturing to miscellaneous repair services and hand trades. This shift
was intended to yield data comparable in scope to those for the 1929—39
biennial censuses, when the Census of Manufactures defined the industry (in
general) to include firms with $5,000 or more output, and for 1947, when the
Census definition of a manufacturing firm was changed to the similar one-ormore employees criterion. The results obtained by the receipts-class cutoffs
for 1947 generally approximated the Census of Manufactures data for that
year on number of proprietors, which were taken as benchmarks.
In general, then, noncorporate net income in the business industries in 1945
and 1947 was derived through adjustment of the tax return data by using
number of proprietors or gross receipts as a measuring rod.

Benchmark estimates, 1939
The coverage of Internal Revenue tabulations for 1939 was limited by
reason of the larger income tax exemptions then prevailing. Nevertheless,
because these tabulations covered the operations of both sole proprietorships
and partnerships in the same year for which Census enumerations were
available, there was opportunity to merge information from the two sources
to obtain a noncorporate income benchmark.
The principal method of preparing the base-year estimates for 1939 consisted of multiplying, separately for sole proprietorships and partnerships,
total receipts reported in the various industrial censuses by profit ratios
derived from Internal Revenue tabulations of data from the business schedule
2. Statistical differences were present here too, as well as in cases where the National Income
Division estimate exceeded the number reported by the Internal Revenue Service. The
Division estimates were subject to estimating errors, and the Internal Revenue data to
sampling errors. In most industries, a high proportion of proprietors was in the receipts
classes sampled. (For net income, however, the proportion of the total sampled was much
smaller.)

80

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

of individual income tax returns and the mandatory informational returns
Interpolations, 1940-44 and 1946
filed by partnerships. The tabulations, published in Statistics of Income—Part 1
For other years of the period 1939-47, estimates of net income of unincorand Supplement were based on a complete count for partnerships and for sole
porated enterprises in the business segment were prepared by interpolation
proprietorships (in general) with income of 55,000 or more, and on sampling
of the 1939, 1945, and 1947 benchmarks. Two factors made the process
for smaller proprietorships. The sole proprietorship tabulations gave for each
difficult and impaired the accuracy of the estimates.
industry the receipts and net income of firms classified by gross receipts
First was the necessity of making sizable adjustments for noncomparability
classes.
of coverage of the data on sole proprietorships reported for several years of
Noncorporate receipts totals were available from the 1939 Censuses of
the period by the Internal Revenue Service. Secondly, the extensive use made
Retail Trade, Contract Construction, Service Establishments, Manufactures,
of corporate data to estimate the movement of noncorporate sales and inand Mineral Industries. The data reported for retail trade, construction, and
come—unfortunate in itself—was complicated by a shift of some firms,
mining were adjusted upward to allow for undercoverage in the manner
because of the tax advantages to be gained, from the corporate to noncorindicated in the discussion of the 1945 and 1947 benchmarks.'
porate form of organization in the war period, and by an opposite shift in the
Adjustment of the Internal Revenue data to take account of the many small
early postwar years after the removal of the corporate excess profits tax.
firms not filing tax returns in 1939 was necessary before they could be used to
1. For 1941 and 1943, the Internal Revenue Service provided tabulations
derive profit ratios to apply against the industrial census receipts figures.
of sole proprietorships' tax returns showing receipts, net income, and number
The principal complication in deriving aggregate profit ratios by industry
of proprietors classified by industry. Because of changes in tax laws and the
stemmed from the strong negative correlation between the profit ratio and
rising level of economic activity, the coverage of sole proprietorships afforded
the size of firm as measured by receipts. Because of this correlation, profit
by these tabulations was much higher than for 1939, though still incomplete
ratios computed from data for firms filing returns were almost always too low
in relation to the comprehensive data for 1945. Expansion of Internal Revenue
to be applicable to the entire noncorporate sector of an industry. The general
coverage over this period was such as to bring in an increasing proportion of
procedure, therefore, was (1) to deduct receipts reported to the Internal
small proprietorships, which, as noted, have very high profit ratios relative
Revenue Service from the census-based receiptsfigures,and then (2) to multito the large proprietorships. In the interpolation procedure, use of the
ply the resulting estimate of receipts of firms not filing tax returns by the
reported data directly, without adjustment for the differing proportions of
profit ratio calculated for one of the smallest Internal Revenue receipts-size
small-sized firms covered, would have imparted a strong upward bias to the
classes. The sum of estimated net income for the firms not filing income tax
estimates.
returns and the net income actually reported on the returns constituted the
aggregate net income estimate (prior to the adjustment for audit discussed
2. Most of our limited knowledge about the magnitude and nature of the
above).
shift in legal-form of organization is confined to its postwar phase.4 It is
evident that it was minor as to the number of firms involved, but appreciable
Two important variants of this general procedure were developed. One was
as to its effects on total sales and net income. The shift was largely between
based on the finding that for sole proprietorships in the nonprofessional comthe corporate and partnership forms, and it was restricted mainly to manuponents of the service industry the ratio of payroll plus net income to receipts
facturing, retail trade, wholesale trade, and contract construction. For these
tended to be constant throughout the reported receipts-size distribution.
industries—in which the relative movements of sales and income from 1939
Thus, the tendency of the profit ratio to decline as the size of firm increased
to 1945 and from 1945 to 1947 differed markedly for the corporate and nonwas just offset by the tendency for the ratio of payroll to receipts to rise.
corporate segments—several special procedures were followed to improve
Where payroll data by size of firm could be obtained from the industrial
the use of available statistical data for interpolation.
censuses, as in services and contract construction, the use of this relationship
The series used to interpolate noncorporate net income in "business"
permitted the estimation of small-firm profit ratios from data for larger firms.
industries between the 1939, 1945, and 1947 benchmarks were prepared as
This was helpful, since the data for proprietorships with receipts under $5,000
reported in Statistics of Incomefor1939 often were fragmentary or appeared to the product of either (1) receipts (sales) and profit ratios or (2) number of
proprietors and average income per proprietor. The first was generally the
be biased.
preferred method, although for some industries the choice was merely a
The second variant—utilized chiefly in retail trade and most of the manumatter of statistical convenience.
facturing industries—relied on relationships between small and large firms
The interpolations between 1939 and 1945 were carried out separately for
shown in the 1945 Internal Revenue tabulations, coverage of which wasi
sole proprietorships and partnerships. The chief purpose was to utilize the
virtually complete. This procedure is described below in connection with the
Internal Revenue Service tabulations of sole proprietorships for 1941 and
interpolations of the 1939 and 1945 benchmarks.
1943. Data for smaller-sized corporations then furnished the basis of the sole
For a few industries in which adequate data on gross receipts or profit
proprietorship interpolations for 1940, 1942, and 1944. The resulting proratios could not be obtained, the net income estimates were prepared as the
prietorship series was usually employed for interpolation of the 1939 and
product of the number of proprietors, taken from the population or industrial
1945 partnership benchmarks. Where small-corporation data (on the ratio
censuses, and average income per proprietor, derived by adjusting tax return
of profits to sales or average income per firm) agreed better with the 1939—45
data for incomplete coverage or from industrial census data on the reported
partnership movement, however, such data were used for the partnership
withdrawals of proprietors. Wholesale trade was the most important industry
interpolations as well. For all industries, the estimates for 1946 were obtained
in which this method was applied.
on the basis of interpolations utilizing data for small corporations, without
The foregoing brief account of the principal methods used to derive 1939
the necessity of making separate estimates for proprietorships and partnernet income estimates in the "business" area is an oversimplification. The
ships.
matching of Census and Internal Revenue receipts tabulations by size classes
The actual methodology for preparing the 1940—44 and 1946 interpolations
and the preparation of adjusted profit ratios and of average income per prowas considerably more diverse than indicated from the foregoing summary.
prietor were generally difficult processes subject to appreciable error. EmHowever, four aspects of the work were of general importance and provide a
phasis on the judgmental factor was necessarily considerable. Moreover,
convenient framework for describing it.
variations of, or departures from, the general methods frequently were
Estimation of number of proprietors.—For most industries, the method of internecessary in particular industries.
polating net income required separate estimates of number of proprietors and
average income. Base-year figures on the number of proprietors were obtained
3. In the case of contract construction, the adjustment was broader in scope than for the
from the industrial or population censuses. For the period since 1939, they
other industries. Labor-force data in the 1940 population census were used to establish the total
of employees and proprietors in contract construction. Subtraction of the National Income
Division estimate of employees in contract construction yielded the total number of proprietors, whether in establishments or own-account workers (such as carpenters and painters)
operating from their own homes. The estimate of average receipts (and of average net income)
assigned to the non-establishment group, which was not covered hi the Census of Construction, was based on fragmentary in'ormation. The total net income of this group was about
two-fifths of the Division's published total for contract construction in 1939.




4. This knowledge derives from a special Internal Revenue Service tabulation showing for
1946 the number of corporations, by industry, which had been sole proprietorships or partnerships in the prior year and, for all industries combined, their distribution by asset-size classes;
some overall annual data for the period 1945-48 from the Bureau of Old-Age and Survivors
Insurance on the number of covered firms undergoing reorganization (believed to comprise
mainly legal-form changes); and direct comparisons of Internal Revenue Service corporate
and noncorporate sales and income data for 1945 and 1947.

NATIONAL INCOME, 19 5 4 EDITION

81

the upper and lower portions of the sales distribution were calculated.
were interpolated by a series measuring the number of noncorporate firms,
The relative difference between the profit ratios (or average incomes)
with adjustment for changes in the ratio of partners per partnership shown
of the two groups was then applied to the reported profit ratio (or average
by Internal Revenue data for 1939, 1945, and 1947. The series was derived
income) of income taxfilersin the food group in 1939 to obtain the estimated
by deducting the number of active corporations (reported by the Internal
profit ratio (or average income) of nonfilers in that year.5
Revenue Service) from the Office of Business Economics estimates of the total
Use of this procedure of "filling in" the sole proprietorship distributions
number of operating firms. For most industries, the separate estimates of sole
thus enabled the calculation of sole proprietorship profit ratios or average
proprietors and of partners required by the method were also prepared by
incomes for 1939, 1941, and 1943 on a consistent basis with the data for
using the noncorporate-firm series as an interpolating index. But in many
1945 reported by the Internal Revenue Service. For retail trade and most of
of the 20 major types of manufacturing and in contract construction, a special
the manufacturing industries, as noted earlier, this procedure directly
procedure was adopted in the attempt to reflect shifts in legal form of organiprovided the data used for the 1939 benchmark.
zation. The number of sole proprietorships derived for benchmark years from
Census reports was interpolated on the basis of the total number of operating
Use of small corporation data.—The preparation of universe estimates of
firms, and the resulting series was subtracted from the estimated total number
receipts and number of proprietors and of profit ratios and average income
of proprietors for the industry in order to derive the number of partners.
per proprietor for sole proprietorships for 1941 and 1943 provided the frameThe rationale of this procedure stemmed from the fact (noted above) that
work for subsequent heavy reliance on small-firm corporation data to fill in
the legal-form changes occurred largely between corporations and partnerthe remaining estimates through interpolation. The relative movements of
ships.
either the ratio of corporate profits to sales or average income per corporation were used for interpolation of the similar averages for unincorporated
Noncorporate business receipts.—For some industries, mainly retail trade
enterprises.
and the personal services, the interpolation procedure required estimates
of noncorporate business receipts. As noted earlier, the receipts series for
For calculation of the corporate series, data from the Internal Revenue
retail trade was derived for benchmark years (1939 and 1948) from the
Service "Source Book", unpublished volumes supplementing Statistics of
Census of Retail Trade and extended to other years by an index computed
Income with more detailed information, were utilized. From the tabulations
by deducting corporate sales from total retail sales; and the series for the
showing the annual corporation income returns classified by asset-size
personal service industries were based on the Census of Service Establishclasses, data for the two or three smallest classes were taken. In order to
ments, with interpolations by means of the National Income Division
approximate the noncorporate income concept, the reported compensation
estimates of consumption expenditures for personal services.
of corporate officers was added to net profit before taxes prior to calculation
of either the ratio of profits to sales or average net income per corporation.
In principle, it may be noted, this procedure of estimating noncorporate
retail sales accounted for sales shifts due to corporations' changing their
Although the basis for generalization is still limited, it would appear that
legal form of organization. The measure, however, doubtless was somewhat
data for small corporations provide a fairly satisfactory index to the moveblurred by errors of estimation in the residual sales figures. In this connecment of noncorporate income—and are decidedly superior in this respect
tion, it may be noted that little could be done to correct for whatever into data for corporations in the aggregate. A striking example is afforded by
consistency was introduced into the corporate sales data by the change
the changes from 1945 to 1947. In nearly all industries, ratios of profits to sales
from the unconsolidated basis of filing through 1941 to permissive conderived from data for corporations in the aggregate were found to be wholly
solidated reporting beginning with 1942.
unsuitable for interpolation of the noncorporate ratios. Noncorporate profit
ratios declined markedly in most industries from 1945 to 1947, whereas
It was necessary to prepare separate receipts estimates for sole proprietorthe corporate ratios tended to be stable. Upon analysis, however, it was
ships and partnerships for the interpolations between 1939 and 1945, so as
found that this divergence reflected size-of-firm, and not legal-form, difto take account of the IRS sole proprietorship data for 1941 and 1943.
ferences. Industry-by-industry compilations of profit ratios (adjusted to inIn the case of retail trade, ratios of proprietorship sales to total noncorporate
clude corporate officer compensation) for corporations in the two or three
sales were obtained for 1939 from the Census of Retail Trade and for 1945
smallest asset-size classes showed the same pattern of decline as exhibited
from the Internal Revenue compilations. These ratios were interpolated
by the noncorporate ratios. The small-firm corporate data, it may be of
along a straight line for intervening years and then applied to the total
interest to note, generally showed stability from 1945 to 1946, with all of the
noncorporate receipts figures. For the personal services industries, similar
1945—47 decline occurring in 1947.
sole proprietorship ratios were computed for 1939 and 1948 from the Census
of Service Establishments and also interpolated in straight-line fashion.
Based on this and other favorable experience with the use of small corExpansion of 1941 and 1943 IRS data to full coverage.—Following the preparaporation data, a principle change in methodology in the revisions of nontion of universe estimates of number of proprietors or gross receipts by incorporate business income for the present report has been the fuller utilization
dustry, the next basic step was to adjust to a full coverage basis the sole
of such data. Their incorporation into the 1939-45 interpolations is the major
proprietorship tabulations provided by the Internal Revenue Service for
instance. As noted earlier, corporation income data by asset-size classes are
1941 and 1943. In order to obtain for each industry an interpolating series
not available in comparable industry detail for the period prior to 1939.
that was comparable over time, it was necessary to add to the reported
Internal Revenue data for 1939, 1941, and 1943 estimates covering the
Extrapolations, 1929-38
nonreporting proprietorships—assigning to them an average net income
or profit ratio appropriate to smaller-sized firms. In this procedure, which
There is comparatively little information for the 1929-38 period relating
required use of the universe estimates of receipts or number of proprietors
directly to the net income of unincorporated enterprises in the business
as a "control," use was made of relationships obtaining between small and
area. Noncorporate receipts data were compiled for 1929 in the Censuses
large proprietorships in the IRS tabulations for 1945. These tabulations,
of Retail Trade, Manufactures, and Mineral Industries. Usefvil data on
it will be recalled, provided practically complete coverage of sole proprietors,
number of proprietors were provided by the Census of Wholesale Trade in
classified by detailed industry and receipts-size classes.
1929, 1933, 1935, and 1939. A compilation of 1936 income tax returns for
The procedure adopted assumed that for 1939,1941, and 1943 the profit ratio
sole proprietorships and partnerships was made available by the Internal
to be assigned to nonreported sales—or the average income to be assigned
Revenue Service, but this could not be used in many industries either because
to nonreporting proprietors—bore the same relation to the reported profit
of large sampling errors in the tabulations or because they could not be
ratio (or average income) as shown by the two segments of the 1945 distribusatisfactorily adjusted to obtain comparability with 1939.6
tion estimated to correspond to the filing and nonfiling segments in those
previous years. As an example, if retail trade food store proprietors reporting
5. In principle, absolute instead of relative differences should perhaps have been applied
to the Internal Revenue Service in 1939 accounted for 40 percent of the
because of the probable tendency for income changes over different phases of the cycle (and
estimated total sales of all retail food store proprietorships, then the 1945
in the short run generally) to be relatively less volatile for small than for large noncorporate
distribution by receipts size was broken into two parts—one composed of
enterprises (chiefly because of the differing relative importance of proprietors' own and
family labor versus hired labor). However, tests showed that the amounts of difference insales in the highest receipts classes covering 40 percent of total sales and the
volved in the application of the two procedures were quite small.
other composed of sales in the receipts classes below that point. The ratios
6. Unlike later tabulations, where sampling was used only for the smaller-sized firms, the
of net income to sales (or average incomes) of the two groups represented by
1936 tabulation was based on a 20 percent nonstratified sample of all returns.



82

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

With direct data on noncorporate business income limited largely to the
foregoing, the estimating methods for the period 1929-38 necessarily relied to
a very large extent on indirect measures. The preparation of estimates for
a substantial number of individual industries provided a basis for utilizing
available detailed information and for taking account of shifts in the importance of industrial components within the business total. The methods
used in a few of the larger industries may be summarized briefly.
In the important retail trade sector, the net income of unincorporated
enterprises was derived as the product of receipts and the ratio of net income
to receipts, each obtained by extrapolations of base-year figures for 1939.
Receipts were extended to 1929 by a series representing (1) Census values
for 1929 and 1939; (2) estimates for 1933 and 1935 derived by preparing
legal-form breakdowns of Census aggregates for those years by interpolating
between such reported breakdowns for 1929 and 1939 by weighted indexes
of chain store sales and independent store sales; and (3) interpolations for
all other years on the basis of the Office of Business Economics estimates of
total retail sales.
Following the derivation of the receipts series, the ratio of payroll plus net
income to receipts for noncorporate retail trade in 1939 was extrapolated to
1929 on the basis of similar data for corporations. Application of the resulting
ratios to the noncorporate receipt series and the deduction of estimated noncorporate retail payroll yielded the 1929—38 estimates of total net income.
For this extrapolation of the 1939 ratio of net income plus payroll to receipts,
the payrolls of corporations and of unincorporated enterprises were estimated
by methods analogous to those followed in obtaining receipts. Profits data
for corporations were those of the National Income Division, based on annual
Statistics of Income reports of the Internal Revenue Service.
The above method of estimating the ratio of net income to receipts was an
application of the relationship, noted above, developed from 1939 Census
Bureau and Internal Revenue Service materials for sole proprietorships in
the nonprofessional service industries—that throughout the receipts-size
distribution the ratio of payroll plus profits to receipts tended to be constant.
This relationship, reflecting the varying proportion of labor performed by
paid employees as against proprietors and their families, was assumed to be
valid with respect to corporate and noncorporate retail trade in a temporal
application. In retail proprietors' income there is a large and comparatively
stable element of labor return corresponding to retail corporations' payroll.
Analysis for 1939, it may be noted, indicated that the difference in profit
ratios between incorporated and unincorporated enterprises in retail trade
was accounted for very largely by differences in the relative importance of
payroll expense.
The income of proprietors in wholesale trade for the years 1929—38 was
estimated as the product of number of proprietors and average net income
per proprietor. The 1939 estimate of number of proprietors was extrapolated
to 1929 by a series comprising Census of Wholesale Trade data for 1929,
1933, 1935, and 1939 and straight-line interpolations for the other years.
Average net income per proprietor was derived for 1936 by adjusting the
Internal Revenue Service tabulations for incomplete coverage, and for all
other years of the period through interpolation and extrapolation by the
average net income estimates for retail trade.
The income of noncorporate enterprises in contract construction was
derived for the 1929-38 period by multiplying estimates of receipts by the
ratio of net income to receipts. Receipts of corporations, obtained from
Statistics of Income, were deducted from the estimated total value of contract
work to obtain a series on noncorporate receipts that was used to extrapolate
the 1939 estimate. Value of work performed was estimated for 1939, by four
major types of construction, from the Census of Construction. The 1939
values for these four types were extrapolated to earlier years by components
of the construction activity series (see notes on New construction), which
covers force-account as well as contract construction, and summated to
obtain total value of work performed. The 1939 noncorporate profit ratio in
contract construction was extrapolated to 1929 by the absolute changes from
year to year in the ratio of profits plus compensation of officers to receipts
for corporations in the industry.
The value of product of noncorporate enterprises in manufacturing was
taken for 1929 and 1939 from the Census of Manufactures. Separately for
the 20 major types of industries, estimates for the intervening odd-numbered
years were obtained by interpolation on the basis of the total value of product
of all enterprises reported in the biennial manufacturing censuses. The even



years were then interpolated by gross sales plus gross receipts from operations
of corporations in manufacturing, as published in Statistics of Income. The
resulting series was then multiplied by estimated profit ratios. These were
derived for each industry for 1936 from the special tax return tabulations
and extended to other years of the 1929-38 period chiefly through use of a
linear regression between noncorporate profit ratios for 1936, 1939, and
1941 and annual (adjusted) corporate profit ratios from 1929 to 1941.

Estimates for 1948 and 1949
After the preparation of the 1947 benchmarks, estimates of noncorporate
business income were next established for 1949. In large measure, these were
derived by extending the 1947 universe by changes shown for sole proprietorships in the Internal Revenue tabulations for 1947 and 1949. As noted,
partnerships' returns were not tabulated for 1949.
The actual use made of the sole proprietorship data for this purpose varied
by industry. In many of the 65 industry groups for which separate estimates
are prepared, the procedure was simply to extrapolate 1947 net income by
the income totals reported for proprietorships alone in the two years. In
certain other industries—again, notably retail trade—the sole proprietorship
figures were used to extrapolate the 1947 combined profit ratio for partnerships and proprietorships, as well as to furnish a check on the requisite
estimates of the sales universe. Analogously, the sole proprietorship data
were also used frequently to project average income per proprietor, which
was then multiplied by independent estimates of the number of proprietors.
Common to these several types of procedure was the application of the
1947-49 relative experience of sole proprietorships to the noncorporate
universe. This was not feasible, however, for industries where partnerships
bulked large in the noncorporate total. In such industries, mainly the 20
groups for manufacturing, separate treatment of the partnership segment was
indicated. The general procedure was (1) to extrapolate the total net income
of sole proprietorships by the reported IRS data and (2) to estimate partnership income as the product of sales and profit ratios. Sales were obtained
by multiplying number of partnerships (estimated from changes in the
number of noncorporate business firms) by average sales per partnership
extrapolated from 1947 by the movement of small-corporation data. Extrapolation of the profit ratios for partnerships relied either on the corporation
data (with absolute changes instead of relative changes utilized where the
size of the partnership and corporation profit ratios differed markedly) or on
the sole proprietorship data. In this and other instances, comparison of the
changes shown by partnerships, sole proprietorships, and small corporations
from 1945 to 1947 was a useful guide as to choice of procedure.
Interpolations for 1948 made extensive use of small-firm data for corporations. In the case of manufacturing, use was made of unpublished tabulations
from the Census of Manufactures of receipts of sole proprietorships and
partnerships. These data referred to the 1947 receipts of manufacturing
enterprises which were in a noncorporate status at the time the census information was collected, in the first part of 1948. As such, the totals for the
relatively small noncorporate segment of manufacturing were appreciably
affected by the shift in legal form of organization still progressing throughout
1947. Particularly for partnerships, the reported receipts figures were
adjudged for national income purposes to be a closer measure of noncorporate manufacturers' receipts for 1948 than for 1947. With timing
adjustments (generally minor) to convert the data for the census universe to
a 1948 basis, it was possible to obtain direct approximations of noncorporate
receipts in manufacturing for that year.7
While the sole proprietorship tabulations for 1949 were thus quite valuable,
lack of comparable data for partnerships made for difficulty and impaired
the accuracy of the estimates by individual industries. It may be noted,
however, that there is reason to believe that the errors by industry were
substantially offsetting in the aggregate. For total nonfarm unincorporated
business income (including the professional services), the relative changes
7. Earlier it was noted that the 1947 Census of Manufactures furnished the basis for benchmark data on number of proprietors. The number of establishments involved in legal-form changes
was relatively quite small, and the number-of-proprietors data were much less affected by
this factor than sales or income. Before utilizing the Census of Manufactures data as 1947
benchmarks, adjustments were made to convert the data from measures of number of proprietors as of early 1948 to the estimated average number during 1947. For this purpose, use
was made of the OBE quarterly data on number of operating firms and of the special Internal
Revenue tabulations of corporations changing their legal form of ownership in 1946.

NATIONAL INCOME, 195 4 EDITION

from 1947 to 1948 and from 1948 to 1949 were found to check closely against
information compiled by the Internal Revenue Service from income tax
returns of individual proprietors and partners. This information consisted of
data compiled from the "face" of the returns, not tabulations of business
income by industry. The data were available only as a single aggregate and
required adjustment for this purpose to eliminate the estimated amount of
net income reported by farmers. It is not believed that errors in this adjustment could affect the results appreciably.

Current estimates
Pending availability of noncorporate tax-return tabulations on an industry
basis, estimates of the net income of unincorporated enterprises in the business segment for current years must be prepared from fragmentary and
indirectly relevant data. At the present time, 1949 is the latest year for which
such tax return information (for sole proprietorships) is available.
For extrapolating the 1949 estimates through 1951, considerable use
was made of information for corporations from Internal Revenue tabulations, available by asset-size classes (for small corporations) through 1950
and for corporations in the aggregate through 1951. For the estimates
beyond 1951, corporate data employed in the estimation of noncorporate
business income were derived from sample compilations of Government
agencies, of private organizations in some cases, and of the National Income
Division from published income statements of individual firms. These
sample data are described in the section on Corporate profits. To obtain
the benefit of appropriate weighting, the extrapolations were carried out
in as much industry detail as feasible.

Exhibit 3.—Comparisons of New and Superseded Series for Income of Business
and Professional Unincorporated Enterprises, 1929-52
[Billions of dollars]
New series
1929
1830
1931
1932
1933

Superseded
series >

11.6
67
5.0

8.1
6.3
4 7
2.!)
3 4

3.1

1934
1935
1936
1937
1938

4.6
5.4
6.6
7.1
6.6

4.3
5.0
6.2
6.7
6.1

1939
1940
1941
1942
1943

8.5
11.5
14.3
17.0

6 9
7.8
10 2
12.9
15.1

1944
1945 .
1946
1947
1948

18.1
19.1
23.0
21.4
22. 1

1949
1950
1951
1952

21.0
24.0
25.1
25.5

17.2
18.8
22.4
21.3
22. 5
21. (
24 t
26. f
26.1

1. As published in the National Income Number (table !) of the July 1953 SURVEV
CURRENT BUSINESS.

For many industries—most of them relatively small—even reasonably
satisfactory corporate information is lacking (or entirely inappropriate) for
current extrapolations of the proprietors' income estimates. This generalization applies to wholesale trade, agricultural services, forestry and fisheries,
and most of the finance, nonprofessional service, and transportation and
public utility industries. For these groups, the movement of net income is
based on one or a combination of such series as number of proprietors
(extrapolated by the Office of Business Economics data on number of operating firms), average earnings of employees, wages and salaries, consumer
expenditures, and estimates of net income in related industries. For contract
construction, principal reference for this purpose is had to estimates of
wages and salaries in the industry and construction activity. For wholesale




83

trade, net income in current years is derived from a sales extrapolation
(based on the Office of Business Economics series) and a profit-ratio extrapolation based on estimates for retail trade.
For current extrapolations of noncorporate business income in the important retail trade area, use has been made for recent years of data contained in the year book issues of "Mail-Me-Monday Barometer of Small
Business," published by the Accounting Corporation of America. This
source summarizes the profit-ratio experience of over 6,000 small enterprises, primarily noncorporate, by line of retail operation. While it has not
yet been possible to check the representativeness of the Mail-Me-Monday
data against Internal Revenue noncorporate data, one indirect test has
been reassuring. It was found that the 1949-50 movements of the MailMe-Monday ratios were in good agreement with those shown for profit
ratios (adjusted to include corporate officers' compensation) reported for
small retail corporations by the Internal Revenue Service. As brought out
above, the profit ratio changes of the smaller-sized corporations accorded
closely with the noncorporate changes from 1945 to 1947.
In the past few years, it has been possible to make two general-type
checks of the current period movement of the estimated total of net income
in the business and professional industries. One, already discussed, is provided by Internal Revenue compilations of net business income reported by
individuals on the face of their Federal income tax returns. At present,
data of this type are available through 1951, and serve to corroborate the
income changes estimated to that point from the 1947 benchmark. A second
summary test is afforded by comparison of the movement of the noncorporate income estimates with that shown by a series derived from Federal
individual nonwithholding tax collections, appropriately lagged and adjusted for tax-rate changes. This series, also available only in the aggregate,
consists largely of taxes against the net income of nonfarm unincorporated
enterprises, but covers also personal income from higher-bracket salaries,
farming, rents, interest, and dividends, as well as capital gains and losses.
Allowance for these other sources of personal income is made in the comparison with the tax series; but because of its broader coverage and the
difficulty of adjusting for changes in tax rates, the measure of the relative
movement in total nonfarm entrepreneurial income which it affords is
far from precise. It nevertheless serves as a useful check, particularly in view
of the weakness of source materials underlying the individual-industry
estimates for very recent years. These estimates are subject to a possible
upward or downward adjustment from the analysis of the tax data.

Revisions of the estimates
In Exhibit 3, the 1929—52 estimates of the total net income of unincorporated
business and professional enterprises prepared for this report are compared
with those published in the July 1953 National Income Number of the
SURVEY OF CURRENT BUSINESS.

The new estimates, unlike previous ones, incorporated results of the 1948
Census of Business, the 1950 Census of Population and Housing, and the 1949
audit study of the Internal Revenue Service. The audit study, as discussed
earlier, was of particular importance since it furnished the first systematic
basis for adjusting the noncorporate business income estimates to the extent
based on compilations from unaudited tax returns.
In addition to use of new source materials, the current revision of the series
reflects improvements in procedures. These included, among others, (1) the
method adopted for expanding the 1941 and 1943 Internal Revenue proprietorship tabulations to a full coverage basis by use of the comprehensive
1945 tabulations as a frame of reference, and (2) the fuller use made of smallfirm corporation data to serve as a guide to the movement of noncorporate
net income.
It will be seen that the overall effect of this statistical reworking has been to
raise the level of the series from 1929 through 1944. Revisions for the earlier
war years are relatively somewhat larger than for the prior period, mainly by
reason of the first change in methodology just noted. The revisions of estimates
for the 1929-38 period stem mainly from the higher 1939 base-year figures
from which they were extrapolated. Although for a few industries the methodology used for this period was changed, the main procedure was to employ
the previous series as indexes for extrapolating the new 1939 estimates.
While, on the whole, the present revisions of noncorporate business income
are not large in the aggregate—particularly when the difficulties encountered

84

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

in this area of estimation are considered—they were quite marked for a number of individual industries. In part, this was attributable to use of actual
audit information on an industry basis in contrast to the previous procedure of
making informal allowances where seemingly appropriate. In addition,
numerous other factors centering in the incorporation of the new Census data
and their integration with tax return materials had sizable effects in the
individual industry series, although operating largely to cancel out in the
all-industry total. This experience in the current revisions influenced the
decision to curtail the extent of industry detail published in table 17.
The estimates of total noncorporate business income published in the July
1953 SURVEY (as shown in the exhibit) had undergone, it may be noted, two
interim statistical revisions of appreciable magnitude. These occurred mainly
in the July 1949 and July 1950 national income reports. In the former report,
the principal change was to raise the war-period level. This revision—from
$16.9 billion to $18.8 billion for 1945—was based on the 1945 Internal
Revenue tabulations of sole proprietorships and partnerships. The earlier
estimates had not taken account of the war-period shift in legal form of
organization.
In the July 1949 report, total noncorporate business income was shown to
increase from $22.7 billion in 1946 to $24.7 billion in 1947. In the following
year—in the July 1950 report—the availability of Internal Revenue proprietorship and partnership data for 1947 resulted in a revision of the estimates to $22.4 billion in 1946 and $21.3 billion in 1947. This reflected a
previous missing of the sharp decline in noncorporate profit ratios in 1947,
noted earlier in this section. Like the legal-form shift, this was a phenomenon
that could not be discerned or measured from the statistical materials available when the preliminary estimates were made.

Net Income of Farm Proprietors
The estimates included in this report for the net income of farm proprietors
were prepared by the Agricultural Economics Division of the Agricultural
Marketing Service. They correspond to that Division's published series on
"total net income of farm operators from farming, including Government
payments." The farm income estimates are described in chapter 20 of The
Agricultural

Estimating

and Reporting Services of the United States Department of

Agriculture (Miscellaneous Publication No. 703), Washington, 1949, and the
description here is therefore limited to very brief compass.
Farming as an industry comprises about 5}i million independent enterprises, most of which do not keep accounts permitting them to calculate net
income on any uniform basis. However, an extensive body of both benchmark
and current statistics on farming operations in the aggregate has been
developed over a long period of time.
The general procedure used by the Agricultural Economics Division in
deriving the net income of farm operators is to make an independent estimate
for each item required for an income-and-expense statement covering all
farms, based on whatever combination of available data best represents that
item. Net income of farm proprietors is then calculated by summing the estimates for components of gross farm income and deducting the sum of the
production expense estimates, as illustrated in Exhibit 4 for the year 1949.
Basic data for preparing the farm income estimates are obtained partly
from the farmers themselves through the quinquennial Census of Agriculture
and a current Crop and Livestock Reporting system designed to include at
least one reporter in every farming township in the country. An important
part of the basic data, however, is derived from the auxiliary industries and
agencies which market, transport, store, process, or otherwise aid or regulate
the production or distribution of farm products. As the assembling and
processing of many of these products are either dominated by large firms or
are subject to sanitary inspection, and the individual products are fairly
homogeneous in nature, it is feasible to obtain from auxiliary industries very
reliable data (with little time lag) on the marketings and market prices of
such important products as meat animals, most dairy products, cotton,
tobacco, and sugar beeis.

Gross farm income
The major components of gross farm income are cash receipts from the
marketing of crops and of livestock and livestock products and the value of



food and fuel produced and consumed on farms. Cash receipts and home
consumption are calculated for each commodity by States from estimates of
the quantities involved and the average prices paid to farmers. The price
estimates are based on monthly mail questionnaire returns from about 30
percent of the 35,000 dealers and farmers to whom such questionnaires are
sent; the same price series are used in deriving cash receipts from marketings
and value of home consumption. The estimates of quantities marketed and
consumed on the farm are made in a framework in which production data
are reconciled with disposition data.
Crop production is estimated as acreage harvested times average yield per
acre. To obtain acreage and yield figures, benchmarks derived from the
Census of Agriculture and several annual State assessors' censuses are extended by sample data on acreage and yield changes indicated by a sample
of close to 100,000 reporting farmers. The changes reported currently are
adjusted on the basis of past relationships between the sample data and the
benchmarks.
Disposition of each farm product covers the respective quantities sold,
consumed by the farm family, used for feed and seed, and added to inventory.
Exhibit 4.—Derivation of Net Income of Farm Proprietors,

1949

[Billions of dollars]
GROSS F A R M INCOME

30.9

Cash receipts from farm marketings
Meat animals
Dairy products
Poultry and eggs
Cotton
Food grains
Feed crops
_
Other marketings...
Value of home consumption
__
Gross rental value of farm homes
Government payments
Value of change in farm inventories—

_

27.9
8.3
3.8
3.1
2.6
2.3
2.3
5. 5

_
_

2.2
1.5
.2
—. 9

LESS: PRODUCTION E X P E N S E S .

18.2

Depreciation
Feed purchased
Hired labor
Operation of motor vehicles
Livestock purchased

3.6
3.0
2.9
1.7
1.6

Net rents and Government payments to nonfarm landlords
Fertilizer and lime
Taxes
Farm mortgage interest
Other...

1.1
.9
8
.2
2.4

EQUALS: N E T INCOME OF F A R M P R O P R I E T O R S

_

12.7

It is generally estimated in percentage terms from annual mail reports filed
by a selected sample of producers—by 15,000 wheat growers, for example.
For most field and vegetable crops, these percentages are applied to production estimates, and the results for quantity sold are generally reconciled
with totals based on annual reports by producers accounting in many cases
for the bulk of the commercial movement. Reports from processors, handlers,
or sanitary inspectors provide such good coverage of certain other major crops
and livestock products that firm estimates of marketings can be based directly
on these. Cotton belongs to this class because of reports made by substantially
all ginners. Meat animals are another example, for which there are Federal
meat inspection reports, returns from a very large stratified sample of
butchering firms, and data from common carriers.
Nonrecourse loans to farmers made or guaranteed by the Commodity
Credit Corporation, net of current redemptions, are considered cash receipts from farm marketings. Such loans and redemptions are reported
currently by the Corporation, and net loans are added to the value of
farm marketings discussed above.
Gross rental value of farm dwellings covers housing on owner-occupied as
well as rented farms. It is derived by (1) calculating a return on dwelling
investment from the estimated value of farm dwellings and the average rate
of interest on farm mortgage loans, and (2) adding to this computed net
value the portion of total farm expenses estimated to be allocable to the
upkeep of dwellings. The basic estimates of the total value of dwellings are
prepared from Census of Agriculture data, with only the 1930 Census

NATIONAL INCOME, 195 4 EDITION
reporting separate data on dwellings as distinguished from other structures.
Value relationships are used to derive the proportions of the various expense
items allocated to dwellings.
Government payments in connection with the various farm programs
are reported from the fiscal records of the responsible agency, the Commodity Stabilization Service.
Change in farm inventories
The value of the change in farm inventories is measured as the difference
between physical quantities of crops and livestock on farms at the beginning
and end of the year, multiplied by year-end prices. It is derived as the sum
of separate State estimates for each crop and livestock item.
This component of gross farm income is entered directly in the gross
national product as a component of "Change in business inventories." No
inventory valuation adjustment is required, as in the case of estimating the
change in nonfarm business inventories, since the farm inventory changes
are computed directly from data on physical stocks and current prices.
Benchmark data on the number of each class of livestock on farms are
obtained every five years from the Census of Agriculture. These census
enumerations are adjusted, where necessary, to obtain complete coverage.
Current estimates of year-to-year changes in the number of livestock on
farms are prepared from surveys made each December. The rate of sampling
varies among States, depending largely on the needs of the individual States
for livestock data. In general, a sample of 3,000 to 6,000 farms is considered
satisfactory for a principal livestock-producing State. Special surveys (especially in States where the size of farm varies greatly); records of marketings,
slaughterings, and rail shipments; livestock tax assessment data; and State
farm censuses are used for checking the inventory estimates derived from
the basic survey data.
Estimates of year-end farm inventories of crops are prepared by one of
two methods. For all major crops except cotton and tobacco, farm stocks are
estimated quarterly. These estimates are based upon the results of mail
surveys covering about 80,000 farmers, from which the usable response is
approximately 30 percent. Each respondent reports the production of each
crop grown on his farm and the quantity on hand at the survey date. For
each crop the reported stocks on farms are expressed as a percentage of
total production for these farms, separately for each State. These percentages
are applied to the estimated total State production of each crop to yield
total stocks on farms at the end of each quarter. Studies made in 1948 substantiated the validity of this method.
The estimates of year-end inventories of crops derived in this manner do
not include crops which farmers own but may have placed in commercial
storage located off their farms. The magnitudes involved are not known,
but believed not to be large. Also, for certain crops a part of the estimated
stocks on farms may have been used to secure a Commodity Credit Corporation loan. As the proceeds of such loans are included in gross farm income,
it is necessary to deduct the quantity of these crops under loan from farm
inventories in order to avoid double-counting. The adjustment is made on
the basis of reported data on Commodity Credit Corporation loans and
redemptions.
For cotton and tobacco and a few relatively minor crops, inventory
changes are calculated in terms of inventories "held for ultimate sale."
Crops held for use as feed or seed on farms where grown are excluded. To
estimate the quantity of each crop still remaining to be sold on January 1 of
each year, the amount of the previous year's production of the crop actually
sold through December is subtracted from the total amount to be sold.
The prices used to value the change in physical quantities of crops are
the December 15 average prices received by farmers for their products as
sold at local markets or at points to which the farmers deliver them. For
livestock, inventory values per head as of January 1 are used, based on farmers' estimates of replacement costs.

85

centage rates of depreciation have been derived for dwellings from a 1934
survey covering 600,000 farms and for most other items from scattered
sample data on length of useful life. These are applied to estimates of current
value based on the Census of Agriculture and/or on cumulation of purchases
(as a measure of current gross increment), estimated for tractors and other
types of machinery and equipment from the annual Farm Machines and Equipment reports of the Census Bureau, adjusted to reflect current values on the
basis of Department of Agriculture price index series.
Purchases of feed are reported at 5-year intervals in the Census of Agriculture. For interpolation and extrapolation, an index calculated from price
and quantity data for 18 important feeds is used. Average price data are
based on monthly mail returns from 4,000 reporting merchants. Quantities
are estimated as production, indicated by reports from farm or factory
producers, plus imports, with the total adjusted for inventory change and
exports and other nonfarm uses.
The estimates of wages paid to hired farm laborers are described in the
section on Wages and salaries.
Farmers' expenditures for livestock, which are prepared by States, cover
purchases from all sources outside the State and from public stockyards
within the State. The estimates of cash receipts from livestock are defined
in a manner consistent with this treatment.
Estimates of livestock numbers shipped across State lines for stacker
and feeder purchases are made from inspection records of State veterinarians,
inspections by the Department of Agriculture at 66 public stockyards, and
data derived from truck and railway movements of livestock. Estimates of
the weight and price of the livestock purchased are obtained from records of
transactions at five important stocker and feeder markets, which handle
more than one-half of the transactions.
Motor vehicle operating costs are estimated separately for automobiles,
trucks, and tractors, and only 40 percent of the total cost for automobiles
is charged to farm production. The basic data on number of vehicles are
obtained from the Census of Agriculture and extended on the basis of sales
data on new tractors, less estimated scrappage, and truck and automobile
registrations in predominantly agricultural states. Costs per vehicle for
some items were obtained from a survey in 1936, and have been extended
on the basis of changes in prices. Consumption of gasoline and oil per vehicle
have also been obtained from surveys, supplemented by occasional spot
checks of mileage and average mileage per gallon in the case of automobiles
and trucks, and the number of days used and fuel consumption per day in
the case of tractors.
The series on net rents to nonfarm landlords is discussed in the section
dealing with Rental income of persons. This series comprises the only
rental figures explicitly included in the farm income statement. The explanation is as follows.
Rental flows within the farm sector do not appear in the statement shown
in Exhibit 4 because it is a consolidated statement in which intra-business
flows cancel out. Net rents earned by farmer landlords become merged
with the net income of farm operators and are not isolated statistically.
However, rents paid to landlords not living on farms must be recorded as
flowing out of the farm sector. Only net rents are shown explicitly. The
difference between them and the gross rents actually paid consists of expenses that are included among the various expense items of the statement.
Taxes on farm property cover all ad valorem taxes levied on farm property by State and local governments. Real estate tax rates are developed
from sample data from local tax officials or from farmers, and from real
estate tax data reported in the Census of Agriculture. The tax rate estimates
are applied to value of agricultural land calculated from the censuses and
interpolated by an index of farm land values constructed by the Department
of Agriculture. Personal property taxes are estimated from the real estate
taxes in conjunction with ratios of the amounts of farm real and personal
property on tax rolls as shown in published reports of State tax commissions,
boards of equalization, or similar bodies.

Production expenses
Estimates are made for each of some 40 categories of production expense.
Among the largest are depreciation, purchased feed, hired labor, purchases
of livestock, costs of operating motor vehicles, net rents to nonfarm landlords,
and taxes. These accounted for more than four-fifths of the 1949 total.
Depreciation estimates are made separately for seven categories of farm
property, on the basis of replacement cost, rather than original cost. Per


Characteristics of revisions
Revisions of the farm income estimates are made periodically by the
Agricultural Economics Division of the Department of Agriculture to incorporate results of the quinquennial Census of Agriculture and of special Department of Agriculture surveys designed to improve certain areas of the esti-

86

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

mates. In addition, when the Agricultural Economics Division publishes
each Fall detailed income estimates for the preceding year, it usually presents also revised estimates for several years back, to incorporate newly
available or improved data on production, marketings, prices, etc. In 1952,
the Agricultural Economics Division introduced a comprehensive revision of
the farm income estimates to incorporate new benchmark data reported in
the 1950 Census of Agriculture.
The farm proprietors' income series shown in this report agrees statistically
for the period 1929 through 1950 with the Agricultural Economics Division's
series on "total net income of farm operators from farming, including
Government payments" presented in The Farm Income Situation, Sept.—Oct.,
1953, and thus includes all revisions made through that date. For 1951 and
1952, the present figures include, in addition, some subsequent statistical
revisions in component income series that were available from the Agricultural Economics Division at the time the present series was prepared. For
1953, the farm income figure is a preliminary estimate developed from
Agricultural Economics Division data.
Subsequent revisions to be made by the Agricultural Economics Division
will be incorporated insofar as possible in the present series. However,
because of differences in the Commerce and Agriculture publication schedules, it will not be feasible to maintain complete statistical conformity in the
series published by the two agencies.

this is a distinct advantage, the necessity of deriving the final figures in this
way has made for complexity in the statistical methodology, with consequent
danger of errors. In particular, it will be noted the final estimates are often
calculated as the difference between much larger items which, in turn, are
also obtained as residuals. Even small errors made at the various stages of the
estimating procedure may significantly affect the final results. Both as to
general level and relative movement the rental estimates must be regarded
as among the least satisfactory of national income statistics.
Exhibit 1 .—Components of Net Rental Income of Persons, 1950

Types of property
1. Rented nonfarm property
a. Rented dwellings
b. Royalty-earning.
c. Business and industrial
2. Owner-occupied nonfarm dwellings
3. Farm realty
Total

Millions of dollars
4,000

1,795
558
1,647
3,379
1,094
8,47*

The property-type components of the net rent estimates may be grouped
as above, to show their relative magnitude and the order in which they are
discussed in the following pages.

7. Rented Nonfarm Property
4. RENTAL INCOME OF PERSONS
Rent items appear on both sides of the national income and product accounts. Net rental income of persons is a distributive share of national income, and space rental value of housing is a component of personal consumption expenditures for services. Space rental value of nonfarm housing
is estimated at an intermediate stage in the derivation of persons' net rental
income arising from nonfarm rental housing or imputed to owner-occupancy
of nonfarm dwellings, and is discussed below in connection with these net
dwelling rents. Rental value of farm housing is estimated by the Agricultural Economics Division of the Agricultural Marketing Service of the
Department of Agriculture, and is described in the section on Income of
unincorporated enterprises.
Primary data on rent income are notably inadequate. Rents are received
by a large number of individual landlords who maintain only fragmentary
accounts. The only major attempt to obtain income statements from these
landlords is in connection with the Federal individual income tax. The rent
data tabulated by the Internal Revenue Service are derived from a sample of
the tax returns and are restricted to the net rent item. Lack of gross rent and
expense tabulations precludes systematic use of collateral source materials
to adjust the Internal Revenue Service data for inconsistencies due to inadequate bookkeeping and incomplete reporting.
The tax return tabulations were used to derive a 1941 benchmark which
covers net rent from property types accounting for about two-fifths of persons' net rents in that year. The remainder of the 1941 estimate and the
totals for other years have been derived by subtracting sample-based estimates of expenses from gross rent receipts computed largely from data on
rent paid as reported by tenants. Tax returns, census enumerations, and
sample surveys—the main sources of rental information—provide much
better coverage of rents paid than of rents received.
The resulting estimates of net rent are considerably less reliable than the
farm income series, which is derived by a somewhat similar formula. Although
the rent estimates pertain to an "industry" comparable in size to agriculture,
public interest in real estate renting as an industry has been insufficient to
justify any broad special program of data collection such as underlies the
farm income estimates. Moreover, the use of data from tenants in estimating
persons' net rental income is complicated by the necessity of excluding rent
paid to landlords other than persons.
Derivation of the net rent estimates via a detailed structural analysis ol
gross rent flows and expenses has the virtue of providing a framework which
facilitates the consistent use of all available data on rents. While on balance



Persons' net rental income from rented nonfarm property is derived in the
course of a rather intricate estimating procedure. The following synopsis will
facilitate understanding of the estimating sequence, details of which are
described later.
An estimate of persons' net rental income from rented nonfarm property
for 1941 was derived from net rents reported on individual Federal income
tax returns. For purposes of extrapolation, the 1941 total was broken down
into three groups: (a) Net rents from rented dwellings, (b) royalties, and (c)
net rents from business and industrial property.
(a) Personal net rents from rented dwellings are calculated annually as
the difference between gross rental receipts and landlords' expenses on all
such dwellings, after nonpersonal landlords' receipts and expenses have been
eliminated from these gross totals.
(b) Personal net royalties are estimated annually by applying suitable
corporate ratios to the series described below on rental receipts of persons
from business and industrial property.
(c) Personal net rents from business and industrial property were obtained
as a residual for 1941. Their extrapolation involves estimating persons'
gross rents from business and industrial property for 1941 and other years
as gross rents paid by business and government less gross rents received by
business and government from business and industrial property; deducting
the 1941 benchmark figure for persons' net rent from such property from the
1941 estimate of persons' gross rent to obtain persons' expenses as a residual;
breaking down this residual into component parts and extrapolating them
individually by use of corresponding expense items for corporate landlords;
and subtracting expenses from gross rents to obtain personal net rents.
Details of these procedures are discussed below.

Benchmark estimates
Total net rents and royalties of individuals and fiduciaries reported on
the rent schedules on forms 1040-41, as tabulated by the Internal Revenue
Service for 1941, were adjusted to cover nontaxable fiduciaries, individuals
reporting on the short form 1040A, and persons and nonprofit organizations not reporting. They were further adjusted to reflect losses on property rentals reported by some filers and to deduct an estimate of net room
rents reported, an allowance for neglected or misplaced expense entries,
and an estimate of net rents reported by farmers. The original total of $1.8
billion was reduced to a figure of $1.4 billion, representing net rent from
nonfarm property and net royalties.

87

NATIONAL INCOME, 19 5 4 EDITION

The taxpayer filing form 1040 may report his net rents either on the business schedule or on the rent schedule. These two schedules are adapted, respectively, to large and complex rental activities and to small or simple
transactions. It is reasonable to suppose, therefore, that most taxpayers who
receive rent as an occupational income fill in the business schedule, and that
rents which are merely a supplemental source of income or accrue to persons
without occupation are generally reported on the rent schedule. This distinction between occupational and nonoccupational income underlies the
distinction in national income measurement between the distributive shares
of income of unincorporated enterprises and rental income of persons.
Statistically, the use of the income tax data has serious drawbacks. Of the
total of $1.8 billion reported by the Internal Revenue Service for 1941,
nearly three-fourths is an estimate (for income classes under $5,000) based
on samples comprising less than 6 percent of the returns. Moreover, reporting
requirements are peculiarly difficult to enforce in the case of rental income,
because such income is received in small amounts by large numbers of landlords and because difficult accounting problems are involved in calculating
net rent. Finally, the various adjustments of the tax data made by the
National Income Division had in most cases to be based on fragmentary or
otherwise unsatisfactory data.
It would be possible to estimate net rents directly from income tax returns
for most other years as well as for 1941. However, the relatively high tax
exemption limits in effect in the earlier years meant that a very substantial
part of persons' rental income was not reported for tax purposes. After 1941,
the movement of net rents indicated by individual income tax return data
diverged substantially, not only from the prewar trend, but also from the
movement indicated by a rather substantial body of other data. These include sample series on net dwelling rents and synthetic income statements for
all rental housing, as well as corporation income tax tabulations showing the
net flow of rent out of the corporate sector of business. As individual income
tax data do not show gross rent or rent classified by property type, it has not
been possible to reconcile the net rent figures from this source with these
other data.

multifamily structures as well as by urbanization. Estimates for 1941 and
later years have been made without a structure-type breakdown, in the
absence of the required information on mean rents and (except for 1950)
number of units.
Next, cell estimates of landlords' total expenses for facility and utility
services included in rent are obtained by an analogous formula. Deduction
of these from the contract rent aggregates yields Space rent of tenant-occupied
nonfarm dwellings, a component of personal consumption expenditure and of
gross national product.
Landlords' other expenses are then estimated and deducted from space
rent to determine landlords' net rental income.
Finally, net rents of nonpersonal landlords are estimated and deducted to
obtain the nonfarm rented dwelling component of the rental income of persons.
The magnitudes involved in these calculations are illustrated by the 1950
values shown below.
Exhibit 3.—Calculation of Space Rent and Net Rent for Nonfarm Rented
Dwellings, 1950
Item
1. Rented nonfarm dwellings (number in millions)..
2. Times: Average annual rent (rounded to dollars),.
Equals: Total contract rent (millions of dollars)...

14,310
504
7,207

3. Less: Landlords' expenses for facility and utility
services included in r e n t . . .
4.
Use of cookstoves, refrigerators, furnishings
5.
Electricity, fuel, water, gas, and miscellaneous.
Equals: Personal consumption expenditure for
space rent

464
1,083

6. Less: Landlords' other expenses
7.
Depreciation...
8.
Taxes.
9.
Mortgage interest..
10.
Other
._
Equals: Net rental income of landlords..
11. Less: Nonpersonal landlord net income.

Exhibit 2.—Derivation and Breakdown of Persons' Net Rents From Rented
Nonfarm Property and of Royalties, 1941

Item
1. Net rents tabulated from tax returns
_
2. Less: Adjustments for coverage and concept
3. Equals: Persons' net rents from rented nonfarm property, and royalties
4. Less: Persons' net rents from rented nonfarm dwellings
5. Less: Persons'net loyalties
6. Equals: Persons' net rents from rented nonfarm business and industrial
property

Millions
of dollars
1,770
393
1,377
664
223
490

The estimating procedure actually adopted for years other than 1941 assumes that corporate tax returns provide a firmer statistical base than tax
returns of individuals. It may be noted that incentives to nonreporting of
rents received by individuals have increased with the requirement since 1941
that landlords file the long form 1040 instead of the much simpler alternative
forms available to most nonlandlords.
The derivation of persons' net rents from rented nonfarm property, and of
persons' royalties, in 1941 is summarized in lines 1, 2, and 3 of Exhibit 2.
For extrapolation the total is broken down as shown in lines 4, 5, and 6, corresponding to items a, b, and c in the synopsis given above for line 1 of
Exhibit 1. This breakdown and its extrapolation will next be considered.

Persons' net rents from rented nonfarm dwellings
The first step in preparing annual estimates of personal net rents from
rented nonfarm dwellings is to derive, separately for urban and rural nonfarm dwellings, midyear annual numbers of rented dwellings and of mean
annual contract rent per dwelling. Annual contract rent for each of the two
cells is then obtained by multiplication. For the period from 1929 through
1940, dwelling unit numbers and mean and total rentals were estimated
separately for each of four cells, being cross-classified between one- and



Rural
nonfarm

Urban

Equals: Net rental income of personal landlords
from nonfarm rented dwellings

3,649
343
1,251

8,458
1,630

1,168
878
1,391
602
1,868

Total

16,828
4,739

2,089
294
1,795

1

The published series for consumption expenditure for space rent also includes allowances, totalling $234 million in 1950, for certain lodging house rents and lodging received
gratis as a perquisite of employment. These allowances are rough approximations based on
a variety of census, income tax, social security and other data.

The derivation of the values in Exhibit 3 is of special importance for two
reasons. In the first place, substantially the same procedures and data sources
used here are used also in deriving estimates for owner-occupied dwellings.
Secondly, the estimated expense series for rented dwellings are used as
adjustment factors in extrapolating the expenses incurred by persons as
lessors of business and industrial property. For these reasons the exhibit will
be discussed line by line. The methodology for owner-occupied dwellings
will be indicated at the same time where convenient.
1. The total number of tenant-occupied nonfarm dwelling units has been
benchmarked on the decennial Censuses of Population and Housing for 1930,
1940, and 1950. Data for 66 cities included in the Commerce Department's
Real Property Inventory of 1934 were used in conjunction with 1940 Census
data from the same cities to derive an estimate for 1934. Census Bureau
data on the size and composition of the housing stock, based on interviews
with scientifically constructed samples of 20,000—30,000 families surveyed
in connection with the Bureau's Monthly Report on the Labor Force, were
used for the years 1944, 1945, and 1947. These benchmarks are interpolated
and extrapolated by estimates of the numbers of rental units, including
vacancies.
The estimated numbers of rental units have been derived by adding to
census- or survey-based benchmark totals the cumulative net number of new
multifamily units provided in each subsequent year by construction or conversion, and deducting an allowance for the number of units demolished.
The construction series rest on Bureau of Labor Statistics estimates of housing
starts, which are based chiefly on reports of building permits issued. (See
section on New construction.) The conversion and demolition series have
been derived, partly by graphic interpolation, in the process of reconciling
the decennial census totals and the Census Bureau's sample-based data with
the estimates for the same dates made by adding new construction cumulatively to the base stock. For the war and early postwar years, such recon-

88

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

ciliations also required allowance for the shift of many dwellings out of the
by the total number sold during that period. Annual value of sales and numrental market to owner occupancy.
ber sold have been estimated basically from Census of Manufactures data,
which have been interpolated, extrapolated, and raised to cover distribuThe sums of the estimates for tenant-occupied units plus a similarly derived
tion cost by use of information from trade sources.
series for owner-occupied units have been checked for 1948 and subsequent
years against annual control totals, for all occupied nonfarm units, based on
For furniture, a sample-based average expense benchmark was derived
the numbers of households reported in the Census Bureau's Current Populafor 1933 in the Financial Survey of Urban Housing. This is extrapolated by
tion Surveys. These population surveys also provided the basis for estimates
Bureau of Labor Statistics price indexes, and the annual results are averaged
of the effects of the Bureau's adoption in 1950 of new criteria for distinguishover the back period without weighting the years.
ing urban, rural nonfarm, and farm dwellings (see its Current Population Re- The numbers of dwellings for which specific facilities are provided at the
ports—Population Characteristics, series P-20, No. 33, of February 12, 1951).
landlords' expense are estimated by applying annual percentage ratios to
The discontinuity in the statistical series between 1949 and 1950 which rethe series on total number of rental units described under (1) above. The most
sulted from the change in criteria has been smoothed out by wedging back
recent general benchmark for each of these ratio series was derived from the
to April 1947. The principal distortion remaining in the final estimates is an
Dwelling Unit Survey undertaken by the Bureau of Labor Statistics in 1949—
exaggeration of the rate of increase in the number of units between April
50 in connection with the revision of its Consumer Price Index. City averages
1947 and April 1950: of an indicated gain of about 5 million units in the total
from this Survey were weighted by an adaptation of the series of population
nonfarm housing stock, just under 1 million represents the smoothed effect
weights developed by the Bureau from preliminary census tabulations, for
of the classification change.
its 91-city Survey of Consumer Expenditures for 1950. The Dwelling Unit
Survey is described in the July and September 1951 issues of Construction, and
Besides tenant-occupied dwellings, line 1 of Exhibit 3 includes an allowance
the Survey of Consumer Expenditures is described in the April 1951 and
for units rented but temporarily unoccupied, based on Census and sample
August 1952 issues of the Monthly Labor Review.
information as to the composition of the stock of vacant units.
The final estimates of the number of rented units are adjusted to represent
Other base-year percentages of dwellings let furnished were determined
the average stock for each year.
from the 1940 Census of Population and Housing and from the 1934 Financial Survey of Urban Housing. These benchmarks are interpolated and ex2. Average annual rent for rented units likewise has been derived from the
trapolated by use of data gathered periodically by the Bureau of Labor
decennial censuses of 1930, 1940, and 1950 and the sample surveys made by
Statistics from families in the sample underlying its Rent Index.
the Census Bureau for 1944, 1945, and 1947. Minor adjustments have been
made in the decennial census data for nonreporting of rent and to convert
Percentages of dwellings let with refrigerator were estimated from the
the final means for April to an annual-average basis. The mean for the 1950
1934 Financial Survey of Urban Housing, the 1935-36 and 1941 studies of
Census date is based on a preliminary sample tabulation of census returns
consumer purchases, and the 1949-50 Dwelling Unit Survey. Interpolation
for about 46 thousand dwelling units in about 14,000 census enumeration
and extrapolation have utilized data analogous to those used in extrapolatingdistricts systematically selected from all enumeration districts throughout
the percentages of dwellings let furnished.
the Nation; arithmetic means for rent were not included in the Census
Most of the source materials for furniture and refrigerators cover only
Bureau's final tabulations of all returns. Values for other years represent
urban housing. Estimates for rural nonfarm units are based on those for
interpolations and extrapolations by reference mainly to the Rent Index
urban dwellings, using differentials indicated by the Census or the 1935—36
component of the Bureau of Labor Statistics Consumer Price Index. Also
Consumer Purchases Study.
used for this purpose were the mean rent data provided by the Financial
For cookstoves, the 1949-50 Dwelling Unit Survey provided the only data.
Survey of Urban Housing for a 12 to 15 percent sample of all families in
The percentages indicated by this source are held constant for all years and
52 cities, covering the years 1929, 1932, and 1933.
for both urban and rural nonfarm units.
The Rent Index is based on returns filed by about 40,000 families in 46
5. The aggregates in line 5 of Exhibit 3 are products cf total number of
cities. The National Income Division has modified the recent-year values in
rented units (described under (1)), percentages of units including the rethe Index to allow for the comparatively high rent of newly constructed
spective utilities in rent, and average cost per unit. The percentages indwellings (as indicated by properties mortgaged under Federal Housing Advolved have been based on the Financial Survey of Urban Housing, the 1949ministration guarantee) and for changes in the common practice of land50 Dwelling Unit Survey, and interim city surveys by the Bureau of Labor
lords with regard to the inclusion of facilities and utilities in rent. Data for
Statistics covering families in the sample underlying its Rent Index.
the latter adjustment are obtained through periodic Dwelling Unit Surveys
The cost averages are based on data for households purchasing utilities
by the Bureau of Labor Statistics.
directly, since data on payments through landlords are not available. The
3. Facility expenses and utility bills, although sometimes paid by the
principal benchmark was provided by the 91-city Survey of 1950 Consumer
resident family directly to the producer, are often included in contract rent
Expenditures mentioned above. The general procedure was to (a) derive a
and paid by the landlord. As outlays for these items are classified in gross
national average outlay, per surveyed family reporting outlay for the given
national product under personal consumption expenditures for household
utility, (b) estimate the implied average for families in dwellings of each
operation, it is necessary to exclude from the housing component the allowstructure type, and (c) combine these to obtain averages presumed appliance which landlords make for them in fixing rents. The total amount to be
cable to units where utility costs were included in rent.
excluded for each item is estimated by multiplying together two factors:
The weights used in (a) for the various cities were those developed by the
the average cost of providing the item for one dwelling; and the number of
Bureau of Labor Statistics. The technique for (£) involved reconstructing a
dwellings for which the item is provided at the landlords' expense. The deritype-of-structure breakdown for the utility costs represented in the survey.
vation of each of these factors is outlined below under (4) for facilities and
It utilized type-of-structure cost differentials from various sample surveys,
under (5) for utilities.
and a 1950 Census classification of units by structure. The latter was modified
to omit units with utilities included in rent, since these units were not repre4. The average cost of providing each type of facility (cookstove, refrigerasented in the survey averages. For step (c), the structure-type averages detor, or furnishings) for one dwelling is calculated for the present purpose as
rived in (b) were weighted together by the estimated number of one-family
the sum of annual depreciation plus maintenance cost, since this sum is the
and multifamily units respectively in which the landlords paid for the utiliextra amount which the rent must cover to make the inclusion of the facilities
ties. In deriving these weights, the total numbers of such units were obtained
in it worth while from the landlords' standpoint.
as percentages of the total numbers of rented units (see above), and disThe estimates of average maintenance cost (except for furniture) are flat
tributed by structure-type on the assumption that multifamily unit rents
rates based on trade opinion. Depreciation averages are calculated from the
rather commonly include utilities and that only a residual would consist cf
estimated original average price of the equipment in use, by straight-line
one-family units.
amortization over the estimated useful life of the equipment.
The resulting cost means pertained only to urban units. Corresponding
The length of useful life is determined for most items from studies made for
averages for rural nonfarm units were derived by applying urbanization
the Internal Revenue Service. The original average price which is amortized
differentials based largely on the 1935-36 Consumer Purchases Study. The
is generally calculated as a weighted moving average of annual prices coverfinal implied aggregates were reconciled approximately with the corresponding a back period equal to the length of useful life. For stoves and refrigeraing values shown in Part V, table 30.
tors, it is derived by dividing the total value of sales during the back period



NATIONAL INCOME, 195 4 EDITION
Annual estimates derived from reports by the Edison Electric Institute
and the American Gas Association could be reconciled fairly closely for 1950
with these benchmark averages for electric and gas utility costs, and are
used to extrapolate them. The 1950 benchmark averages for heating fuel
cost, and secondary benchmarks derived from the Financial Survey of
Urban Housing of 1934, are interpolated and extrapolated by an indicator
representing the product of the Consumer Price Index fuel component
times an index of heating season degree days as reported by the Weather
Bureau. Average water bills, estimated for 1950 by the procedure described
above and for 1932 and 1942 from water utility revenues reported to the
Census of Governments, are interpolated and extrapolated by city water
utility revenues per capita of population.
A 1940 average for the cost of such minor utilities as rubbish removal and
telephone service included in rent, based on data from the Federal Housing
Administration, Office of Price Administration, and other sources, is moved
from year to year by the weighted average combined cost of the other utilities
considered abeve.
6. The other expenses deductible from contract rent may conveniently be
described here for owner-occupied as well as for rental units. Most of these
estimates are initially made for tenant-occupied, owner-occupied, and
vacant units separately. Expenses on vacant units are subsequently allocated
between rental housing and home ownership in proportion to the rental
value of the vacant units of each type.
7. Depreciation is derived by applying a flat rate of 2 percent to the estimated original cost value of all nonfarm dwellings. The result is allocated
by tenure groups in proportion to aggregate value in each category.
The depreciation rate is based on several surveys of average length of useful life of dwellings, including in particular that made by the National Association of Real Estate Boards for the Internal Revenue Service.
Original cost value of all nonfarm dwellings is estimated by cumulating
annual estimates of original and subsequent construction outlays on units
still existing. A basic distribution of the stock of housing by year of construction was taken from the 1940 Census of Population and Housing. Unit mean
original cost of dwellings in each age group was estimated chiefly from the
building permit data described in the section on New construction, and was
multiplied by the number of units found by the Census in that age group to
obtain a census-year aggregate. The corresponding aggregates for noncensal
years are derived by adding subsequent-year construction outlays to (or subtracting previous-year outlays from) the benchmark.
The depreciation totals have been allocated between tenure classes in
proportion to assessed value rather than original cost, in order to allow for a
somewhat lower depreciation rate on owner-occupied than on tenant-occupied units. (The assessment ratios have been lower for owner-occupied than
for rental units.) To estimate assessed values for the years 1929—41, mean
rents or rental values were capitalized by use of a regression derived from
the 1940 Census of Population and Housing, and multiplied by assessment
ratios indicated in State and local government fiscal reports. The resulting
averages were then multiplied by the numbers of units involved. The value
aggregates so derived have been extrapolated forward from 1941 by originalcost values. These are calculated by cumulating construction outlays as described above, but separately for one-family and multifamily units and with
an adjustment for the shifting of units among tenure classes.
8. Taxes are estimated as a variable fraction of the property tax series
described in the section on Government receipts and expenditures. This tax
series is allocated among nonresidential property, rental housing, and
owner-occupied dwellings.
For the period from 1929 through 1941, the allocation was derived from a
detailed study of State and local government fiscal reports. For 1949, average
taxes per owner-type unit and per rental unit were based on frequency distributions from the Survey of Residential Financing taken in conjunction
with the 1950 Census. These 1949 benchmark means were consistent with
figures derived from the 1950 Survey of Consumer Expenditures, applying
to owner-occupied units, and with figures for rental units in a few cities surveyed by the Office of the Housing Expediter. They were multiplied by the
corresponding numbers of units to obtain benchmark totals. Interpolators
for 1942-48 and extrapolators from 1949 have been derived by allocation of
the overall property tax series: between nonfarm residential and other realty
by reference to 1941 taxable values as changed by new construction, depreciation, and trends in assessment practice; and within nonfarm residential
by aggregate assessed values.



89

9. Mortgage interest for 1950 was estimated primarily on the basis of the
Survey of Residential Financing mentioned above. Extrapolation utilizes the
products of estimated mortgage debt multiplied by estimated effective interest rates.
For 1950, amounts of debt and of interest liability for mortgage loans on
rental units and on owner units separately are estimated from distributions
of debt by interest rate classes reported in the Survey of Residential Financing. The Survey totals for debt held by savings and loan associations were
reconciled with similar totals based on lenders' reports to the Federal Savings
and Loan Insurance Corporation. The details of the reconciliation provided
a basis for adjusting the Survey debt aggregates to the timing, coverage, and
concepts appropriate for the present purpose. Together these adjustments
added about 9 percent to the aggregate debt figures reported in the Survey.
The Census of Population and Housing provided the basis for estimates of
owner-occupants' total mortgage debt and average interest rate paid in
1940. Interpolation (between 1940 and 1950) and extrapolation utilize the
Federal Savings and Loan Insurance Corporation's estimates of debt on oneto-four family properties, its tabulations of mortgage loans and interest income
of building and loan associations, and a variety of sample data. Census data
for 1940 corresponding to those used for owner-occupied housing were not
compiled for rental units. Debt and interest rates on mortgage loans secured
by rental housing are extrapolated from the 1950 benchmark by use of
the building and loan association data and corporate and other sample
information.
10. Other property expense consists of estimates for repair and maintenance, insurance, and miscellaneous costs.
Repair and maintenance cost accounts currently for two-thirds of the sum
of these three components. For each of a large number of regional, rentlevel, and city-size cells, benchmark averages derived from surveys by the
Office of Price Administration, from the 1935—36 Study of Consumer Purchases and from a similar but much smaller survey made in 1941 were multiplied by 1930 and 1940 census data on number of dwellings by tenure in each
cell to obtain cell subtotals. These, summed by tenure groups, represent
benchmark aggregates for repair and maintenance cost in 1930 and 1940.
For owner units, the 1950 Survey of Consumer Expenditures supplied the
basis for an additional benchmark. State sales tax and sample data on sales
of building material retailers and (since 1947) tabulations from the 3,000family Surveys of Consumer Finances made for the Federal Reserve Board
are used for interpolation and extrapolation.
Expenditures on rental units are extrapolated from 1940 through 1946 and
estimated for 1950 by use of survey data provided by Federal rent control
agencies. Estimates for other years have been interpolated or extrapolated
by reference to space rent and to average family income.
The foregoing estimates of facility, utility, and other expenses incurred on
rental housing are deducted from the estimated rentals of such dwellings to
obtain the figure for net dwelling rents of all landlords.
11. Rental housing, however, is owned by business enterprises and government agencies as well as by persons. In the absence of any direct information,
contract rentals and expenses on holdings of persons are derived by eliminating the business- and government-owned portions from the totals. In line
11 of Exhibit 3 only the net adjustment is shown.
For government, the estimates of rental receipts are derived chiefly from
reports and records of the Public Housing Administration.
Business dwelling-rent receipts are calculated as the sum of (a) real estate
corporation apartment building rentals; (b) other dwelling rentals received
by real estate corporations, and dwelling rentals of insurance companies; (c)
dwelling-rent receipts of corporations in banking, brokerage, and finance,
n. e. c ; (d) for all other corporations, an arbitrary five percent of their total
rent receipts as reported for income tax purposes; and («) for each of these
groups of industries, the same percentage of noncorporate as of corporate
total rent receipts. A special tabulation of about 500 returns selected at
random from the Survey of Residential Financing lent some support to the
total so derived, by confirming the indicated relationship of business to nonbusiness rentals from housing.
(a) For corporation-owned apartments, the base-year estimate was obtained by a reconciliation of total rent receipts shown on real estate corporation income tax returns for 1937, which excluded apartment buildings,
with those shown on returns for 1938, which included them. This estimate
was extrapolated over the 1929—41 period by total rent on multifamily dwellings. For 1942 and later years, it has been extrapolated in combination with

90

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

the remainder of real estate corporation dwelling rentals (see (£)), as a variable percentage of total rental receipts of all real estate corporations filing
income tax returns. The percentage distribution of such corporations' rentals
between dwelling and nondwelling properties has been moved forward from
the 1941 benchmark by use of data reported in Moody's Manuals of Finance,
Insurance and Real Estate for corporate lessors of each type of property.
(b) The estimates of insurance company dwelling rentals and, for 1929-41
only, those of nonapartment dwelling rentals of real estate corporations utilize
tax return data to determine total rent receipts, which are distributed between residential and nonresidential by use of reported figures on value of
life insurance companies' investment holdings of dwelling and nondwelling
real estate.
(c) An analogous method is applied for banking, brokerage, and finance,
n. e. c , using reported data on types of bank-owned realty. Part (d) requires
no comment, and the noncorporate totals required for (e) are benchmarked
on Federal tax return data for sole proprietorships and partnerships and
extrapolated by the corporate series for the respective industries.
The resulting series for dwelling rental receipts of business is used in two
connections—as a deduction from total nonfarm dwelling rents in arriving at
dwelling rental receipts of persons, and as the basis for the estimates of
business firms' expense on rental housing, which are deducted from total
nonfarm dwelling landlords' expenses in arriving at dwelling rental expenses
of persons. For the latter purpose, the business receipts series is multiplied by
suitable corporate expense ratios. Separate ratios are calculated for interest,
depreciation, and the sum of all other expenses.
These expense ratios, and the corresponding ratios mentioned below for
the nondwelling properties of realty corporations, are based on a special
analysis of the corporate data, as follows.
Tabulations from the income tax returns of corporate owner-operators and
lessors of real property permit estimation of these corporations' rental income and associated expenses, but for dwelling and nondwelling properties
combined rather than separately. To distribute their rental income by property type, the relevant estimates described under (a) and (b) above are taken
as measuring the residential portion; and the difference between these and
the tax-return-based total is taken as a measure of the nondwelling portion.
Detailed expense ratios are derived, for dwelling properties and for nondwelling properties separately, from annual income statements appearing in
Moody's Manuals for a sample of several hundred concerns specializing in
one or another of the two types. These ratios are applied to the respective
rental receipts series to obtain estimates, for dwelling and for nondwelling
properties separately, of the dollar amounts of each expense item incurred
by realty corporations. Finally, the expense estimates so obtained are adjusted pro rata to the corresponding item totals from the tax return data for
both sorts of property combined. Expense ratios implied by these final adjusted estimates are attributed to business-enterprise lessors generally.
Expenses on public housing, like rental receipts, are estimated from fiscal
reports and accounting records.

Net royalty receipts
These are estimated annually by use of Federal income tax return data, on
the assumptions that (1) the ratio of royalty receipts to nonfarm nonresidential rental receipts is the same for persons as for corporations, and (2) the
ratio of net to gross royalty income is the same for persons as for corporations
specializing in royalty-yielding investments. Preliminary estimates for the
two most recent years, pending the availability of the tax return tabulations,
are made by extrapolation using a suitably weighted index of mineral production and an index of the net/gross ratio for nonfarm business and industrial
property.
Since the basic use made of these estimates of net royalties is in the allocation and extrapolation of the net rents and royalties total derived from the
individual income tax returns, any error in the 1941 level is balanced by an
opposite error in the benchmark level of persons' net rent from business and
industrial property. The net effect of any such error is therefore to mis-weight
the extrapolator series.

Rent on business and industrial property
In estimating net rental income of persons from nonfarm business and
industrial property, individual income tax return data for 1941 were adjusted



for coverage and concept and to eliminate net rent from nonfarm dwellings
and net royalties, as shown in Exhibit 2 above. The resulting benchmark has
been extrapolated as the difference between annual estimates of gross rental
receipts and of expenses. The calculations involved are illustrated in Exhibit 4,
Exhibit 4.—Persons' Net Rents from Business and Industrial Property,
and 1950

1941

[Millions of dollars]
Item

1941

1950

Rent on nonfarm business and industrial property:
3,681
1,360

4. Less: P E R S O N S ' E X P E N S E S as landlords of nonfarm business and industrial property
5. Equals: PERSONS' N E T R E N T S from such property-.

1

6,907
2,654

2,321

1. Paid by business and government
2. Less: Received by business and government-.
3. Equals: GROSS R E N T A L R E C E I P T S OF P E R S O N S from
nonfarm business and industrial property.
-

4,253

1,831

2,606

s 490 i

1,647

1. Derived as line 3 less line 5, in 1941 only.
2. Benchmark estimate based on data from Federal individual income tax returns. See
Exhibit *.

Gross rental receipts of persons from nonfarm business and industrial
property (line 3) are estimated by a residual method: business enterprise and
government rent receipts other than from dwellings and farm property (line
2) are subtracted from business enterprise and government rent payments
(line 1). Of the series used in this calculation, total business enterprise rent
receipts and payments are estimated primarily from Federal income tax
return data for corporations, partnerships, and sole proprietorships. Total
government rent receipts and payments are estimated from the Federal
Budget and fiscal reports of sample States and municipalities. Dwelling rent
receipts by government and business are derived as indicated above (in the
discussion of line 11 of Exhibit 3) and netted out of their total receipts. Rents
on farm property, likewise netted out, are estimated partly from government
fiscal records and partly by use of percentages from a 1945 study of a sample
of 150,000 farm owners.
Persons' expenses (line 4) were estimated for 1941 as persons' gross receipts
(line 3 obtained as described above) less their net rents (line 5 derived from
tax data as shown in Exhibit 2). This benchmark is broken down by categories
of expense, which are extrapolated by the corresponding expenses of corporate lessors of such property, adjusted in four steps as outlined below. The
estimates for persons' rental housing and for realty corporations' dwelling
and nondwelling property, which are introduced at various stages of the
procedure, have already been described. The sum of the extrapolated expenses is then deducted from persons' gross rental receipts to yield annual
net rental income of persons from nonfarm business and industrial property.
(1) To allow for the difference in expense patterns between personal and
corporate landlords, the itemized estimates of expenses for corporations
handling business and industrial property are multiplied for each year by
the ratios of persons' to corporations' expense for the same items in connection with rental housing.
(2) The resulting values were summated for 1941, and adjusted pro rata
to the global benchmark total for persons' expenses on business and industrial
property shown in Exhibit 2, line 4. The adjusted amounts were taken as
representing a breakdown of this global benchmark total.
To extrapolate them to other years, the unadjusted series obtained in the
first step are (3) adjusted for all years by the same percentage found necessary
for 1941 in the second step and (4) corrected for the difference between the
movements of personal and corporate rental receipts.
Adjustment (4) is necessary because the original series tended to vary
directly over time with the ratio of personal to corporate rental receipts from
housing, and did not reflect variations in the ratio of personal to corporate
rental receipts from business and industrial property. Each expense series
is accordingly divided by the former ratio series, based on 1941 as 100, and
multiplied by the latter, similarly based.
Tax return data become available at intervals of varying length for the
various items used in estimating net rent on business and industrial property.
Annual tabulations from corporate income tax returns are completed only
after a lag of 3 years; and periodic tabulations from individual and partnership tax returns must be extrapolated over periods ranging up to 4 or 5 years.

91

NATIONAL INCOME, 195 4 EDITION
The extrapolations necessitated by these lags have generally been based for
corporations upon samples of published corporate financial statements and
upon indexes of employment, payrolls, or industrial production; and for
noncorporate business, upon the rent series for corporations in the same
industries.

2. Owner-Occupied Nonfarm Dwellings
Imputed income derived from owner-occupancy of nonfarm dwellings
made up nearly two-fifths of total net rental income of persons in 1950, and
a somewhat smaller fraction in most of the earlier years. This imputed income
item is defined as the gross return which the owner-occupants of nonfarm
dwellings could theoretically have realized had they offered their houses for
rent, less their expenses. The implied option is considered from the standpoint of the individual owner-occupant, but without regard to any special
treatment accorded under rent control legislation to dwellings previously
owner-occupied.
In general, the methods used to estimate the imputed item of rental income
as well as the corresponding consumer expenditure component, Space rental
value of owner-occupied nonfarm dwellings, parallel those discussed above for

rental housing in the use of population census benchmarks, periodic sample
surveys, and the Rent Index. Even the statistical problem of facilities included in rental value occurs with owner-occupied as well as with rented
dwellings, since the 1940 Population and Housing Census, which provides
the sole benchmark for rental value of owner units, reported such value
inclusive of cookstove, refrigerator, and (for multifamily structures) any
utilities included in rent of other dwellings in the same structure.
Accordingly a separate description of the methodology underlying the
imputed rent estimates is not given. Only two points of difference between
imputed and contractual rent estimation need be noted. First, in the case of
owner-occupied dwellings it is unnecessary to allow for ownership by landlords other than persons. Secondly, the estimation of mean contract rental
value per unit is complicated for owner-occupied dwellings both by a relative paucity of data and by the more subjective character of such data as
exist.
The benchmark data for mean rental value of owner-occupied units are
estimates for the individual dwellings reported in the 1940 Census of Population and Housing. Enumerators were instructed to base the estimates on
rents actually being charged for similar dwellings in the neighborhood. In
deriving the benchmark means from these data, a systematic downward adjustment, amounting to about 2 percent, was made to eliminate the apparent
effect of enumerators having occasionally used a rule-of-thumb ratio to
market value in estimating rental value.
A direct estimate was also developed for 1930, based on Population Census
value data for that year used in conjunction with value-rent relationships
indicated by the 1940 Census. Estimates for 1932 and 1933 are extrapolations
from 1930 by data for 22 cities from the Financial Survey of Urban Housing.
They reconcile very closely with the results of extrapolation back from 1940
by series underlying the Rent Index compiled in the Bureau of Labor
Statistics.
Estimates of rental value for 1944 and 1945 were obtained by extrapolation
from 1940, as follows: (a) The mean rental value of all occupied units combined (rented as well as owned) was extrapolated by mean rent for rented
units, which in turn had been estimated from Census Bureau sample data;
(b) the means for all occupied units and for rented units were multiplied
respectively by the corresponding numbers of units, and (c) the products
were differenced to secure aggregate rental value of owner-occupied units.
This procedure was adopted primarily to take account of the effect of the
shifting of units from rental to owner-occupancy.
Mean rental values for 1947 and 1950 were estimated by extrapolation
from 1945, separately for urban and rural nonfarm owner-occupied units,
using the corresponding means for tenant-occupied units as extrapolators.
For other years, compilations made by the Bureau of Labor Statistics for
its Consumer Price Index have served as the chief basis for interpolating
and extrapolating mean rental value. The compilations used for 1935-39
were indexes of rent by rent ranges. Other interpolation estimates are based
in most cases on the published Rent Index.
For all years, the aggregates relating to owner-occupied units have been
increased by small allowances for summer homes and other dwellings temporarily unoccupied but held for the owners' use.
901«Q9°

nd

7




3. Farm Realty
The basic series used in estimating net rent from farm property are prepared by the Department of Agriculture, using the data-collecting system
and some of the specific estimates described in the section on Income of unincorporated enterprises. In conformity with the Department of Agriculture
treatment, all farm net rents received by or imputed to landlords living on
farms are regarded as incident to the business of farming, and hence are
included in national income under the heading of net income of unincorporated (farm) business rather than under the heading of rental income of
persons. The magnitudes involved in the 1950 rent estimates are as follows:
Exhibit 5.—Persons' Rentals from Farm Property,

1950

Item

Millions
of dollars

1. Gross farm rent and Government payments to landlords

2,920

2. Less: Rent payable to farmer landlords, nonfarm business, and government -

1,053

Equals: Gross rental income of persons from farm property
3. Less: Expenses of personal landlords
4. Equals: Net rental income of persons from farm property

1,867
773
1,094

1. Gross rent payable to landlords includes crop share, livestock share,
and cash rents. It is estimated by the Production Economics Research Branch
of the Department of Agriculture from acreage and production statistics,
using relationships from Agriculture Census reports, Crop Reporter data,
and a special survey for 1936 of a sample group of 15,000 reporting landlords.
As used by the National Income Division, the series also includes income from
Government payments made to landlords qua landlords (as distinguished from
payments made to farm operators as such).
2. The distribution of the total by landlord groups is based on data from
the 1936 survey mentioned above and from over 150,000 returns to a 1945
survey. A primary distribution between landlords living on farms and other
landlords is calculated by the Department of Agriculture from these surveys,
and ratios from the 1945 survey are used by the National Income Division to
break out nonfarm business and government tent receipts for that year from
the total going to landlords not living on farms. For other years, the business
and government receipts are extrapolated by government fiscal and corporate tax return data.
3. Landlords' expenses represent those shares of farm overhead and production expense items such as taxes, depreciation, repairs, insurance, fertilizer, and seed (see discussion of farm proprietors in the section on Income of
unincorporated enterprises) which are borne by personal landlords. The total
paid for each item is allocated first between rented and owner-occupied
farms, generally by use of acreage or property value ratios derived from the
quinquennial Census of Agriculture. The major expense items involved for
rented farms can be assumed to be paid by the landlord rather than by the
tenant. Seed, fertilizer, and binder twine, however, illustrate a different
treatment, lessors being estimated to pay a fraction of these determined by
the ratio of share-cropped acreage to total crop acreage. The amounts of expenses ascribed to landlords are prorated between landlords living on and
off farms in proportion to the land owned by each, as indicated by the 1936
and 1945 surveys. Finally, expenses of landlords living off farms are prorated
between persons and nonpersonal landlords by use of the corresponding distribution of rental receipts described above.

Characteristics of the Revisions
As has been indicated above, the rent estimates are based upon a wide
variety of sources, many of which are nonrecurrent, or become available infrequently and irregularly or with substantial delay.
To the extent that the estimates of rents rely on corporate income tax data,
the figures for the two most recent years published each July are subject to
revision. However, revisions may extend further, for instance when new
benchmark data lead to the substitution of interpolations for previous extra-

92

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

polation procedures over a number oi years. Important changes of this type
have resulted from the incorporation of data from the 1950 Census of Population and Housing, the Survey of Residential Financing, and the 1949-50
Dwelling Unit Surveys and Survey of Consumer Expenditures in 1950 by
the Bureau of Labor Statistics.
In addition, Part V of the present supplement incorporates substantial revisions made when preliminary estimates based on a more summary methodology were replaced by the results of the detailed calculations described
above. Further details of the new estimates were published in the June 1953

For a number of industries (particularly in manufacturing), the comparability of the 1929—47 estimates with the estimates for later years is impaired
as a result of changes in the industrial classification of basic data supplied by
the Internal Revenue Service. (See introduction to Part III.) Where quantitatively important discontinuities are involved, their nature and magnitude
are indicated by footnotes to table 18 of Part V of the present report.
In the following pages the base year estimates and their current extrapolations will be discussed in turn.

issue of the SURVEY OP CURRENT BUSINESS.

Base Year Estimates
All of the adjustments made in the tax return data are indicated in table
38, Part V. Exhibit 1 below lists for the year 1950 those which involve no
estimation, the required values being shown separately in the tax return data
or reported elsewhere on a complete basis. Further adjustments which must
5. CORPORATE PROFITS
be made using estimated values are shown in Exhibit 2.
The "Totals as reported" from tax returns are tabulated by the Internal
The basic data underlying the estimates of corporate profits are the annual
Revenue Service as Compiled net profit, Total tax (consisting of Federal intabulations of corporate income tax returns compiled by the Internal Revcome and excess profits taxes), and Dividends paid in cash and assets other
enue Service. The tabulations are available, for the most part, from its
than own stock, respectively. These totals provide the starting points for
annual Statistics of Income—Part 2, or are taken from the "Source Book", un- the estimates described in the present section.
published volumes supplementing Statistics of Income with more detailed information.
These data are sufficiently complete and reliable to overcome many of
Exhibit 1.—Adjustments Not Requiring Estimation in the Derivation of
the difficulties which are inherent in the estimation of profits. Filing of deCorporate Profits Before Taxes, Profits Taxes, and Net Dividend Payments, 1950
tailed returns is mandatory, and the returns are prepared in the knowledge
[Millions of dollars]
that they are likely to be audited. Although the measurement of profits involves many conceptually difficult problems, the imposition of administrative
Profits
Profits
Dividend
Item
rules and regulations during the long time period over which the Internal
taxes
payments
before taxes
Revenue Service reporting system has been in operation has gone far toward
Amounts derived from tax returns:
standardizing accounting practice for corporate income tax reporting.
17,317
Totals as reported
_ .
42,831
11,553
On the other hand, certain important problems still arise in the translation
Adjustments:
of the tax return data into estimates of corporate profits for national income
Domestic dividends received
-2.460
-2,460
-644
-644
Foreign dividends received_
purposes. The data must be adjusted in various ways to secure comparabil+1,709
ity with other entries in the national income and product tables. One of these
Depletion
_. __ _.
..
adjustments, in particular, involves a considerable amount of estimation—
Net capital gain
-1,129
-639
Net gam, sales other than capital assets
.
allowance for the additional profits disclosed by auditing. The auditing
+223
Net loss, sales other than capital assets
process to which returns accounting for the bulk of profits in any year are
Mutual nonlife insurance companies
-38
-16
+10
subjected requires about 15 years to complete, and its ultimate, full effect
-467
Foreign income taxes
_.
in revising the profits figures originally tabulated must be estimated from
reports of its progressive effect on tax liability as the auditing proceeds.
+197
+195
Amounts reported by Federal Reserve System.
+4
There is a time lag before the tax return data become available. This neces17,031
40,148
Totals from tax returns, as adjusted.
8,463
sitates reliance on extrapolation procedures to obtain estimates for the two
most recent years. As is indicated at the end of this section, such extrapolations have at times led to errors of significant size, particularly on an industry
Dividends received are deducted from profits and dividends to obtain unbasis. The situation in this respect has improved, however, due to progress
duplicated totals reflecting income originating in United States corporate
in the current reporting of corporate profits through the joint surveys of the
production; depletion is added to profits since it is not regarded as an element
Securities and Exchange Commission and the Federal Trade Commission.
of capital consumption in the national income and product accounts; and
These surveys have provided improved data on manufacturing since 1947.
capital gains and losses are eliminated from profits as not measuring incomes
The corporate profits estimates are classified industrially on the basis of
arising in current production.
companies, or firms, rather than establishments. As discussed briefly in the
The adjustment for mutual nonlife insurance companies will be commented
Introduction to this Part, this results in non-comparability with the income
upon in connection with a similar adjustment in Exhibit 2 for mutual life
shares that are classified on an establishment basis. The most serious pracinsurance companies. The adjustments for foreign dividends received and
tical limitations in this connection arise in the comparison of payrolls and
foreign income taxes paid also will be discussed under Exhibit 2, in connecprofits in certain industries.
tion with related international adjustments.
Use of the company, as against the establishment, as the unit of classificaFinally, reported data for the Federal Reserve Banks are added, since they
tion leads to an industrial distribution further removed from a product, or
are not included in the basic Internal Revenue Service tabulations. Reserve
activity, basis. This is so because the operations of companies are generally
Bank profits are measured by current net earnings, exclusive of capital gains.
more heterogeneous than those of establishments. Consequently, with comTaxes are measured by payments to the United States Treasury in lieu of
panies assigned to specific industry groups on the basis of their major activity
franchise tax, at a rate of about 90 percent on a base consisting of net earnas measured by receipts, there is more likelihood that movements in the
ings (including capital gains) less dividends paid.
industry series will reflect changes in classification of firms due to marginal
In addition to these adjustments, there are others which are made by use
changes in the composition of their activity, rather than substantive changes
of estimated rather than reported values. These are itemized in Exhibit 2.
in the type of industrial activity indicated by the industry designation. Special cases of shifts of this type resulted from Federal tax legislation affecting
the filing of consolidated and unconsolidated returns. In this respect, the inAudit adjustment
dustry breakdown of corporate profits in the periods 1929-33 and 1942 to
date is not strictly comparable to that for the years 1934-41, when, with cerThe tax return data shown in Exhibit 1 are compiled from the returns
tain exceptions, filing on an unconsolidated basis was required.
before audit. The first adjustment indicated in Exhibit 2 is the addition of the



NATIONAL INCOME, 195 4 EDITION

estimated profits and tax liability disclosed through subsequent audits by the
Internal Revenue Service. It should be noted that not all returns are audited,
and that even in the audited returns there may remain some understatement
of profits.
Furthermore, the national income estimates of the actual results of audit
are subject to error because they must be based on data rather unsatisfactory
for the purpose. In the first place, the auditing process is far from complete
for many of the years for which estimates must be made. Secondly, the results
of audit are reported in terms of the effect on tax liability only, and the effect
on total profits must be estimated from this.
The estimating procedure is carried out in three steps. The first is to determine the effect on tax liability to date of the auditing process for the given
tax year. Increases in tax liability are estimated by cumulating the gross
additional tax assessments made on that year's income by the end of the following June, by the end of the second following June, and so on up to the
latest reported. Since in some cases the audit process leads to a reduction in
tax liabilities, the indicated reductions must be derived and cumulated in
like manner. Their annual total is derived as the sum of refunds, credits,
and abatements.
The second step is to add the expected results of further auditing of returns
for the given tax year. These are estimated by applying experience-based
ratios to the totals obtained in the first step. The case of the tax year 1948 will
illustrate the method. At present (July 1954), data available for the first step
provide measures of the additional taxes assessed and original overassessments
found against 1948 income as a result of about 3 years of work by the auditors. Study of the auditors' rate of progress in auditing returns for earlier tax
years indicates that the fourth year of the process is likely to add about 24
percent to the gross additional assessments made in the first 3 years; the fifth
year is likely to add about 17 percent to the initial 3-year total; and so on.
Cumulating the percentages forward suggests that future auditing of 1948
tax returns may add about 76 percent to the gross additional assessments
already reported due to audit. The estimate obtained in step 1 is accordingly
raised 76 percent. The initial estimate of tax liability reductions is similarly
raised by the application of experience-based ratios. The net balance of additional assessments over reductions is taken as the audit adjustment applicable
to the tax liability total which was originally tabulated for Statistics of Income
from the 1948 tax returns.
The third step is to calculate the additional profits implied by the estimated
net additional tax assessments. The latter total is divided by the statutory
tax rate (representing a marginal effective rate) to derive the corresponding
figure for additional profits.

Exhibit 2.—Adjustments Requiring Estimation in the Derivation of Corporate Profits Before Taxes, Profits Taxes, and Net Dividend Payments,
1950
[Millions of dollars]
Profits
before
taxes

Item
Totals from tax returns, as adjusted per Exhibit 1...

...

Foreign income tax on dividends
Mutual life insurance companies

40,148

17,031

+556

Adjustments:
Profits disclosed by audit
"Rest of the world" adjustment

Profits
taxes

Dividend
payments

+207

+376
—226

8,463

+426
+241

-1,654

-59

+770

+770

39,970

17,829

+77

-120
Final estimates

.

9,207

Besides revising upward the gross total of positive profits, the audit process
affects the all-corporation net total by revising downward the deficits of some
corporations. An allowance for this effect—amounting to about $100 million
a year in recent years—is made by including in the audit adjustment (1) an
estimate of the total of deficits originally reported by companies found upon




93

audit to have had taxable net income, and (2) an equal amount for deficits
reduced but not eliminated. Estimate (1) is based on tabulations by deficit
size class for selected years. Estimate (2) is arbitrarily made equal to (1).
Special complications have occasionally been introduced into the first and
third steps described above, mainly as a result of tax legislation in connection
with World War II and the Korean war. For the first step, adjustments
were made in the basic tabulations to eliminate the effects of recomputing
emergency amortization charges, and the effects of the carrybacks of net
operating loss and of the unused excess profits tax credit. The adjustments
applied to profits and/or taxes for these items are derived independently of
the audit adjustment, as indicated later in this section. Similarly, the effects
of the credits for debt retirement and postwar refund (under the World War
II excess profits tax) were eliminated, since tax liability is measured net of
these credits in the national income accounts.
For the third step, it was generally necessary to utilize separate estimates
of income tax assessments and excess profits tax assessments, and separate
estimates of marginal tax rates.
The results obtained by the auditors during the first three or four years
after the returns are filed are believed to provide a fairly adequate basis for
applying the method that has been outlined. For more recent years, the audit
adjustment must be based largely on judgment.

Adjustments for international flows
The international adjustments listed in Exhibits 1 and 2 are designed collectively to yield: for profits, an aggregate which includes profits and dividends
received from abroad but excludes profits and dividends paid to abroad; for
dividends, an aggregate which includes dividends received from abroad but
excludes dividends paid to abroad; and for taxes, a measure of liability to the
United States Government.
For profits, the tax return data as adjusted in Exhibit 1 exclude (1) the
dividends received by United States stockholders of foreign corporations. On
the other hand, these data include (2) dividends accruing to foreign stockholders of United States corporations and (3) profits to foreign corporations
from their branches in the United States. The "Rest of the world" adjustment added to profits in Exhibit 2 is a correction for all three of these characteristics of the data.
It is calculated as the total inflow of dividends (item 1 above) less the sum
of dividends paid to foreigners (item 2 above) and profits of foreign corporations' United States branches (item 3 above). All of the items in the adjustment are measured net of income taxes levied by the payer country, in accordance with the definitions used in calculating net foreign investment. The same
definitional principle requires that foreign taxes be netted out of the income
earned by United States corporations from their branches abroad; this is not
done in the tax return data and is accomplished by the second international
adjustment to profits shown in Exhibit 2.
For profits taxes, foreign taxes on both branch profits and dividends are
deducted (in Exhibit 1) from reported Federal tax liability, against which
they constitute a credit.
For dividends, the tax return data as adjusted in Exhibit 1 have three inappropriate characteristics. They exclude dividends received by United
States stockholders from foreign corporations. They include dividends accruing to foreign stockholders of United States corporations. Finally, they are
understated by the amount of foreign taxes on United States corporations'
dividend receipts, which have been netted out along with the dividend receipts themselves since the latter are reported (and deducted as in Exhibit 1)
gross of such taxes. In Exhibit 2, the "Rest of the world" adjustment (calculated as the total inflow of dividends less dividends paid to foreigners) and the
adding back of foreign income taxes paid by United States corporations or>
dividend receipts from abroad together correct for these characteristics.
Of the series entering into these international adjustments, foreign income
taxes are measured by the tax credits claimed on the corporate tax returns.
The series composing the rest-of-the-world adjustment are prepared in connection with the balance-of-payments estimates of the Office of Business
Economics. (See section on Net foreign investment.) The distribution of
foreign income taxes between taxes on branch profits and taxes on dividends
is based on the relative proportions of branch profits and dividends earned
from abroad.

94

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Exclusion of mutual insurance companies
Mutual companies are not considered part of the corporate universe for
national income purposes, and it is therefore necessary to remove from the
tax return data the values included for mutual insurance companies. For
mutual nonlife companies, the amounts to be removed are reported separately
by the Internal Revenue Service and are used as shown in Exhibit 7. Tax return data for mutual life insurance companies are tabulated in combination
with the data for stock life insurance companies, and must be isolated statistically to permit the adjustments for them shown in Exhibit 2.
The amount of the adjustment to corporate profits before tax is calculated
by subtracting an estimate of stock life company profits from the total of life
insurance company profits given in the tax return data. The required series
for stock life company profits is estimated as the sum of cash dividends paid
and Federal income tax liability less stock life company receipts of dividends.
The dividend payments and tax liability of life insurance companies as reported to the Internal Revenue Service have generally been applicable to
stock companies only. In the most recent years, reported tax liability has included amounts for mutual life companies as well. Estimates of these have
been excluded from the reported tax liability of life insurance companies in
this adjustment, and are also deducted from total corporate tax liability as
shown in the second column of Exhibit 2. The dividend receipts of the stock
companies are computed as 20 percent of the total dividend receipts reported
on life insurance company tax returns, the ratio being based on a special
tabulation from reports compiled by the Spectator Company for its Insurance
Yearbook.

The adjustment shown in the final column of Exhibit 2 for insurance companies represents the remaining 80 percent of the reported dividend receipts
of life insurance companies. These are ascribed to mutual life carriers, and
added back to net dividends to cancel the effect of their inclusion in the corporate dividend receipts previously deducted. The original overall deduction
and the corresponding adding back of dividend receipts of mutual nonlife insurance companies are shown in Exhibit 1.

Adjustments not applicable to 1950
A number of adjustments not applied for 1950 had to be made in the Internal Revenue Service data to derive profit estimates for other years. (See
table 38.) These included corrections for gross renegotiation refunds, emergency amortization acceleration, war losses, and the unjust enrichment and
Vinson Act excess profits taxes.
Renegotiation of war contracts led to refunds by the contracting corporations which reduced their profits and their tax liability from the amounts
originally reported, mainly for the war years. The totals tabulated from the
tax returns originally filed were adjusted accordingly, by use of special tabulations published in Statistics of Income for the years in which the item was
quantitatively important.
When World War II ended and emergency amortization charges on defense-connected facilities were recomputed retroactively (see section on Capital consumption allowances), both profits and tax liability for the war years
were reduced by comparison with the amounts originally reported. The effect
on tax liability for each of these years was estimated by use of data and procedures similar to those described above for carryback tax refunds. To determine the corresponding effect on depreciation charges and consequently
on before-tax profits, the annual adjustments in tax liability were divided by
the marginal tax rates as in the third step of the audit adjustment.
The tax return data on profits, for 1942 particularly, were net of deductions claimed by some corporations for losses due to enemy capture of properties they owned abroad. A minimum allowance for such losses was estimated
from the published income statements of United States corporations believed
to have such foreign holdings. This allowance was added back to profits as
reported in the tax return data, to bring the treatment of the losses involved
into conformity with the treatment of capital losses generally.
The unjust enrichment and Vinson Act excess profits tax collections reported by the Treasury Department were allocated to the years in which the
liabilities were estimated to have accrued, and were added to the profits tax
series for those years.

Industry breakdown of estimates
State income taxes
State income taxes are among the deductions made by the taxpayers in
arriving at the profits figures shown in the Internal Revenue Service tabulations, but are not shown separately from other taxes in these tabulations and
must therefore be ascertained from other source materials. For 1937 and subsequent years use has been made of statistics compiled by the Governments
Division of the Census Bureau on collections reported by the various State
governments under their corporation income tax laws. The total so reported
for each year has been taken as a measure of the tax liability incurred in the
preceding year, to obtain the amount of the adjustment illustrated in Exhibit
2. For the years prior to 1936, the adjustment was estimated from a sample
including States which in 1939—42 assessed about half of the national total of
such taxes. The raising ratio applied to the sample data was varied annually
to allow for the adoption of corporation income taxes by additional States.

Carryback tax refunds
These have been occasioned by statutory provisions allowing unused excess
profits tax credits and net operating losses for a current year to be carried back
and applied in recomputing the tax liability of a previous year or years.
The adjustment of tax liability to allow for reduction from this cause represents the reported amount of "tentative adjustments" claimed by the taxpayers, plus an estimate of the additional carryback refunds found by the auditors to be due. (As noted above, these refunds are not considered in making
the regular audit adjustment.) The latter estimate is based on Internal
Revenue Service tabulations for selected years, and is derived by a formula
similar to the first two steps of the audit adjustment described above. The
effect of auditors' disallowance of tentative claims is reflected in the regular
audit adjustment.
For the latest years covered by the tax return tabulations, the total of
"tentative adjustments" claimed to date provides the basis for preliminary
estimates of the carryback adjustment.



In the derivation of corporate profits, taxes, and dividends by industry,
such of the adjustments in Exhibits 1 and 2 as are applicable to more than one
industry are made separately to the respective Statistics of Income industry
benchmarks. The corrections illustrated in Exhibit 7 are generally available
in the required industry detail, with the exception of the item for foreign
income taxes which appears also in Exhibit 2. For those in Exhibit 2, the
relevant all-industry totals are distributed as follows: the audit adjustments,
proportionately to income and excess profits taxes as originally reported;
foreign income taxes, proportionately to foreign dividends and branch profits
received; and State income taxes, proportionately to Federal corporate income taxes.
Further adjustments are made in the industry estimates so derived to improve their industrial comparability. They deal with changes made by the
Internal Revenue Service in its classification system from time to time (mainly
in 1938), and, less important, with categories of firms whose precise industrial
attachment is not determined by the Internal Revenue Service.
Changes in the classification system cause minor industries previously tabulated under a major industry heading to be transferred to another heading.
Generally the resulting discontinuity is eliminated by transferring the prioryear values for the minor industry to the new heading. If the minor industry
values to be transferred are not reported separately for all years, those for the
missing years are estimated by extrapolation using the broader industry series
which contain them. Such statistical procedures, however, could not effectively be applied to correct the noncomparability introduced into several of
the manufacturing series by adoption of the new classification code in 1948.
To eliminate the "not allocable" categories in the industrial breakdown of
the tax return data, the values in these categories are generally prorated by
the corresponding values in the industries among which the "not allocable"
totals are to be distributed.
With the principal exception of industry series affected by the 1948 industrial classification change, formal comparability over time in the individual
industry estimates has thus been secured by reasonably satisfactory methods.
The more fundamental difficulties affecting comparison, for which no adjustment can be made, have been noted earlier.

NATIONAL

INCOME,

1954

EDITION

95

The measures of profits ("net income before Federal income taxes"),
Federal tax liability, and dividends in this survey are based on definitions
As indicated above, the tax return tabulations do not become available
similar to those used by the Internal Revenue Service except for the effects of
until more than 2 years after the year to which they refer. It is accordingly
(1) consolidated reporting and (2) more liberal rules in the survey for the
necessary to use extrapolation procedures to obtain the estimates of profits,
expensing of current additions to reserves. The chief variations between the
taxes, and dividends for the two most recent years. Extrapolators are applied
national income definitions and those used for the survey are therefore due to
to industrial benchmark estimates uncorrected for audit results, and an adthe items listed in table 38. The impossibility of adjusting for any of the
justment for audit results is then introduced into the resulting estimates.
latter (apart from the expected results of audit) inevitably introduces some
error into the estimates.
Although reporting corporations are classified by industry each year, the
Exhibit 3.—Components of Current Corporate Profits Estimates, by Basis
industry breakdown has some shortcomings, chiefly because of the general
for Extrapolation, 1950
use of consolidated returns for affiliated corporations.
In spite of these limitations, the Financial Report series provides extrapolaProfits before taxes
tors for these manufacturing industries substantially more reliable than those
Data used in extrapolation
Industry
(compiled from privately published financial statements as described under
Millions Percent
of dollars of total
3 below) used prior to the initiation of the Financial Report series in 1947.
2. Banking, transportation, and communications and public utility corAll industries, total, excluding Rest-of38,970
100
porations accounting for the bulk of the profits in their industries are subject
the-world.
1. Manufacturing, except newspaper Federal Trade Commission
23,025
69
to Federal regulation, and submit regular financial statements to the respecpublishing.
and Securities and ExchangeCommission, Quartive regulatory agencies—the Federal Deposit Insurance Corporation, the
tet ly Financial Repoit for
Interstate Commerce Commission and the Civil Aeronautics Board, the FedUnited States Manufacturing Corporations,
eral Communications Commission, and the Federal Power Commission. The
2. Banking; railroads, highway passen- Data reported to Federal
6,463
14
ger transportation, highway freight
regulatory agencies.
coverage of the data used in extrapolation for most of these industries is
transportation and warehousing,
consequently very good. On the other hand, neither the industry classificaair transportation, pipeline transportation; telephone, telegraph and
tion scheme nor the report forms can be made to match exactly the definirelated services; radio broadcasting
tions used in national income estimates; in some cases, there are important
and television; utilities: electric
and gas.
differences in concept or coverage. These discrepancies affect the level and
8. Mining; newspaper publishing; retail Tabulations of sample data
5,645
16
movement of the extrapolator series, and result in errors in the estimates
trade; real estate; hotels and other
from n o n g o v e r n m e n t
based on them.
lodging places, motion pictures.
sources.
3. For a number of industries covered neither by the Financial Report
4. Agriculture, forestry, and fisheries; Sales data and similar indi4.837
12
contract construction; wholesale
rect evidence.
series nor by reports to regulatory agencies, the estimates are extrapolated
trade; finance, excluding banking
and real estate; water transportawith the aid of sample data compiled from published financial reports of
tion, services allied to transporindividual companies. These samples include substantially every domestic
tation; local utilities and public
services, n. e. c ; services other
nonmanufacturing corporation for which the requisite data are published in
than motion pictures and hotels.
Moody's Manual of Industrial Securities. Nevertheless, their coverage is less
adequate in general than that of the data obtained from regulatory agencies^
and they fall short in coverage also of the manufacturing sample. Moreover,
The source materials available for the construction of extrapolator series
they are much less representative than the latter, since in general only large
may be classified into four categories. These are listed in Exhibit 3 together
corporations can be included. Their definitional comparability with the
with the industries for which they are used. The results of the individual
benchmark estimates is subject to the same general limitations as the extrapoindustry extrapolations are checked by, and occasionally modified in the
lators described under (1) and (2) above. Capital gains and losses, as well as
light of, independent estimates of total corporate profits derived from tax
charges to special reserves not allowable for tax return purposes, are elimcollection data. The estimates obtained by extrapolation and those based
inated to the extent permitted by the published details, in accordance with
on the tax data will be discussed in turn.
national income definitions. The results are tabulated and weighted in as
Estimates for the Rest-of-the-world industry, which is not included in
fine an industrial breakdown as the sample reports and the basic tax return
Exhibit 3, are discussed in the section on Net foreign investment.
tabulations allow.
Retail trade is by far the most important industry in this group, accounting
for about 60 percent of the profits listed in line 3 of Exhibit 3. Sample data on
Industry extrapolations of corporate profits
profits, dividends, and taxes are compiled for nine lines of trade (general mer1. The Quarterly Financial Report is by far the most important source noted chandise, food, apparel, furniture, autos and trucks, auto accessories, filling
in Exhibit 3, being used for manufacturing industries which accounted for
stations, drugstores, and restaurants) which together account for about fourmore than one-half the all-industry total of corporate profits in 1950. It is
fifths of total retail profits. For three other groups (auto repair shops; hardbased on regular reports from about 9,000 corporations in a sample drawn
ware, building materials, fuel and ice dealers; and a miscellaneous category),
primarily from among companies which filed 1949 Federal income tax rethe estimates are extrapolated by reference to sales series, with rough allowturns. The sample is stratified, and the sample data are expanded, in terms
ances for changes in profit-sales ratios.
of asset-size classes within each industry.
Published financial reports are available for about 200 corporations, which
Estimates for all manufacturing corporations with securities listed on an
together earned approximately 40 percent of total retail corporate profits in
exchange, and certain other registered concerns, are prepared in the Securi1950. Sample survey data compiled by various trade groups now provide
ties and Exchange Commission from reports filed by nearly all of such comsubstantial additional coverage for several lines of trade including general
panies. To adjust for any current nonreporting, these companies are crossmerchandise, apparel, furniture, and auto and truck dealers.
classified by industry and asset size, and the reported figures for each cell are
The coverage of the sample data varies widely among the nine lines of retail
multiplied by the ratio of the preceding-year assets of all registered corporatrade listed, depending in part on the prevalence of large corporations in
tions in the cell to the assets of reporting corporations in that cell. Estimates
each line. In general, benchmark profits for each line are extrapolated by
for unregistered manufacturing corporations are prepared in the Federal
the product of sample-based profit-sales ratios and sales indexes based on
Trade Commission, by multiplying the data for reporting corporations in each
estimated retail sales in each line of trade. (See section on Personal consumpasset-size class in each industry in the sample by the ratio of the estimated
tion expenditures for commodities.)
total number of unregistered corporations currently in that cell to the total
Adequate sample data on profits are not available for the other three trade
number reporting. A sample of nonrespondents for each quarter is queried
groups. Sales estimates are obtained, separately for each group, by extrapoto determine the effect of bias introduced in the statistics because of nonlating benchmark sales, as tabulated in the corporation income tax returns,
response.
by the corresponding components of total retail sales. Profit-sales ratios are

Recent Year Estimates




96

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

obtained by moving the benchmark ratios by reference to the movement of
sample-based ratios in other lines of retail trade, and on the basis of available
limited information from trade sources.
It may be noted that the inclusion of these residual parts of retail trade in
line 3 of Exhibit 3 is somewhat arbitrary and dictated largely by expositional
convenience. The methodology is very similar to that applied to wholesale
trade, which is listed in line 4 and discussed below.
Retail (and wholesale) trade profits estimates for the year 1952 shown in
Part V were extrapolated from 1951 with the aid of sample series, unfortunately not continued in 1953, from the Quarterly Financial Report.
Other major components of the total in line 3 of Exhibit 3 are mining and
real estate, which account respectively for about 25 percent and 10 percent
of this total. The sample used in extrapolating each includes more than 100
companies. Both samples are unavoidably biased toward corporations
whose stock is not closely held.
Similar comments apply to the sample for newspaper publishing and in
lesser degree to that for motion pictures. Tabulations by the accounting firm
of Horwath and Horwath, based on reports from about 100 hoteb in more
than 50 cities, provide the indicators used for the hotel industry.
4. The final category of industries distinguished in Exhibit 3 consists of
those for which little or no current data on profits are available. For these
industries, base year profits estimates are extrapolated by tenuous procedures
involving, in general, indicators of total sales adjusted to allow for probable
changes in profit ratios.
Wholesale trade is by far the most important of these industries, accounting
for about three-fifths of the total in 1950. The basic estimate of corporate
sales is extrapolated by the sales of merchant wholesalers. The sources and
methods for estimating this sales series are the same as those for wholesale
inventories, summarized in the section dealing with the change in business
inventories. A basic ratio of profits to sales for corporate wholesalers is derived
from the income tax return data and extrapolated by reference to various
indicators including the corresponding sample-based ratio series for retail
trade. The resulting ratio series is applied to estimated corporate wholesalers'
sales to derive corporate profits.
For other industries in this group sales data are in general less adequate
than for wholesale trade and use is sometimes made of other indicators of
gross business volume such as payrolls. Applicable profit ratios are estimated
in principle by means of regressions of profits against these indicators based
on past experience.

Industry extrapolations of profits taxes and net dividends

than that of the corresponding profits series, because many corporations
report their dividends but not their profits.

All-industry estimate based on tax collections
The profits estimates obtained for the two latest years through industryby-industry extrapolation as described above are summed to obtain a tentative aggregate for all industries combined. This aggregate is checked, and
in some cases modified, by reference to an independent estimate of total
corporate profits based on collections of current Federal corporation income
taxes. These data are available with little time lag. The procedure followed
in deriving this estimate may be summarized in terms of five steps.
The first step is to estimate tax liability for the given year from tax collections for the following year, by use of an experience-based ratio. Liability for
the most recent past year must be estimated from tax collections made in
the current year, which are still incomplete. Collections reported to date in
the current year are raised to a full-year basis by use of past patterns.
The second step is to divide the estimated tax liability by an effective tax
rate, to obtain a figure for taxable net income of all corporations having net
income. The effective tax rate is determined from tax return data for the
recent years for which such data are available, and is projected forward with
adjustments to take account of changes in statutory tax rates.
The third step is to estimate and to deduct the total net deficit of corporations having no net income, since for national income purposes corporate
profits are calculated net of corresponding losses. The net deficit is estimated
by projection from recent years for which tax return data are available, its
movement being determined in the light of the total obtained in step (2) and
the relationships between the two series shown in the tax return data for
earlier years.
Taxable income of all corporations, as derived in step (3), is next increased
by estimates of tax-exempt interest income, the amount of prior-year operating loss carried forward, and the credit for dividend receipts. These three
items together represent the difference between taxable income and the
current Internal Revenue Service definition of "compiled net profit".
The interest item is estimated by projecting the trend shown in the tax
return data for recent years. The allowance for loss carried forward is likewise projected from tax return data, its year-to-year movement being based
on that of estimated (or reported) prior-year losses appropriately lagged.
The dividend receipts credit is extrapolated from the latest tax-return-based
estimate by the net dividend component of national income.
The final step is to proceed from the estimate of compiled net profit, obtained in step (4), to an estimate of corporate profits as defined for national
income purposes. The adjustments required are those indicated in Exhibits
7 and 2 (and in Part V, table 38). Some of them must be made on a more
or less arbitrary basis, in the absence of current data.
The procedure outlined is subject to significant error at several stages:
in the ratio of part-year to full-year tax collections; in the ratio of tax liabilities to collections; in the ratio of taxable profits to tax liabilities; in the
magnitude of corporate deficits; and in the adjustment items (notably for
capital gains and losses) between taxable profits net of deficits and profits as
defined for national income purposes. Consequently, a rough range of overall error is calculated. If the sum of the industry estimates falls outside this
range, their detail is reexamined and adjustments are made in the direction
indicated by the tax data.

For the industries listed in lines 1, 2, and 3 of Exhibit 3, the latest base-year
data on corporate profits tax liabilities are extrapolated in general by the
movement of series on "provision for Federal income taxes" taken from the
same sources. In instances in which such data are not available, or in which
they are subject to erratic movements, effective tax rates are estimated by
reference to base-year data with allowance for changes in tax rates. These
estimated rates are then applied to the estimates of profits before tax described
above. This latter procedure is used also for the industries listed in line 4.
It may be noted that the tax provision for carryback of operating losses
(and formerly of unused excess profits tax credit) prevents the determination
of actual corporate tax liability until the following year's operating experience
has been established. In the meantime, it is possible only to base adjustments
on the apparent trend of compiled net deficits.
Characteristics of the Revisions
Net dividends (dividends paid minus dividends received) are estimated for
each industry by extrapolating base-year figures by total dividends paid.
Since the estimates published three years after the event are based upon
Dividends paid by the industries listed in line 1 of Exhibit 3 are shown in the
much more nearly complete information (from the tax return data) than are
same source used for the extrapolation of profits before tax and taxes.
the initial preliminary estimates published with a lag of less than a year,
Dividends paid by industries listed in lines 2, 3, and 4 are extrapolated by
comparisons between the two cast some light on the reliability of the latter.
the corresponding industry components of the National Income Division's
Such comparisons are shown in the first four lines of Exhibit 4.
series on publicly reported dividend payments, compiled monthly from data
The preliminary estimates shown differed by a maximum of about 4 perin Moody's Dividend Record and published regularly in the Survey of Currentcent from those which replaced them when the tax return data became availBusiness. For these industries, the coverage of this dividend series is broader able.




NATIONAL INCOME, 195 4 EDITION

97

for large areas of the economy. Corporate interest transactions are covered in
income tax return tabulations published by the Internal Revenue Service in
its Statistics of Income—Part 2, and those of mutual financial institutions in reports of regulatory agencies or in other comprehensive summaries of basic
data. Information on government interest transactions is also, in large part,
comprehensive.
Personal landlords, unincorporated enterprises (farm and nonfarm), households and institutions, and the rest-of-the-world sector are the major remainExhibit 4.—Preliminary and Revised Estimates of Corporate Profits Earned
ing groups whose interest transactions must be taken into account in calcuin 1948-50
lating interestflowsfor national income purposes. The quality of the informa[Billions of dollars
tion with respect to them varies widely. In general, the estimates must rely on
distinctly less adequate and regular sources. The largest single item in this
Estimates for year—
group is mortgage interest paid by personal landlords.
Published in the Survey
Although the interest flows in national income measure amounts accruing
of Current Business
1950
to United States persons and government, these accruals cannot be estimated
directly, because of lack of information from individual recipients. Instead
July 19»_
2 27.6
they are measured as the payments less the receipts of relevant payer groups.
2 41.4
•28.3
1951 supplement.
27.1
139.6
July 1952
This residual method of estimating has some advantages, mainly because it
27.1
41.0
July 1953
leads to entries consistent with the measurement of corporate profits in the
40.0
1954 supplement.
26.2
income and product accounts. But it is subject to the shortcomings of all
residual estimating procedures, in which small errors in the minuend and
1. Extrapolated forward 1 year from tax return data base.
subtrahend may lead to significant ones in the remainder. The fact that the
2. Extrapolated forward 2 years from tax return data base.
recording of interest by creditors and corresponding debtors may differ both
as to coverage and timing introduces special hazards into the procedure.
special tabulations among the source data for the audit adjustment outlined
The above evaluation refers to years for which Internal Revenue Service
above, as well as associated refinements in the method of estimate.
corporate tax return tabulations are available. For the two most recent years,
In general, the industry extrapolations of corporate profits have been less
for which information is much less adequate, the reliability of the estimates
accurate than the sums of these. That is, errors in the individual industry
is reduced, particularly on an industry basis.
estimates have tended to compensate in the all-industry total.
As will become apparent, the definition of interest flows in the national
For the manufacturing division as a whole, recent year estimates have in most
income is a complex matter, mainly with respect to the imputed interest
instances required only comparatively minor changes when the tax return
flows arising in connection with financial institutions. In this section, the
data became available. Extrapolations for some of the groups within this dioperational procedures used in the determination of theseflowsare set down
vision have proved considerably less accurate for the years that can be checked,
precisely, with but little attempt to explain their basic rationale, which is
but the quality of such extrapolations for 1952 and later years is believed significantly improved. In the joint SEC-FTC Quarterly Financial Report upon which discussed in Part II.
The actual measurement of imputed interestflowsis based, in general, upon
the extrapolations are based, the Federal Trade Commission sample has
data of a high order of reliability, but accuracy is somewhat impaired by the
been strengthened and the technique employed was modified in 1951 to
necessity in some instances (as explained later) of substituting statistical assecure better figures for individual industries.
sumptions where the precise data called for by the definitions are lacking.
Among the industries for which extrapolation is based on data from regulatory agencies, differences of coverage and concept between the national
income estimates of railroad profits and the Interstate Commerce Commission series used as extrapolators have necessitated revisions of significant size.
Relation of Major Interest Flows
Among the nonmanufacturing industry estimates for recent years based on
nongovernmental sample data, those for retail trade have been subjected
The interest component of the national income, "net interest", measures
to percentage revisions somewhat larger than those for the group as a whole.
total interest (monetary and imputed, private and government) accruing to
The extrapolations for wholesale trade, the most important based on maUnited States persons and governments minus total interest paid by United
terials considered definitely inadequate, have likewise been revised by comStates governments. Government interest (Federal and State and local) is
paratively large percentages when the tax return data became available.
deducted because it is not considered income arising in current production.
Revisions in the preliminary estimates of taxes have been similar to those
It is necessary not only to exclude the portion of it paid directly to persons
in the profits estimates, since both are based largely on a common set of
and governments, but also to deduct the portion of it paid to business, because
source materials. Sample data for dividend payments are more adequate
the latter is reflected in the incomes paid out or retained by the business
than for profits or taxes. Revisions in the preliminary estimates of net divisystem.
dends have in general been smaller percentage-wise than the revisions in the
Since accruals to persons cannot be ascertained directly, net interest is obprofits series.
tained by estimating its algebraic equivalent—the difference between (a)
total interest paid by United States business and persons plus total interest
paid to the United States by foreigners, and (b) total interest received by
United States business plus total interest received by foreigners from the
United States.
The interest component of personal income, "personal interest income,"
measures total monetary and imputed interest paid to United States persons.
It is obtained algebraically by adding to the interest component of national
6. INTEREST
income the excess of interest payments by United States governments over
their interest receipts.
This section contains a discussion of the interest components of national
income and personal income, as well as of the three components of personal
The nature of these algebraic relationships can be understood from the
consumption expenditures for services which are estimated in an interrelated
following schematic presentation. Since for the economy as a whole total
statistical procedure. These are Interest on personal debt, the interest element of payments must equal total receipts, then the sum of total interest:
Brokerage charges and interest and investment counselling, a n d Services furnished with(1) Paid by United States business,
out payment by financial intermediaries except life insurance.
(2) Paid by United States persons,
The basic accounting data underlying the interest estimates and the re(3) Paid by foreigners to the United States, and
porting systems by means of which they are summarized are quite satisfactory
(4) Paid by United States governments

Revisions, usually of minor amounts, have been made in the estimates
from time to time after incorporation of the tax return data. These have
occurred primarily as a result of changes, based on progressively accumulating data for the given year, in the estimate of the effects of audit. Sizable
revisions made this year (shown for 1948—50 in the last line of the exhibit)
stem partly from this cause, but reflect chiefly the incorporation of certain




98

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

must equal the sum of total interest:
(5) Received by United States business,
(6) Received by United States persons,
(7) Received by foreigners from the United States, and
(8) Received by United States governments.
By deducting (4), (5), and (7) from both sides of this equation, the receipts
side is converted into (6) + (8) — (4), the initial formula for the interest component of national income as defined above; and the payments side is converted into the equivalent formula by which this component is actually
measured using available data, (1) + (2) + (3) — (5) — (7).

The bulk of all interest stems from a comparatively small number of industries. Exhibit 2 shows these industries, and their respective contributions to the interest share of the national income, for selected years. The
problems of industrial classification and their treatment are substantially the
same for interest as for profits (see preceding section and the Introduction
to this Part). They have little effect, however, on the estimates of interest
shown by broad industry division in Part V or, because of the smallness of
the corporate interest series, on the industrial distribution of the national
income.

Exhibit 1.—Derivation of the Interest Components of National Income and
Personal Income, 1950

Monetary Interest Paid

Millions
of dollars

Item
Monetary interest paid
_
Imputed interest paid
Loss: Monetary interest received
Imputed interest received

9,879
5, 870
8,352
1. 48f>

Equals: Net interest (component of national income)

5,912

Plus: Net interest paid by government (excess of interest payments by United
States governments over interest received by United States governments) ._

4,716

Equals: Personal interest income (component of personal income) _

10,628

The interest component of personal income (6) differs from that of national
income in the formula (6) + (8) — (4) by including (4) and excluding (8). It
can therefore be calculated from the national income share by adding (4)
and deducting (8).
The derivation of the two interest series is shown in summary form in
Exhibit 1, and will be discussed in the same order in the text. The derivation
of the related components of personal consumption expenditures will be explained at appropriate stages of the discussion.
Exhibit 2.—Chief Industrial Sources of thu Interest Component of National
Income, 1929, 1939, 1945, 1950
[Millions of dollars]
1929
Agriculture

1945

1950

833

All other....

455

301

500

822
11

_

Farms.-

448
7

298
3

496
4

_

_

2,381
2,853
-472

597

414

333

480
79

516
81

379
35

286
47

422

Transportation

1,979 J 1,513
2,083
-104

599

Real estate
Allother

2,360
2,407
-47

Finance, insurance, and real estate..

Railroads

1939

350

446

349
73

289
61

311
135

Allother
Communications and public utilities
Utilities: electric and gas

323
71

1,768
-255

576

1,997

1,535
140

734
134

490
86

1,868

Private households

624

283

31

Allother—

577
47

127
156

130
-99

248

6,445

4,604

3,185

5,912

Allother
Services

Other industry divisions
Rest of the world
Allother
Total (Part V, table 23)

1,675

Corporations
Monetary interest paid by corporations accounted for about one-third of
total monetary interest paid in 1950. With the single exception of Federal
Reserve Banks, this item is based on Federal corporation income tax returns,
tabulated by the Internal Revenue Service in Statistics of Income—Part 2.
Where greater industrial detail is needed, it is obtained from the supplementary, unpublished "Source Book."
Adjustment of the tax return aggregate to the corporate universe as defined
for national income purposes involves for all years the deduction of estimates
for mutual insurance carriers and the addition of data for Federal Reserve
Banks. In both cases, adjustments have been negligible for the entire period.
For the latest two years, for which Internal Revenue data are not yet available, total corporate monetary interest paid is obtained by adding separate
estimates for banking and for the total of all industries except banking.
Banking interest is obtained by extrapolating the last base year estimate by
interest payments of insured commercial banks on time and savings deposits,
raised to the all-commercial bank level on the basis of yearly asset ratios
(Federal Deposit Insurance Corporation data).
Estimates for all corporations except those classified in banking are extrapolated by means of an index representing the regression of annual interest
paid against National Income Division estimates of corporate gross long-term
debt plus notes and accounts payable as of the beginning of each year. The
debt figures are based on Statistics of Income balance sheet data, and estimated
for current years by adding increments derived from Securities and Exchange
Commission, Federal Deposit Insurance Corporation, and Interstate Commerce Commission reports.
To obtain a breakdown of the nonbank total, direct estimates are made for
several industries. For farms, the latest Statistics of Incomefigureis extrapolated
on the basis of interest paid by unincorporated enterprises in that industry
(described below). For contract construction, extrapolation is by the value
of new private construction activity (described in the section on New construction). For railroads, telephone and telegraph, and electric and gas utilities, interest paid data (reported, respectively, to the Interstate Commerce,
Federal Communications, and Federal Power Commissions) are used as
extrapolators.
For each individual industry not listed above, the latest base year figure is
extrapolated by the regression-based total for all industries less the industries
for which specific estimates are made.

129
255

..

While the concept of national income calls for measurement of interest
flows on an accrual rather than cash basis, this distinction cannot be maintained in statistical practice. In Exhibit 7 the major components of the interest flows are labeled uniformly on a cash basis—as "paid" or "received."
In the subsequent detailed discussion the terms "paid" and "payable" and
"received" and "receivable" are used to indicate the exact nature of the
flows whenever possible.



The composition of total monetary interest paid in 1950 is shown in Exhibit 3.

Sole proprietorships and partnerships
In 1950, monetary interest paid by sole proprietorships and partnerships
amounted to about 12 percent of total monetary payments. The 12 percent
was almost evenly divided as between farm and nonfarm proprietors.
Farm.—The series for long- and short-term interest payable by farmers
(exclusive of that payable by nonfarm landlords) is obtained from the Department of Agriculture. In general, the computation of these interest charges
is based upon multiplication of the amounts of different types of debt outstanding by relevant interest rates. The long-term debt and interest rates
are estimates by the Agricultural Research Service of the Department of
Agriculture, based on Census of Agriculture benchmarks and sample reports
from lending agencies. The short-term debt to institutional lenders and

NATIONAL INCOME,

interest rates are based on reported information from Federal farm lending
agencies and from commercial banks. The short-term interest payable to
noninstitutional lenders (amounting to $193 millions in 1950) has been
described by the Department as "merely a rough approximation".
Nonfarm.—Total monetary interest paid in this sector is obtained as the
sum of industry estimates. For the years 1929 through 1941, 1939 is the
general benchmark. For many industries this benchmark was obtained by
multiplying estimated total receipts of all proprietorships and partnerships
in the industry (see the section on the Income of unincorporated enterprises)
by the ratio of interest paid to total receipts, taken from 1939 informational
partnership tax returns published by the Internal Revenue Service in the

195 4

EDITION

99

Interest paid by mutual nonlife insurance carriers, which is small, is based
on Internal Revenue Service data.
For savings and loan associations, interest paid and dividends paid to shareholders are together considered interest. The estimates are based on Federal
Home Loan Bank Board reports from member associations raised to cover
nonmembers. Dividends paid by credit unions (assumed to measure the whole
of interest payments on deposit and share accounts) are also estimated on the
basis of reports to official agencies.
Nonprofit organizations servicing business.—These organizations (mutual utility
companies, farmers' cooperatives, etc.) have been required to report their
operations to the Internal Revenue Service. Their combined statements were

Supplement to Statistics of Income—Part 1. In other areas, where total receipts

published in Supplement to Statistics of Income for 1943—Part 2. Reporting was

are not available, a frequently used method was to multiply interest paid per
partner (as calculated from the partnership returns) by the total number of
active proprietors.

incomplete by an indeterminate amount, but the total unreported was undoubtedly small. In general, the method used is to extrapolate the 1943
reported figure over the entire period by other relevant interest series.
Personal landlords {nonfarm).—Monetary interest payable by nonfarm individual property owners (other than professional real estate operators) represented 25 percent of total monetary interest paid in 1950. This component
includes interest payable on mortgages against farm property owned by nonfarm landlords, owner-occupied nonfarm dwellings, and other nonfarm residential and nonresidential property owned by individuals. The farm mortgage interest series is prepared by the Agricultural Marketing Service of
the Department of Agriculture, using sources and methods outlined above.
The two series on nonfarm mortgages are derived in connection with the
estimates of the rental income of persons. Revisions in these series, based on
the new methodology described in the section dealing with the latter estimates, are reflected in the back-year interest totals shown in Part V as compared with those previously published.

Exhibit 3.—Components of Monetary Interest Paid, 1950
Item
Corporations..
Corporations reporting to Internal Revenue Service
Less; Mutual insurance carriers (life and nonlife)- _
Plus: Federal Reserve Banks

Millions
of dollars
3,215

Percent

32.5

3,215
0
0
1,189

Mutual financial institutions.
Savings and loan associations

12.0

620
869

63

3,218

Farm
Nonfarm

32.6

734
367
0
344
23
25
2,459

7.4
3.7
.0

5.8

3.6
2

Households and institutions

Monetary interest paid entered under the heading "Households and institutions" amounted to about 20 percent of the total in 1950. The major share
of this represents nonmortgage interest paid by individual consumers.
1,956
19.8
Consumers.—Nonmortgage interest payable by individual consumers is of
Consumers. . . . ___
1,868
18.9
several types. The first embraces interest on (1) single-payment loans, (2) inNonprofit organizations servicing individuals88
.9
stallment credit held by financial institutions and (3) installment credit held
Rest of the world
301
3.0
by automobile dealers. The amount of each of these classes of credit is esti9,879
100.0
mated by the Federal Reserve Board, on the basis of monthly reports from a
sample of lending agencies and with a periodic adjustment to more comprehensive benchmarks. The portions of (2) and (3) consisting of automobile
sales credit extended to business purchasers are deducted, in line with the
Internal Revenue Service tabulations itemizing interest paid by industry
consumer-business allocation of automobile purchases described below in the
for both partnerships and sole proprietorships in 1945 were used directly to
section on Personal consumption expenditures for commodities, to obtain a
establish nonfarm benchmarks for that year. The coverage of these tabulaseries for personal debt in each category. To derive the corresponding amounts
tions was much higher in 1945 than in 1939. Also, partnership tabulations
of interest, these series are multiplied by the average effective interest rates
for 1947 permitted benchmarks for that sector and fairly reliable extrapolaindicated by fragmentary sample data from lending agencies. The estimates
tion from 1945 of the estimates for sole proprietors.
for 1929-39 are less reliable than those for later years, since the basic series
The general method of extrapolation and interpolation used to obtain inavailable for use in the formula described were somewhat less suitable in
dustry estimates is to employ the corporate interest paid series, adjusting to
detail for this use.
the three benchmark years. In the mining and manufacturing sub-groups,
No interest on charge account debt, service debt, or installment sales credit
value-of-product data for 1929 and 1939 from the Census of Mineral Inheld by dealers other than automobile dealers is included, for reasons of condustries and Census of Manufactures permitted a further adjustment for the
sistency with the data on interest received by these dealers. It is believed that,
changing relative importance of the corporate and noncorporate segments.
in general, business creditors do not enter in their books an explicit interest
item as received on such loans, but rather include the amount in the sales
price or in "other income."
Other private businesses
The second type of consumer interest payments arises in connection with
Mutual financial institutions.—Mutualfinancialintermediaries in 1950 ac- borrowings against life insurance policies. The interest payable to life insurcounted for about 7 percent of total monetary interest paid. For each of the
ance carriers is estimated by applying an average interest rate to the average
components the estimated series is based on reported information from a subamount of policy loans outstanding, both averages being based on data from
stantial portion of the industry and is reliable both as to level and to movethe Institute of Life Insurance. The third class of consumer payment is that
ment.
made to United States Government life insurance and adjusted service certificate funds.
Interest payments by mutual savings banks from 1943 forward are based
on data compiled by the Federal Deposit Insurance Corporation for insured
The total of these three categories is entered in personal consumption exbanks, raised to the universe level by asset ratios for all banks to insured
penditures as Interest on personal debt.
banks in each year. Prior to 1943, the series was estimated on the basis of
A fourth type of interest payments by individuals arises in connection with
Federal Deposit Insurance Corporation and Comptroller of Currency tabubrokers' loans. This segment was large in 1929, but in recent years has been
lations, reports on savings banks by the Commissioner of Banks in New York
of minor importance. The method of estimation here is to apply averages of
State, reports by the Commissioner of Banks in Massachusetts, and reports
quoted short-term money rates (New York Federal Reserve Bank) against
by the American Bankers Association.
average brokers' loans to customers (members only, New York Stock ExNonprofit organizations servicing business
Personal landlords (nonfarm) _




.2
24.9

100

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

change). This item is also entered directly in personal consumption expenditures, as part of Brokerage charges and interest.
Nonprofit organizations servicing individuals.—Informational

returns by labor

unions, social groups, etc., to the Internal Revenue Service have been tabulated for 1943 and 1946. Undercoverage in these data is large percentagewise,
but the amounts involved are negligible. Larger in amount, but still relatively
minor in comparison with the monetary interest paid total, are the payments
made by other types of organizations such as churches, hospitals, and private
schools. Here only fragmentary data have been available and the estimates
take the form of average interest rates applied against estimated mortgage
indebtedness, or interest-to-receipts ratios applied against estimated total
receipts.

Rest of the world
Interest paid to United States residents by foreigners is estimated in connection with the United States balance of payments estimates of the Office of
Business Economics. The series is described in the section on Net foreign investment.

Imputed Interest Paid
In addition to monetary interest, national income and personal income include imputed interest flows. These arise in connection with the transactions
of financial intermediaries. In national income accounting commercial banks
are conceived of as paying out to their depositors the entire amount of property income received, so that imputed interest paid by commercial banks
equals their property income received less interest actually paid on deposits.
Correspondingly, they are conceived of as making a charge for the services
rendered to their depositors in excess of the monetary service charges actually
made. These imputed service charges are numerically equal to imputed
interest paid.
Exhibit 4.—Components of Imputed Interest Paid, 7950
Item
Corporations
Commercial banks
Federal Reserve Banks
Finance, n. e. c
Stock life insurance carriers-..
Other private businesses
Mutual savings banks
Mutual life insurance carriers.
Savings and loan associationsCredit unions
Total imputed interest paid

Millions
of dollars

Percent

3,624

61.7

2,927
67
207
423

49.9
1.1
3.5
7.2

2,246

38.3

292
1,594
32S
35

5.0
27.2
5.6
.6

5,870

100.0

To the extent that these imputed flows represent intrabusiness transactions,
they cancel out in the aggregate and have no effect on the size of national
income and product, although they do affect its industrial distribution. To the
extent that they reflect transactions between commercial banks and persons,
they result in matching entries in (a) the interest components of national and
personal income and (b) personal consumption expenditures for services.1
A similar procedure is applied to corporate financial institutions of the investment trust type.
An interest imputation is made also in connection with life insurance. Imputed interest is measured in this instance by property income received, which
in national income accounting is treated as accruing directly to policy-holders.
In turn, policyholders are assumed to make payments to life insurance companies to cover their operating expenses. Mutual financial institutions other
than life insurance are given a similar treatment.
Imputed interest paid is described immediately below, and imputed interest
received is described following the discussion of monetary interest received.
Imputed charges made by financial institutions (analytically distinct from
1. In tracing the detail of the imputed flows, it will be noticed that imputed interest and
service charge transactions between commercial banks and government are not recognized
and are instead treated as occurring between commercial banks and personal recipient*.




the imputec interest flows) are also discussed, to the extent that they are
derived in a statistically interrelated procedure. The charges for life insurance, for which the methodology is different, are covered in the section on
Personal consumption expenditures for services.
In general, imputed interest paid by financial intermediaries is measured
as the excess of property income received over property income actually returned in monetary form to owners of the funds entrusted to the intermediary.
As will be noted in the following discussion, the precise content of this formula
varies among the several types of financial institutions involved. The composition of imputed interest paid in 1950 is shown in Exhibit 4.

Corporations

Most of the imputed interest paid by corporations originates in commercial
banking.
Commercial banks.—In estimating imputed interest paid by commercia
banks, basic data are drawn from annual reports of the Federal Deposit Insurance Corporation for the period beginning in 1935 and from the Board of
Governors of the Federal Reserve System for the 1929-34 period. This departure from the general use of the Internal Revenue Service data in the
corporate area was occasioned by shortcomings in the industrial classification
of the Statistics of Income banking industry.2
Imputed interest paid by commercial banks is measured as the excess of
interest and dividends received over interest paid on demand and time deposits. Reported data for member banks (Federal Reserve System) in the
1929-34 period and for member banks (Federal Deposit Insurance System)
in later years are raised to all-commercial-bank levels on the basis of asset ratios
derived from the above sources and from Comptroller of the Currency data.
Federal Reserve Banks.—Imputed interest paid by Federal Reserve Banks
(measured as interest received less interest paid and profits) is calculated from
aggregate financial statements published by the Federal Reserve Board.
Finance, not elsewhere classified.—Imputed interest paid by corporations classified in the "Finance, n. e. c." group, mainly investment trusts, holding companies, and long-term and short-term credit agencies, is measured as property
income in the form of interest and dividends received less the sum of interest
paid and profits (before income and excess profits taxes and without deduction for dividends received).
When income accounts are not in such a form as to permit an isolation of
loan and investment activities from other operations, profits as reported
reflect both elements, and a statistical isolation of profits earned in loan and
investment activities is needed, in order that the imputation procedure be
confined to these activities. In view of the fact that operations other than loan
and investment activities play a significant role in the "Finance, n. e. c." industry, thi; refinement seems desirable.
To effect the separation, investment trust companies of the management
type are assumed to be institutions in which operations are substantially
limited to loan or investment activities. Relationships between imputed
interest paid and property income received for investment trust companies are
computed and applied to the total property income in "Finance, n. e. c." to
estimate total imputed interest paid by this group.
These ratios were based for years prior to 1938 upon a report of the Securities and Exchange Commission, Investment Trusts and Investment Companies,

Part Two (March 1939), and for 1938 and subsequent years on Statistics of
Income—Part 2, including unpublished detail from the "Source Book".
Series for interest and dividends received by the "Finance, n. e. c." group
generally are based upon Statistics of Income—Part 2 data and unpublished
detail from the "Source Book." Only data for the 1929-33 period and after
1941 have been used directly, because of the break in the series introduced by
corporate reporting for tax purposes on an unconsolidated basis from 1934
through 1941. For these years, estimates were interpolated by total monetary
interest and net dividends paid by all corporations. For the two most recent
years, the Statistics of Income series is extrapolated by the sum of (a) monetary
interest payments by all corporations except those in the finance industries,
and (b) publicly reported cash dividend payments by all United States corporations except those in finance. (The series on publicly reported dividend
payments is compiled by the National Income Division from data in Moody's
Dividend Record and published regularly in the Survey of Current Business.)
2. It may be noted that, as a consequence, net interest originating in commercial banks,
as estimated for Part V, differs from dividends received by banking, to which by definition
it should be numerically equal.

101

NATIONAL INCOME, 195 4 EDITION

Stock life insurance carriers.—The property income (monetary and imputed The estimates for banking (excluding Federal Reserve Banks) are obtained
by extrapolating the last base year estimate by income from loans and investinterest, dividends, and net rents) received by life insurance carriers is rements reported by insured commercial banks to the Federal Deposit Insurance
garded as being withheld to the account of policyholders and is treated as
Corporation (raised to the all-commercial bank level). Interest received by
though it were actually disbursed. Accordingly, a payment is imputed for
Federal Reserve Banks is available from the Board of Governors of the Federal
life insurance in Exhibit 4.
Reserve System. Current estimates for monetary interest received by stock
Basic data for the measurement of dividend and monetary interest receipts
insurance carriers are prepared jointly for life and nonlife carriers. Interest
of both stock and mutual life insurance carriers combined are taken from
Statistics of Income—Part 2, supplemented by unpublished detail from the receipts of all stock insurance companies (life and nonlife) are assumed to
vary with the corresponding series obtained for total life insurance. (See "Im"Source Book". Reported dividend figures from 1929 through 1939 were adputed interest paid, stock life insurance carriers.")
justed to include dividends received from foreign corporations. Receipts of
mputed interest are derived by procedures described below under Imputed
The recent-year estimates for all corporations except banking and insurance
interest received. Series for gross rents received, also taken from Internal
are prepared separately for interest received on holdings of taxable United
Revenue Service sources, are converted to net rents realized by means of netStates Government securities and all other investments. This break is availgross rent ratios. The break between stock and mutual life insurance carriers
able in the Statistics of Income data. Interest received on taxable United States
is estimated on the basis of data published in The Spectator Company's InGovernment securities is extrapolated by a series calculated as United States
surance Tear Book.
securities held by corporations and associations (excluding banks and insurExtrapolation of the series based on Statistics of Income to current years is ance companies) times the computed midyear average interest rate on the
Federal debt, both available in the Treasury Bulletin. All other interest received
accomplished by means of Institute of Life Insurance data. An extrapolating
is extrapolated by the series for monetary interest paid by all corporations
series is constructed by multiplying security and mortgage asset holdings
except banks and insurance companies.
(averages of year-end figures) by net earning rates.

Other private businesses

Exhibit 5.—Components of Monetary Interest Received,

1950

Mutual savings banks.—Imputed interest paid by mutual savings banks is
Millions
Item
of dollars Percent
measured as (a) property income received (interest and dividends) less (b)
interest and dividends paid depositors and interest paid on capital notes and
63.4
Corporations5,291
debentures. The nature of the series for (a) and (b) is described under
6,505
Corporations reporting to the Internal Revenue Service-.
"Monetary interest received" and "Monetary interest paid", respectively.
1,489
Less: Mutual insurance carriers (life and nonlife)
275
Mutual life insurance carriers.—The procedure for estimating imputed interest Plus: Federal Reserve Banks
paid by mutual life insurance carriers has been described above under "Stock
Sole proprietorships and partnerships..
108
1.3
life insurance carriers".
52
Security and commodity brokers..
.6
56
Finance, n. e. c .
Savings and loan associations.—Imputed interest paid by savings and loan
.7
associations is measured as total interest income less the sum of interest and
2,900
Other private businesses..
34.7
dividends paid. The twoflowsare identical to those entered under "Monetary
Mutual financial institutions
2,883
34.5
659
7.9
Mutual savings banks
interest received" and "Monetary interest paid", except for the addition to
1,513
18.1
Mutual insurance carriers (life and nonlife) .
the former of imputed interest received from commercial banks, estimated by
653
7.8
Savings and loan associations
58
.7
Credit unions
procedures described below.
17
.2
Nonprofit organizations servicing business
Credit unions.—Imputed interest paid by credit unions is measured as inter53
.ft
Rest of the world
est received less dividends paid (the latter taken as interest payments on both
Total monetary interest received.
8,352
100.0
shares and deposits). Interest received by credit unions is assumed to equal
interest paid to them by individuals. (See above "Monetary interest paid,
Households and institutions".) Dividends paid are obtained from compilaTo obtain further industrial breakdowns, direct estimates are made for
tions of annual reports to the Bureau of Labor Statistics.
railroads, pipeline transportation, telephone and telegraph, and electric and
gas utilities, in general by extrapolating the base year estimates by interest
received as reported to the respective regulatory commissions. The sum of
Monetary Interest Received
the estimates for these industries is deducted from the figure derived above
for all corporations except banking and insurance. The residual, amounting
to about one-tenth of the all-industry total, is used to extrapolate the latest
In the calculation of monetary interest received by business and foreigners,
tax-return-based figure for each of the remaining industries.
business recipients are defined to include all corporations (63 percent of the
total in 1950), unincorporated security and commodity brokerage firms and
miscellaneous proprietors in the finance, n. e. c. category (together about 1
Sole proprietorships and partnerships
percent of the total), mutual financial intermediaries (35 percent), and
nonprofit organizations servicing business (negligible). Interest received by
Monetary interest received by unincorporated security and commodity
the rest of the world from the United States accounts for less than 1 percent
brokers and certain loan companies classified under "Finance, n. e. c." is
of the total. Further details are shown in Exhibit 5.
assumed to be received as an integral part of business operations. It is deducted
in arriving at the interest share and correspondingly included in the receipts
of these enterprises in calculating their net income.3 Source materials are
Corporations
similar to those noted under monetary interest paid. Except for security and
commodity brokers in 1929, amounts involved are small.
The general nature of the basic data covering interest received by corporations has already been discussed under "Monetary interest paid". Interest
received by mutual insurance carriers (measured in conformance with the
Other private businesses
Internal Revenue Service definition of interest received) is removed by use of
the data described above, under "Imputed interest paid, stock life insurance
Mutualfinancialinstitutions.—In general, the source materials used for esticarriers". Interest received by Federal Reserve Banks is tabulated by the
mating monetary interest receipts of mutual financial institutions are identical
Board of Governors of the Federal Reserve System.
to those employed in the "Monetary interest paid" estimates.
For the two most recent years, for which tax return data are not available,
3. In general, monetary interest received by owners of unincorporated enterprises Is not
total monetary interest received by corporations is obtained by adding sepadeducted in computing net interest because it is assumed to be received by them in a personal
rate estimates for the banking and insurance industries and for all industries
rather than business capacity; and, correspondingly, it Is excluded in estimating incomes of
except banking and insurance.
unincorporated enterprises.



102

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Monetary interest received by mutual savings banks is measured as interest
and dividends on securities plus interest and discount on loans. The method
of estimation is similar to that used for monetary interest paid.
Monetary interest receipts of mutual insurance carriers are measured as
the sum of interest and dividends for national income purposes. (This accounts for the difference between the two entries for these institutions in
Exhibit 5.) The method of estimate for life insurance carriers has been described above, under "Imputed interest paid, stock life insurance carriers".
Receipts of mutual nonlife carriers are obtained from Statistics of Income tabulations, extrapolated by the method described under "Imputed interest paid,
stock life insurance carriers."
Monetary interest received by savings and loan associations is measured
on the basis of reports by member associations to the Federal Home Loan
Bank Board. Receipts by credit unions are assumed equal to payments by
individual borrowers, and are estimated by methods described above under
"Monetary interest paid, households and institutions".

These distributions are computed from Federal Deposit Insurance Corporation deposit data from 1934 forward, and from Federal Reserve Board deposit
data from 1929 to 1933. The sources are the Annual Reports of the Federal
Deposit Insurance Corporation a n d Banking and Monetary

Statistics.

2. The further distribution of the total received by persons and businesses
is based primarily on Derivation of Liquid Asset Distribution Estimates (mimeo-

graph), Board of Governors of the Federal Reserve System. This publication
gives the distribution of demand and time deposits by selected groups of
holders, as of the end of December, from 1939 forward. For the most part,
estimates presented in Solomon Shapiro's article on "Distribution of Deposits
and Currency in the United States, 1929-1939", in the Journal of the American
Statistical Association, December 1943, were used to extrapolate the Federal
Reserve figures back to 1929, using a December 1939 link.

Exhibit 6.—Imputed

Interest

Paid by Commercial

and Federal

Reserve

Banks, by Major Recipients, 1950

Nonprofit organizations servicing business.—The basis for estimation has been

described above, under "Monetary interest paid."
Millions
of dollars

Item

Percent

Rest of the world
Interest received by foreigners from United States residents is described in
the section on Net foreign investment.

Imputed Interest Received
The outflows from financial intermediaries of imputed interest paid become
imputed interest received when viewed from the standpoint of recipients to
whom such imputed interest accrues. It is necessary, consequently, to determine what groups receive the imputed interest that arises in each of the financial intermediaries. The underlying procedure in making this determination
is based upon the ownership of the funds by use of which financial intermediaries obtained property income.
Data on ownership are not generally available except for commercial banks,
for which there is indication of ownership by broad categories. It is necessary,
therefore, to solve the problem of ownership in most cases by use of assumptions.
A review of the several financial intermediary groups indicates the substantial validity of assuming that persons own all the funds entrusted to
mutual savings banks, life insurance carriers, savings and loan associations,
and credit unions. The corporate component of the finance, n. e. c. group,
comprising investment trusts, holding companies, and both long-term and
short-term credit agencies, is less clear-cut. It is quite possible for business as
well as persons to own securities of these companies. Lacking ownership data,
it is nevertheless assumed that for this intermediary group also ownership is
vested in persons. However, the consequent error of assuming that all imputed transactions in this area are with persons should be appraised in the
light of the fact that imputations among affiliated companies are eliminated
by the statistical estimating methods employed. (See the reference to the use
of consolidated returns under "Imputed interest paid, finance, n. e. a")
In summary, for all financial intermediaries except commercial banks the
flows of imputed interest paid by intermediaries are treated as going entirely
to persons. For commercial banks (including Federal Reserve Banks), imputed interest is allocated among recipients by use of estimates of the ownership of deposits. The main sources of data for the allocation to principal
classes of recipients are the published banking statistics of the Board of Governors of the Federal Reserve System and the Federal Deposit Insurance
Corporation, together with the Federal Reserve liquid asset surveys. The
distribution of imputed interest paid by commercial and Federal Reserve
Banks is shown in Exhibit 6.
This distribution is accomplished in three major steps:
1. Imputed interest received is estimated for three broad groups: (a) Federal Government, (b) State and local governments, and (c) persons and businesses. The procedure consists of allocating the elements of imputed interest
paid by commercial banks (total property income, interest paid on demand
deposits, and interest paid on time deposits) on the basis of percentage distributions showing the ownership of demand deposits and of time deposits by
the three groups, and then deducting the sum of the two interest paid estimates from total property income.



Imputed interest paid by commercial and Federal Reserve Banks
(see Exhibit 4); to be distributed among major recipients

2,994

Step 1:
United States Government
State and local governments..,
Persons and businesses

85
219

2.8
7.3

2,690

89.8

2,690

100.0

1,205

44.8

__

Step 2

-

_

Persons
Trust funds for individuals
-Nonprofit organizations servicing individuals
Individuals. __
_
Businesses (see Exhibit 1)

_.

_.

100.0

38
43

1.4
16
.

1,124

41.8

1,485

55.2

Corporations
.
Finance, insurance and real estate
Stock life insurance carriers
Stock nonlife insurance carriers
Other finance, insurance and real estate
Other industries
_

923
114
8
26
80
809

34.3

30.1

Proprietorships and partnerships

509

18.9

42
13
29
467
128
339

1.6
.5
11
.

Finance, insurance and real estate
Security and commodity brokers
Other finance, insurance and real estate.
Other industries
.
Farm
Nonfarm
Other private businesses
Mutual life insurance carriers
Mutual nonlife insurance carriers
Savings and loan associations

4.2
.3
10
.
3.0

17.4

4.8
12.6

53

2.0

28
Q
y
18

1.0
«]

.
6

Step 3 >
1 Consists of a further detailed industrial breakdown of allocations shown under step 2

3. This step consists of a further industrial breakdown, within the industry
groups specified under step (2), thus completing the allocation by major and
minor industries under the National Income Division industrial classification.
In general, the method is to distribute imputed interest to industries on the
basis of relative holdings of cash and deposits, as there is no further information on deposit ownership by industry.
Cash and deposits held by corporations filing balance sheets are tabulated
by industry in Statistics of Income—Part 2. The reported data are raised (by
industry) on the basis of ratios of total compiled receipts of all corporations
to total compiled receipts of those submitting balance sheets, and are adjusted
to the National Income Division industrial classification.
For sole proprietorships and partnerships, only indirect methods of estimating cash and deposit holdings are available. In general, ratios of cash and
deposits to total receipts for proprietorships and partnerships are assumed to
equal similar ratios for corporations with assets under $50,000 in the respective industries. These corporate ratios were obtained from Statistics of Income—
Part 2, for 1941, and extrapolated to all other years by similar ratios for all
corporations.
For both corporations and unincorporated firms, the relative distributions
of cash and deposit holdings are assumed to remain constant in the two years
prior to availability of the tax return data.

NATIONAL INCOME, 195 4 EDITION

Services furnished without payment by financial intermediaries, except life insurance
It is convenient to describe at this stage the derivation of the series, "Services furnished without payment by financial intermediaries, except life insurance", a component of personal consumption expenditures. This component
includes entries for commercial banks, corporate finance, n. e. c , mutual
savings banks, savings and loan associations, and credit unions.
Services furnished without payment by commercial banks to persons are
numerically equal to imputed interest received by persons from commercial
banks. For corporate finance, n. e. c. the entry equals the imputed interest
paid item whose derivation has been explained above, under "Imputed interest paid." For mutual savings banks, savings and loan associations, and credit
unions the entry equals imputed interest paid, calculated as explained above,
less income taxes and retained profits (before deduction of dividends). In
general, the series on retained profits and taxes are developed from the same
sources as the series on interest and dividend flows.

Net Interest Paid by Government
A breakdown of the government interest calculation is shown in Exhibit 7.
It should be noted that the transactions covered comprise not only those of
general governments (including trust funds) but also those of government
enterprises.
Exhibit 7.—Net Interest Payments by United States Governments, 1950
Millions
of dollars

Item
Excess of interest payments by United States governments over
interest received by United States governments

Monetary interest paid

-

100.0

5,804
624

State and local governments

State and local governments

4,716
6,428

Monetary interest paid by governments

Monetary interest received by governments

Percent

90.3
9.7

1,712

100.0

1,373
339

80.2
19.8

103

Monetary interest received
Federal Government.—Interest received by the Federal Government comprises interest income of the social insurance funds, interest income covered
into miscellaneous receipts of the Treasury, and interest received by government enterprises, as well as trivial amounts received by several other funds.
It may be noted that the bulk of these receipts is intragovernmental, and
offsets identical amounts under interest paid.
Interest received by the social insurance funds is reported in the Daily
Treasury Statement. Other miscellaneous interest receipts of the Federal Government are derived from annual analyses of miscellaneous receipts of the
Treasury as detailed in the Budget document, fiscal year totals being averaged
to obtain calendar year estimates. Interest received by Federal Government
enterprises is, in general, obtained on the same basis as the corresponding
interest paid component.
State and local governments.—Estimates of interest received by these governments, like those of their interest payments, are drawn from the various
Census releases on Government Finances.
Fiscal year data for State governments are available for the years 1929-32,
1937-40, and from 1952 on. Estimates for the 1933-36 period were obtained
by straight-line interpolation, while those for 1941-51 were interpolated by
reported investment earnings (not entirely confined to interest) of State pension, sinking, and trust funds. The fiscal year totals are averaged to obtain
calendar year estimates throughout the period.
The Census reports of interest received by local governments are fragmentary in most years. A comprehensive benchmark for 1931 can be established
from the 1932 Census of Governments, however, on the assumption that the
interest component of reported revenue from "highway privileges, rents, and
interest" for local units other than cities of more than 30,000 is proportionately the same as for cities of 30,000 to 100,000 (for which, along with the
larger cities, a more detailed breakdown is available). Sample-based Census
Bureau estimates are available from 1952 on, with a 1-year lag. Interpolations were governed primarily, through 1942, by reported data on investment
earnings of public trust, pension, and sinking funds of cities with populations
over 100,000. The basic annual interpolator series for 1943—51 cover only
cities of 250,000 or more.
Because of the time elapsed in compilation of the basic Census reports,
interest receipt estimates for the most recent year, both for States and for
local governments, have to be extrapolated in more or less arbitrary fashion,
with due regard for the trend of the series and for miscellaneous collateral
information.

Federal Government.—The largest component of interest paid by the Federal
Government is that paid or payable on the public debt. Data for all years are
available in the records kept by the United States Treasury Department. The
published source is the Daily Treasury Statement. Small amounts of interest are
also paid on several types of government-administered trust funds. Calendar
year estimates of the latter are obtained by averaging fiscal year data from the
7. PERSONAL CONSUMPTION EXPENDITURES FOR
Budget of the United States Government. Estimates of interest paid by Federal
COMMODITIES
Government enterprises are prepared in connection with the derivation of
the enterprise surplus or deficit. (See section on Government receipts and
Personal consumption expenditures for commodities—like wages and
expenditures.)
State and local governments.—Estimates of interest paid by State and local salaries—represent transactions that can best be estimated from business,
rather than individual, records. Unlike wages and salaries, however, congovernments, prepared separately for several types of government units, are
based upon data drawn from the Government Finances publications of the sumer commodity purchases involve indirect estimation. The business sales
Census Bureau, particularly the releases on Governmental Debt in the United data underlying the latter series neither generally nor uniformly distinguish
sales to consumers or to the other broad purchaser groups (such as business
States a n d t h e Summary of Governmental Finances in 1952.
and government) relevant to national income measurement.
Interest payments by State governments are available for fiscal years 1929The general problem, then, is to derive consumer sales, valued at prices
32 and 1937 forward, and for intervening years were interpolated on the basis
paid by consumers, from total sales. This qualification regarding prices is
of the reported gross debt of State governments. The latest fiscal year estimate
important. If the estimating procedure starts with sales (or output) at the
is obtained by extrapolation on the same basis. Adjacent fiscal years are
producer level and then adjusts them to cover sales of finished consumption
averaged to obtain calendar year estimates.
commodities only, there still remains the substantial task of raising their
Comprehensive coverage of local government interest payments can be obvaluation to a consumer-price basis.
tained from the Census reports for the years 1931, 1936, and 1940 forward.
Since the reported data represent diverse fiscal periods, minor adjustments of
timing are required to place them on a calendar year basis.
Nature of commodity flow method
In the years of incomplete coverage, total municipal interest payments are
assumed to move proportionately to those reported for the larger cities only,
This approach of starting with producers' sales (or output) is, in fact, the
while interest payments by other types of local government units are interbasic one employed in this report. Termed the "commodity flow" method, it
polated or extrapolated on the basis of reported gross debt statistics.
was used for the 1929-39 period and 1947 to derive estimates for consumer



104

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

commodities making up over four-fifths of total consumer commodity
expenditures. (See Exhibit 1.)
First developed by Simon Kuznets of the National Bureau of Economic
Research and detailed in his National Bureau volume on Commodity Flow
and Capital Formation (1938), the procedure involves numerous estimating
steps. In broad outline, these entail securing commodity data at producers'
prices on the output of factories, farms, and fisheries; segregating for each
commodity the portion of total output not requiring further processing,
Exhibit 1.—Methods of Estimating Consumption Expenditures for Commodities,
1939 and 1947
1939

Item

Estimated by commodity flow method.._

Millions
Millions
of dollars Percent of dollars Percent
34,928

83.6

94,769

83.4

Estimated by other methods

6,873

16.4

18,901

16.6

Retail valuation method

5,148

12.3

14,652

12.9

3.869
•' 4, 408

3.4
3.9
3.2

Tobacco products—
New cars and net purchases of used cars
Gasoline and oil
Other fuel and Ice except fuel produced and
consumed on farms
Imputations
Food produced and consumed on farms
Fuel produced and consumed on farms
Standard clothing issued to military personnel
Miscellaneous
Flowers, seeds, and potted plants
Tips in purchased meals and beverages
Lightin
Other !
Total..

(')
1,679
2,181
1,288

4.0
5.2
3.1
3.0

3,630
2,745

2.4
2.5

2.7

2,895
2,532
134
229

.2

485

1.1

1,354

1.2

191
117
86
91

.4
.3
.2
.2

475
429
(')
450

.4
.4

41,801

100.0

113,670

100.0

1,240
1,114
110
16

2.2
.1

.4

1. Estimated by the commodity flow method.
2. Excludes expenditures for housing type trailers, which were estimated by the commodity
flow method.
3. Comprising the following: In 1939, paintings and art objects, products of custom establishments, n. e. c , and expenditures by United States Government personnel abroad; in
1947, expenditures by United States Government personnel abroad.

and destined for personal consumption; and then passing from finished
output at producers' prices to final costs to consumers by tracing the commodities through the various stages of the distributive system.1
The Census of Manufactures, with its vast commodity detail, is the basic
statistical source for the commodity flow method. This census was available
biennially for odd-numbered years of the 1929-39 period and for 1947.
The most important supplementary sources utilized by the method are the
Censuses of Retail Trade and Wholesale Trade, which, during the span
covered by this report, were taken for the years 1929, 1933, 1935, 1939,
and 1948. These censuses—together with the Censuses of Distribution of
Manufacturers' Sales (taken for 1929, 1935, and 1939 in conjunction with
the manufacturing censuses)—provide most of the basic information for
segregating consumer commodities from total output, tracing their flow,
and measuring their costs of distribution.
Taken at frequent intervals, the manufacturing and trade censuses furnished a comprehensive basis for estimating consumer commodity purchases
during the 1929-39 period. The gaps during this period left by the absence
of one or more of these censuses were filled, in generally satisfactory manner,
by an interpolation process carried out within the commodity flow framework. For this purpose, a wide variety of statistical series was assembled
from government and trade associations and other private sources, and some
reliance was placed on relationships developed from the census materials.
For the period after 1939, the infrequency of industrial census materials
has restricted use of the commodity flow method to the single year 1947.
Benchmark estimates for the commodity flow segment of consumer commodities for that year are presented for the first time in this report.
The commodity flow method is admittedly "roundabout" and complex.
It was adopted mainly because of the very detailed commodity classification
1. The commodity flow procedure was also used in the estimation of producers' purchases
of durable equipment for this period. This is covered more particularly in the section on
Producers' durable equipment.




and the comprehensive coverage of output afforded by the Census of Manufacturers. For both of these reasons, the method was distinctly to be preferred
to the principal alternative method employing the Census of Retail Trade as
the basic source. This method, the retail sales approach, involves adjusting
retail sales to eliminate all sales by retail establishments not made to consumers and then estimating independently, and adding in, that part of
consumer expenditures on commodities not made in retail establishments.

Use of retail sales for interpolation and extrapolation
While it has not been feasible to employ the simpler, more direct retail
sales approach in preparing benchmark estimates of consumer commodity
expenditures, retail sales data have furnished the principal basis of an interpolation and extrapolation procedure used in deriving estimates for the
years 1940-46 and subsequent to 1947. For commodity groups estimated by
the commodity flow procedure for 1939 and 1947, relative changes in retail
sales data have been used very largely to interpolate between the 1939 and
1947 benchmarks and to extend the latter forward into the current period.
Basing the movement of consumer commodity expenditures on retail sales
has significant limitations. In itself, the method cannot take account of
year-by-year changes in the relative importance of business purchases at
retail and consumer purchases outside retail trade—allowances for which are
difficult to make. Moreover, with the exception of department store sales
figures, the retail sales data are not available by commodities, but according
to a not very detailed classification by type of store. Thus, in the extrapolation
of consumer commodity expenditures from one census period to the next,
the general assumption is made for the various commodity groups that
benchmark relationships of consumer purchases to the sales of selected types
of retail stores remain valid.
Although estimates for the recent period continue to be derived through
extrapolation by retail sales, the feasibility of basing commodity flow estimates upon product data obtained in the Census Bureau's Annual Survey of
Manufacturers is receiving thorough study. This is a sample survey, begun in
1950, of manufacturers' shipments (and other data), and is utilized in the
producers' durable equipment estimates. (See Section 10.) Sampling variability in the manufacturers' shipments data is substantial, however, for some
consumer expenditure groups. Moreover, product information in the annual
surveys is much less detailed than in a full census, increasing the difficulties
of making consumer-nonconsumer allocations. Among other related problems, the most important has been that of obtaining current information on
retailers' markups. But given continued improvement in the annual manufacturing surveys and increased availability of auxiliary data, it may be
possible ultimately to institute annual commodity flow estimates.
In short, then, the bulk of all consumer commodity expenditures was
estimated by the commodity flow method for the years 1929—39 and 1947
and by interpolation and extrapolation based on retail sales for other years.
For certain commodities, different methods were considered preferable. For
still others, such as food and fuel produced and consumed on farms, the
commodity flow method was not appropriate.

Other estimating methods
The main alternative to the commodity flow method (with interpolation
and extrapolation by retail sales) was the direct approach of multiplying
estimated quantities purchased by consumers by appropriate average retail
prices. This procedure, which may be termed the "retail valuation"
method, was used in estimating consumer expenditures for passenger cars
and for gasoline and oil for all years; for tobacco products from 1940 to the
present; and for "other" fuel and ice (except fuel produced and consumed
on farms) for the years 1929-39 and 1947, with a retail sales interpolation
and extrapolation being employed for other years. These four groups accounted for 13 percent of the $114-billion commodity total in 1947.
The remaining estimates are of minor importance except for the imputed
item covering food produced and consumed on farms. This accounted for
about 2 percent of the 1947 total.

General considerations of reliability
The subsequent description of methodology will afford, at various places,
judgments about particular aspects of the estimates of consumption expendi-

NATIONAL INCOME, 195 4 EDITION

tures on commodities. Several propositions regarding their general character
may be worth noting at this point, with, however, the realization that any
appraisal of over-all reliability is necessarily indirect and cannot be definitive.
Concerning the estimates for 1929-39 derived by the commodity flow
method, there is the primary fact that they entailed lengthy procedures
processing large masses of data and resting at times to an uncomfortable
degree on judgment. Counterbalancing this are several considerations pointing toward the general appraisal that the final estimates of the totals are not
markedly in error.
The first is the wealth of census information on which the estimates for
this period were founded—information which, if not ideal, was adequate for
the requirements of most phases of the estimating process. Secondly, certain
evident sources of error in the individual commodity estimates—such as the
necessity of using type-of-store data in computing retail markups—-may be
assumed to have a tendency to compensate, or offset, in the total. Thirdly,
substantial temporal comparability in the estimates—that is, reliability of
movement over time—seems indicated from the use of census data having
a high degree of comparability and the application of uniform procedures
and assumptions.
Exhibit 2.—Comparison of Commodity Flow and Provisional Estimates for 1947
1939
Benchmark

Group

estimates

1947

Benchmark
esti-

mates

1947
Provi- Percent disional
vergence
estiof provimates ' sional from
benchmark
estimates

Millions of dollars

Food excluding food produced and consumed on
farms and tips
Shoes and other footwear
Clothing and accessories except footwear

6
1
3

17, 933
1,226
5,893

51,210
2,955
15,610

' 54.172
2,975
16, 105

Jewelry and watches
Toilet articles and preparations
Furniture
Kitchen and other household appliances
China, glassware, tableware, and utensils

355
486
949
774
475

1.463
1,245
2 552
3.179
1.348

1,348
1,208
' 2, 521
1
3, 012
1,442

Other durable house furnishings
Semidurable house furnishings
Cleaning and polishing preparations, and miscellaneous household supplies and paper products »

908
681

2,453
2,135

2,965
1,868

5
-13

508

1,523

1,146

-25

Stationary and writing supplies
Drug preparations and sundries
Ophthalmic products and orthopedic appliances.
Tires, tubes, accessories and parts

149
612
172
484

440
1,313
400
1,674

391
1.358
386
1,626

-11
3
-4

Books and maps
Magazines, newspapers, and sheet music 3
Nondurable toys and sport supplies
Wheel goods, durable toys, sport equipment,
boats, and pleasure aircraft
Kadio and television receivers, records, and
musical instruments

226
554
285

536
1,243
910

747
1,118
1,006

39
-10

228

972

761

Total«..

420

1,429

33,318

94,590

1.724
97,4'

O

-3
2

-5
7

11
-22
21
3

1. With exceptions indicated In footnotes 2 and 3, these were published in the July 1950
SURVEY OF CURRENT BUSINESS and the 1951 NATIONAL I N C O M E supplement and were

derived by extrapolating the 1939 benchmark estimates by retail sales.
2. Based on 1947 estimates published in July 1949 because adjustments influenced by the
1947 and 1948 censuses were made in the subsequently published estimates for these groups.
3. Retail sales extrapolation is not used in these groups.
4. The benchmark total for 1939 differs from the 1939 commodity flow figure in Exhibit 1
principally because of the omission above of the tobacco estimate amounting to $1,767 million.
In additioa, three small non-commodity flow items totaling $157 million are included in the
1939 total above for comparability with 1947. The benchmark total for 1947 is less than the
1947 commodity flow figure in Exhibit 1 because of the omission above of housing-type trailers,
amounting to $179 million.

The estimates for 1947 obtained by the commodity flow approach are o*
the same general character as the prewar benchmarks. The 1947 Census of
Manufactures, it may be noted, provided even greater commodity detail
than the biennial censuses of the 1929-39 period—a factor making for more
accuracy in the derivation of consumer product values at the manufacturers'
level. On the other hand, the postwar industrial census materials were somewhat deficient for purposes of the commodity flow work, notably in the lack
of the information provided in earlier censuses on the distribution of manufacturers' sales and on retailers' operating expenses. Certain changes in the
basic commodity flow method were therefore required, with some probable
impairment of reliability.
Because of the long intervals between the industrial censuses used for
benchmarks, the procedure of basing the movement of consumer commodity



105

expenditures on retail sales has assumed signal importance in the period
since 1939. A basis for judging the reliability of this procedure is provided by
Exhibit 2. This compares the new commodityflowbenchmarks with previously
published estimates for 1947. The last column of the exhibit—showing the
percent difference between the provisional and revised estimates—directly
affords an appraisal of retail sales extrapolations over the eight-year period
through 1947 marked by extensive war and postwar changes and a near
tripling in total dollar expenditures. (To the extent, of course, that the 1939
and 1947 benchmark estimates could not be made entirely comparable, this
factor also would be reflected in these percentages.)
The results are broadly satisfactory. Of first note is that the projection by
retail sales of 1939 consumer expenditures for commodities missed the commodity-flow benchmark total in 1947 by only 3 percent. The showing by
commodity groups, it will be observed, is mixed, although for most of them
the divergence is 7 percent or less. The series showing the widest percentage
discrepancy, moreover, are generally those for which retail sales data are
weakest for measuring consumer expenditure changes—usually because the
products involved are least closely identified with a particular type of store.
The estimates for the war years, it may be added, may have been subject
to special bias. It is possible that the substantial shifts which occurred in the
retail spending pattern were accompanied by temporary changes in the relationships of consumer expenditures for specific commodities to sales in the
lines of retail trade used for interpolation. For lack of data, little could be
done in the commodity estimates to allow for such changes.
The above record of revisions for the commodity flow portion of consumer
commodities might be broadened to note that the downward revision for
total consumer commodities for 1947 also amounted to 3 percent. Further,
consumer expenditures for services, as explained in that section, also underwent statistical revision—notably to incorporate results of the 1948 Census of
Business and the 1950 Census of Population and Housing—and, in total,
were raised by about 4}i percent for 1947. The net result was that the new
1947 total for commodities and services combined differed by less than one
percent from the previous total.
Of the commodity groups derived by methods other than the commodity
flow, it may be noted that the estimates for passenger cars, gasoline and oil,
and purchased household fuels are subject to a significant qualification.
The allocation of these commodities to consumers is based on limited information and may be considerably in error.
Though indirect and inconclusive, some evidence bearing on the general
reliability of the consumer commodity totals is afforded by the size of the
statistical discrepancy between the income and expenditure sides of the
national income and product account. For a component series as large as
consumer commodities—comprising in most years nearly half of the gross
national product—the small size of the discrepancy throughout the period
of the estimates may be taken as some external indication that the general
level of the series is not markedly in error. This is particularly the case since
considerable reliability can be attached to certain other large income and
product components, such as wages and salaries, corporate profits, and
government purchases of goods and services.
Additional indirect evidence on the reliability of total consumer commodity expenditures can be adduced. This stems chiefly from the relatively
close agreement between the Commerce and SEC estimates of personal
saving. (See table 6, Part V.) These two saving series, in very large degree,
are based on independent statistical sources. This is also true of disposable
income and the two consumer expenditure series (commodities and services)
which are subtracted from it to derive the Commerce personal saving estimates. Given these facts—plus the heavy statistical influence of the large
consumer commodity series on the calculation of personal saving—it seems
reasonable to conclude that the commodity series is relatively accurate. For
that not to be true, either of two seemingly unlikely situations would have to
obtain. Either (a) the level of personal saving as indicated by the Commerce
and SEC series would have to be substantially in error or (b) any marked
error posited for the consumer commodity totals would have to be offset
regularly in personal saving by errors in disposable income and/or consumer
expenditures for services.
The following, main section of the notes gives a description of methodology
in terms of the principal types of methods, as summarized in Exhibit 1. The
description is particularly limited for the large commodity flow segment,
since notes of this scope cannot deal with the numerous departures from

106

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

general procedure or with the individual commodity groups, for which the
relative importance of the various steps in the estimating procedure (and
hence reliability) often varies significantly.

Commodity Flow Method, 1929-39
The individual steps in deriving total personal consumption expenditures
estimated by the commodity flow method are summarized for 1939 in
Exhibit 3. Each of these steps will be described briefly.
1. Distribution offinishedand mixed manufactured commodities.—The basic source

of manufactured commodities is the Census of Manufactures. Detailed output
data—for about 4,000 commodities in 1939—were reported there for odd
years of the 1929-39 period.
For the most part, the census data were comparable from year to year.
There were, however, several minor deficiencies. One was that census coverage was slightly less complete in 1933 than in other years. The 1933 census
data were raised whenever some indication was given of the degree of incompleteness in an individual industry—usually less than 2 percent. Another
minor deficiency stemmed from the fact that the reported commodity information was somewhat less detailed for some years, notably 1933, than others.
It was often necessary to break down combinations of commodities on the
basis of the details reported for proximate years.
Still other limitations lay in the difficulty of establishing strictly comparable
series over the 1929-39 decade for many commodities made in the textile
and apparel industries, and in the differing degrees of coverage of some of
these industries in particular census years. Aside from careful attempts to
achieve proper classification, little could be done to overcome these limitations. It is possible that, especially for 1935 and 1937, the figures for some of
the component commodities of the clothing and semidurable house furnishings
groups are slightly too low.
Following these and other minor adjustments of the reported census commodity values, the first major task was to classify the full array of commodities
into the categories of finished, unfinished, and mixed.
Finished commodities were defined to include (a) commodities that had
reached the stage at which they would be taken over by individual consumers
without further processing, and (b) durable equipment intended for multiple
use in production and with an average life of one or more years.2
Classification under (a) did not depend solely on the degree of processing; it
was based also on the use to which a commodity is put. Flour, for example,
was classified as finished if consumed in households but as unfinished if
consumed by a factory making bread or other products for which flour is a
raw material.
Unfinished commodities were defined to include (a) all commodities entering further into the productive process other than those with an average life
of one year or more; and (b) in the context of this statistical study, finished
commodities purchased by government and by nonprofit institutions serving
persons, as well as construction materials—all of which are accounted for in
other segments of the gross national product and estimated by methods other
than the commodity flow.3 Exports, counted under net foreign investment,
also are unfinished in this context. They were not classified as such initially,
but were eliminated at a subsequent step in the estimating process.
The mixed category is a tentative grouping for commodities having appreciable diversity in use, and which therefore could not be classified wholly as
2. The formulation under (a) covers the bulk of the consumer items. A more precise definition of a finished consumer commodity would add to the phrase "without further processing"
the qualification in manufacturing or the services. This qualification is needed to take account
of marginal cases in the application of the commodity flow method. If a commodity is further
processed outside manufacturing and the services, it is counted as finished in the present
context. For instance, clotn bought by custom tailoring establishments is classified as finished
because the output of such establishments is not listed in the Census of Manufactures, and
hence would be missed if purchases of materials by custom tailors were not classified as
finished. However, if a commodity is purchased by a service establishment and covered in
its service receipts it is treated as unfinished in the commodity flow method to avoid duplication, as it is accounted for in personal consumption expenditures for services. For instance,
spark plugs purchased and installed by automobile repair service establishments and covered
by their service receipts, rather than sold separately to consumers, are classified as unfinished
as they are reflected in consumer purchases of services.
3. One complicating element, which is explained in the section relating to producers'
durables and need only be mentioned here, is that for statistical reasons government purchases
of durable equipment were not eliminated as "unfinished" in the process of commodity classification, but at a later stage.




finished or unfinished. The mixed commodities belonged in part to the
unfinished grouping and in part to the producer durable and/or consumer
commodity groupings, and required allocation among them. In the strictest
sense, it is realized, most commodities are mixed, but it would not have been
feasible to allow for very small fractions of finished or unfinished use. Therefore, only when there was reason to believe that the secondary use of a
commodity was appreciable was it assigned to the mixed category for
allocation.
Exhibit 3.—Derivation of Total Personal Consumption Expenditures Estimated
by the Commodity Flow Method, 1939
[Millions of dollars]
1. Distribution of finished and mixed manufactured commodities, before deduction of government purchases of durable equipment
a. Finished
1. Producers' durables
_
2. Consumer commodities
3. Combined, allocated to
a. Producers' durables.
b. Consumer commodities
..
b. Mixed, allocated to
1. Producers' durables
2. Consumer commodities3. Unfinished_

_
_

_

_.

25,978
12,327
1,714
9,188
1,425
206
1,219
13,651
1,112
8,210
4,329

2. Manufacturers' production of finished consumer commodities [la(2)-(-la(3b)-f18, 617

lb(2)]_

43

3. Subtract: Change in manufacturers' inventories.

18, 574

4. Equals: Manufacturers" sales of finished commodities

2,641

5. Producers' sales of finished nonmanufactured commodities

21,215

6. Producers' sales of finished commodities (4+5)

993

7. Add: Transportation charges

22,208

8. Equals: Producers' sales, inclusive of transportation charges, distributed to
a.
b.
c.
d.

Exports
WholesalersRetailers. _
Consumers

_

_.
_

_

_

_

9. Imports

504
12,669

10. Total purchases by wholesalers (8b+9).

106

11. Subtract: Change in wholesalers' inventories

12, 563

12. Equals: Cost of goods sold by wholesalers

2,767

13. Add: Wholesalers' markups

15,330

14. Equals: Wholesalers' sales of finished commodities.-

308

15. Subtract: Wholesalers' exports

15,022

16. Wholesalers' domestic sales, distributed to:

14,140
882

a. Retailers
b. Consumers
17. Total purchases by retailers (8c+16a)

170
12,165
8,778
1,095

22,918
_

231

18. Subtract: Change in retailers'inventories

22,687

19. Equals: Cost of goods sold by retailers

9,874

20. Add: Retailers' markups

32,661

21. Equals: Retailers'sales of finished commodities

34,538

22. Consumers'purchases exclusive of general retail sales taxes (8d+16b+21)
23. Add: General retail sales taxes

-

24. Equals: Personal consumption expenditures estimated by the commodity flow
method

34,928

The classification of commodities as finished, unfinished, or mixed was
greatly facilitated by the extent of commodity detail in the Census of Manufactures. This detail made it possible to classify as finished or unfinished many
product categories which, if combined, would have required allocation. For
example, soap chips, flakes, washing powder, cleansers, and similar products
were reported in 1939 broken down between packaged and bulk. While
consumers buy some of the latter and businesses and other "nonconsumers"
some of the former, the reported breakdown undoubtedly furnished a
satisfactory basis for distinguishing between consumer and nonconsumer uses.
On the other hand, the equally important granulated, powdered, and
sprayed soaps were reported without a breakdown between packaged and
bulk and therefore had to be allocated.

NATIONAL INCOME, 195 4 EDITION
The greater number of the commodities reported could be classified
directly as either finished or unfinished. An overwhelming proportion of
commodities in this initial finished classification could be assigned immediately to either the producers' durable or the consumer commodity category.
However, some finished commodities—household furniture and tools are
examples—fell into both of these categories, and so required allocation. This
combined group of finished commodities was not relatively large—as shown
for 1939 by the data in Exhibit 3—and its allocation was not a likely source of
appreciable error, particularly with respect to the gross national product
total. Misallocation between consumer commodities and producers' durables
would not lead to error in that total except for the addition of an inappropriate distributive cost element in passing from the producer value to the final
value.
Commodities which could not be classified directly as wholly finished or
unfinished—the "mixed" commodities—required special study to allocate
them among the categories of producers' durables, consumer commodities,
and unfinished. Fortunately, census reports provided two types of commodity
information that were very useful for this purpose: (1) Data in the biennial
Census of Manufactures on the quantity and cost of principal materials consumed in certain industries; and (2) data on the distribution of sales by class
of purchaser in the Distribution of Manufacturers' Sales for 1929, 1935, and
1939, and in the Census of Wholesale Trade for 1929, 1933, 1935, and 1939.
These census data were checked and supplemented by use of numerous
special commodity studies from both government and trade sources, and by
correspondence with business firms and commodity experts.
The materials consumed data were used to estimate the unfinished part of a
number of mixed commodity items in the large food group. Special commodity data were used in the allocation of durable equipment items. In some
cases, with adequate data lacking, the allocations had to be based on judgment, including outside expert opinion. But the method most generally
followed in estimating the nonconsumer portion of mixed commodities
involved application of the sales distribution data.
In the manufacturers' and wholesalers' sales distributions provided by the
censuses, sales to "industrial, commercial, professional, and institutional
users"—the 1939 designation, termed "sales to industrial users" for brevity—
were taken to indicate nonconsumer use. Included in the coverage of this
designation were private institutions and governmental bodies. In addition to
sales to industrial users by manufacturers and wholesalers, interplant transfers by manufacturers were included in the measurement of nonconsumer use.
Nonconsumer use of mixed commodities was estimated in two parts. First,
the percentages that manufacturers' interplant transfers and sales to industrial
users constituted of manufacturers' total transfers and sales were calculated
for industries representing the individual commodities most closely. In some
instances the industry sales and transfer data were adjusted by commodity
data to make their coverage more representative, and thus to prevent bias.
Secondly, the percentages of manufacturers' total interplant transfers and
sales formed by their sales to wholesalers were multiplied by the percentages
of wholesalers' sales (adjusted for duplication) made to industrial users, with
the lines of trade given in the wholesale census having to be matched with
the commodities involved.*
The two percentages were then added and applied to the detailed commodity totals. For mixed commodities allocated by sales distribution data,
the unfinished portion thus represented sales by manufacturers to "industrial
users" either directly or through the channels of wholesale trade. The
allowance for nonconsumer use represented by wholesalers' sales to industrial
users could have been estimated and deducted later, but was handled at the
manufacturers' level to simplify the estimating procedure.
By using the sales distribution data, materials consumed data, and special
commodity information, it was possible to achieve fairly reliable breakdowns
for most of the mixed commodities. Nevertheless, by their very nature these
4. Wholesaling in this analysis is defined to cover service and limited function wholesalers,
manufacturers' sales branches and offices, assemblers, and, in 1929 and 1933, chain store warehouses (classified in retail trade by the 1935 and 1939 censuses). Agents and brokers, included
'n the census definition oi wholesale trade, are generally excluded; their operations are covered
by sales distribution and other data reported by their principals, manufacturers and wholesalers. Also to avoid duplication, the sales distributions were adjusted to eliminate sales to
other wholesalers from the total sales and sales distributions of the trade making them. For
most trades, wholesalers' reporting of sales distribution data to the census was substantial
but not complete; in each such trade the data reported by each type of wholesaling (service
and limited function wholesalers, manufacturers' sales branches, etc.) were blown up to total
sales reported by that type and added to give the sales distribution of the trade.
29:1692°—54- - 8



107

breakdowns were approximations, and undoubtedly were subject to errors.
As computed from the data shown in Exhibit 3, allocations from mixed commodities formed 44 percent of the estimated value of manufacturers' production of finished consumer commodities in 1939.
At this point the procedure of estimating consumer commodities had
arrived at the manufacturers' value of finished production for odd years
of the 1929-39 period. This was computed, with reference to Exhibit 3, by
summing the values of consumer commodities (1) assigned directly as finished, (2) allocated from the combined total of finished output of consumer
commodities and producer durables, and (3) allocated from the mixed
commodity total.
The detailed commodity figures were next combined into groups (in
general, those shown in table 30, Part V). Further steps in the commodity
flow procedure related to commodity group totals, rather than to individual
products within the groups. In these steps, the various sales, output, inventory, foreign trade, and mark-up data—whether for commodities, industries, lines of wholesale business, or types of retail store—were always
first combined so as to correspond best with the commodity groups.
For the even years of the period, the estimates of manufacturers' output
of finished consumer commodities were of necessity based on partial information, which was utilized for interpolation between the census-based
estimates. Nevertheless, the intercensal figures are believed to be of a fairly
high order of reliability.
With respect to the actual procedure, either manufacturers' production or
manufacturers' sales were interpolated, depending on whether the best
available interpolating data related to production or sales. In instances of
the latter, the interpolation was carried out, of course, at the state of the
estimating procedure indicated by line 4 of Exhibit 3—where manufacturers'
sales had been derived by subtracting changes in inventories from manufacturers' production.
From a wide variety of sources were assembled as many different interpolating series for each commodity group as were available. These series, representing various product and industry data, were tested against one another
by examination of coverage and of relative movement from one census year
to the next. Reasonably good intercensal interpolations were obtained for
every commodity group.
2. Manufacturers' production of finished consumer commodities.—-Following t h e
lengthy process of classifying the biennial census data and then making
interpolations for other years, estimates of manufacturers' production of
finished consumer commodities were available for all years of the period
1929-39. (The procedural exception noted above regarding interpolations of
sales may be mentioned again.) The estimates represented manufacturers'
output for export and consumer use.
3. a n d 4. Subtraction of change in manufacturers' inventories to derive manufacturers' sales.—Annual changes in the inventories of finished consumer commodities held by manufacturers were estimated and subtracted (algebraically)
from the production figures to obtain manufacturers' sales of finished commodities. Much of the 1929 census data, however, already represented manufacturers' sales, and no adjustment was necessary.
Changes in manufacturers' inventories of consumer products for the years
1937, 1938, and 1939 were estimated from values of finished product inventories reported in the manufacturing censuses for 1937 and 1939. These
censuses obtained beginning and ending inventories broken down into
finished product inventories and all other inventories—materials, work in
process, merchandise, etc. The ratios of finished product inventories to total
products of establishments in each appropriate industry were applied to the
individual commodity values to obtain beginning and year-end inventories
of finished consumer commodities for 1937 and 1939. These were summed by
commodity groups, and the resulting inventory totals were differenced to
obtain the annual changes for 1937, 1938, and 1939.
For the years 1929-36 the inventory adjustments were based on inventory
and sales data compiled by the Internal Revenue Service from corporate
income tax returns, as given in Statistics of Income—Part 2 and the underlying
"Source Book," containing further industrial detail. Inventory changes were
not actually calculated for this period. Rather, the procedure for translating
the commodity output data to sales was to (1) compute sales-production
ratios for the commodity groups for the year 1937; (2) extend them to earlier
years by similar ratios computed from the corporate industry data; and (3)
multiply the commodity group production values by the resulting ratios.
A limitation of the 1937-39 inventory-change estimates stems from the

108

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

fact that the census finished-product data on inventories were too broad in
soope for the purpose at hand, covering finished nonconsumer, as well as
consumer, goods. The corporate data used for the 1929-36 period were
much less satisfactory. They represented total corporate inventories (the
unfinished part generally being large and relatively volatile), and were
available in an industrial detail too limited to provide appropriate representation of many of the individual commodity groups.5 Little confidence can
be placed in the inventory adjustments for the earlier period. Fortunately,
however, they do not, for the most part, form an appreciable element of the
final consumer commodity values.
5. Producers' sales of finished nonmanufactured commodities.—Producers' sales of
foods reaching consumers without undergoing manufacture were added at
this stage. Since the initial basic data used represented sales, rather than production, no inventory adjustment was required.
Statistics on agricultural products, available annually, were secured from
the Department of Agriculture. Because substantial amounts of fruits, vegetables and other farm products are used in the manufacturing process, the
estimates of farmers' cash receipts from marketings had to be allocated
between finished and unfinished. The allocations for the many individual
products—believed, in general, to be fairly reliable—were derived largely
from data of the Department of Agriculture.
Nonmanufactured foods also included the products of commercial fisheries.
Estimates of the value of edible fish other than that canned, dried, or otherwise preserved (already covered under manufactured foods) were derived
chiefly from compilations of the Fish and Wildlife Service.
6. Producers' sales of finished commodities.—With the addition of producers'
sales of nonmanufactured foods to manufacturers' sales, there was obtained
a complete measure of finished commodity sales (f. o. b. at producers' prices)
for export and consumer use. In 1939, this measure accounted for 61 percent
of the final market value of consumer commodity purchases estimated by the
commodity flow method. Manufactured commodities represented 53 percent
of that value, and nonmanufactured foods 8 percent.
7. Aidition of transportation charges to producers' sales.—The transportation
allowances added to producers' sales covered transportation from producer
to the first buyer, whether retailer, wholesaler, or consumer, and to port
of export. Costs of transporting consumer commodities beyond the first
buyer were assumed to be included in wholesalers' and retailers' markups.
While this treatment was substantially correct, it very probably resulted in
some incompleteness of coverage of the transportation cost element.
The estimates of transportation charges were based almost wholly on data
from the Interstate Commerce Commission. The procedure followed was to
multiply producers' commodity sales by annual ratios of freight revenue
per ton to value per ton at point of production for most nearly appropriate
Interstate Commerce Commission commodity classifications. Ratios were
calculated for 89 of the Interstate Commerce Commission's 157 commodity
classifications found to be related to one or more of the consumer commodity
classifications.
Freight revenue per ton of freight carried was computed for each year of
the 1929-39 period from freight revenue and tonnage data published annually
by the Interstate Commerce Commission. Values per ton at point of production were obtained from Interstate Commerce Commission publications
for the years 1928, 1930, 1933, 1936, and 1939. Data on value per ton for
other years were derived for the relevant Interstate Commerce Commission
commodity classifications by interpolation on the basis of 1928-39 price
indexes constructed principally from wholesale price series of the Bureau of
Labor Statistics.
For lack of data on other forms of transportation, rail freight charges
formed the sole basis of the specific transportation allowances in the commodity flow estimates except for a rough supplementation to cover trucking in
the case of nonmanufactured foods. Truck transportation is important for
many consumer commodities, and to the extent that truck rates differed
from rail rates the procedure would lead to error in the estimated cost of
moving the commodities from the producer to first buyer.
8. Distribution of producers' sales, inclusive of transportation charges.—The
distribution of manufacturers' and other producers' sales among exports,
wholesalers, retailers, and consumers is a vital step in the commodity flow

5. Failure to take account of the differential movement of inventories in the small noncorporate sector, for which data for the 1929-39 period are fragmentary, was probably only a very
minor source of error.



method. Producers' exports are eliminated at this point since they are accounted for, by a different methodology, in the net foreign investment component of the gross national product. The three remaining channels are
differentiated because varied markup treatments must be applied to the
finished consumer commodities flowing through them.
No markups are applied to producers' sales direct to consumers, which
become immediately a part of the consumption expenditure estimates.
Wholesalers' markups are applied to commodities sold by producers to wholesalers and then either exported or sold directly to consumers. Cumulative
wholesalers' and retailers' markups are required for consumer commodities
sold by producers to wholesalers and then sold by them to retailers. Lastly,
only retailers' markups must be added to sales made by producers directly to
retailers.
The distribution of manufacturers' sales, inclusive of transportation charges,
was estimated for 1929, 1935, and 1939 very largely from data provided in
the census reports on Distribution of Manufacturers' Sales. With two qualifications noted below, appropriate percentage distributions were derived directly
from the reported data for detailed industry groups and applied to the consumer commodity groups. For intercensal years, the derived census-year
percentages, which did not change appreciably, were interpolated along a
straight line. The distributions of other producers' sales—nonmanufactured
foods—were based chiefly on estimates of farmers' sales distributions made by
commodity specialists in the Department of Agriculture.
In two particular aspects, sales distributions as reported by the censuses
required modification. First, since wholesalers' sales to industrial users had
been eliminated in arriving at manufacturers' sales of finished consumer
products, census data on manufacturers' sales to wholesalers were reduced
by the proportions of wholesalers' sales to industrial users which had been
employed in estimating the nonconsumer-use portion of mixed commodities.
Secondly, exports required special estimation since the census sales distributions showed manufacturers' exports separately only in 1939. The export
estimates derived for 1939 by applying the industry sales distribution data to
the commodity values were extrapolated to the years 1929—38 by selected
commodity data on exports obtained from Foreign Commerce and Navigation of
the United States, published by the Bureau of Foreign and Domestic Commerce.8 The 1929 and 1935 sales distributions and the annual estimates of
producers' commodity sales were then adjusted to eliminate exports, and
the distribution estimates were made in terms of domestic sales to wholesalers,
retailers, and consumers.
It would appear that the most significant limitations of the estimated
distributions of producers' sales stemmed from the necessary application of
industry and line-of-trade data to product data. Although every effort was
made to achieve appropriate matchings of these data, the distributions for
some commodities could have been appreciably in error.
9 and 10. Addition of imports to produces' sales distributed to wholesalers to arrive
at total purchases by wholesalers.—Imports are a source of commodities sold to
consumers and must be added to producers' domestic sales. In the commodity
flow process imports are handled entirely as purchases by wholesalers. In the
main, this accords with fact, as only a minor portion of consumer product
imports is made by manufacturers, by retailers, or by consumers directly.
Imports were estimated for the various consumer commodity groups for the
years 1929—39 from annual commodity data on "imports of merchandise for
consumption" (with "calculated duty" added) published in Foreign Commerce
and Navigation of the United States by the Bureau of Foreign and Domestic
Commerce. Nonconsumer use was not allowed for explicitly—that is, by
carrying out the sort of allocation process employed in classifying the domestic output of manufactured products. Instead, commodities shown in the
import statistics were selected by inspection to correspond with the various
consumer commodity groups, and reliance was placed on compensating
errors of inclusion and omission.
The estimated value of consumer imports, amounting to less than 2
percent of the final value of consumption commodities in 1939, was somewhat understated. Because of the nature of the basic data, the estimates
omitted ocean transportation charges from country of export to the United
States. Also, they did not cover handling, transportation, and related
charges between port of import and purchaser, although this factor is
minimized by the fact that to a considerable extent importers are located
in ports of import.
6. The derivation of the export series is discussed in step 15, deduction of wholesalers'
exports.

NATIONAL INCOME, 195 4 EDITION
11 a n d 12. Subtraction

of change in wholesalers' inventories to obtain cost of

goods sold by wholesalers.—The next stage in the commodity flow procedure
calls for estimation of changes in wholesalers' inventories of finished consumption commodities. They are subtracted from wholesalers' purchases
in order to derive the costs of commodities sold. To these costs can then be
added wholesalers' markups to arrive at their sales of consumer commodities.
The general procedure for estimating changes in wholesalers' inventories
of finished consumer commodities was to prepare inventory-cost of goods
sold ratios and to apply those ratios to the annual estimates of wholesalers'
purchases.' The procedure may be outlined as follows.
(1) Ratios of wholesalers' year-end inventories at cost to annual sales
(adjusted for duplication) were computed for the years 1929, 1933, 1935,
and 1939 from data contained in the Census of Wholesale Trade.
As the first element of these computations, total inventories reported by
type of operation—service and limited function wholesalers, manufacturers'
sales branches and offices, and, in 1929 and 1933, chain store warehouses—
in each relevant line of trade were prorated among sales to other wholesalers, sales to industrial users, and other sales (exports and sales to retailers
and consumers). The data did not permit of any other apportionment of
reported inventories, but it is believed that the assumption of a uniform
rate of inventories to sales within a single trade and type of operation was
substantially correct.
Inventories held against sales to other wholesalers were next prorated
between sales to industrial users and other sales. This was done by commodity groups, combining all trades and types of operation assigned to each
group.
In combining trades and types of operation to obtain rates of inventories
to sales for the consumer commodity groups, the weights used were the
appropriate census sales totals—exports and sales to retailers and consumers,
plus the prorated portion of sales to other wholesalers covering these categories.
(2) The ratios of inventories to sales for the census years 1929, 1933,
1935, and 1939 were converted to ratios of inventories to cost of goods sold
oy use of wholesalers' markup ratios (described below). With S = sales, M =
markups, C=cost of goods sold, and I=inventories, then C-f-M = S and p
, I /
,M\
was computed as 5 ( 1 -f—g- ) •
(3) The ratios of inventories to cost of goods sold so derived were multiplied
by wholesalers' purchases to obtain year-end inventories of consumer commodities for census years.
(4) In the estimation of year-end inventories for intercensal years, one
of two procedures was followed. For some commodity groups, the censusbased inventory-cost of goods ratios were interpolated and multiplied by
purchases. But the more general procedure followed, because of the nature
of available data, was to interpolate the census-year inventory figures directly.
The information on wholesalers' inventories by line of trade available
for interpolations was quite limited. The wholesale data used were sometimes not directly representative of the individual commodity groups, and
in many instances it was necessary to resort to use of retail inventory data.
For the years 1929-34, department-store inventory data from the Controllers' Congress of the National Retail Dry Goods Association were most
generally used for interpolation. For the 1935-39 period, the principal
interpolating data were sample series on wholesalers' closing inventories
compiled by Dun and Bradstreet and by the Bureau of the Census from its
monthly Wholesale Survey. For commodity groups for which other information was lacking, use was made for all or part of the 1929-39 period of
corporate wholesale trade data from Statistics of Income. These related, for
all wholesale corporations combined, to the ratio of total inventories to
cost of goods sold.
13 a n d 14. Addition

of wholesalers'

markups

to obtain wholesalers'

sales of

finished commodities.—Wholesalers' markups form an appreciable element
of the final market value of consumer commodity purchases, amounting
to 8 percent in 1939.
In the estimation of wholesalers' markups, ratios of operating expenses
to sales (adjusted for duplication) in 1929, 1933, 1935, and 1939 were de7. In deriving inventories, inventory-cost of goods ratios should be applied to costs of goods
sold, not purchases. Application to the latter yields only first approximations of inventories,
which could have been adjusted. This was not done because the quality of the data did not
warrant the labor involved.




109

rived for each commodity group from wholesale census data for appropriate
lines of trade. Because of the lack of a relevant breakdown of the reported
operating expense figures, the same sort of procedure was required as that
applied to census inventory data in estimating changes in wholesalers'
inventories of consumer commodities. In brief, total operating expenses for
each relevant type of operation in each trade were first prorated among
sales to other wholesalers, sales to industrial users, and other sales (exports
and sales to retailers and consumers), after which the operating expenses
prorated to sales to other wholesalers were divided by commodity groups
between sales to industrial users and other sales. The weights used in combining selected lines of trade and types of operation into a single operating
expense-sales ratio for each commodity group were the relevant sales totals,
covering exports and sales to retailers and consumers and a prorated portion
of sales to other wholesalers attributed to these categories as against industrial
users.
Principal sources used to interpolate census-year expense ratios for
intercensal years were the series of wholesale surveys made by Dun and
Bradstreet and the 1941 report on Distribution Costs, An International Digest
of the Graduate School of Business Administration, Harvard University.
When appropriate wholesale data were lacking, the movement of the comparable group expense ratio for retail trade was used.
It was necessary to add profit ratios (ratios of profits to sales) to the operating expense ratios to obtain gross margin ratios. The available information
on wholesale profits by line of trade was scanty. Some of the special wholesale
surveys were helpful, but frequent use was made of gross margin-operating
expense ratios developed for comparable retail trade groups. And for some
commodities the selection of the profit ratios had to be largely arbitrary.
Still, it should be noted, even large errors in the profit element of wholesalers'
gross margins would have little effect on the consumer expenditure estimates.
The annual wholesale gross margin ratios were converted to markups by
use of the equation M = J J p

where M is the markup percentage of cost,

and G is the gross margin percentage of sales.
15. Subtraction of wholesalers' exports.—The basic commodity data, representing production or sales for both exports and consumer use, were adjusted
at the producer level for domestic nonconsumer use but not for exports.
Exports were deducted in part from manufacturers' sales of finished commodities. Remaining at this stage of tracing the commodity flow was the
elimination of wholesalers' exports from their sales.
Wholesalers' exports were derived by deducting previously estimated
manufacturers' exports from an independent estimate of total exports. The
totals were obtained by a process of commodity selection. The selection was
made from the detailed commodity classification of exports (valued at port
of exportation) given annually in Foreign Commerce and Navigation of the United
States. Where the export classification combined products included in the
commodity groups with other products not included, the omission or inclusion of the classification was based on rough appraisal of the relative
errors involved.
The procedures for deriving wholesalers' and manufacturers' exports may
lead to error on several scores. The error involved, however, is less than
would obtain from treating total exports as either manufacturers' or wholesalers' exports. This is because of the difference in markup treatment accorded
commodities eliminated from producers' sales and wholesalers' sales.
16. Distribution of wholesalers' domestic sales.—Wholesalers' domestic sales
of finished consumption commodities are broken down into sales to retailers
and sales to consumers. The latter, as final sales, are segregated, whereas
sales to retailers are still subject to inventory and markup adjustments.
The Census of Wholesale Trade provided the basis for the sales breakdowns
for 1933, 1935, and 1939. These censuses collected data on sales to retailers
for resale and on sales to consumers. Only limited use could be made of the
1929 wholesale census, which obtained information on sales to consumers,
but not to retailers, and in certain other respects was deficient for this purpose.
The sales distributions as reported in the 1933, 1935, and 1939 censuses
generally were somewhat incomplete. They were raised to reported total
sales by line of trade and type of operation (service and limited function
wholesalers, manufacturers' sales branches and offices, and, in 1933, chain
store warehouses). The resulting figures were summed for retailers and consumers, and the relationship of the two employed to break down the commodity estimates of wholesalers' domestic sales. The trades selected for

110

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

application to each commodity group were, to the extent possible, the same
as those utilized in deriving wholesale inventory and operating expense
ratios. Changes in classification of lines of trade were ratio-adjusted so far as
possible, but in some cases it was necessary to use 1935 retailer and consumer
proportions for 1933.
For some commodity groups, analysis indicated the appropriateness of
utilizing for 1929 the proportion of consumer sales to total sales less sales to
"industrial consumers" computed from census data. But for most groups it
was necessary to carry the 1933 or 1935 proportions back to 1929. In a few
instances, where fluctuations in the consumer and retailer proportions
appeared to be cyclical, 1939 proportions were used for 1929.
For intercensal years, percentage breakdowns of wholesale sales to retailers
and consumers were obtained (except where 1933 or 1935 percentages were
held constant back to 1929) by interpolating the census-based percentages
along a straight line. The smallness of the changes between census years
suggested that this procedure was not markedly in error.
17. Total purchases by retailers.—After the distribution of wholesalers'
domestic sales is accomplished, total purchases by retailers are secured by
summing wholesalers' sales to retailers and (from 8c) producers' sales to
retailers.
18 and 19. Subtraction of change in retailers' inventories to obtain cost of goods

sold by retailers.—Estimates of changes in retailers' inventories of finished
consumer commodities were computed from inventory totals derived from
a combination of two distinct procedures. (1) Tentative estimates were first
prepared, for individual commodity groups, by a procedure paralleling that
used in making the wholesale inventory adjustment. (2) Modifications of
these estimates for certain years were made after comparing the relative
movement of the commodity inventory totals with that of the series on yearend book values of inventories in retail trade derived in the process of calculating tht inventory component of the gross national product. (See section
on Change in business inventories.)
(1) Year-end inventories of consumer commodities for census years were
derived by (a) computing ratios of retailers' year-end inventories at cost to
annual sales for the appropriate type-of-store classifications, (6) converting
these ratios to ratios of inventories to cost of goods sold by applying retaileis'
markup ratios (described below), and (c) multiplying the inventoiy-cost of
goods ratios by the estimated retailers' purchases of the various commodities.
Inventory change estimates by commodity groups for other than census
years were arrived at by differencing totals computed either by interpolating
the census year inventory-cost of goods ratios and multiplying them by purchases, or by interpolating the census year inventory figures directly. In the
former procedure, the most important source used was the Internal Revenue
Service tabulations of inventories and cost of goods sold from retail corporations' income tax returns. The latter procedure of direct inventory interpolation utilized mainly the department store data published by the Controllers' Congress of the National Retail Dry Goods Association in Depart-

The estimates of retailers' markups in the consumer commodity series,
like those of retailers' inventories, represented an integration of the results
of two methods. (1) Estimates were first prepared for individual commodity
groups, and (2) these were modified for the years 1930-38 on the basis of
relevant retail trade industry data.
(1) The first step involved the estimation of retailers' markups in a manner
similar to that employed for wholesalers' markups. Census-based ratios of
operating expenses to sales were interpolated by noncensus source materials,
raised on the basis of these materials to gross margin ratios by the addition
of profit and loss allowances, and then converted to markup ratios.
Operating expenses as a percentage of sales for comparable types of stores
most closely related to the various commodity groups were derived for
1929, 1933, 1935, and 1939 from the retail censuses. For 1939, the census
reported only payrolls; allowances for other operating expenses were based
on the 1935 relationship of total operating expenses to payrolls. Since
only the 1933 census included a satisfactory allowance for the services of
proprietors and firm members of unincorporated establishments, a similar
adjustment to the expense data had to be made for the other census years.
This was done on a basis comparable with that for 1933.
Data from numerous sources were used to obtain, by interpolation,
operating expense-sales ratios for intercensal years. An important source
providing commodity data was the annual reports on Departmental Merchandising and Operating Results of Department Stores and Specialty Stores pub-

lished by the Controllers' Congress of the National Retail Dry Goods
Association. Studies made by Dun and Bradstreet, the Federal Trade
Commission, the Harvard University Graduate School of Business Administration, and various trade groups provided additional ratios for many types
of stores. These various sources also provided the basic information on profitsales relationships necessary to translate the operating expense ratios into
gross margin ratios.
Several factors bearing on the validity of these estimates may be noted.
First, it is likely that errors stemming from the use of type of store data to
estimate retailers' markups for commodities have a tendency to be offsetting,
or compensating, in the total—a type of assumption which is much less
valid when data on manufacturing industry or wholesale line of trade are
applied to commodity data. This is because the preponderant part of retail
sales is made to consumers; that is, the nonconsumer element of sales is less
in retail trade than in wholesale trade or manufacturing.
Running counter to this consideration, however, is the fact that for all
years of the 1929-39 period the data for estimating the volatile profit-andloss component of retailers' markups for individual commodity groups were
partial and generally inadequate. Moreover, the available information for
making the interpolations of operating expense-sales ratios between census
years was far from comprehensive.
These several considerations were taken to indicate that the commodity
totals of retailers' markups were probably not satisfactory in cyclical movemental Merchandising and Operating Results of Department Stores and Specialty
ment over the period.
Stores. The departmental breakdown of inventories given was detailed and
(2) This movement was checked against, and modified by, a series which
adaptable to the commodity groups. This source could be used only for the
may be termed "adjusted income originating in retail trade." It was derived
1929-34 period, as the inventory figures after 1934 were presented as stock
by two types of adjustment of the National Income Division's 1929-39 estiturns, unsuited to estimation of year-end inventories. For the later years of
mates of income originating in retail trade and automobile services. The first
the period, inventory series from a variety of sources, chiefly Dun and
consisted of subtracting the inventory valuation adjustment and adding
Bradstreet, were utilized.
depreciation, gross rental payments, and indirect business taxes.
(2) The relative movement of the total of commodity group inventories
The resulting series, while more nearly representing retail gross margins,
derived by the foregoing procedure was similar to that of the retail trade
was too broad in scope for the purpose at hand. It included certain lines of
industry data in the 1929-33 period but diverged in the subsequent period.
trade not covered by the commodity flow estimates (such as motor vehicle
The industry data were used to interpolate between the census-based comdealers, service garages, and filling stations) and certain others in which the
modity totals for 1933, 1935, and 1939. The differences between the revised
nonconsumer element was large (such as lumber dealers). Separate income
totals and those obtained in step (1) as the summation of commodity groups
data for these various lines of trade were not available. Accordingly a corwere prorated among the latter.
rection to reduce the scope of the adjusted income total was made by use of
operating expense data in the following manner.
This over-all check and modification on the basis of the retail industry
estimates was indicated by the weakness of the available data for interpoFor the census years 1929, 1933, 1935, and 1939 total operating expenses
lating census-based estimates for many of the commodity groups. This type
for retail trade and automobile services were divided into operating expenses
of check had not been feasible in the case of the wholesale inventory adjustexclusive of the various lines of trade noted above. These ratios, with straightment because of the large role of unfinished commodities in the operations
line interpolations for intercensal years, were then multiplied by the adjusted
of the wholesale trade industry.
income originating totals for retail trade and automobile services.
20 and 21. Addition of retailers' markups to arrive at retailers' sales of finished
The income series obtained by this second adjustment was used to intercommodities.—Retailers' markups are a very sizable element of consumer com- polate between the 1929 and 1939 estimates of retailers' markups obtained
modity purchases estimated by the commodity flow method. In 1939 retail
by summing the individual commodity groups. The relationship of 1939 to
markups formed 28 percent of the final consumer value.
1929 was similar in the two series, but the movement within the period




NATIONAL INCOME, 19 5 4 EDITION
differed appreciably, the income series showing a wider amplitude. The
absolute differences between the revised estimates of total retailers' markups
for the years 1930-38 and the original estimates obtained as the summation
of individual commodity groups were prorated among them.
This adjustment, though rough, probably improved both the estimates of
retailers' markups in the consumer commodity series and the statistical consistency between national income and gross national product. A similar
adjustment was not made for wholesalers' markups because of the greater
importance of unfinished products in the wholesale trade industry.
22.

Consumer commodity purchases exclusive of general retail sales

taxes.—The

final market value of consumer commodity expenditures, except for the
addition of general retail sales taxes, is arrived at by summing sales of finished
commodities to consumers by producers (8d), wholesalers (16b), and
retailers (21).
Exhibit 4.—Derivation of Total Personal Consumption Expenditures Estimated
by the Commodity Flow Method, 1947
[Millions of dollars]
Distribution of finished and mixed manufactured commodities, before deduction of government purchases of producers' durable equipment
a. Finished..

81,993
42,840

(1) Producers' durable equipment,.
(2) Consumer commodities
(3) Combined, allocated to

10,076
29,053
3,711

(a) Producers' durable equipment.
(b) Consumer commodities.._
b. Mixed, allocated to..

734
2,977
39,153

(1) Producers' durable equipment.
(2) Consumer commodities
(3) Unfinished

3, 657
18. 494
17,002

Manufacturers' shipments (or production) of finished consumer commodities
(la(2)+la(3b)+lb(2)]
Subtract: Changes in manufacturers' inventories

50, 524
187

Equals: Manufacturers' sales of finished commodities

60, 337

Producers' sales of finished nonmanufactured commodities

8,550

Producers' sales of finished commodities (4+5)

58,887

Add: Federal manufacturers' excise taxes

2,592

Add: Transportation charges

_

1,696

Add: Imports

_

975

Subtract: Changes in wholesalers' inventories

258

Add: Wholesalers' markups

_.

6,699

Add: State excise taxes paid by manufacturers and wholesalers..

424

Subtract: Exports

2,613

Equals: Sales to retailers and consumers

68,402

a. Producers' sales directly to consumers
b. Wholesalers' sales directly to consumers-.
c. Retailers' purchases (14—14a—14b)
15. Subtract from retailers' purchases: Changes in retailers' inventories.
16. Equals: Cost of goods sold by retailers

_

3, 346
1, 325
63, 731
1,037

__

17. Add: Retailers' markups

25,886

18. Equals: Retailers' sales of finished commodities

88. 580

10. Consumers' purchases exclusive of retail taxes (14a+14b+18)

93, 251

20. Add: Federal retail excise taxes

521

21. Add: General retail sales taxes

997

22. Equals: Personal consumption expenditures estimated by the commodity flow
method

94, 769

23 and 24. Addition of general retail sales taxes to obtain personal consumption
expenditures estimated by the commodity flow method.—General retail sales taxes

and part of the alcoholic beverage taxes levied by State and local governments were the main taxes not covered by the trade markups in the 1933—39
period. Fiscal year data for that period were obtained from the surveys of
State and local government finances by the Bureau of the Census and shifted
to a calendar year basis. Adjustments of the tax totals to make them conform
in scope to the consumption expenditure estimates were made by using tax



111

collection data, by type of store, reported for a small number of sample
States. The adjusted tax totals then were prorated by the expenditure estimates among applicable commodity groups, with allowance for the exemption of food in certain jurisdictions.
After arrival at thefinalfigures,inclusive of taxes, the individual commodity
groups were classified as durable or nondurable, as shown in table 30 in
Part V. Durable commodities were generally defined as those having an
average life of three years or longer.

Commodity Flow Method, 1947
The commodity flow method used in making the benchmark estimates for
1947 followed the general approach outlined above for 1929-39. Basic values
were those of the 1947 manufacturing census, classified and allocated product
by product. These products, together with allowance for nonmanufactured
foods, were combined into consumption commodity groups and then raised
from producers' to consumer expenditure levels by adding estimated taxes,
transportation costs, and trade markups with lesser adjustments for exports,
imports and inventory changes. In the estimation of markups and inventory
changes, the 1948 wholesale and retail censuses were utilized, together with
retail margin data obtained from a sample of Federal income tax returns
for 1948.
The central procedure which gives the commodity flow method its name,
however, had to be modified because of the omission of sales distributions
from the 1947 manufacturing census. In the procedure for the prewar series,
as already described, manufacturers' sales to their various classes of customers
together with similar distributions of wholesalers' sales had been the basis
for tracing the flow of commodities through wholesale and retail channels,
as well as for estimating direct factory sales to consumers.
For wholesaling, it was necessary to give up the commodity flow approach
for 1947, since data for estimating wholesalers' purchases—the core of this
approach—were not available. The substitute procedure was to estimate,
by commodities, total wholesalers' markups and inventory changes and then
to determine the amounts applicable to consumers. This alternative, which
had been available at previous business censuses, may seem more direct,
but in actual practice (described below) proved complex and probably
less reliable.
For retailing, however, continuation of the commodity flow approach
proved possible. In the earlier estimates, retailers' purchases had been
derived by summing sales to retailers by manufacturers and sales to retailers
out of wholesale commodity flows—both unavailable for 1947 in the absence
of manufacturers' sales distributions. However, retailers' purchases of consumption commodities for that year could be obtained somewhat indirectly.
The basic commodity flow procedure with modifications dictated by lack
of manufacturers' sales distribution data yielded a composite total of sales
to retailers and consumers; and from this were then deducted estimates of
direct sales to consumers by (a) producers and (b) wholesalers. For the former
item, which accounts for only a small part of all consumer purchases, it was
necessary to follow the general practice of applying to 1947 manufacturers'
shipments percentages based on the 1939 sales distributions. Wholesalers'
direct sales to consumers were derived as part of the substitute procedure for
wholesaling noted above.
Exhibit 4 summarizes the 1947 commodity flow estimates. Following the
selection and allocation of basic product values, as well as minor adjustment
for manufacturers' inventory changes, sales offinishedconsumer commodities
by manufacturers and other producers appear as item 6. To these producers' sales were added manufacturers' and wholesalers' excise taxes (items
7 and 12), transportation charges (item 8), imports (item 9), and wholesalers' markups (item 11), while deductions were made for changes in wholesalers' inventories (item 10) and exports (item 13). Item 14 (derived from
items 6 through 13) consists of total sales to retailers and consumers. From
this total, direct sales to consumers by producers and wholesalers (items
14a and 14b) were deducted in order to derive retailers' purchases (item
14c). This served as the commodity flow base for estimating changes in
retailers' inventories (item 15) and—after this inventory adjustment to
obtain cost of goods sold by retailers (item 16)—for computing retailers'
markups (item 17). The addition of retail taxes (items 20 and 21) to the
total of items representing sales to consumers (14a, 14b, and 18) completed
the derivation of total consumer expenditures by the commodity flow
method (item 22).

112

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Except for those already discussed, procedural changes indicated by comparison of the 1939 and 1947 commodity flow exhibits were minor. Additional
taxes had to be estimated because the 1947 census excluded excise taxes.
Federal retail excise taxes, moreover, were adopted after 1939. State excise
taxes paid by manufacturers and wholesalers are shown separately in the
1947 exhibit (item 12), but in the 1939 exhibit were included as a matter of
convenience with wholesalers' markups (item 13). On the other hand, the
need for estimating manufacturers' inventory changes was narrowed to a
few industries, chiefly in food and clothing, by the fact that the 1947 census
generally collected shipments (sales) data and not production data, as in
the case of the 1939 census.
The effects of omitting manufacturers' sales distributions from the 1947
census were not limited to the procedural changes indicated by the exhibits.
They extended also to the underlying product allocations summarized in
item 1, Exhibit 4. In the earlier commodity flow work, data on sales to
industrial users had been employed extensively in allocating individual
commodities between consumer and nonconsumer uses. With such data
not available, it was necessary in the 1947 estimates to have recourse
largely to trade information and expert opinion concerning commodity uses.
Use was also made, as in the earlier period, of materials consumption data
collected in the census and of available commodity studies.
Following is a brief discussion of each of the steps in the 1947 commodity
flow procedure, as shown in the exhibit. This discussion is presented in light
of the preceding, lengthier description of the 1929-39 work, which contains
considerable information about the basic commodity flow procedure that is
not repeated here.
1. Distribution of finished and mixed manufactured commodities.—The problem
was to convert the product data in the 1947 manufacturing census, supplemented by other sources, into the consumer product base for the commodity
flow estimates. As in preparing earlier benchmarks, the first task was to
classify the full census list of commodities into the basic categories of consumer,
producers' durables, and unfinished—according to the definitions given in
the description of the 1929-39 procedure. For this, the 1947 census provided
data on about 6,000 separate products, compared with the 4,000 shown in
the 1939 census. Such additional detail serves to break up otherwise mixed
product items into products which can be assigned individually, thus reducing
the problem of allocation. It also strengthens the classification process in
respect to product items which may have appeared to belong entirely in
one use category but are shown by the additional census detail to include
products of another category or categories, which can then be properly
classified.
Another aspect of the 1947 classification of commodities concerned the
levels of consumer or nonconsumer use which were considered to require
allocation. Unavoidably, there are practical limits to the minor allocations
that can be made. Products subject to less important consumer use are
omitted altogether, while those with lesser degrees of nonconsumer use are
included fully. In the 1947 work, allocations were carried somewhat further
than had been attempted in earlier benchmarks.
The absence of manufacturers' sales distributions for 1947 has already been
noted as detrimental to the 1947 allocations. Except for a few foods, these
allocations had to be made without explicit assistance from the 1947 census.
The principal types of allocations employed in the 1947 estimates were as
follows:
(1) Adjustment of shipment values to allow for quantities used in further
manufacture—obtained from the 1947 manufacturing census for a few
products and estimated for others. Allocations of this type were employed
mainly in food, but were developed also for some other products. In the 1947
census, "net" shipment figures obtained for the important meat products
group represented a very considerable improvement over previously available
data, but the only other manufacturing use data reported were for sugar, the
bulk of wheat flour, and a few other food products in part.
(2) Deduction of purchases made by governments, institutions, physicians,
and others for nonconsumer use. These purchases in some cases were based on
reported information; in other cases, they were estimated from sample
surveys and other data. Important applications were in allowing for government purchases of food and clothing, for institutional purchases of food, and
for the nonconsumer use of pharmaceutical preparations and related
products. The allowance for business-use purchases of meals and beverages
was necessarily somewhat arbitrary.



(3) Use of sales distribution data as the principal basis for the allocation of
textile piece goods to consumers. The largest element in this allocation was
wholesalers' sales of piece goods to retailers and consumers, treated as the
consumer share of these products. Such sales were estimated in connection
with wholesalers' markups (see the description of item 11), by use of wholesalers' sales distributions from the 1948 census. Also included in this allocation were manufacturers' sales of piece goods to retailers—approximated by
applying manufacturers' distributed sales proportions for 1939 to their 1947
fabric shipments.
(4) Allocations in accordance with (a) expert opinion or (b) personal
judgment as to proportions of various uses. These types of allocations predominated outside of foods, clothing, and piece goods. In the "expert
opinion" category, use patterns were often suggested by industry or commodity specialists, obtained by them from trade papers, industrial studies, and
personal associations in the industries involved. Where such information was
not available, but allocation nevertheless required, recourse to judgment was.
necessary. This drew heavily on the 1939 percentage allocations, which
tended to be carried over to 1947 barring knowledge of changed conditions or
other considerations dictating different treatment.
Throughout the procedure of classifying manufacturers' shipments data
into the various use categories, the assignments and allocations made were
checked against those prepared by the Division of Interindustry Economics
of the Bureau of Labor Statistics in a task of generally similar nature. The
work of the BLS staff was of material assistance in this phase of the commodity
flow work.
2. Manufacturers' shipments (or production) of finished consumer commodities.—This is a simple summation of finished consumer products derived in step 1
through specific assignment, allocation of commodities combining consumption and durable equipment uses, and allocation of mixed commodities. The
equivalent producers' durable equipment items become the basis for coordinate estimates of expenditures for producers' durable equipment described
under that heading.
3 and 4. Subtraction of changes in manufacturers' inventories to derive manufacturers''
sales of finished consumer commodities.—This item of estimated changes in
manufacturers' inventories is much the smallest in Exhibit 4. The 1947 census
called generally for shipments (and interplant transfers) instead of production as in previous censuses. In the few industries for which production
data instead of shipments were obtained, it was necessary to make inventory
adjustments in order to derive product shipments. The changes in inventories
of finished products reported by these industries in the census were used to
estimate the adjustments.
5 and 6. Addition of producers' sales of finished nonmanufactured commodities to
obtain total producers' sales of finishei commodities.—Nonmanufactured products
entering the commodity flow estimates consist entirely of nonmanufactured
foods. As in the 1929—1939 estimates, data sources for nonmanufactured
foods in 1947 were the Department of Agriculture and the Fish and Wildlife
Service. Required allocations between fresh or unprocessed consumption and
use in manufactured foods were more largely available from published data
of the Department of Agriculture than for the 1939 and earlier estimates.
Combining nonmanufactured foods (item 5) with manufactured products
(item 4) completed the estimates of consumer products at producers' prices
shown in item 6.
7. Addition of Federal manufacturers' excise taxes.—These tax estimates were
introduced in the commodity flow procedure because the 1947 census
generally excluded excise taxes from shipment values. Internal Revenue
collections data were appropriately lagged to shift them to an accruals basis.
Where these adjusted excise taxes could not be assigned directly to a consumption commodity group, the general procedure was to use manufacturers'
shipments of the taxed products as the principal basis for any necessary
allocation between consumer and nonconsumer categories and apportionment
among commodity groups.
8. Addition of transportation charges.—Transportation charges were estimated
for 1947, as in 1939 and earlier estimates, by multiplying producers' commodity sales by ratios of freight revenue per ton to value per ton at point of production for most nearly appropriate Interstate Commerce Commission
commodity classifications. Changes in ICC classifications in 1947, however,
limited the applicability of the Commission's study for 1946—there was none
for 1947—of values per ton of products shipped. For many commodity
groups, it became necessary to estimate values per ton on a basis comparable
with the new ICC commodity classification for freight revenue per ton.

NATIONAL INCOME, 1954 EDITION

113

(2) Gross margins were estimated for merchant wholesalers, manufacturers' sales branches, and farm assemblers in each line of trade. Operating
expense ratios from the 1948 census were applied to the 1947 sales estimates,
and profits were added for merchant wholesalers and farm assemblers, by
use of profit ratios (ratios of profits to sales) based on "Statistics of Income"
data of the Internal Revenue Service and other available sources on wholesalers' profits.
Neither gross margins for manufacturers' sales offices nor profits for manufacturers' sales branches were allowed for, because these were believed to
have been covered by product values reported to the manufacturing census.
Commissions of selected types of agents and brokers, obtained by applying
commission rates from the 1948 census to the 1947 sales estimates, were
included in the gross margin calculations, while commissions of other types
were excluded on the grounds that they already had been covered in the
expenses of manufacturers or other wholesalers.
(3) Wholesalers' sales and gross margin ratios were adjusted for duplication—to eliminate sales to other wholesalers. Percentage distributions of
wholesalers' sales by classes of customers intbe 1948 census were used for this
purpose. In general, the nature of the adjustment was to reduce sales to a net
basis excluding sales to other wholesalers, while margins remained unchanged although at a higher rate to the reduced net sales.
The implicit assumption of such an adjustment was that the sales in question were made to other wholesalers in the same line of trade and type oi
operation. It was possible to improve upon this, however, in the case of
manufacturers' sales branches. Their sales to other wholesalers were assumed
to have been made to merchant wholesalers in the same line of trade, and a
proportionate share of their margins (actually, only operating expenses) was
transferred to margins of the latter.
(4) The estimates of wholesale sales and margins for lines of trade were
next transformed to a commodity basis. For sales, this was accomplished by
applying commodity percentages for each line of trade and type of operation
from the 1948 census to the estimated 1947 sales. Commodity margins were
derived by applying to the sales of individual commodities margin rates for
the lines of trade handling the major part of the commodity (separately by
types of operation). A single rate was selected if possible, but generally the
rates in two or more of the principal lines of trade handling the commodity
were averaged, using their sales of the commodity as weights. As a check on
the detailed estimates, the resulting margin rates on commodities were
compared with the margin rates of lines of trade in broad groups, and found
11. Addition of wholesalers' markups.—Estimating wholesalers' markups was to be in close agreement.
(5) As the final step preparatory to allocating wholesalers' margins, the
the area of most innovation in the 1947 benchmark procedure. As noted
commodity sales and margin estimates were broken down by customer classes
earlier, the lack of manufacturers' sales distributions for 1947 required a new
[exports, industrial users, retailers, and household consumers; other wholeapproach to wholesalers' markups.
salers had been eliminated in (3) and were no problem here] by use of
The principal problem in estimating such markups is to derive the volume
wholesalers' sales distributions from the 1948 census. Within each type of
of consumer commodities handled by wholesalers, because wholesalers'
operation, customer class percentages of lines of trade handling the major
markup rates cannot be applied to consumer commodity totals, but only to
part of each commodity were weighted by their sales of the commodity, and
the flow through wholesalers. Manufacturers' sales to wholesalers, estimated
the resulting pattern applied to total sales of the commodity. (Minor adjustfor individual consumption commodities from manufacturers' sales distriments were required to bring export sales thus derived from 1948 percentages
butions and combined into groups, had provided the solution in 1939 and
into conformity with actual 1947 exports.) Commodity margins were disearlier estimates. With the addition of imports and after inventory adjusttributed within types of operation in proportion to these customer class sales
ments, they had furnished the base to which wholesalers' markup rates were
of each commodity.
properly applied. The 1939 sales distribution patterns conceivably might
have been utilized again for 1947—in fact, were utilized in a limited way—
As noted, these estimates of the Bureau of Labor Statistics were the basis
but the degree of error that would have been involved was considered defifor deriving wholesalers' markups for the consumption commodity groups.
nitely greater than in the alternative adopted.
In general, the margins on sales to retailers and "household consumers" of
The key to this alternative 1947 procedure was allocation. Following the
selected commodities—as obtained from step 5—were classified as consumpderivation of wholesalers' sales and gross margins, the latter were broken
tion expenditures. In a few instances, where the customer class distributions
down by commodities. These commodity margins were then further broken
seemed inappropriate, other procedures were necessary. These included use
down between consumer and nonconsumer uses. In these breakdowns, or
allocations, of wholesale margins, it was practical to utilize the detailed
of the same pattern of allocations already made of manufacturers' shipments
estimates of wholesaling in 1947 by the Interindustry Economics Division of
of the products involved and development of markup estimates of the comthe Bureau of Labor Statistics. The preparation of these BLS estimates will
modity flow type on the basis of 1939 manufacturers' sales distributions.
be described below in five steps.
After this basic procedure of consumer allocation, it is to be noted, whole(1) Wholesalers' sales reported in the 1948 census were extended to 1947
salers' markups were still in terms of the wholesale commodity classification
by reference principally to the OBE published estimates of merchant wholeused in the 1948 census. Where this did not match the consumer commodity
saling. Although the extrapolating series were generally quite broad, the
classification of products, wholesalers' markups needed to be apportioned
extrapolation was carried out in full census detail with respect to lines of
among consumer commodity groups. This was based on producers' sales
trade (such as groceries, confectionery, and meats) and also types of operation
values of the individual products included in the wholesale commodity classes
(merchant wholesalers, manufacturers' sales branches, etc.). This was
under apportionment.
necessary for utilization of other census data available in similar detail.
The value-per-ton estimates were based, so far as possible, on quantities
and values in the 1947 manufactures census. Export statistics and other
shipping weight sources also were used. Due to data limitations, it generally
was not possible to develop values per ton for all products in a group. In such
cases, producers' sales of products for which values per ton could be obtained
were converted into tonnages, and the latter multiplied by freight revenue
rates for appropriate ICC commodity classifications. The ratio of freight
revenue to producers' sales from the summation of these individual product
items was applied to producers' sales for the entire commodity group to derive the transportation charges.
9. Addition of imports.—Imports for 1947—accounting for only about one
percent of the final value of consumption commodities—were handled in
virtually the same fashion as in the 1939 and earlier estimates. Consumer
products were selected from the full listing of commodity imports tabulated
by the Bureau of the Census. Consumer allocation of individual imports was
undertaken in only a few cases. "Calculated duty" figures, which were not
available for 1947 for the full product detail, were prorated among finer
categories of consumer imports. The area coverage of the import data
included offshore territories as well as the continental United States, but
adjustments to approximate the latter basis were made for the more important
foods involved.
10. Subtraction of changes in wholesalers' inventories.—In the estimation of
wholesalers' inventory change for each commodity group, the first step was
to compute the ratio of inventory change to sales for a single wholesale trade
selected as appropriate or representative. The selection was from among
trades included in the published merchant wholesaling estimates of the
Office of Business Economics—representing combinations of more detailed
trades in the wholesale census, on which the estimates are benchmarked.
Ratios of inventory change to sales of merchant wholesalers in 1947 from the
OBE series were modified to cover manufacturers' sales branches and farm
assemblers, for which 1947 estimates were not available, on the basis of
differences in ratios of beginning inventories (partly estimated) to sales in the
1948 census. These adjusted ratios of inventory change to sales then were
converted into ratios of inventory change to wholesalers' markups. For this
conversion, the markups by line of wholesale trade used were consistent with
the wholesalers' markups by consumption commodity groups (described
immediately below), to which the derived inventory change ratios were
applied to obtain commodity group estimates of wholesalers' inventory
changes.




114

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

12. Addition of State excise taxes paid by manufacturers and wholesalers.—As dis- the cost of goods sold, using for this purpose the retailers' markup rate for
tinguished from general retail sales taxes (item 20), also largely imposed by
each group described under item 17. Finally, the base of reference was
the States, these are taxes on specified commodities payable by manufacshifted from the cost of goods sold to purchases, the resulting inventory
turers or wholesalers, or upon purchase by retailers, and hence properly
change ratios being directly applicable to retailers' purchases (item 14c).
included in costs of commodities to retailers. Data on State tax collections
Computed inventory changes were subtracted, yielding the cost of merwere obtained from the Census Bureau's Compendium of State Governmentchandise (finished consumer commodities) sold by retailers.
Finances and adjusted to the 1947 calendar year by prorating 1947 and 1948
17. Addition of retailers' markups.—The estimation of retailers' markups
fiscal year collections according to the number of fiscal year months in
by applying markup rates, largely from selected lines of retail trade, to the
calendar 1947.
cost of goods sold in each commodity group followed earlier procedure in the
13. Subtraction of exports.—Consumer items were selected from the detailed main. However, retail operating-expense information was not collected
in the 1948 census, and the computation of markups was based instead
Census Bureau listing of domestic merchandise exports and were allocated
largely on data from a sample of Federal tax returns for 1948 tabulated
between consumer and nonconsumer use in the same proportions as proby the Bureau of the Census and the Internal Revenue Service. The followducers' sales of the products. Such allocation of exports—a minor departure
ing five numbered paragraphs describe briefly the construction of retail
from 1929-39 procedure, where exports were based only on a process of
markup rates from these sample data, obtained for 20 selected lines of retail
product selection—was made necessary by the fact that domestic use patterns
trade.
had been applied to producers' total sales which covered exports as well as
domestic sales. As in the case of imports, the area covered included offshore
(1) The markup was defined in the first instance as the difference between
territories, but adjustments were made approximating a continental United
total receipts and cost of goods sold as measured by merchandise bought
States basis for the principal foods involved.
for sale less inventory change. The cost of goods sold reported as such on
14. Sales to retailers and consumers.—Following the various adjustments the returns was not used because it might include wages, materials, and
(items 7-13) of producers' sales (item 6), there were available at this stage of
similar costs which generally belonged in the commodity flow retail markup
the commodity flow work estimates of producers' and wholesalers' sales of
If these other costs in the reported cost of goods sold were large enough to
finished consumer commodities to retailers and consumers. In Exhibit 4, the
suggest the presence of manufacturing activity which might have been
overall total (item 14) is shown broken down into three parts: producers'
included in the manufacturing census, however, the individual firm was
sales directly to consumers (item 14a), wholesalers' sales directly to conremoved from the sample. This factor was important only for the apparel
sumers (item 14b), and retailers' purchases (item 14c), derived as a residual.
store sample (used for the clothing markup).
The first two are final expenditures except for the later addition of retail
(2) The sample was stratified into four classes: single unit firms in three
taxes (in items 20 and 21), but retailers' purchases require further adjustment
annual sales classes, divided at $100,000 and $1,000,000, and multiunit
for retailers' inventory changes (item 15) and retailers' markups (iteml7).
firms with four or more units. To obtain the average margin rate for each
line of trade, margin rates for these various classes were weighted by 1948
14a and 14b. Producers' and wholesalers' sales direUly to consumers.—For most
census sales data. (The sales of two and three unit firms, which had been
consumer commodity groups, manufacturers' direct sales to consumers were
omitted from the sample, were included in the sales weights for single unit
estimated by applying the same commodity group proportions as had been
firms.) The legal form pattern—a random element in the sample—was
used for 1939 (based on 1939 census product data and manufacturers' sales
found in the weighted sample sales figures to approximate closely that
distributions). For some groups (manufactured food, clothing, and a few
shown in the 1948 census.
others) for which suitable group proportions were not available from the
1939 estimates, direct sales estimates were prepared by applying detailed
(3) The margin rates thus obtained for individual lines of trade then were
manufacturers' sales distribution patterns for 1939 to 1947 data on products
adjusted to eliminate services. Service receipt percentages shown by the
in the commodity group. Manufacturers' direct sales of books and producers'
1948 census were used in this adjustment, the effect of which was to deduct
direct sales of nonmanufactured foods, however, were approximated entirely
service receipts in equal amount from both sales and margins, and reduce
from 1947 data. The manufacturers' and other producers' direct sales estithe margin rate accordingly. The assumption underlying this adjustment
mates, it may be added, include proportionate amounts of Federal manuwas that service costs had been included entirely or largely in margins as
facturers' excise taxes and transportation charges.
based on the original sample data.
Wholesalers' sales directly to consumers represented largely selected com(4) A further adjustment in these margin rates was made for sales to other
modity sales by wholesalers to household consumers from the wholesale
retailers. Such sales introduce a dupJication in sales but not in margins, and
estimates already described in reference to wholesalers' markups (step 5 of
thus depress margin rates for present purposes. To correct for this, sales
item 11). In the few instances in which this procedure did not yield the
were reduced to a net basis by 1948 census percentages of sales to other
allocation required for measuring personal consumption, resort was had to
retailers, and margin rates advanced accordingly. Such adjustment necesthe allocation patterns previously established for producers' sales. Where
sarily assumed the other retailers to have been in the same line of trade,
wholesale commodity classes did not match consumer commodity groups,
possibly entailing some errors which, however, should tend to offset in the
moreover, direct sales in these classes were apportioned on the basis of
aggregate.
producers' sales of individual products included in them.
(5) Finally, since the sample data were for 1948, it was necessary to
14c. Retailers' purchases.—As previously indicated, retailers' purchases were extrapolate the results for individual trades back to 1947. This was done on
obtained as a residual, representing purchases from producers and wholethe basis of margin rates computed from regularly published corporate
salers and also presumably some direct purchases of imports. They were the
returns to the Internal Revenue Service, departmental margin rates of the
base for calculation of retailers' inventory changes, and after inventory
Controllers' Congress of the National Retail Dry Goods Association, and
adjustment became the cost of merchandise sold to which retailers' markup
margin data of various other trade associations.
rates were applied.
In addition to the sample data, the sources just noted also were employed
15 and 16. Subtraction of changes in retailers' inventories to obtain cost of goods in the estimation of retail markup rates for some commodity groups. In the
sold by retailers.—Ratios of inventory change to sales in 1947 were obtained
case of the Internal Revenue statistics, however, proprietorship markup
for lines of retail trade from regularly published retail estimates of the Office
data were not tabulated for 1947 and had to be projected from 1945 by the
of Business Economics. To derive ratios for consumer commodity groups,
movement of corporate or partnership rates. The markup rates thus comthe inventory change ratios of the principal lines of trade handling computed for 1947 for certain lines of retail trade were weighted by 1948 census
modities in each group were averaged, using as weights the sales of these
sales of these legal forms of organization, and the same adjustments described
commodities by lines of trade as shown in the 1948 census. (This procedure
in (3) and (4) immediately above were then applied. In the case of the
was in contrast with the reliance in the 1929—39 estimates upon the indepartmental data of department and specialty stores published by the
ventory experience of single lines of retail trade selected for each group.)
Controllers' Congress, "workroom net costs" were added to the indicated
gross margin rates before conversion to a markup basis.
The inventory change ratios for commodity groups thus derived for 1947
were not directly usable because they related to sales rather than purchases.
For each commodity group, when appropriate, the markup rate of a single
However, they were readily transformed into inventory change ratios to
line of retail trade or department (in the department store data) was selected.




115

NATIONAL INCOME, 195 4 EDITION
In other instances, rates were combined by use of producers' sales of individual products in the group as weights. To obtain retailers' markups for the
various commodity groups, the estimated markup rates were applied to retailers' cost of finished consumer commodities sold (item 15).
18 a n d 19. Derivation

of retailers' sales and consumers' purchases exclusive of

retail taxes.—The addition of retailers' cost of goods sold and markups constituted retailers' commodity sales to consumers (item 18). To these were
added sales made directly to consumers by producers and wholesalers (items
14a and 14b) to derive total consumer purchases before retail taxes (item 19).
20—22. Addition of Federal retail excise taxes and general retail sales taxes to
obtain total personal consumption expenditures estimated by the commodity flow method.—

Retailers' excise and general sales taxes were the last adjustments necessary.
Federal retail excise taxes, which were treated as entirely chargeable to
consumers, were based on Internal Revenue collections lagged two months.
The necessary apportionment among commodity groups was made on the
basis of manufacturers' shipments of the taxable commodities. State and
local general sales, use, and gross receipts taxes required somewhat more
extended treatment, in conformity with earlier procedure. Collections of
these taxes by States, California cities, and other cities with populations of
25,000 and over were obtained from annual reports of the Bureau of the
Census entitled Compendium of State Government Finances and City Finances,
and from City Sales Taxes in California and More City Sales Taxes issued by

the California State Board of Equalization. Processing of the data included
conversion of fiscal year collections to the 1947 calendar year in accordance
with thefiscalyear months included; elimination of estimated taxes on services, nonconsumer commodities, and consumption commodities not in the
commodity flow estimates; and apportionment of the remaining taxes
among consumer commodity groups pro rata, but with approximate allowances for certain commodity exemptions in some tax jurisdictions. Addition
of these retail taxes completed the 1947 estimates of personal consumption
expenditures by the commodity flow method, totaled in item 22.

Extension of Commodity Flow Benchmarks
As indicated in the introductory remarks to this section of the notes, the
commodity group estimates established for 1939 and 1947 have been interpolated and projected into the current period primarily on the basis of
relative movements in retail sales. Retailers' excise taxes and applicable
general retail sales taxes have been estimated separately and added, except
in cases where the retail extrapolators covered such taxes (mainly the Census
Bureau retail sales series beginning in 1951).
The principal retail sales series that have been used in the consumer commodity interpolations and extrapolations are (1) retail sales, by type of store,
prepared through 1951 by the Office of Business Economics and since 1951
by the Bureau of the Census, and (2) department store sales, by type of
department, compiled by the Board of Governors of the Federal Reserve
System. In addition, some specific use has been made of unpublished compilations of State sales tax data, sales data from trade associations and other
private organizations, quantity and price data, and Federal retail excise tax
collections.
(1) The Office of Business Economics retail sales estimates represented
interpolations and extrapolations of sales totals by type-of-store groupings
given in the 1939 and 1948 Census of Retail Trade. The annual interpolations
and extrapolations of these benchmarks were based primarily on sales tax
collection data from a group of 10—20 States, accounting for about onefourth to one-half of all retail sales, depending on the period and kind of
business. Other data used included the Federal Reserve Board department
store sales estimates; Internal Revenue Service income-tax tabulations showing sales of corporate and noncorporate retail businesses by kind of business;
and tabulations of reports from the Bureau of the Census constant sample of
independent retail stores corrected by the Office of Business Economics for
changes in the retail store population.
The Bureau of the Census monthly retail sales estimates, starting with
January 1951, are based on sales data for all large organizations and a sample
of other stores in sample areas. Since May 1953, the large organizations
covered have included all those operating 11 or more retail stores in 1948 and
all department stores with 1948 sales volume in excess of $5 million, while the
sample areas have numbered 230. Small stores in these areas are sampled in
a way providing for inclusion of new stores.



(2) The department store sales data, by major departments, of the Federal
Reserve Board are based on departmental data reported by about 365
independent department stores located in various cities throughout the
country and accounting for about 50 percent of the estimated sales of all
department stores, including the national chains. Not all stores report data
for all of the departments asked; consequently, the sample for individual
departments is not so comprehensive as that for the major departmental
groups.
Under the method of interpolating and extrapolating consumer expenditures for 1939 and 1947, estimates for a particular commodity group often
have been moved by sales of one or more relevant types of stores (on the
assumption that sales of store groups tend to reflect the movement of sales of
commodities). However, when a commodity group is not handled almost
exclusively by specialized stores, or when data for these stores were not available separately, the movement of total sales of the commodity has been
estimated from the experience of departments handling that line in department stores or by use of combined relevant data for specialized stores and
department stores. Weights for combining these series usually were derived
from the commodity distributions of sales by kinds of business in the 1939 and
1948 Census of Retail Trade. Since departmental sales by type of department
more nearly represent commodity sales, they appear to provide a more
appropriate index than those specialized stores whose sales cover more than
one commodity group. Consequently, the department store data have been
given more than their proportionate weight in certain instances.
A significant shortcoming of retail trade data in these applications is that,
with the exception of department store sales, they cannot be broken down
by commodity groups. The type-of-store classification found in the basic
data assigns a store to a single classification on the basis of the commodity
accounting for the principal part of its business. But food stores, for instance,
sell goods other than food, and food is sold also in other kinds of stores,
including drug and department stores. With diverse movements in the sales
of various commodities, use of sales data classified by type of store to move
commodity groups will give rise to errors. This is borne out by the differences,
shown earlier, between the extrapolated 1947 estimates and the 1947 benchmarks for the various commodity groups. The presumption, however, that
these errors tend to offset in the total is also borne out, since adjustment to
the 1947 commodity flow aggregate amounted to only 3 percent for an
extrapolation covering a period of eight years.
Another limitation of retail sales data for the extrapolation of consumer
commodities arises from the fact that consumers buy from producers and
wholesalers as well as from retailers. In the general absence of data for estimating purchases from producers and wholesalers separately, they have been
assumed to vary with retail sales. Errors stemming from this assumption,
however, would necessarily tend to be minor because of the relatively small
volume of non-retail purchases—about 5 percent—involved in consumer
expenditures.
In the interpolation and extrapolation of the commodity flow benchmarks
by retail sales, it has been necessary, but difficult, to make some allowance
for the changing proportion of business purchases at retail. This has been
restricted largely to the war period. Downward adjustments were made in
that period for some commodities to allow for the apparent increase in the
relative magnitude of business purchases. There was very little quantitative
information on which to base the adjustments.

Retail Valuation Method
The main alternative to the commodity flow procedure (with retail sales
interpolation and extrapolation) is to multiply by an appropriate average
retail price the estimated quantities of consumer commodities purchased.
This approach—the retail valuation method—was used for passenger cars
and gasoline and oil from 1929 to the present; "other" fuel and ice (except
fuel produced and consumed on farms) for 1929-39 and 1947, with estimates
for other years based on the movement of retail sales data; and for tobacco
products beginning with 1940. In both 1939 and 1947, the groups for which
the retail valuation method was employed comprised about one-eighth of
total commodity expenditures.

116
Passenger cars

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS
consultation with specialists in the Office of Price Administration, the consumer allocation was reduced to 50 percent during the January 1942—May
1945 period, except that for April and May of 1945 all new cars were counted
as producers' durables.
The allocation of the passenger car estimates must be regarded as an
uncertain element. Errors in the allocation would affect both consumption
expenditures and producers' durables, but not the gross national product
total.
(4) The 1947 benchmark for housing trailers, which was a commodity flow
estimate, was extrapolated back to 1945 (when postwar trailer production
began) and forward from 1947 by a series based on data from the Trailer
Coach Manufacturers Association, and representing trailer coach production
valued at prices derived from the distribution of manufacturers' shipments
by price class. Housing-type trailers were omitted from the earlier estimates
because of their relative unimportance.

Preparation of the consumer expenditure series on "new cars and net
purchases of used cars" was carried out in four steps: (1) multiplication of
data on number of new cars by an average retail price; (2) addition of gross
margins on used car sales ("net purchases of used cars"); (3) allocation of the
totals of new passenger cars (after a deduction for government purchases)
and of gross margins between personal consumption expenditures and
producers' purchases of durable equipment; and (4) addition of housing-type
trailer coaches.
(1) The Automobile Manufacturers Association reported for the years
1929-41 the number of passenger cars sold at retail and an average retail
price at the factory. The latter was raised to cover estimated transportation
costs and also incidental charges by retail dealers. State and local retail
sales taxes were not included in the price, but were added separately.
With the discontinuance by the Automobile Manufacturers Association
of this retail sales series, the number of new passenger automobiles was
Gasoline and oil
estimated for the years 1942-45 from Office of Price Administration rationing statistics. Average prices for this period were obtained by extrapolation
For the period 1929—39, estimates of purchases of gasoline and oil—prior
of the 1941 price by use of data based on O. P. A. price regulations.
to consumer allocation—represent the sum of separate series for gasoline and
For the years 1946—50, the number series was new passenger car registraoil, each obtained by multiplying quantities by an average price.
tions compiled by the R. L. Polk Co., with a small upward adjustment for
The quantities of gasoline and lubricating oil used in passenger cars were
the estimated difference between registrations and dealers' sales. Beginning
estimates published by the Bureau of Mines in its Minerals Yearbook. The
with 1951, the number series on new passenger cars has consisted of dealers'
gasoline price was the 50-city service station average compiled by the
sales as reported to the Commerce Department's Business and Defense
American Petroleum Institute and published in American Petroleum Mews. The
Services Administration. For the period since 1946, the average retail price
50-city price excluding taxes was used, and these were estimated separately
of new cars has been based on the Bureau of Labor Statistics composite price
and added. Lubricating oil was valued at the Department of Agriculture
for Chevrolet, Ford, and Plymouth, adjusted upward to cover other makes
average of prices paid by farmers for medium grade motor oil.
of cars and factory installed extras.
For years subsequent to 1939, the procedure of estimating purchases of
(2) In the 1929—39 period the estimates of gross margins on used car sales
gasoline and oil has been changed somewhat. With the discontinuance of the
covered only used car dealers. Used car margins of new car dealers were
Bureau of Mines quantity series, the quantity of gasoline used in passenger
assumed to be covered by the average price series used in estimating purcars has been based since 1940 on estimates by the Bureau of Public Roads.
chases of new cars. The price series, being an average of list prices, was
The American Petroleum Institute average gasoline price inclusive of taxes
believed to have exceeded actual prices largely because it did not reflect
has been used. Finally, with the quantity series for lubricating oil no longer
excessive allowances for trade-ins. Used car margins of new car dealers were
available, a roughly equivalent quantity estimate was worked out for 1947,
introduced in successively larger proportions in 1940 and 1941, and were
carried back to link with 1939 and forward to 1953 by the movement of the
included entirely, along with the margins of used car dealers, from 1942,
gasoline quantity series, and valued in terms of the Department of Agriculture
when the full retail list price for new cars was assumed to have been realized,
average price series noted above.
through 1951. Beginning in 1952, a procedure of adjusting used car margins,
The figures so derived for expenditures on gasoline and oil were allocated
when necessary, for excessive trade-in allowances was adopted.
70 percent to personal consumption expenditures in the periods from 1929
Data for estimating gross margins on used car sales have been, for the most
through 1941 and beginning with 1946. During the war period, the following
part, inadequate. For the 1929-39 period, gross margins of used car dealers
percentages were used: 65 for 1942, 55 for 1943 and 1944, and 60 for 1945.
were based on retail census data on operating expenses and proprietors'
Errors stemming from this thinly based allocation, which was generally
compensation, with rough interpolations for intercensal years. The estimates
the same as that used in the passenger car series, would lead to error in the
for 1940-41 were obtained by extrapolating used car sales from the 1939
gross national product through their effect on consumer expenditures. The
retail census on the basis of sales finance company data and multiplying sales
business-use portion of gasoline and oil expenditures, unlike that of passenger
by gross margin ratios which had been adjusted upward moderately from
car purchases, is an intermediate product not included in the gross national
the 1939 gross margin ratio. Sales and gross margin ratios also have been
product, and therefore cannot counterbalance any error in the consumer-use
used for years since 1946. Used car sales from the 1948 retail census have
portion.
been extrapolated over this period by used car purchase estimates of the
Federal Reserve Board's Survey of Consumer Finances. Gross margin ratios have
Household fuels and ice
been developed from corporate tax returns of automobile and truck dealers
(1951 being the latest year for which this information is presently available
Expenditures for "other" fuels and ice (except fuel produced and confrom the Internal Revenue Service), partnership returns of automotive
sumed on farms) were estimated for the years 1929-39 and 1947 chiefly by
dealers for 1947, the Census Bureau—Internal Revenue Service sample of
multiplying quantity consumed data from the Bureau of Mines by retail
tax return data of used car dealers for 1948 (described above under commodprices selected from series maintained by the Bureau of Labor Statistics.
ity flow 1947 retail markups), and data of the National Automobile Dealers
Expenditures in other years represent interpolations and extrapolations on
Association for the period beginning with 1950. The used car margin estithe basis of Office of Business Economics estimates of retail sales of fuel and
mates for 1942-45 were interpolations on the basis of registrations of new
ice dealers. These are unpublished estimates developed from reported retail
cars and other cars weighted to provide a rough index of used car sales.
sales tax collections in 10-17 States.
The Bureau of Mines quantity data underlying the 1929—39 and 1947
(3) The totals of new passenger cars (after a small deduction for Federal,
estimates covered anthracite and bituminous coal, coke for domestic use,
State, and local government purchases) and gross margins on used cars were
fuel briquets, packaged fuel, kerosene, fuel oil No. 1 sold as range oil, distillate
allocated 70 percent to personal consumption expenditures and 30 percent
heating oils, and (in 1947) residual heating oils and liquified petroleum gases.
to producers' durable equipment except during the war period from January
The necessary allocations to consumer use were made after consultation with
1942 through May 1945. These proportions were derived mainly from surveys
specialists in the Bureau of Mines. The retail price data used to value conmade by the Bureau of Public Roads in 1934-37 to determine road use in
sumer quantities were obtained from the Bureau of Labor Statistics except
terms of mileage. The 70-30 allocation was also applied during the war
for kerosene, which was valued in part at average prices paid by farmers,
period to gross margins on used cars, since the distribution of used cars was
from the Agricultural Marketing Service. To these components derived
not controlled by rationing. For the new passenger car series, however, after




117

NATIONAL INCOME, 195 4 EDITION
mainly from data of the Bureau of Mines and Bureau of Labor Statistics were
added estimates for purchased firewood, based on Department of Agriculture data, and for ice. For the period 1929-39, the ice estimates were based
upon manufacturing census data marked up by an arbitrary 50 percent; for
1947, upon quantity and price data in the trade magazine Ice and Refrigeration.
In line with suggestions of the industry association, allocation to consumer
use was placed at 60 percent in the period 1929-39 and 20 percent in 1947.

None of the other miscellaneous items has been of sizable magnitude in
the period since 1929 exceptforexpenditures by U. S. Government personnel
abroad, which rose sharply during the war and reached a peak of $1.4
billion in 1945. This component is described in the section on Net foreign
investment.

Tobacco products
Expenditures for tobacco products for years after 1939 were derived from
quantities multiplied by average prices for each of three components: (1)
cigars, (2) cigarettes, and (3) other tobacco. The sum of the three series for
1939 agreed very closely with the commodity flow estimate for that year.
(1) Based on Internal Revenue Service reports, the number of large cigars
removed tax-paid (including those manufactured from both domestic and
imported tobacco) in each taxable price class was multiplied by estimated
average prices for the class. The estimated retail values of small cigars, taxfree large cigars, and imported cigars were minor special adjustments worked
out for 1947 and applied percentagewise to the cigar series throughout.
(2) The reported number of cigarettes removed tax-paid, lagged two
months, was valued at average retail prices of the Bureau of Labor Statistics
and the Agricultural Marketing Service. Purchases of tax-free cigarettes—
such as in PX's abroad and on merchant vessels—were estimated for 1947
and added for other years in the same proportion.
(3) The pounds of chewing and smoking tobacco and snuff removed taxpaid, lagged two months, were valued at average prices based on the pipe
tobacco retail price series of the Bureau of Labor Statistics and Agricultural
Marketing Service, with adjustment by 1947 manufacturing census data to
cover other forms of tobacco.

Other Methods
Imputation
Three imputed items, comprising 3 percent of consumption expenditures
for commodities in 1939 and 1947, are listed in Exhibit 1. Food and fuel produced and consumed on farms are discussed in the section on Income of
unincorporated enterprises. The third item, standard clothing issued to
military personnel, is covered in the section on Wages and salaries.
The consumer commodity estimates include one other imputation—
"food furnished government (including military) and commercial employees"—which is not shown in the breakdown of the estimates for 1939 in
Exhibit 1. An explanation of this omission may conveniently be used also
as a means of discussing an aspect of the commodity estimates not covered
in the notes to this point. That concerns the nature of the breakdown of the
large food group provided in table 30 in Part V.
For the 1929-39 period and again for 1947, food expenditures were estimated in total by the commodity flow method except for food produced and
consumed on farms, which is an independent imputation, and tips, which
were estimated and added separately. Purchased meals and beverages and
"food furnished" were components estimated for purposes of the breakdown
given in table 30, and food purchased for off-premise consumption was
derived as a residual by subtracting these two items from the commodity
flow food aggregate (exclusive of tips and imputed farm food). For the years
1940—46 and after 1947, however, these three types of food expenditures—
off-premise food, purchased meals and beverages (excluding tips), and "food
furnished"—represent separately estimated components of the food aggregate, with the last named assuming considerable importance as an imputation in some years.

Miscellaneous
The remaining, "miscellaneous" category, accounting for one percent of
consumer commodity expenditures in 1939, warrants only brief comment.
Expenditures for flowers, seeds, and potted plants and for lighting supplies
have been estimated primarily from retail censuses and other sources of
retail sales data, such as State tax collections. The tips series is described in
the notes on Wages and salaries.



8. PERSONAL CONSUMPTION EXPENDITURES FOR
SERVICES
In descriptions of national income methodology, resort is sometimes had
to the phrase, "constructed from a great variety of source materials." This
easy generalization is probably nowhere so apt as in the case of personal
consumption expenditures for services. The 55 service items for which expenditures are shown in table 30 of Part V are comprised of several hundred
separate series of estimates; and these represent the incorporation of numerous
types of data from many government and private sources, processed by
procedures virtually running the gamut of those used in national income
estimation. A primary factor in the detailed extent of estimation has been the
desire to take advantage of available sources of information, however piecemeal, and to minimize errors stemming from the estimation of broad components on the basis of data differing in scope or internal composition.
Perhaps a meaningful way of classifying the service estimates to facilitate
a general explanation of methodology is according to the broad types of
underlying statistical sources listed in Exhibit 7. This shows that for 1950
items comprising 23 percent of the $65.0 billion total were founded on comprehensive annual reports by government agencies and private sources;
that components forming 55 percent of the total were benchmarked on
periodic comprehensive sources, mainly the Censuses of Business and Population; that another 13 percent stemmed from sample information; and that
items accounting for 9 percent of all consumer service expenditures fell into a
"Miscellaneous" category.

Exhibit 1.—Personal Consumption Expenditures for Services, 1950
No sub- SubstanNo sub- Substanstantial tial allostantial tial alloTotal allocation cation Total allocation cation
problem problem
problem problem
Percent

Billions of dollars
14.8

13.5

22.8

20.8

2.0

Government agencies

5.8

4.5

8.9

6.9

2.0

Private sources

9.0

9.0

13.9

13.9

55.2

51.2

Comprehensive annual reports

1.3

2.6

4.0

35.9

33.3

Census of Population and Housing

21.9

21.9

33.7

33.7

Census of Business

10.1

7.5

15.5

11.5

Census of A griculture

1.4

1.4

°2

2.2

Census of Religious Bodies

1.6

1.6

2.5

2.5

.8

.8

1.3

1.3

Sample information

8.4

6.9

1.5

12.9

10.6

Miscellaneous

5.9

4.3

1.6

9.1

6.6

2.5

G5.0

58.0

7.0

100.0

89.2

10.8

Periodic comprehensive sources

__

Biennial Survey of Education..

Total

4.0

2.3

The exhibit affords some insight into aspects of both strength and weakness
of the estimates of consumer service expenditures. Thus, it is evident that—
contrary to an apparent general belief that statistical source materials on
consumer services are sparse and inadequate—a substantial body of data
exists for estimating this segment of the gross national product. Even so, the
reliability of the series is impaired by the necessary heavy reliance on periodic
(and sometimes rather infrequent) source materials. This consideration has

118

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

particular relevance to estimates—comprising one-half of the service expenditure total—which are benchmarked on the Censuses of Population and
Business. In the present edition of the NATIONAL INCOME supplement are
published revised estimates (adjusted to the 1950 Census of Population and
1948 Census of Business) replacing previous provisional figures based on
extrapolations over a 12- or 13-year period.
Another key factor bearing on the reliability of the consumer service
estimates is brought out in Exhibit 1—that components adding to only 11
percent of the total in 1950 constituted a "substantial allocation problem."
First to be noted in this connection is that expenditures for consumer services,
like those for consumer commodities, are generally estimated from data on
producers' sales. In the case of services, however, the estimating procedure is
simplified (and, hence, reliability is enhanced) by two facts: (1) There are
no intermediary enterprises between the producer and consumer; and (2) in
very large degree, consumer-type services are by nature items normally
purchased only by persons. The area of estimation centering in the allocation
of producers' total sales between consumers and others is relatively small.
To quantify this latter aspect of the consumer services series, and to arrive
at the information given in the exhibit, the 1950 estimates were classified
according to those (1) representing items purchased wholly, or very nearly so,
by persons and thus not requiring an allocation of sales between persons and
business and government; (2) representing items purchased in appreciable
degree by other groups as well as consumers, but for which reliable information on the nonconsumer element was available; and (3) requiring a sizable
allocation for which the available data, if any, for making it were regarded as
inadequate.
This classification—in some measure a matter of judgment—resulted in the
finding that in 1950 components amounting to 89 percent of consumer service
expenditures did not involve any substantial problem of allocating producers'
sales to arrive at consumer purchases only. More specifically, components
forming as much as 69 percent of the total were classified in the first category,
as being normally purchased only by persons and thus presenting no problem
in this regard. Items comprising another 20 percent were assigned to the
second category, for which the problem of statistical allocation was taken
care of adequately (although, for the most part, only periodically) in the
basic data. Individual series constituting only 11 percent of consumer services
fell into the third category with the label of "substantial allocation problem",
and may be viewed as subject, on this score, to sizable error.
The following description of the estimates of personal consumption expenditures for services is outlined according to the four types of statistical sources
shown in the exhibit. Probably no broad statistical classification of the numerous and heterogeneous service items would be entirely appropriate, and the
present one is no exception. Some items, though relatively few, did not fall
readily into any one grouping. The difficulty was usually resolved by assigning them to "Miscellaneous."

The annual reports of the Federal Communications Commission contain
statistics on telephone operating revenues and excise taxes. Additional
supporting statistics in these reports are used to reduce total revenues to an
estimate of consumer payments for telephone service. This allocation,
regarded as satisfactory, is accomplished by means of formulae devised with
the assistance of the American Telephone and Telegraph Company.
Exhibit 2.—Consumer Expenditures for Services Based on Comprehensive
Annual Reports by Government Agencies, 1950
Millions of
dollars

Item
Services furnished without payment by financial intermediaries except insur^
ance companies

2,028

Telephone '

1,943

Railway (including commutation) and sleeping and parlor car

525

Intercity bus

315

Airline

__

__

__

___

177

Pari-mutuel net receipts

239

Bank service charges, trust services, and safe-deposit box rental

341

Athletic and social clubs—dues and fees'

182

Money order fees3
Total

_
_

51
_

5, 801

1. Component of Telephone, telegraph, cable, and wireless, item V-9 in table 30, Part V.
2. Component of Clubs and fraternal organizations except insurance, item IX-9 in table 30,
Part V.
3. Component of Other personal business, item VII-7 in table 30, Part V.

Passenger revenues are compiled annually by the Interstate Commerce
Commission for railways and intercity bus lines and by the Civil Aeronautics Administration for airlines. However, the statistical basis for adjusting
these comprehensive revenue data to measures of consumer expenditures is
limited.
For the railroad series, travel fares paid by the Federal Government are
first excluded from passenger revenues. Little direct information on Federal
outlays for rail travel has been available except for that regularly furnished
the National Income Division by the Pullman Co. For years since 1942,
railroad coach passenger revenues from sources other than the Federal
Government have been estimated by an indirect, though fairly adequate,
method utilizing data on collections by the Government from the tax levied
on the various types of transportation of persons. Of total railroad revenues
from sources other than the Federal Government, in all years the allocation
to consumer expenditures has been placed at 65 percent for railroad coach
transportation (excluding commutation) and 40 percent for Pullman service
(including both seat and berth charges).
These percentages were drawn from a study made in 1942 by the Office of
Comprehensive Annual Reports
Defense Transportation, which furnished similar information also with
respect to intercity bus transportation. Passengers arriving at major railroad
Regular comprehensive reporting systems exist for most of the servi ce
and bus terminals during the survey week were interviewed to determine the
requiring large aggregations of capital to produce. In some cases, generally
purpose of travel. The tabulated results of the study required some adjustwhere the producing industry is subject to government regulation, the basic
ment so as to reflect the source (consumer or business) rather than the
data for estimating consumer service expenditures are furnished by Federal or
purpose of expenditure.
State agencies. In others, the data are provided directly by private industries
with excellent facilities for the collection of statistical information.
The Interstate Commerce Commission data on passenger revenues of
motor carriers from intercity schedules have been reported since 1937.
These were extended to earlier years by data on total revenue from operation
Government agencies
of intercity motor bus companies compiled by Bus Transportation magazine.
Over the period since 1929, receipts from Federal Government travel have
The service items for which comprehensive annual data—before deduction
been estimated at 2 to 5 percent of the total. Seventy percent of the remainder
of any allowance for nonconsumer use—are collected by Federal and State
has been apportioned as receipts from consumers, as indicated by the 1942
agencies are listed in Exhibit 2, together with the 1950 consumer expenditure
survey of the Office of Defense Transportation.
value for each.
The reported passenger revenue data for the airlines, covering the period
The estimates of services furnished without payment by financial intersince 1935, were extrapolated to 1929 by data on passenger mileage flown.
mediaries are developed mainly from data furnished by the Federal Deposit
For the years 1939-41, the consumer source of passenger revenues was
Insurance Corporation, Board of Governors of the Federal Reserve System,
estimated as one-third on the basis of sample data presented at hearings
and Internal Revenue Service (from tabulations of corporations' income
tax returns, as published in Statistics of Income—Part 2). This series is described before the Civil Aeronautics Board and collected in questionnaire surveys
conducted by several of the airlines. This proportion has been modified for
in the section on Interest. As pointed out there, the necessary allocations of
other years as indicated by scattered information relating to business and
reported totals to derive measures covering only consumers can be made on
government use of airlines.
a generally satisfactory statistical basis.



NATIONAL

INCOME,

In addition to these several transportation components, estimates of (1)
bank service charges, trust services, and safe-deposit box rental and (2)
money order fees lack adequate data for making the sizable nonconsumer
allocations involved in their derivation.
The series on pari-mutuel net receipts includes the tracks' and States'
shares of amounts wagered and of breakage. Data are obtained from the
Census Bureau, as summarized in State Finances, and from reports of the
individual State racing commissions. The data are entered as consumer
expenditures without any nonconsumer allowance.
Estimates of dues and fees paid to athletic and social clubs are based on
Federal tax collections data. Only very minor estimation, for the nontaxable
portion, is required.

Private sources
Comprehensive annual data are available from private sources for estimating the types of consumer expenditures for services shown in Exhibit 3.
Vfc*ualiy complete data relating to the two insurance items in the
listing are collected annually by the Spectator Company and published
in its Insurance

Yearbook.

Data on the income and expenditures of hospitals are collected in annual
censuses of these institutions by the American Hospital Association. These
data, with minor adjustments, have supplied measures since 1946 of personal
consumption expenditures on privately controlled hospitals and sanitariums.
This component consists of payments by patients to proprietary hospitals
and of current expenditures (including depreciation) of nonprofit hospitals
in providing care of patients.
Prior to 1946, however, considerable estimation was required in the
absence of comprehensive annual data. Benchmark figures were derived
for 1930 from a study by the Committee on the Costs of Medical Care and
for 1935 from the Census of Hospitals. The movement for other years was
estimated, for the most part, as the product of the average daily number
of patients, published by the American Medical Association, and the hospital
price component of the Bureau of Labor Statistics Consumer Price Index.
For the entire period covered by the estimates, the Edison Electric Institute and the American Gas Association have reported annual data relating,
in effect, to sales to consumers separately from sales to other customers.
Thus, data of the Edison Electric Institute on revenues from residential or
domestic sales and from rural sales at distinct rural rates have had to be
adjusted only to eliminate revenue from farm business use to furnish a
measure of consumer expenditures for electricity (inclusive of landlords'
expenditures). And a similar measure of expenditures for gas has been
secured annually from data reported by the American Gas Association on
revenues from sales to residential consumers.

1954

relatively small numbers of large and responsible business enterprises keeping
adequate accounting records. The estimates of consumer expenditures
derived from them, involving only minor problems of estimation, may be
presumed to be subject to comparatively little margin of error.

Periodic Comprehensive Sources
The Census of Population and Housing yields benchmark data for items
aggregating 34 percent of the consumer service total in 1950; the Census of
Business for components amounting to 16 percent. Other periodic comprehensive sources furnish the underlying data for items forming 6 percent
of the 1950 services total. The expenditure components based on these various
sources are listed in Exhibit 4.
For most items classified in this category, periodic universe values of
consumer expenditures can be estimated with considerable confidence.
The accuracy of the benchmark estimates is somewhat affected, however,
by the fact that many of the reporting units are individuals or small business
enterprises and tend to have less adequate accounting records. Also, estimating adjustments of the reported data often are necessary.
Exhibit 4.—Personal Consumption Expenditures for Services Based
Periodic Comprehensive Sources, 7 9 5 0

Source and item

Space rental value of owner-occupied nonfarm dwellings
Space rent of tenant-occupied nonfarm dwellings
Domestic service

Millions of
dollars

Expense of handling life insurance

2,035

Accident and health insurance and mutual accident and sick benefit associations—premiums less claims '
_

497

Privately controlled hospitals and sanitariums

1,975

Electricity

1,955

Gas.

1,177

Street and electric railways and local bus lines
Total

_

1,388

Automobile repair, greasing, washing, parking, storage, and rental...
Cleaning, dyeing, pressing, alteration, storage, and repair of garments including furs (in shops), not elsewhere classified.
_
_
Motion picture theater admissions
Barbershops, beauty parlors, and baths
Laundering in establishments
Funeral and burial service '
Maintenance services for appliances and house furnishings a
Transient hotels and tourist cabins'
Photographic studios and photo developing and printing *
Radio and television repair
_
_
Shoe cleaning and repair
Other'

2,138
1,466
1,394
1,049

854
642
437
' 403
269
324
201

Census of Agriculture:
Rental value of farm houses.

1,448

Census of Religious Bodies:
1,630

Biennial Survey of Education:
Higher education (private)

822
35, 911

1. Included with cemeteries and crematories and monuments and tombstones in Funeral
and burial expenses, item VI-8 of table 30, Part V.
2. Consists of care of electrical equipment except radios and of stoves, upholstery and furniture repair, and rug, drapery, and mattress cleaning and repair, and is a component of Other
household operation, item V-ll of table 30, Part V.
3. Component of Other housing, item IV-4 of table 30, Part V.
4. Component of Other recreation, item IX-12 of table 30, Part V.
5. Includes the following recreation items: billiard parlors, bowling alleys, dancing, riding,
shooting, skating, and swimming places, amusement devices and parks—components of item
IX-10 of table 30, Part V; "other" commercial amusements, component of Item IX-12; legitimate theaters and opera admissions, component of item IX-8b; professional baseball and
football admissions, components of item IX-Sc. Also includes watch, clock, and jewelry
repairs, miscellaneous personal services related to clothing, and costume and dress suit rental,
components of item II-8; employment agency fees, component of item VII-7; and miscellaneous household operation services, component of item V-ll.
6. Component of Religious and welfare activities, item X I of table 30, Part V.

9,027

i Component of Medical care and hospitalization insurance, item VI-7 in table 30, Part V.

The American Transit Association publishes annual data on passenger
revenue receipts from operations of street and electric railways and local bus
lines. To them are added estimated taxes on electric railway fares to obtain
the consumer expenditures series.
The comprehensive annual data supplied by the several private sources
noted above are gathered through well-established reporting systems from



12,195
7,062
2,668

Census of Business:

Total—.
Item

Millions of
dollars

Census of Population and Housing:

Religious bodies 8
Exhibit 3.—Consumer Expenditures for Services Based on Comprehensive
Annual Reports by Private Sources, 1950

119

EDITION

Only to a small degree are the estimates in this grouping impaired by
serious difficulties of consumer allocation. Items amounting to 4 percent
of the 1950 services total were classified as constituting a "substantial allocation problem."
A major factor bearing on the reliability of the series in this group is that
the availability of comprehensive data only periodically necessitates the
development of interpolating and extrapolating indexes. In a number of
cases, these indexes must be based on data that are partial and unrepresentative or of only indirect relevance.

120

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

It is thus to be expected that many of the individual series included in
Exhibit 4—mostly items not published separately in table 30, Part V—are
subject to appreciable error in relative movement, particularly for years
estimated by the projection of benchmark estimates over a considerable span.
Such errors, however, cannot be expected to be cumulative. On the contrary, it may be presumed on the basis of experience in estimating individual
components from masses of diverse data that errors in the individual series
will have a tendency to offset in the total. As the latest evidence on this
score, discussed later, recent adjustment of the relevant service expenditure
series to the 1948 Census of Business and the 1950 Census of Population and
Housing revealed an error in the extrapolating series amounting in the
aggregate to about 5 percent, with the range of error for the individual
series considerably larger.

Census of Population and Housing
The decennial Census of Population and Housing provides data on the
number of dwelling units in each tenure class, on the average monthly rent
of rented dwellings, and on the fair market value of owner-occupied units.
The census taken in 1940 also included a question on the rental value of
owner-occupied units. These basic data have been used to derive benchmark
estimates of the space rental value of owner-occupied nonfarm dwellings and
of space rental payments on tenant-occupied nonfarm dwellings. Their
derivation, along with the construction of estimates for intercensal years, is
explained in the section on Rental income of persons.
Expenditures for domestic service (cash payments and value of meals
furnished) are equal to wages paid in the Private households industry. These
wage estimates, based very largely on benchmarks derived from population
census data, are described in the section on Wages and salaries.

Census of Business
Service establishments were canvassed in connection with the 1933 Census
of American Business and the 1935, 1939, and 1948 Census of Business. In
addition, hotels (both year-round and seasonal) of 25 rooms or more and
most types of automobile servicing establishments were covered by censuses
in 1929. Power laundries and cleaning and dyeing plants, first included in
the Census of Business in 1939, were covered in conjunction with the Census
of Manufactures in odd years from 1929 through 1935.
Census coverage of service establishments was accepted as complete
throughout in making estimates for automobile repairs. It was necessary,
however, to make an estimated adjustment of the data on service receipts of
motor vehicle dealers given in the Census of Retail Trade for 1935, 1939,
and 1948, so as to include the value of parts used in making repairs—an
unidentified element of the reported sales totals. The census data on amusements for 1935, 1939, and 1948 also were taken as reported and used, with
the exception of motion pictures in 1935. The 1933 census data on amusements were used in only a few cases. Census coverage of hotels was considered
complete except for seasonal hotels in 1933, and for seasonal hotels in the
New England and Middle Atlantic States in 1939. (However, for 1929 it
was necessary to add estimates for hotels with less than 25 rooms, not canvassed by the Census.) Data on tourist courts were used as given by the
Census in all years.
For the consumer expenditure estimates, census data in the category of
personal services (apart from laundering and dry cleaning) were accepted as
reported except for 1948. For that year, moderate upward adjustments of
reported census figures were made in preparing estimates for barber shops
and beauty parlors, shoe cleaning and repair, and miscellaneous personal
services. These adjustments were based on information available from the
Census Bureau—chiefly results of its "Post Enumeration Survey", a sample
check on the completeness of the 1948 Census of Business—and on comparison of census data on wages and employment in personal service industries
with corresponding data reported under the social security laws.
In the preparation of consumer expenditure estimates relating to laundering
and dry cleaning, census data were used as reported except for 1933 and
1948. Upward adjustments of 8.7 percent and 17.1 percent, respectively,
were applied to all data for 1933 utilized from the Census of Power Laundries
and the Census of Cleaning and Dyeing Establishments. These adjustments
were based on information as to coverage reported by the Census Bureau.
For 1948, the census data were raised moderately before incorporation into



the service expenditure estimates—on a basis similar to that noted above for
personal services.
For most consumer service items based on the Census of Business, it has
been necessary to make some adjustments for differences in the definition
and type of data reported in the various years. These adjustments, based on
census data themselves, generally have been of minor magnitude.
Where appropriate, the estimates of consumer service expenditures derived
from the Census of Business generally represent the service receipts of retailers
(reported in the retail censuses) as well as of service establishments. Sales
data are tabulated by line of business, with a general distinction between
commodity sales and service receipts in each line. In general, the classification by line of business approximates the desired classification by type of
service. The instances in which the line-of-business receipts data could not
be taken to indicate expenditures by type of service have been relatively few.
The necessary allocations of receipts data in such instances did not affect
the consumer services total.
The series derived from Census of Business information that involve substantial problems of consumer allocation are as follows: automobile K5_pc:r,
greasing, washing, parking, storage, and rental; transient hotels and tourist
cabins; and miscellaneous household operation services.
The following summary for a number of the larger series based on the
Census of Business may serve to indicate more specifically the uses and
adjustments made of census data and the nature of the methods—sometimes
diverse and indirect because of the inadequacy of available source materials—that are employed for interpolation and extrapolation.
Automobile repairs
Expenditures for automobile repair, greasing, washing, parking, storage
and rental were based for 1929 (partly estimated), 1933,1935, 1939, and 1948
on Census of Business receipts data covering service and retail establishments.
As noted above, data on the service receipts of motor vehicle dealers were
raised to cover parts used in making repairs.
For the period prior to 1935, interpolations were made by the Motor and
Equipment Manufacturers' Association index of shipments of service parts
to wholesalers; for the period from 1935 to 1939, they were based on an index
of sales of parts and accessories stores computed from sales tax reports in
Illinois, Iowa, and Indiana. The estimates for years after 1939 represent
interpolations and extrapolations by an index based largely on payrolls in
the automobile repair services industry, as compiled by the Bureau of
Employment Security of the Department of Labor from summations of
employers' reports filed by the various State unemployment insurance
agencies. This index was extended from 1942, when the payroll data became
available, back to 1939 by a series representing the product of passenger car
gasoline consumption (from the Bureau of Public Roads) and the official
index on automobile repair prices (from the Bureau of Labor Statistics).
The estimates of total expenditures so derived have been allocated by the
proportion of consumer to total highway use of gasoline. This proportion
has been estimated as the product of ratios of (1) consumer to total passengercar use of gasoline and (2) passenger car to total highway use of gasoline.
The first ratio is the same as that used in the allocation of expenditures for
gasoline and oil (see section on Personal consumption expenditures for
commodities). The second has been derived from annual data of the Bureau
of Mines and Bureau of Public Roads.
Motion picture admissions
The 1939 and 1948 consumer expenditure figures for admissions to motion
picture theaters were taken from the Census of Service Establishments for
those years. Other years have been estimated chiefly by using gross receipts
from operation of motion picture theater corporations for interpolation and
extrapolation, with allowance for changes in the admissions tax, not reflected in the data reported to the Internal Revenue Service. For the two
most recent years of the series, when such data are not available, the estimates are made as follows: (1) total admissions to motion pictures and other
(largely spectator) amusements covered by the Federal admissions tax are
extrapolated by an index obtained by dividing the Federal admissions tax
rate into the total collections from this tax and adding the tax collections,
to the quotient, and (2) the independently estimated admissions to the other
amusements are deducted from the total to arrive at the figures for motion
pictures.

NATIONAL INCOME, 195 4 EDITION

Cleaning and dyeing
The consumer expenditure series on cleaning, dyeing, pressing, alteration,
and repair of garments is based on census values for the years 1929, 1931,
1933, 1935, 1939, and 1948 (adjusted for undercoverage in 1933 and 1948
as noted above). The census figures required numerous, but minor, adjustments for comparability from year to year. The problem of consumer
allocation was obviated by the type of data given in the census reports.
Most important in this respect was the provision of data on receipts from
cleaning and dyeing at retail (separately from wholesale) of cleaning and
dyeing plants, rug cleaning establishments, and power laundries. Estimates for intercensal years prior to 1935 were obtained by interpolation by
the American Institute of Laundering index of cleaning plant sales of its
member power laundries. The 1936-38 estimates were made by using the
Bureau of Labor Statistics index of payrolls of cleaning and dyeing plants
for interpolation. This index for 1939—42 extended to later years by State
unemployment insurance wage data has also been employed for interpolation
and extrapolation of the 1939 and 1948 census benchmarks. The unemployment insurance wage figures used for this purpose are the combined
totals for cleaning and dyeing plants and cleaning, pressing, alteration, and
garment repair shops.
Barber shops, beauty parlors, and baths
Census data, with 5 to 8 percent (the latter in recent years) added for
tips, provided the basis for estimating consumer expenditures for barber
shop and beauty parlor services for 1933, 1935, 1939, and 1948. Estimates
for 1930 were obtained by multiplying the number of employed persons in
these industries, as derived from the 1930 population census, by average
receipts per person estimated by reference to census data for later years.
For other years of the 1929-35 period, estimates were prepared by using
for both the barber shop and beauty parlor series consumer expenditures
for laundering as an extrapolating and interpolating index. To obtain
interpolations for the two series between 1935 and 1939, the barber shop
and beauty parlor service components, respectively, of the Consumer
Price Index were multiplied by an output index for laundry services, computed by dividing the Bureau of Labor Statistics index of payrolls in power
laundries by its index of average hourly earnings for the same industry.
For the 1940-47 period, it was possible to establish benchmark values
for 1945 and 1947 by use of receipts data for barber shops and beauty parlors
tabulated by the Internal Revenue Service from income tax returns of corporations, partnerships, and individual proprietors. Interpolations for other
years of the period were based largely on the BLS price indexes. For later
years, the 1948 Census-based total for barber shops and beauty parlors
has been projected by the product of (a) employment for the two types of
establishments combined as reported from unemployment insurance data
by the Bureau of Employment Security and (b) the barber shop and beauty
parlor price indexes of the Bureau of Labor Statistics weighted by the 1948
consumer expenditure values.
The portion of the estimate relating to baths and masseurs is quite small
—1.3 percent in 1950. It has been derived principally from data reported
in service censuses for 1935, 1939, and 1948, with interpolation and extrapolation by means of the series on consumer expenditures for barber shop
and beauty parlor services.
Laundering in establishments
As derived for the years 1929, 1931, 1933, 1935, 1939, and 1948 (adjusted
for undercoverage in 1933 and 1948 as noted above) from the Censuses of
Power Laundries, Cleaning and Dyeing Establishments, and Service Establishments, the figures on consumer expenditures for laundering in establishments include receipts from bundle work and family services of power
laundries, laundry receipts of cleaning and dyeing plants, and 80 percent
of the service receipts of hand laundries. (Payments from hand laundries to
power laundries were taken at 20 percent of the formers' receipts.) Receipts
of hand laundries for 1929, 1931, and 1933, not given in the census reports,
were estimated by the movement of the other components.
Estimates for intercensal years of the pre-1939 period were obtained by
interpolation by the American Institute of Laundering index of laundry
receipts of power laundries. For the later period, the 1939 and 1948 Censusbased estimates have been interpolated and extrapolated by an index of
laundry payrolls, based on Bureau of Labor Statistics sample data for the
years 1939—42 and on Bureau of Employment Security tabulations of unemployment insurance wage data for subsequent years.



121

Funeral and burial service
Consumer expenditures for funeral and burial service were obtained for
1935, 1939, and 1948 from census receipts data covering funeral directors'
and embalmers' services and coffins, plus an estimate of the funeral receipts
of combination furniture and undertaking establishments derived from the
1929 Census of Retail Distribution. Odd-year estimates prior to 1935
were obtained by extrapolating the 1935 figure by data on the value of
coffins and funeral supplies produced (from the biennial Census of Manufactures). These were interpolated for 1930, 1932, and 1934 by billings of
funeral directors and embalmers, as estimated by Rolf Nugent, Consumer
Credit and Economic Stability (Russell Sage Foundation, New York, 1939).
Estimates for other years (1936-38, 1940-47, and 1949-53) are interpolations and extrapolations by data on collections of sales taxes levied on funeral
directors' receipts in a number of States.

Census of Agriculture
Estimates of the rental value of farm houses, described briefly in the
section on Income of unincorporated enterprises, are prepared by the
Agricultural Economics Division of the Department of Agriculture, and
based on value data reported in the quinquennial Census of Agriculture.
The basis of estimation, however, is not a direct one. Census-based figures
on value are multiplied by an annual interest rate, and the product is then
converted to a gross measure of rental value by adding an estimate of farm
expenses allocable to the upkeep of farm dwellings.

Census of Religious Bodies
The "religious bodies" component of consumer services represents all
current expenditures by churches except for local cash relief and charity,
plus estimated depreciation of buildings. It is included in group XI of table
30, Part V.
Estimates of cash current expenditures were prepared for 1926 and 1936
from data reported in the Census of Religious Bodies, with a small adjustment for churches not reporting. The value of clergymen's imputed income
(see section on Wages and salaries) was added. For all other years of the
period since 1929, estimates of total current expenditures (cash and imputed)
have been obtained by interpolation and extrapolation on the basis of the
National Income Division series on total payroll of religious organizations,
as described in the section on Wages and salaries. Depreciation, estimated at
about 1.5 percent of the value of church edifices, was derived by a downward
modification of hospital building depreciation rates (see section on Capital
consumption allowances).

Biennial Survey of Education
Office of Education data reported in the Biennial Survey of Education, covering
school years ending in even numbers, form the principal basis of the consumer
service series on "higher education" and "elementary and secondary schools."
The estimated depreciation of educational buildings and equipment (see
section on Capital consumption allowances) is added to figures on current
expenditures by these institutions for educational services, as derived from
the Biennial Survey.

Estimates of current expenditures for intervening school years not covered
by the Biennial Survey are obtained by straight-line interpolation. For the
recent years of the series before the Biennial Survey becomes available, the
latest reported data are projected on the basis of partial information furnished by the Office of Education. The resulting school-year figures are
converted to calendar years by averaging. To obtain the estimate for calendar
year 1950, for example, the 1949-50 school-year data were weighted by 2
and the 1950-51 school-year data by 1.
It should be noted that the series for elementary and secondary schools is
less firmly based than that for higher education. The data published in the
Biennial Survey of Education for private elementary and secondary schools
include capital outlays, and these must be estimated and deducted. Moreover, the published data represent estimates based on reported enrollment
in these private schools multiplied by average expenditure per pupil in
public elementary and secondary schools. Accordingly, this series is rot
regarded as being based on periodic comprehensive sources, as listed in
Exhibit 4, but is included in the "Miscellaneous" category discussed below.

122

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS
Sample Information

Consumer service expenditures whose levels have been estimated from
sample data comprised 13 percent of the service total in 1950. These series
include expenditures for professional services (7.3 percent); interest on personal debt (2.8 percent); automobile insurance, net payments to labor
unions, postage, and telegraph, cable, and wireless (combined, 1.9 percent);
and foreign travel and remittances (1.0 percent).
The professional services category noted above includes separate expenditure estimates for the services of physicians, dentists, lawyers, osteopathic
physicians, chiropractors, chiropodists and podiatrists, private duty trained
nurses, miscellaneous curative and healing professions, and veterinarians.
The sample information used in making these estimates is obtained very
largely from periodic questionnaire surveys by the National Income Division,
generally with the cooperation of the professional associations. This information covers (1) average gross income from independent practice, (2)
percentage of persons in the profession engaged in independent practice,
and (3) the percentage of gross income (in the case of physicians, dentists,
lawyers, and veterinarians) received from government or welfare agencies
or business organizations, as contrasted with individuals.
These data are used in conjunction with available information on the total
number of persons in each profession to secure the consumer expenditure
estimates. The main sources of information on numbers in the various professions are the occupational data of the Census of Population, compilations
of the individual professional associations (such as the American Medical
Association, American Dental Association, and the American Osteopathic
Association), and the Fisher-Stevens, Inc. lists.
The methods of estimating consumer expenditures for professional services
are directly analogous to those used in deriving the total net income of independent practitioners in the various professions. These professional income
series are described at some length in the section on Income of unincorporated
enterprises.
Estimates of interest on personal debt, as noted in the section on Interest,
are based very largely on consumer credit data (compiled by the Federal
Reserve Board from sample reports adjusted periodically to more comprehensive sources), multiplied by fragmentary sample information on interest
rates applicable to the various types of credit.
Data from the 1935—36 Consumer Purchases Study (given in reports by
the National Resources Planning Board and the U. S. Departments of Labor
and Agriculture) provided benchmark materials for estimating the consumer
service series on automobile insurance (premiums minus claims paid), net
payments to labor unions (dues and fees paid minus cash benefits received),
postage, and telegraph, cable, and wireless. Little confidence can be placed
in these series. Apart from sampling or reporting errors in the basic data,
considerable estimation was required in utilizing the data to obtain the
benchmark consumer expenditure values; and the available information for
extrapolation is not adequate. Except for the labor union item, the presence
of large nonconsumer elements in the extrapolating data is an added difficulty.
The consumer service series relating to foreign travel and remittances
(net) are components of group XII in table 30, Part V. The estimation
of these series, which involves in substantial degree the problem of consumer
allocation, is based largely on questionnaire surveys. (See section on Net
foreign investment.)

For the most part, these seven series incorporate a great deal of relevant
statistical information (notably in the cases of social welfare and foreign relief
agencies, elementary and secondary schools, and brokerage charges and
interest and investment counseling). Difficulty of consumer allocation is
present for the series on moving and warehouse expenses, brokerage charges,
etc., and taxicab fares and tips, as well as for the estimates of water expenditures with respect to the data used for extrapolation of the 1940 base-year
figure.
Of the remaining items in the "Miscellaneous" category, nearly all are
relatively small. Source materials vary widely in adequacy among this group
of service items but are not satisfactory for many of them. It often is necessary
to use indirect estimating methods, and to rely on occasional and incidental
information obtained from public and private research studies, correspondence with trade associations, and informal sources such as magazine and
newspaper accounts.

Nature of Recent Revisions
As noted before, availability of data from the 1948 Census of Business, the
1950 Census of Population and Housing, and the 1950 Census of Agriculture
has permitted preparation of new benchmark estimates for numerous series
on consumer expenditures for services, as listed in Exhibit 4. In addition, a
review has been made of the data and procedures utilized in the estimates for
all other service items.
As a result of the incorporation of census materials and this additional
review work, about one-half of the items comprising the 55 service series
published in table 30, Part V have undergone statistical revision. The net
effect was a raising of the series. This amounted in 1950 to 5 percent for the
total of items that were revised and to 4}i percent for the overall consumer
services aggregate.
Adjustment to the postwar censuses revealed that, in total, the 30-odd
service items which had been extrapolated from the 1940 Census of Population and the 1939 Census of Business were low by 7 percent in 1950. About
5 percentage points of this revision may be attributed to deficiencies in the
extrapolators. The remainder occurred in the automobile repair series as a
result of correcting it to include parts used in making repairs.
The revision in this area is a modest one when viewed against the problem
of making 10-year extrapolations of series which, in total, increased by 145
percent. Still, it should be noted that this result reflected a netting of errors in
the individual components—a tendency for them to offset in the total. While
the degree of revision was tolerable for most components in this group, it
ranged up to 25 percent or more for some of them. Consideration of these
revision results, while gratifying in the overall, influenced the decision to
compress the extent of published detail on consumer services.
It may be added that not since the census of 1936 have benchmark expenditure data been available for the series on religious bodies, a component
of group XI in table 30. However, this series has been revised upward
considerably because of revision of the payroll estimates used as an extrapolator. These revisions, in turn, mainly reflected the incorporation of employment data from the 1950 Census of Population. (See section on Wages and
salaries.)

Miscellaneous
Consumer service expenditures for items classified under "Miscellaneous"
sources comprised about 9 percent of the 1950 total. A less rigid basis of
classification, as already indicated, might have considerably reduced the
proportion of expenditures in this category.
Of the items in this category, only five are shown separately in table 30,
Part V. These are elementary and secondary schools, taxicab fares and tips,
household expenditures for water, brokerage charges and interest and investment counseling, and bridge, tunnel, ferry and road tolls. Together with two
other relatively sizable items—social welfare and foreign relief agencies
(included in group XI of table 30) and moving and warehouse expenses
(included in item V—11)—-these series accounted for about three-fifths of the
1950 "Miscellaneous" total.



9. NEW CONSTRUCTION
The estimates of construction activity (see table 31) have been prepared
jointly by the Bureau of Labor Statistics and the Commerce Department
Building Materials and Construction Division. These estimates differ from
the totals used in the gross national product only because the latter include
petroleum and gas well drilling (in private construction) and exclude work
relief (from public construction). The value of work relief construction appears in gross national product as government expenditures for payrolls,
purchases of materials, etc.
New private construction is an independent element in the statistical determination of cross orivate domestic investment. New Dublic construction

NATIONAL INCOME,

is a component of government purchases of goods and services, which, however, are estimated as a total independently of the construction series. (See
the section on Government receipts and expenditures.)
New construction represents the value of progress made during the given
year in the production of fixed works and structures. The value of progress
made, or "work put in place," is defined as equivalent to the value of labor
and materials used plus overhead costs and profits accrued on operations
during the given period. It includes the installed value of equipment generally considered an integral part of a structure and commonly included in
the construction contract price.
Exhibit

1.—Components of New Construction Activity,
Principal Data Source, 1950
Private

Data source and class of construction

Diiect reports of work done or paid for
Public utility
_ Highway
Military facilities
Public residential building .
Public nonresidential building (Federal
and Federal-aid).
Other types of public construction (Federal and Federal-aid) _

Classified

Public

Millions
of dollars

Percent

Millions
of dollars

3,330

15

4,249

3.330

by

2,272
177
345
541

Percent
61

1954

EDITION

123

in these notes, are therefore generally somewhat better founded than those
for earlier years.
Modern techniques of systematic reporting and sampling are difficult to
apply to the direct statistical measurement of construction activity. Many of
the producers that should be covered are hard to identify. Much construction work is done by firms only intermittently attached to the industry; many
construction firms have no fixed and readily recognizable place of business;
and any firm in any industry may undertake force account projects. (The
several p?st censuses of construction, it may be noted, omitted by intent the
substantial volume of force account construction.) The alternative of covering the purchasers of construction work is generally subject to the same sort
of difficulties.
This problem of identifying the units to be covered affects both the enumerative and the sampling approaches to data collection. In particular, it
makes enumeration expensive. Sampling, the theoretical validity of which
rests on the homogeneity and continuity of the universe sampled, must deal
here with large, unstandardized, nonrepeating projects, many arising out of a
special local need. Under these circumstances, estimation relies heavily on
indirect evidence of construction activity, as is explained below, and is subject to considerable uncertainty as to coverage, valuation, and timing.

Summary of sources

The construction estimates are developed from several different types of
statistical sources and methods. These may be summarized very briefly as
follows.
Contract awards.
4,064
2,751
39
18
From certain special classes of buyers, regular current reports are received
Private nonresidential, except farm and
on the actual progress of, or expenditure for, their construction work. In genrmblic utility
3,777
eral, Federal Government and Federally-aided projects are covered by curPrivate residential, except farm—nonhousekeeping
175
rent reports on progress, and construction done by or for public utility com"All other private"
112
Public nonresidential building (State
panies is regularly reported from accounting records.
and local)
1,843
Other nonfarm nonresidential construction is generally estimated from
Other types of public construction (State
and local)
908
monthly data on value of contract awards.
Other nonfarm residential building is estimated chiefly from monthly reBuilding permits
10,605
46
ports of building permits issued.
Private residential, except farm—new
For some types of construction, the foregoing kinds of information are indwellings in permit-issuing areas l and
additions and alterations
10,605
adequate or unavailable, and estimates must be made from a variety of data
of varying appropriateness and reliability. A small part of nonfarm resiOther sources
4,734
21
dential construction, all farm construction, and oil and gas well drilling are
Private residential, except farm—new
included in this miscellaneous category.
dwellings in nonpermit-issuing areas ' . .
1,820
The relative importance of these various sources and methods in deriving
Farm
1,635
Oil and gas well drilling
1,279
the new construction estimates is illustrated in Exhibit 7. Each component cf new construction (as given in tables 2, 3, 5, and 24 of the May
Total
- ..
100
100
7,000
22,733
1953 Statistical Supplement to Construction and Building Materials) is here
classified according to the principal type of source material used in estimating
1. Breakdown of new dwellings between permit-issuing and non-permit-issuing areas is
rough; based on unpublished data.
its value.
The four classes of source materials distinguished in Exhibit 1 may be comFixed works and structures include not only dwellings and other buildings
pared briefly with one another in terms of coverage, quality of value data,
but also dams, bridges, roads, canals, and the like. Certain types of works,
and the indication given as to timing of the construction work done.
such as mine tunnels and farm ditches, which might be classified as construcDirect reports of work done or paid for are obtained through formal reporting
tion, are not included. They are probably unimportant quantitatively, and
systems involving the regular and sometimes mandatory cooperation of
a substantial portion of them is likely to be charged to current expense.
parties to the construction contracts. Coverage of work done by or for the
Hence it is believed that their omission from gross national product does not
categories of purchasers included in these systems is generally excellent. The
result in any appreciable discrepancy in the national income and product
value information reported, being taken on a fairly standardized basis from
accounts. Also, the estimates of private residential nonfarm construction
accounting records appropriate to the purpose, is also quite satisfactory.
exclude certain "speculative" profits, because these are regarded as reflecting
Finally, the reported timing of contract construction work is based largely
sales rather than construction activity. These profits should be part of the
on engineering inspections made to check contractors' claims for progress
new construction component of gross national product; they are omitted
payments, and therefore accords very closely with the conceptual requirebecause there is no satisfactory way of estimating them.
ments of the estimates.
The factual content of the present notes is largely limited to a summary of
Contract award reports (which take into account subsequent cancellations)
descriptions published in the Building Materials and Construction Division's
provide reasonably good value information for projects covered, particularly
May 1953 Statistical Supplement to Construction and Building Materials andin periods when escalator clauses or other similar arrangements are not
in the Bureau of Labor Statistics bulletin on Techniques for Preparing Major important. As will be seen, however, the coverage of these data is far from
BLS Statistical Series, revised. No attempt is made to indicate differences complete, and an estimate for the noncovered areas must be made. Furtherbetween the estimating methods underlying back-year values and the
more, there is a varying time lag between the reported date of a contract
methods currently in use; only the latter are described below. Although the
award and the start of actual construction, and there is further variation in
basic sources and methods have been the same throughout the period since
the rate of progress after the start. The allocation of contract values to the
1929, a number of important improvements have been made. It should be
particular time periods during which the work is presumed to have been done
noted, however, that most of them did not cast light on back-year moveis based on past activity patterns and cannot be precise.



914

124

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

ports and they are therefore classified under this heading. For some other
of not indicating the time period during which the work is done and undergo
years their classification under "other sources" (see below) would have been
a similar timing adjustment in the course of deriving the estimates. The
more appropriate.
valuations entered on building permit applications are generally Jess reliable
Direct Federal construction of housing as carried on during the depression,
than are values to which the parties are committed by contract. An adjustdefense, and war periods under the Lanham Act was estimated from progment for undervaluation, and an allowance for lapsed permits are made on
ress reports by the Public Housing Administration on the number and cost
the basis of sample studies. The coverage of the permit data is very high in
of units built. Construction of low-rent units and slum clearance expenditures
the areas where building permits are required.
The miscellaneous sources used in estimating the remainder of construction by State and local agencies with Federal loans and grants under the Housing
Act of 1949 are estimated from progress reports made available by the Public
activity vary widely in quality. In general, they yield results less reliable
Housing Administration. Estimates of the value of work accomplished on
than those derived by the use of the other three types of source materials.
the locally financed New York City projects have been based on the progress
reports used to determine payments to contractors.
Of the 1950 total value of public nonresidential building, about one-fourth
Estimates Based on Direct Reports of Work
was Federal or Federally aided; all but a small part of this was covered
Done or Paid For
by progress reports.
In hospital and institutional construction, a major item under this heading,
Direct reports are the chief basis of components accounting in 1950 fcr
two large programs are involved: that of the Veterans Administration and
15 percent of estimated private construction and 61 percent of estimated
the National Hospital Program. Estimates of the value of veterans hospital
public construction. Exhibit 2 shows the components estimated wholly or
building are based on monthly progress reports supplied to the Office of the
in large part from materials of this sort. They consist primarily of projects
Chief of Engineers and to the Veterans Administration by project engineers
in which the Federal Government is involved—as a contracting party, as a
in the field. The National Hospital Program is one of Federal aid. Estimates
source of aid funds, or through agencies which regulate the purchaser's indusof the value of work done under this program are prepared from progress
try. However, private corporations and trade associations also perform an
reports on individual projects submitted to the Hospital Facilities Division
important service in data collection, as is indicated by the list ofreporting
of the Public Health Service by State agencies administering the program.
systems included in the exhibit.
The rest of public hospital construction, not aided by the Federal GovernThe estimates for privately owned public utilities generally involve only
ment (included in the total shown in Exhibit 2, since its value is not published
minor problems with respect to coverage, valuation, and timing. Totals from
separately), is evaluated by applying activity patterns to contract award
annual reports by railroads, electric utilities, oil companies, and members of
values compiled from F. W. Dodge Corp. reports (see below) and other
the Bell Telephone System are adjusted to reflect construction by or for
sources. The value of other construction included in public nonresidential
small concerns which do not report. The values reported for petroleum pipe
building, consisting chiefly of work done for the Public Buildings Administralines and electric and gas utilities include some expenditures for the purchase
tion, is estimated from progress reports of the supervising agency.
of existing facilities or producers' durable goods; statistical adjustments are
made to eliminate these. With respect to timing, the annual reports for public
Exhibit 2.—Components of New Construction Activity Estimated Chiefly
utilities are somewhat less appropriate for these estimates than are the
from Direct Reports of Work Done or Paid For, 1950
progress reports upon which most of the other values shown in Exhibit 2 are
based. However, there is no reason to expect any important timing discrepMillions
ancy on an annual basis.
Reporting agency
Component
of dollars
For the most recent year, financial data on construction outlays are generally not yet available for public utilities except the large railroad systems.
3,330
Private construction..
Preliminary estimates are based mainly on the utility companies' previously
Public utility:
Federal Power Commission
1,268
Electric light and power
announced plans for construction and checked against current sample series
1,102
Manufactured and natural gas.. American Gas Association
on plant and equipment expenditures, such as the quarterly CommerceAmerican Telephone & Telegraph
435
Telephone
Co.
Securities and Exchange Commission estimates. Construction plans are
315
Interstate Commerce Commission,
Railroads
Association of American Railroads.
ascertained from reports compiled either by trade sources or by the same
165
Interstate Commerce Commission..
Petroleum pipe-line..
agencies which later tabulate the financial data.
40
American Transit Association
Local transit
5
Western Union Telegraph Co
Telegraph
More than one-half cf all public construction covered by reports of
4,249
Public Constructionwork done or paid for consists of road-building. Monthly estimates of
Highway, by source of funds:
the value of work accomplished under the Federal-aid highway program
Federal
46
Bureau of Public Roads
_
are prepared from reports of the Bureau of Public Roads showing State
397
Federal-aid
....do
State and local
Bureau of Public Roads (through
1,829
"earnings," which are based primarily on reports by engineers of the progress
State highway departments).
made on individual projects. The Bureau of Public Roads also compiles
Military facilities l
Department of Defense
177
annual reports of highway construction on Federal lands. State, county,
Public Housing Administration;
345
Public residential building i
New York City Housing Authority.
and municipal highway, road, and street construction outside the Federal541
Public nonresidential building (Fed- Federal agencies responsible; Bureau
eral and Federal-aid).
of the Budget.
aid program is estimated mainly from special financial reports submitted
914
Other types of public construction Federal agencies responsible; Bureau
annually through State highway departments to the Bureau of Public
(Federal and Federal-aid).'
of the Budget.
Roads.
1. Public residential construction by the Department of Defense is included under "MiliFor military facilities, the expenditures reported represent the volume of
tary facilities." Housing at the sites of reclamation and flood control projects is included in
all construction, regardless of type, at Federal military installations. The
the "conservation and development" category among "Other types of public construction."
relatively small amounts of military construction by the States (armories,
rifle ranges, and the like) are included with other public construction cateA minor part of the total shown under public nonresidential building in
gories according to type of construction. The data for the two major conExhibit 2 represents construction for the National Advisory Committee for
struction agencies of the Department of Defense—the Office of the Chief of
Aeronautics, the Bureau of Prisons, and similar agencies which engage in
Engineers and the Bureau of Yards and Docks—are based on monthly progconstruction only occasionally or on a small scale. These agencies usually do
ress reports for all construction projects by service engineers.
not have sufficiently large construction staffs to warrant setting up regular
progress reporting systems. Estimates for them are made by applying activity
Public housing construction progress reports are gathered for Federal and
patterns to contract award data from the agencies, supplemented by annual
Federal-aid housing by the Public Housing Administration and are prepared
statistics from the Budget of the United States Government.
for locally financed projects in New York City by the local Housing Authority.
The remainder of public residential construction is based on local government
The "Other types of public construction" distinguished in Exhibit 2 conestimates of the cost, duration, and starting date of projects reported to
sist largely of outlays to conserve, develop, or control the Nation's water
regional offices of the Bureau of Labor Statistics. In 1950 the preponderant
resources. The bulk of these expenditures is estimated from monthly progshare of the public residential building estimates was based upon direct reress reports by Government engineers on projects of the Bureau of Recla


125

NATIONAL INCOME, 195 4 EDITION
mation and the Civil Works Division of the Office of the Chief of Engineers.
The Tennessee Valley Authority also provides monthly summaries of actual
cost of its construction activities. The small balance of conservation and development work included in Exhibit 2 is carried on by the International
Boundary and Water Commission and similar agencies, and is evaluated
from annual fiscal data shown in the Budget. Civil airport construction activity is estimated primarily from monthly progress reports to the Civil Aeronautics Administration; and Federal construction of a few other nor.conservation items of minor size is approximated from Budget data.

Estimates Based on Building Permits
Building permits are the most important data source for private nonfarm
residential construction. This category of construction is estimated in four
parts: construction of dwelling (housekeeping) units in areas in which building permits are required; construction of dwelling units in other areas; construction of nonhousekeeping units; and additions and alterations. The following tabulation illustrates the relative importance of each of these parts.
Exhibit 3.—Components of Private Nonfarm Residential Construction,

Estimates Based on Contract Awards
Monthly reports of the value of contracts awarded are used to evaluate
construction projects which collectively account for 18 percent of the private
construction and 39 percent of the puHic construction included in the 1950
gross national product. (See Exhibit 7.)
The chief source of data on contract awards (including information on subsequent cancellations) is the F. W. Dodge Corp., whose local correspondents
keep informed of new projects through personal contacts in the construction
and related industries, press reports, permit-issuing offices, and a variety of
other sources. Values of projects thus located are generally ascertained directly from the contracting parties. It is obviously difficult to avoid the omission of a considerable amount of construction activity when data are gatheied
in this way. The system is considerably more effective than the difficulty of
the task would suggest, however. Its relatively good coverage is due to the
long experience of the Dodge organization, which initiated this work in 1901,
and to financial support from contractors and suppliers who keep in touch
with (heir markets through such information.
To derive estimates of construction work dene, the Building Materials and
Construction Division adjusts the contract awards data for coverage and
timing.
The first adjustment of coverage is required because the F. W. Dodge
Corp. data apply only to the 37 States east of the Rocky Mountains. Western
State figures comparable to the Dodge tabulations are estimated chiefly
from building permit information and reports on construction contract
awards appearing in various trade journals.
A further coverage adjustment is applied to the national totals so obtained,
because neither the Dodge materials nor those available for the West purport
to cover all construction projects in the categories estimated from contract
awards. For private construction in each of these categories, a raising ratio
is computed by use of estimates based on the 1939 Census of Construction and
trade sources. Allowances for undercoverage of public construction are based
on extensive correspondence with State and local government officials.
After these adjustments, the data represent estimates of the value of construction work of each type for which contracts have been let during the
month. Some projects undoubtedly are started within the same month in
which contracts are awarded; others will not be started until 2 or 3 months
later. In the absence of definite information on the patterns of these delays,
it is more or less arbitrarily assumed that their average is one month; the
estimated value of contracts let in the given month (excluding those subsequently cancelled) is accordingly taken to measure the total value of projects which will be started in the following month.
Most construction projects take several months to complete after they are
started. Through surveys of thousands of actual projects, the Building Materials and Construction Division of the Department of Commerce has established typical activity patterns for various types and sizes of projects, showing
percentages of value "put in place" in successive months. These patterns are
used to translate the value of starts into construction activity.
Modifications of these patterns were made during the war years from data
collected by the War Production Boai d, and also during the immediate postwar years, when materials shortages delayed construction, on the basis of
data collected by the Commerce Department for the Civilian Production
Administration. The patterns are subjected to constant revision as additional
information becomes available and on the basis of judgment gained through
experience over a decade of investigation.
A final adjustment in the resulting estimates of work put in place is made to
eliminate offices, warehouses, and other buildings constructed by public
utilities, all of which are classified as public utility construction.



Item
Dwelling units built under permit
Other dwelling units >
Nonhousekeeping units
Additions and alterations
Total

Millions
of dollars

1950

Percent

9,705
1,820
175
900

77
15
1
7

12,600

100

1. Breakdown of the total of these items is rough; based on unpublished data.

Estimates of the third item are based upon the contract awards data which
have already been discussed. Estimates of the second item are based upon
procedures explained below. Those of the first and fourth items are based
upon information on building permits, and will be described next.

Nonfarm dwelling units built under permit
Most nonfarm residential construction is carried on under building permits. The permit-issuing officers of the responsible local governments compile
totals for number of dwellings and estimated value from the records of permits issued, and they report these totals to the Bureau of Labor Statistics on a
monthly questionnaire form.
Coverage of permit-issuing localities is not far from complete—about 90
percent of all such localities report each month. To allow for the remainder,
the individual localities are first cross-classified into cells according to type
of jurisdiction, location, urbanization, and population size. A raising ratio
for each urban cell is then computed by dividing the number of localities reporting for the month into the total number of localities in the cell. For each
rural nonfarm cell, the raising ratio is computed from 1940 population census
figures, as the ratio of dwellings in all localities to dwellings in lociditi^s reporting.
Reported values raised by these ratios are next adjusted to allow for permits
lapsing, and then for the understatement of ultimate construction values
which is generally characteristic of estimates entered on building permits.
Periodic field studies are made by the Bureau of Labor Stati tics to determine
the appropriate percentage allowance for each of these factors.
The timing adjustments next applied—of the same general nature as those
of the contract awards data—utilize activity patterns based upon Bureau of
Labor Statistics investigations. Specifically to be noted is that the estimates
of the lags which occvir between the issuance of permits and the start of the
projects are based upon Bureau of Labor Statistics field studies instead of the
summary assumption used in the case of the contract awards data.

Additions and alterations to nonfarm dwellings
Most local jurisdictions with building-permit systems require permits not
only for new structures but also for additions and alterations to existing
structures. However, compliance with such regulations is much Jess complete
in the case of additions and alterations, particularly in rural nonfarm areas.
Special problems are therefore met in translating the permit data on additions and alterations into estimates of the value of work done. For the urban
United States, the procedure is substantially the same as for new dwelling
units except that no specific adjustment is made for lapsed permits. The
resulting estimates of value of work in place are then expanded to cover
rural nonfarm activity, using relationships between rural nonfarm and
urban construction of new units. Studies of family expenditures—notably
the Consumer Purchases Study of 1935-36—provide the basis for a final
adjustment which allows both for understatement of value in permit applications and for undercoverage of projects.

126

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS
Estimates Based on Other Sources

Nonfarm dwelling units not built under permit
As is suggested by Exhibit 3, a significant fraction of residential construction
occurs in rural nonfarm areas where no building permit is required. Representatives of the Bureau of Labor Statistics cover a sample (currently, nonpermit-issuing parts of 96 counties) of such areas, and report numbers of
units started as well as prevailing average construction costs. Reported numbers are raised to full coverage by a method similar to that applied to building
permit data for rural nonfarm areas and multiplied by average construction
cost. The results are then distributed forward over succeeding months by
normal activity patterns.

Farm construction
Annual estimates of construction on farms are prepared by the Agricultural
Economics Division of the Agricultural Marketing Service, Department
of Agriculture, and described in Agricultural Estimating and Reporting Services

(Washington, December 1949). They are based chiefly on data from sample
surveys of construction expenditures of farm operators in 1934—37, 1939,
1946, and 1949. Estimates for other years are made by interpolation and
extrapolation, based in part on inferences from data on farm electric lighting
systems, windmills, silos, etc., reported in the annual Farm Machines and
Equipment reports of the Census Bureau. The bulk of the dollar amounts
involved, however, for other than benchmark years represents approximations
based on changes in indices of farm construction costs and in indicators such
as estimated consumption of lumber, sales of building materials in rural
areas, and nonfarm residential construction. The Commerce Department's
farm construction series represents the expenditure series of the Agricultural
Economics Division adjusted to exclude estimated expenditures for building
repairs.

Oil and gas well drilling
Oil and gas well drilling is not classified as new construction in the classification system employed by the Building Materials and Construction Division of the Department of Commerce, which prepares the estimates, but is
included as such in gross national product. All costs of drilling are covered,
including the cost of casings (but not the cost of installed production equipment).
Estimates for the base year 1939 were derived for each State from data
gathered in the 1939 Census of Mineral Industries and compilations of
reports to trade publications. These estimates have been projected from 1939
by using figures on the number of wells completed, as reported in trade
sources, multiplied by the average cost per well. The latter is determined by
adjusting the 1939 average cost in each State to take approximate account
of changes in labor costs and efficiency, material costs, proportion of wildcat
ventures, and average depth of wells.

Characteristics of Revisions
New data and methodological or conceptual changes have necessitated
occasional major revisions in the estimates of certain components of construction. For the most part, these revisions have not been of such a nature
as to cast light on the degree of reliability attained in the currently published
series. There is evidence, however, that some of the components of new
private construction may be somewhat understated.

10. PRODUCERS' DURABLE EQUIPMENT
The estimates of producers' durable equipment shown in table 32, Part V
represent statistical revisions for the entire period since 1929. These reflect
the incorporation of new data, mainly from the postwar censuses of manufactures and business. Also, in accordance with 1947 Census of Manufactures



practice, the Standard Industrial Classification of 1945 was adopted as the
basis of commodity classification. (See Exhibit 1.) Use of new basic data
and a different classification scheme entailed a reworking of the estimates
back to 1929. In the course of this, the opportunity was taken to review all
of the earlier estimates and to introduce a number of improvements in
methodology.

Exhibit 1.—Commodity Classification of Producers' Durable

Content in
terns of the
Standard
Industrial
Classification •

Product group

Furniture and
fixtures
_
Cutlery and hand tools.
Fabricated metal products (except cutlery and hand tools)
Engines and turbines
Tractors
_
_
Agricultural machinery (except tractors)
_
Construction machinery
Mining and oil field machinery
Metalworking machinery
Special-industry machinery n. e. c
General industrial machinery

_
_

Trucks, busses, and trailers
Passenger cars
Aircraft
Ships and boats.
Railroad equipment
Instruments
Miscellaneous equipment

--

25
3421-25
34
351
3521
3522
3531
3531,3532
354
355
356,3591

-

Office and store machines
--Serrfce-industry and household machines
Electrical machinery

Equipment

357
358
36

-

—

W

371
371
3721
373
374
3S

1. This column refers to code numbers of group? listed In the J045 Standard Industrial
Classification for Manufactures (Office of Statistical Standard?, Bureau of thp Budget).
2. Includes producers' share of the following: Miscellaneous manufactures (Group 39);
Motorcycles (Group 37511); Transportation equipment, n. e. c. (Group 3790); Motor vehicle
heaters (no code): Textile mill products (Group 22); Miscellaneous fabricated textile products
(Group 239): Lumber and wood products, except furniture (Group 24); Sadd'.ery, harness,
and whips (Group 3192); and Stone, clay, and glass products (Group 32).

The commodity flow method, summarized in the section on Personal
consumption expenditures for commodities, is the principal one used in
estimating producers' purchases of durable equipment. It accounted for 66
percent of the total in 1939 and 88 percent in 1947. (See Exhibit 2.)
For the 1929-39 period and for 1947, the availability of requisite data,
mainly from the manufactures and trade censuses, made it possible to carry
out the method in detail. The numerous estimating steps entailed segregating finished producers' durables from total manufacturing output and
then tracing their flow and measuring their distributive costs so as to arrive
at the final costs to purchasers. (See Exhibits 3 and 4.)
For years since 1940—except 1947—lack of industrial census materials
has prevented a detailed application of the commodity flow method. For the
period 1942 through the first half of 1946, the method was used in an abridged
form to establish "secondary" benchmarks, based mainly on information,
arising out of production-control programs, on the commodity breakdown of
manufacturers' sales in the metal fabricating industries. Government purchases and exports were eliminated from the sales data, and imports were
added. It was not possible, however, to trace the flow of manufacturers'
domestic sales of finished equipment through the various distributive channels. Adjustment for inventory change was ignored and transportation
allowances, wholesalers' markups, and retailers' markups were added to
manufacturers' domestic sales by application of percentages obtained by
straight-line interpolation of those computed from the 1939 and 1947
benchmarks.
Similar "secondary" benchmarks were made possible for the period 1950—
52 by commodity data on manufacturers' shipments collected by the Bureau
of the Census in its annual sample Survey of Manufactures. Government
purchases and exports were deducted from manufacturers' sales, and adjustment was made for inventory change. Estimates for other steps in the procedure were based, for the most part, on relationships derived from the 1947
benchmark.
Estimates for the commodity flow segment of producers' durables for the
remaining periods—1940-41, the second half of 1946, and 1948-49—were
also prepared within the framework of the commodity flow method. With

127

NATIONAL INCOME, 195 4 EDITION
some variation among these periods, the method accounted separately for
manufacturers' commodity sales, government purchases, transportation
allowances, exports, imports, inventory changes, and wholesalers' and
retailers' markups. The estimates for these periods are to be distinguished
from the "secondary" benchmarks in that it was necessary to rely upon
interpolations to derive manufacturers' commodity sales as well as the other
major elements of the commodity flow buildup.
As described in the section on consumer commodities, the commodity flow
portion of personal consumption expenditures was estimated for the years
1940-46 and 1948-53 by interpolating and extrapolating the 1939 and 1947
commodity group estimates very largely on the basis of sales of retail stores.
This was feasible because a very high proportion of consumer commodities
is purchased at retail (and purchases by nonconsumers do not bulk large in
the retail sales totals).
For the producers' durables groups, however, the procedure of utilizing a
single series for interpolation or extrapolation would not be valid. These
goods are purchased in large volume both from manufacturers and wholesalers. Also, substantial quantities of durable equipment are channeled
through manufacturing and wholesale trade for government use and export, and these (particularly government purchases) have fluctuated widely
in relative importance. Continuation of the commodity flow approach for
non-censal periods was necessary to provide the basis for separate measurement of these major elements.
For a few of the producers' durable groups, as shown in Exhibit 2, it has
been found preferable to prepare estimates by methods other than the commodity flow. The largest such group is business passenger cars, which
formed 17 percent of the 1939 total and 11 percent of the 1947 total. The
general method used for this group consists of multiplying the estimated
number of units purchased by producers by an appropriate average price.
A similar procedure was used to estimate truck purchases for all years except
1947, 1951 (second half), and 1952—for which benchmark values were
computed by the commodity flow method.
Expenditure data reported by the Interstate Commerce Commission
were the basic source for estimating producers' outlays on railroad and transit
equipment prior to adoption of the commodity flow method in 1947. The
ships and boats series, of minor quantitative importance, was estimated by
the commodity flow method in the 1929—38 period. For subsequent years,
it has been derived mainly from 1939 and 1947 Census of Manufactures
data and various information provided by the Maritime Administration
and the Bureau of Customs.

Considerations regarding reliability
In the section describing the sources and methods of estimating consumption expenditures for commodities, the conclusion is reached that the Censusbased benchmark totals for the large commodity flow segment of consumer
commodities are "not markedly in error." Analysis of the essential differences in the application of this method to the consumer and the producer
durables series indicates that even higher reliability may be attached to the
latter. Two points, amplified later, are of main relevance.
First, the problem of allocation was less in the producer durable series.
In larger degree than in the case of consumer commodities, manufacturers'
sales of finished producer durables (apart from the special problem of government purchases) could be derived from detailed product data in the
Census of Manufactures through (a) selection of items in their entirety or
(b) allocation of "mixed" items having relatively little other use. In general,
estimation and judgment—and hence possible error—are involved to a
greater extent in allocation than in the selection offinisheditems.
Secondly, and more important, the successive estimating adjustments to
manufacturers' sales required by the commodity flow method, as illustrated
in Exhibits 3 and 4, were relatively smaller for producers' durable equipment than for consumption commodities.
As would be expected, commodity flow estimates of producers' durables
other than for 1929-39 and 1947 are of lesser reliability. Nonetheless, the
availability of substantial amounts of requisite data (particularly on manufacturers' sales) and the satisfactory record with regard to revisions, as set
forth later, warrant the conclusion that the totals for these non-benchmark
periods are generally adequate.
The estimates for 1940 and 1941 are least firmly based. For these years, it
was necessary to interpolate manufacturers' commodity sales by industry



sales data, which have significant limitations in such use, and statistical
information on government purchases was sparse. In addition, the 1940-41
estimates are affected by weaknesses in the 1942 secondary benchmarks.
The secondary benchmarks prepared for 1942-46 are subject to error
mainly from two sources: (1) elimination of parts and other unfinished
products from sales at the manufacturers' level and (2) the deduction of
government purchases, which it is believed may err in the direction of being
somewhat too large. For most of the commodity groups, purchases of durable
equipment by government during the war exceeded those by business and
were difficult to estimate. The quality of source materials for the 1942-46
period varied, the data for 1944 being most adequate. The estimates for this
period are believed, in general, to be considerably more reliable than those
for 1940 and 1941, but less reliable than those for 1947 or 1950-52.
The estimates for 1948 and 1949 (as well as for the second half of 1946)
represent interpolations carried out within the commodity flow framework.
They benefit materially from the high quality of the benchmarks for proximate years, and may be regarded as generally satisfactory.
The 1950-52 secondary benchmarks, which rest principally on the Census
Bureau's special annual surveys of manufacturing, are subject to sampling
variability at the manufacturers' sales level—although for most producer
durable groups this is small—and to errors stemming from the use of 1947
relationships to determine the major elements of distributive costs. GovernExhibit 2.—Producers' Durable Equipment, 1939 and 1947
1939

1947

Millions
of
dollars
Estimated by commodity flow method
Estimated by other methods:
Tracks, busses, and trailers
Passenger cars
Ships and boats
Railroad equipment
Total producers' durable equipment

_ _- _

Percent

Millions
of
dollars

2,749

66

14,542

88

489
715
57
170

12
17
1
4

1,889
236

11
1

4,180

100

16,667

100

Percent

ment purchases were of increased relative importance in this period, particularly in 1951 and 1952, but the available data on such purchases are
believed to be reasonably accurate.
Concerning the commodity groups based on methods other than the commodity flow, it is to be noted that the largest one—business passenger cars—
depends on an allocation between consumers and business which is based on
limited information. However, errors in this allocation do not affect the gross
national product total. The accuracy of the estimates for trucks and busses is
impaired by the necessity of making several sizable adjustments on the basis
of partial information.
The series on railroad and transit equipment is regarded as relatively
accurate. The same generalization applies to the ships and boats series for
1929-39 and 1947—for which Census of Manufactures data furnished the
principal basis of estimation—but the figures for other years are interpolations and extrapolations derived from not very satisfactory information.
A rough check on the postwar estimates of producers' durable equipment is
possible through comparison with the results of the New Plant and Equipment Survey conducted jointly by the Office of Business Economics and the
Securities and Exchange Commission. This series differs in definition from
the producers' durable estimates (it is confined to nonagricultural industries,
includes plant as well as equipment, but excludes purchases of equipment
charged to current expense, to mention the major differences), and adjustments to secure comparability cannot be made in a fully satisfactory way.
However, to the extent that comparisons are possible, they have been
broadly reassuring for most of the postwar period.
The 1953 estimate of producers' durable equipment is preliminary. Pending
the availability of the Census Bureau's Annual Survey of Manufactures, it
has been based on an overall extrapolation technique utilizing New Plant and
Equipment Survey results in conjunction with a variety of information for a
few important commodities.
The balance of this section describes the estimates of producers' durable

128

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

equipment prepared by the commodity flow approach—separately for
1929-39, 1947, 1940-46, and 1948-52—and by the other methods.

Commodity Flow Estimates, 1929-39
As already noted, the commodity flow method was used to derive estimates
for the 1929-39 period for producer durable groups comprising 66 percent
of the 1939 total. The method was employed extensively for this period also
in the estimation of personal consumption expenditures for commodities
(yielding 84 percent of the 1939 total), and was explained at some length in
the section describing that series.
The application of the method in the estimation of producer durables and
consumer commodities was substantially similar, so that no attempt has been
made here to repeat the description in the earlier section. However, there
were several particular aspects in which the methodology differed. Further,
individual steps in the estimating procedure varied widely in relative importance between the two series and have marked bearing on their statistical
reliability. These two types of differences may be summarized by reference to
Exhibit 3, which outlines for 1939 the individual steps in the derivation of
total producers' durables estimated by the commodity flow method.
Distribution of finished and mixed manufactured commodities, before deduction of
government purchases of durable equipment.—(Line 1.) As noted in the section on

consumer commodities, producers' items are defined as durable equipment
for multiple use in production and with an average life of one year or longer.
Even if the latter criterion were changed to as much as three years, it may be
added, the statistical measure of producers' durables would be affected only
slightly.
From product detail presented in the biennial Census of Manufactures, it
was possible to select and assign in their entirety commodities which comprised 56 percent of the 1939 total (line 2) of manufacturers' production of
finished producers' durables (before deduction of government purchases).
Seven percent represented allocation from the combined group of finished
commodities having appreciable use both by producers and consumers. The
remaining 37 percent was derived from allocations of the "mixed" category.
This category included commodities which could not be classified directly as
wholly finished or unfinished. They belonged in part to the unfinished category and in part to the producer durable and /or consumer commodity
categories, and required allocation among them.
The foregoing percentages were sharply affected by revision work undertaken for the 1954 edition of the NATIONAL INCOME supplement. In the

earlier series shown in the 1951 edition, direct selection and assignment of
values reported in the Census of Manufactures accounted for 84 percent of
the 1939 factory value of producer durables (before the government deduction). Reduction of this figure to 56 percent stemmed from revisions of assignments and allocations based mainly on detailed product relationships
developed for the 1947 benchmark study. In large degree, it reflected the
fact that small allocations were made for many commodities which previously had been assigned wholly to producers' durables or to one of the
other categories.
The changes in classification appreciably improved the estimates of manufacturers' output for producer durables of the type customarily charged to
current expense, rather than amortized through depreciation allowances. On
the basis of the greater product detail in the 1947 Census of Manufactures
and additional research into product uses, many such items formerly regarded
as producers' durable equipment were reclassified wholly or in part as intermediate products. In the segment of producers' durables represented by
products commonly charged to current expense—such as miscellaneous
types of small equipment—the problem of classification as between parts and
finished equipment is at best difficult. Apart from this fringe segment—that is,
for the preponderant part (well over 90 percent) comprised of items capitalized by business—the classification changes (mainly a general resort to more
allocation) were a refinement which sharpened the accuracy of individual
commodity groups but had little effect on the total.
Earlier, the generalization was made that, because of the greater emphasis
on estimation and judgment, allocation from "mixed" groups tends to involve
more error than does the direct selection offinisheditems. When the producer
durable and consumer commodity series are assessed from this standpoint,
comparison of the percentages of manufacturers' value derived from "mixed"



groups—37 and 43 percent, respectively—may be somewhat misleading. It
should be supplemented by the consideration that to a larger extent the
producer-durable percentage was based on product values for which the
allocation—and hence possible error—was relatively small.
There were two related statistical differences between the producer durable
and consumer commodity estimates with respect to the classification of
finished and mixed manufactured commodities.
(1) In the allocation of consumer commodities primary use was made of
information on the distribution of sales by class of purchaser given in the
Census of the Distribution of Manufacturers' Sales. Data reported under the
designation of "Sales to industrial, commercial, professional, and institutional
users" were taken to indicate the unfinished part. This could not be done in
the case of producers' durables since finished producers' durables as well as
unfinished commodities are included in sales to these user groups. It was
necessary, therefore, to allocate the mixed category for producer durables on
the basis of various types of specific commodity information, with some resort
to judgment and outside expert advice.
(2) Unlike the procedure that could be followed in the consumer commodity
estimates, where the government deduction was made jointly with all other
nonconsumer elements except exports, this deduction was carried out as a
subsequent, separate step. Since the available data for making the adjustment
related to government purchases, the deduction was made after manufacturers' sales had been derived from manufacturers' production by adjusting
for changes in inventories.
Deduction of government purchases of durable equipment.—(Line 5.) Apart from

the fact that Exhibit 3 refers only to commodities estimated by the commodity
flow method, it is to be noted that line 4 is not a complete measure of government purchases of durable equipment. Rather, it is a special-purpose measure which, in general, is restricted in scope to business-type items.
Comprehensive information on government purchases of durable equipment was lacking. Federal purchases were estimated by drawing on a
number of different sources, which generally pertained to one or more of
the years 1937-39 only, and required adjustment for differences in definition
or timing. For earlier years, it was frequently necessary to employ indirect
methods, which for the most part were not satisfactory. Such methods were
necessary for estimating State and local purchases in all years of the period.
The annual (1929—39) information on Federal aircraft purchases supplied
by the Departments of the Navy and the Air Force was quantitatively the most
important source of data on government purchases of durable equipment
Aircraft purchases alone accounted for two-fifths of the estimated government
total and considerably enhanced its accuracy.
The source most generally used in estimating Federal purchases of other
items was the tabulations by the Bureau of Labor Statistics of orders under
the Walsh-Healey Act, available on a quarterly basis beginning with 1937.
These tabulations excluded orders classified as secret or confidential or
amounting to less than $10,000. No adjustment for this incompleteness of
scope was attempted. The reported data were always shifted ahead, generally
by one quarter, to time them more nearly with actual deliveries. It fre-'
quently was necessary to estimate, and deduct, the amount of parts and other
unfinished products included in the Walsh-Healey data, and also to adjust
for differences in commodity definitions underlying these data and the
producer durable series.
For a number of the commodity groups, data on Federal purchases were
available for 1938 from the report of the Temporary National Economic
Committee on Study of Government Purchasing. These data, too, generally
required adjustment for unfinished products or for differences in commodity
definition.
Data on Federal purchases of durable equipment furnished by the Treasury
Procurement Division could be utilized for a few of the commodity groups.
For two groups—electrical apparatus and equipment and office machinery—•
data obtained for part of the period from several large manufacturing
concerns formed the basis for estimating Federal purchases.
The estimates of government purchases of durable equipment for the 1929—
39 period are probably conservative. Walsh-Healey data, as mentioned, are
not complete, and State and local purchases were estimated for only those
commodity groups in which such purchases were presumed to be of appreciable magnitude.
In this period, however, government purchases were not large. Even an
error of 20 percent—which seems unlikely, particularly in view of the relative

NATIONAL INCOME, 195 4 EDITION
accuracy of the aircraft item—would have affected the 1939 estimate of
manufacturers' sales of finished producers' durables (line 6, Exhibit 3) by
only 2 percent.
The estimated deduction for government purchases was negligible or less
than 5 percent of manufacturers' sales before this deduction (line 4) in special
industry machinery, mining machinery, general industrial machinery, metalworking machinery, agricultural machinery, tractors, fabricated metal products and instruments. It exceeded 15 percent only in construction machinery
and aircraft.
Exhibit 3.—Derivation of Tofal Producers' Purchases of Durable Equipment
Estimated by the Commodity Flow Method, 1939
[Millions of dollars]
1. Distribution of finished and mixed manufactured commodities, before deduction
of government purchases of durable equipment

-

25,978

1. Producers' durables
2. Consumer commodities
3. Combined allocated to
a. Producers' durables
b. Consumer commodities
b. Mixed, allocated to

--

-.

1,714
9,188
1,425
206
1,219
13,651

1. Producers'durables
2. Consumer commodities
3. Unfinished

1,112
8,210
4,329

2. Manufacturers' production of finished producers' durables, before deduction of
S, 032

government purchases [la(l)+la(3a)+lb(l)]

-23

3. Subtract: Change in manufacturers'inventories.
4. Equals: Manufacturers' sales of finished producers' durables, before deduction
of government purchases

2,790

6. Equals: Manufacturers' sales of finished producers' durables
7. Add: Transportation charges

59
2,849

C. Equals: Manufacturers' sales inclusive of transportation charges, distributed to.
Exports
Wholesalers
Retailers
Businesses and nonprofit institutions, for own use
. _

3,055
265

5 Subtract: Government purchases of durable equipment

9. Imports

outline, the deduction of manufacturers' sales flowing into exports was
approximately offset by the additions for wholesalers' and retailers' markups
on commoditiesflowingthrough trade channels. Only these three adjustments
were of any quantitative importance; even unexpectedly large errors in them
would not have had substantial effect on the final total.
Two further, minor differences in the application of the commodity flow
method to producer durables and consumer commodities in the 1929-39
period may be noted.
First, in the consumer commodity series exports were eliminated partly at
the producer level and partly at the wholesale level. In the producers'
durable estimates, exports were eliminated in their entirety at the producer
level. They were deducted (in step 8a) as part of the distribution of manufacturers' sales, inclusive of transportation charges.
The second difference is that the consumer commodity estimates, but not
the producer durable estimates, required a final adjustment for the addition
of general retail sales taxes.

12,327

a. Finished

a.
b.
c.
d.

129

___ .

17

_

1,131

10. Total purchases by wholesalers [8b+9]_
11. Subtract: Change in wholesalers'inventories
12. Equals: Cost of goods sold by wholesalers

499
1,114
1«8
1,068

6
..

1,125
249

13. Add: Wholesalers' markups

267
1,107

About nine-tenths of the 1947 total value of producers' durable equipment
was estimated by the commodity flow procedure. For the producer durable
series, the 1947 statistical basis for this procedure was an improvement over
that for the 1929-39 period. The principal factors were the greater extent of
product detail furnished by the 1947 Census of Manufactures and the availability from specific commodity studies of more information on product uses.
These factors facilitated and strengthened the process of selecting and allocating individual product items to arrive at manufacturers' sales of finished
producer durables. As in the 1929-39 period, the adjustments necessary to
convert manufacturers' sales to final market value did not bulk large in the
procedure and were virtually offsetting.
Lack of data on distribution of manufacturers' sales and on retailers'
operating expenses which had been provided in previous censuses handicapped the producer durable estimates comparatively little. As discussed
earlier, the sales distribution data cannot be employed in these estimates for
the allocation of "mixed" commodities in the classification process. Further,
distributors' markups—for which both the sales distribution and retail operating expense data would have proved useful in the 1947 work—are relatively small elements of the producers' durable totals.
The following description of the 1947 commodity flow estimates of producers' durable equipment is geared to Exhibit 4. Like the description just
given for 1929—39, it does not repeat numerous aspects of the basic commodity
flow method already covered in the section on Personal consumption expenditures for commodities.

1,374

14. Equals: Wholesalers' sales, distributed to.

Commodity Flow Estimates, 1947

a. Retailers

Classification of manufactured products

In the classification of manufactured product items, stricter criteria with
respect to commodity allocation were employed for producer durables for
7
1947 than in the 1929—39 estimates published in earlier national income
15. Total purchases by retailers [8c-H4a]
reports. In those estimates, simple assignments had been the general rule; and
428
16. Subtract: Change in retailers' inventories
possible allocations of commodities involving durable equipment use in either
146
17. Equals: Cost of goods so!d by retailers
very large degree or very minor degree were passed over. While reflecting
574
18. Add: Retailers'markups
importantly a lack of more detailed information on product composition and
19. Equals: Retailers' sales
use, this procedure rested on the general assumption that the inclusions and
2,749
exclusions entailed by such resort to direct selection and assignment would
20. Producers' purchases of durable equipment estimated by the commodity flow
tend to offset.
method [8d+14b+19]
Other steps ofprocedure.—Manufacturers' sales offinishedproducers' durables In the commodity flow benchmark estimates for 1947, greater use of small
(line 6) account for all producers' sales of finished durable equipment estiallocations was facilitated, and indicated, by the finer product detail in the
mated by the commodity flow method; there are no nonmanufactured pro1947 census—about 6,100 separate products as compared with about 4,000
ducer durables. In the derivation of consumer commodities by this method,
in the 1939 census—and by the extensive amount of specific commodity
the value of nonmanufactured foods (agricultural and fishery products) had
information that was available from both public and private sources. Asto be accounted for and added to manufacturers' sales.
signment and allocations made by the National Income Division staff
It has been noted that the problem of commodity allocation was less for
were checked and often modified on the basis of special commodity studies
producers' durable equipment than for consumer goods. Of further releand the advice of outside experts. The most important such check was
vance in gauging the comparative reliability of the producers' durable series
furnished by the Division of Interindustry Economics of the Bureau of Labor
is that the estimates in it which had to be made to convert producers' value
Statistics, which for a study of similar nature drew not only on published
to final market value were relatively small. The central fact here is that only
information but also on data obtained from special inquiries and other
a comparatively small volume of producers' durables flows through the retail
unpublished sources. For the producer durables work, an important inchannel, whereas retailers handle the large bulk of all commodities purchased
stance of the latter was a sample survey undertaken by the Bureau of the
by consumers. In 1939, manufacturers' sales of finished durable equipment
Census on the types of items included as "materials" (in contrast to plant
were virtually the same as the final value of producers' durables. In broad
and equipment) on reports to the 1947 Census of Manufactures.
b. Businesses and nonprofit institutions, for own use.




435

130

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

As already noted, relationships developed from the more intensive 1947
classification process were used for review and revision of the manufacturers'
sales estimates for 1929-39 (as well as for other, "secondary benchmark"
periods). By this revision, the 1939 proportion of manufacturers' sales of
finished producers' durables (before deduction of government purchases)
obtained through allocations from the combined or mixed categories was
increased from 16 percent to 44 percent. Although, it may be added, the
corresponding figure for 1947 is only 30 percent, the reduction from the
revised 1939 percentage resulted largely from the inclusion of railroad
equipment and trucks in the 1947 commodity flow table.
From the commodity analysis summarized in line 1 of Exhibit 4, manufacturers' sales of finished producers' durables (before deduction of government
purchases) were obtained by summing three cf the listed categories. These
include producers' durables (1) assigned directly as finished, (2) allocated
from the combined category of finished consumer and producer durable
goods, and (3) allocated from "mixed" commodities belonging in part to the
unfinished category.
The detailed commodity figures were next combined into broad categories
as listed in table 32, Part V. Further steps in the commodity flow procedure—
from the elimination of government purchases shown in line 3 through the
derivation of the final purchase value in line 16-—were carried out in terms of
the broad commodity groupings.

Government purchases of durable equipment
The estimates of Federal and State and local government purchases of
producers' durables utilized in the 1947 commodity flow work were prepared
largely by the Division of Interindustry Economics of the Bureau of Labor
Statistics. Since they were based for the most part on reported expenditure
data and constituted (for all commodity groups combined) only a 3 percent
adjustment of manufacturers' sales, it is very improbable that error in the
government-purchase estimates was the source of any appreciable inaccuracy
in the producer durable benchmarks.
In deriving Federal Government purchases, which formed about half of the
total government purchases for which an estimate was required, the Bureau of
Labor Statistics obtained detailed expenditure data, by commodities, for the
Army (including Air Force), Navy, Post Office, War Assets Administration,
Bureau of Public Debt, Bureau of Engraving and Printing, and Government
corporations (in total). In addition, a single control total of expenditures for
all other civilian agencies combined was furnished by the Treasury Department. This was distributed by commodities according to the pattern shown
by a sample of five agencies: Justice Department, Public Health Service,
Bureau of Old-Age and Survivors Insurance, Patent Office, and Food and
Drug Administration.
For a few commodities, where adjudged more reliable, estimates of Federal
Government purchases based on tabulations of contracts awarded under the
Walsh-Healey Act were substituted by the Bureau of Labor Statistics for those
derived from the agency expenditure data. The series were also checked
with BLS industry analysts, and modified in a few cases on the basis of
their independent estimates relating to Government purchases of durable
equipment.
In deriving the series on State and local government outlays for durable
equipment, the Bureau of Labor Statistics first prepared separate estimates
for States, cities, counties, and townships, towns, and special districts according to afive-wayfunctional breakdown: highways, hospitals, schools (primary
and secondary), higher education, and "all other." Equipment expenditures
under these five functions were next summarized for all State and local
governments and allocated to the various commodity groups. In the data
sources utilized, noted below, the commodity breakdowns either were available along with the totals by types of governmental unit and function or
could be estimated satisfactorily from collateral information.
The estimates of equipment expenditures for highway purposes, which
comprised about three-fourths of the State and local aggregate for 1947, were
based on data from the Bureau of Public Roads. The necessary data for
hospitals were compiled from directories of the American Hospital Association. The Office of Education—both through its regular reports and special
compilations of unpublished data—was the principal source of the estimates
relating to primary and secondary schools and to higher education. Basic
data for the estimates in the "all other" category of functions were taken from
reports by the Bureau of the Census on State, city, and county government



finances, as well as the decennial census publication on Government Finances in
the United States: 7942. The most recent equipment data tor counties referred
to 1946; they were moved forward to 1947 by data for cities. Equipment
purchases by townships, towns, and special districts were available only for
1942; they were extrapolated to 1947, separately for broad functions, by
means of estimates for counties.
Exhibif 4.—Derivation of Total Producers' Purchases of Durable Equipment
Estimated by the Commodity Flow Method, 1947
[Millions of dollars]
1. Distribution of finished and mixed manufactured commodities, before deduction of government purchases of durable equipment
a. Finished

81,993

__

42,840

1. Producers' durables.
2. Consumer commodities
3. Combined, allocated to
a. Producers' durables

_

10,076
23,063
3,711
734

_

39,153

b. Consumer commodities
b. Mixed, allocated to

2,977
-

1. Producers' durables
3,657
2. Consumer commodities
_ 18,494
3. Unfinished..
._
_ 17,002
2. Manufacturers' sales of finished producers' durables, before deduction of government purchases [la(l)+la(3a)+lb(l)]
3. Subtract: Government purchases of durable equipment

14,467
__

4. Equals: Manufacturers'sales of finished producers'durables

451
14,016

5. Add: Federal manufacturers'excise taxes.

137

6. Add: Transportation charges...

339

7. Add: Imports

46

8. Subtract: Changes in wholesalers'inventories
9. Add: Wholesalers'markups..

139

_

1,437

10. Subtract: Exports

2,154

11. Equals: Sales to retailers and final users

13,682

a. Manufacturers' and wholesalers' sales to final users (11-llb)

10,385

b. Eetailers' purchases.

3,297

12. Subtract from retailers'purchases: Change in retailers'inventories
13. Equals: Cost of goods sold by retailers

144
—

3,153

14. Add: Retailers'markups

1,004

15. Equals: Retailers'sales

4,157

16. Producers' purchases of durable equipment estimated by the commodity flow
method (lla+15)

14, 542

For most of the commodity groups the estimates of government equipment purchases required allocation between finished and unfinished products.
This was done (for Federal and State and local government purchases combined) on the basis of the proportions used to allocate factory sales.

Other steps in the procedure
Deduction of government purchases of durable equipment yielded manufacturers' sales of finished producers' durables, as shown in line 4 of Exhibit 4.
Up to this stage, the commodityflowprocedure for producer durables differed
from that for consumer commodities in respect to (1) the elimination of
government purchases as a subsequent, separate step instead of in the classification of Census of Manufactures data, (2) the omission of any adjustment
for manufacturers' inventory changes (all data relating to durable equipment
were reported on a shipments, or sales, basis in the 1947 Census of Manufactures), and (3) the omission of any allowance for production outside of
manufacturing since there are no nonmanufactured types of producer durable
commodities. Line 4 of the exhibit, therefore, corresponds directly to line 6 of
the 1947 commodity flow exhibit for personal consumption commodities.
With the exception that inventory adjustments were, in general, required
for the biennial 1929-39 benchmarks, the 1947 procedure for deriving
manufacturers' sales offinishedproducers' durables was generally the same as
in the earlier work. Thereafter, the procedure—like that for consumer commodities—required modification principally because of the unavailability of

NATIONAL INCOME, 1954 EDITION
1947 census data on the distributions of manufacturers' sales. These data,
together with corresponding distributions for wholesalers, had been used in
the basic commodity flow method for measuring the volume of producer
durables moving from manufacturers through wholesalers, and then through
retailers. In essence, they provided the basis for (1) immediately segregating
manufacturers' direct sales of finished durable equipment to final users
(businesses and nonprofit institutions), (2) applying wholesalers' markups to
goods sold by manufacturers to wholesalers and then sold directly to final
users, (3) applying cumulative wholesalers' and retailers' markups to goods
sold by manufacturers to wholesalers and then sold by them to retailers, and
(4) adding retailers' markups to sales made directly by manufacturers to
retailers.
In the 1947 method, omission of sales distributions from the manufacturing
census required foregoing the commodity flow method for wholesaling, since
data were lacking for estimating the amounts of finished durable equipment
purchased by wholesalers. Instead, as in the parallel case of consumer goods,
the resort was to estimate total wholesalers' markups (and inventory changes)
for each commodity and then to determine the amount allocable to finished
producer durables. On the basis of this innovation, and given the inability to
segregate manufacturers' direct sales to final users, a combined total consisting of sales to final users and to retailers was derived. {Exhibit 4, line 11.)
The latter sales were then estimated directly as the sum of retailers' purchases
from (1) manufacturers (by reliance on 1939 sales distribution patterns) and
(2) wholesalers (largely as a byproduct of the substitute procedure for
deriving wholesalers' markups). Through this estimation of retailers' purchases it was thus possible to (1) continue the commodity flow approach for
retailing and (2) obtain as a residual manufacturers' and wholesalers' sales
of finished durable equipment to final users.
By comparison of the 1939 and 1947 commodityflowexhibits, it can be seen
that the modifications of method were basically similar for producers' durables
and consumer commodities. For both series, necessary abandonment of the
commodity flow approach to wholesaling and the estimation of wholesalers'
markups (and inventory changes) through a special substitute procedure
yielded a composite total of sales to (1) retailers and (2) final users. And for
both series, direct estimation of one of these elements permitted the commodity flow approach (as modified) to be carried through to completion. A point
of specific note is that the element of direct estimation—retailers' purchases
in the case of producer durables and producers' direct sales to final users in
the case of consumer commodities—was chosen because it was very much the
smaller. The alternative of deriving it as a residual between the total and a
direct estimate for the larger element would have been a weaker procedure.
There follows, with reference to Exhibit 4, a brief outline of the various steps
entailed in converting manufacturers' sales to the final value of producers'
purchases of durable equipment estimated by the commodity flow method.

131

cation used in the producer durable series, and the producers' share (as
distinct from the consumer and intermediate) was allocated by proportions
computed for wholesalers' aggregate markups.
Except for this final step of allocation, the 1947 estimates of wholesalers'
inventory changes were prepared by the Bureau of Labor Statistics.
Addition of wholesalers' markups.—Basic to this step in the 1947 method
was the estimation of wholesalers' margins by wholesale commodity groups.
This was a four-step procedure—performed by the Bureau of Labor Statistics—which has been described in the section on consumption expenditures
for commodities. The fifth step in the BLS analysis of wholesaling, which
broke down the total margins by customer classes, could not be employed
for the producer durable series to indicate the proportion of the total that
was allocable to finished commodities. (This was chiefly because the "industrial users" category included sales of parts and other unfinished products
as well as finished equipment for the manufacturers' own use.) Instead,
after the wholesale commodity margins had been arranged according to
the producer durable classification, they were allocated by the National
Income Division to the producer and other (consumer and unfinished)
categories chiefly on the basis of the original allocations used for manufacturers' sales.
Subtraction of exports.—Export commodity data were obtained from the
Census Bureau's detailed tabulations. The same allocation factors generally
were applied to each specific commodity as had been applied to manufacturers' sales.
Distribution of sales to retailers and final users.—At this point, the 1947 com-

modity flow procedure for producers' durables arrived at the composite
total of sales to retailers and final users discussed above. Sales to retailers
("Retailers' purchases") were estimated separately, and sales to final users
thereby became available as a residual estimate. Unlike retailers' purchases,
these were at final market value and required no further adjustment.
Retailers' purchases, as noted, were estimated as the sum of purchases
from manufacturers and from wholesalers. In the estimation of the former,
it was assumed that the same proportion of a given type of producers' durable equipment was sold by manufacturers to retailers in 1947 as had been
shown for 1939 by the sales distributions data in the Census of Manufactures.
Manufacturers' sales to retailers thus computed were converted to retailers'
purchases by adding excise taxes, where applicable, and a transportation
allowance. Retailers' purchases from wholesalers were derived from the
BLS study (noted above) that formed the basis of the estimates of wholesalers' markups. In this study, step 5 obtained breakdowns of wholesalers'
sales (by wholesale commodity groups) according to customer classes. The
amounts of wholesalers' sales to retailers for the various wholesale commodity groups were arranged according to the producer durable classification, and then were allocated to the producer durable and other categories
Addition of excise taxes, transportation charges, and imports.—It may be seen that by use of the same allocations that had been employed at the manufacturers' sales level. This procedure was thus directly analogous to that
the first three adjustments of manufacturers' sales of finished producers'
followed for wholesalers' markups.
durables consisted of the addition of Federal manufacturers' excise taxes,
transportation charges, and imports. For these three estimates—all of minor
Subtraction of retailers' inventory changes.—The amount of retailers' inventory
magnitude—procedures were similar to those already described for consumer
change to be subtracted from retailers' purchases so as to derive the cost
commodities. It may be noted, however, that the computations of transportaof goods sold by retailers was estimated by (1) computing ratios of inventory
tion charges were not based solely on Interstate Commerce Commission data;
changes to retailers' purchases by line of trade and (2) applying these ratios
for three of the commodity groups (Electrical machinery, Trucks and buses,
for relevant lines of trade to the various producers' durable groups. Paralleland Railroad equipment) estimates obtained from the Bureau of Labor
ing the procedure for consumer commodities, inventory change-sales ratios
Statistics took account of means of transport other than rail. Also, the
based on Office of Business Economics data were converted for this purpose
estimates of imports relied mostly on allocation (on the basis of the producerto inventory change-purchase ratios.
durable share of manufacturers' sales of the same products), rather than on a
Addition of retailers' markups.—The cost of goods sold for the various proselective matching of import data against final product data, as was done for
ducer-durable commodity groups was multiplied by markup rates computed
the consumer commodity estimates.
for 1948 by weighting selected subsamples of Internal Revenue Service
returns by Census of Business sales data, and then extrapolated from 1948
Subtraction of wholesalers' inventory changes.—Ratios of inventory change to
to 1947 by means of Internal Revenue and trade association data. In this
sales were computed for a detailed list of wholesale commodities by extrainstance, too, the procedure was closely similar to that described for perpolating beginning-year inventories as shown in the 1948 Census back to the
sonal consumption commodities.
beginning of 1947 on the basis of Office of Business Economics data for
mei chant wholesalers, modified (as in the consumption commodity estimates)
Producers' purchases of durable equipment estimated by the commodity flow method.—
to include manufacturers' sales branches. The ratios of inventory change to
Addition of retailers' markups to the cost of goods sold by retailers yielded
sales were converted into ratios of inventory change to markups, and these
retailers' sales of finished producers' durables. (Line 15.) There remained
were applied to wholesalers' aggregate markups (derived in the step below)
only to combine retailers' sales and manufacturers' and wholesalers' sales
to obtain aggregate inventory changes by wholesale commodity groups. The
(line lla) in order to obtain the final total of producers' durable equipment
wholesale commodity estimates were next arranged according to the classifiestimated by the commodity flow method. (Line 16.)




132

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Commodity Flow Estimates, 1940-46
As noted in the introductory remarks, a modified commodity flow method
was followed in estimating producers' durable equipment for 1940—46 for
those groups to which this method was applied in detail for the 1929-39
period. In terms of the basic step of estimating manufacturers' sales, the
figures for 1942-46 may be characterized as "secondary" benchmarks and
those for 1940-41 as interpolations.
The same commodity flow steps, except for omission of the adjustments for
inventory change, were followed in deriving estimates for 1940-46 as for the
earlier period. Because of the nature (or lack) of available source materials,
however, the methods, described briefly below, were different.

Derivation of manufacturers' sales
For most of the producer durable groups, comprehensive and detailed
commodity sales estimates (before deduction of government purchases and
exports) could be derived for the years 1942—45. The primary source was the
reports submitted on Form WPB-732 by the largest private and governmentowned plants engaged in fabricating or assembling metal products beyond
the primary stages. These reports—made quarterly from the fourth quarter of
1943 through the second quarter of 1945 and then monthly through the
second quarter of 1946—were collected and tabulated by the Bureau of the
Census for the War Production Board, and published in the Census
Facts for Industry releases. The "732" tabulations accounted for an
estimated 90 percent of sales in the metal-products industries covered. These
were raised to full coverage on the basis of similar commodity information for
the first quarter of 1945 compiled by the Census Bureau for the Smaller War
Plants Corporation.
The detailed commodity data derived from the "732" reports as supplemented by the Smaller War Plants Corporation data, were extrapolated from
the fourth quarter of 1943 back through the third quarter of 1942 on the
basis of unpublished commodity sales figures collected and compiled by the
Census Bureau for the War Production Board. These compilations covered
about 3,100 large metal-products plants accounting for four-fifths of all
fabricated metal products in the first quarter of 1945.
The individual commodity data were then extended from the third quarter
of 1942 through the first quarter of that year by less detailed commodity data
compiled from an earlier and smaller Census-War Production Board sample.
The number of individual commodity items which could be derived from
the "732" and related data was substantially less than that given in the
Census of Manufactures. The necessary product assignments and allocations
were made almost entirely on the basis of relationships developed for 1947,
although in some instances the availability of specific information from either
Government or private sources permitted a modification of the postwar ratios.
Apart from failing to account for changed composition of the commodity
groups between the war and postwar periods, the general use of 1947 relationships in the classification of manufacturers' sales may have resulted in
some understatement of the allowance for parts and other unfinished products (and an overstatement of finished commodity sales) because of their
probably more widespread use during the war.
Estimates of manufacturers' sales of producers' durables (before deduction
of government purchases and exports) were obtained for 1940 and 1941 by
interpolation between the 1939 and 1942 estimates on a commodity group
basis. For this purpose, commodity data were almost wholly lacking.
As much information as possible was gathered from specifically relevant
sources. For example, the mining machinery series was interpolated by an
index based on capitalization of the Oklahoma tax on sales of oil field equipment and of the California tax on sales of oil well and refining supplies. As
part basis for interpolation of the metal working machinery series, sample
figures on machine tool sales were obtained from the War Production Board.
For tractors, comprehensive sales data were available from the Bureau of the
Census. However, it was also necessary to make considerable use of less direct
indicators, such as industry payroll and sales data, chiefly the annual corporate sales tabulations of the Internal Revenue Service.

categories—Federal military agency purchases, Federal civilian agency
purchases, Federal Government purchases for Government-owned defense
plants, and State and local government purchases. Federal military purchases were computed for the years 1942-45 as it was believed that such
purchases during 1946 were negligible.
Navy Department purchases of selected items of equipment were estimated
largely on the basis of unpublished data on contracts awarded by fiscal years.
As a rough allowance for the lag between contract awards and deliveries,
these data were treated as deliveries during the calendar year in which the
fiscal year ended. For the most part, the data used to derive Army purchases
for the years 1942—45 (exclusive of Government defense plants) were those
available on a calendar-year deliveries basis in Statistical Review, World War II,
A Summary of ASF Activities, Statistics Branch, Control Division, Hq., A. S. F.,
War Department. For both the Navy and Army, the producer-durable share
of total purchases for a given type of commodity was derived either by the
application of producer-total ratios observed in manufacturers' sales allocations or by the selection or omission of items in their entirety.
For some commodity groups, the largest of which was office and store
machines, the Navy and Army data were not available in sufficient product
detail to use as the basis for estimating military agency purchases. Instead,
the estimates were derived from "claimant agency" data shown in CensusWar Production Board Facts for Industry reports. However, these data, which
recorded the value of products allocable to military claimants, were not
strictly appropriate for the purpose. It is believed that they might also have
covered to some extent products allocable to final users (such as privately
owned manufacturing plants engaged in war output).
Federal purchases of equipment for Government defense plants were
estimated in the aggregate for the entire period 1942-45 from data furnished
in Report on Government Owned Industrial Plants As Of September 30, 1947, War

Assets Administration. This aggregate was distributed by years on the basis of
the equipment value of Federally-financed facilities put in place, as shown by
the Census-Civilian Production Administration report on Facilities Expansion, July 1940-June 1945. It was then necessary to distribute the total for each
of the years 1942—45 among the producer durable commodity groups. This
was done mainly on the basis of the claimant agency data noted above—
after, however, they had been adjusted (largely on the basis of the Army and
Navy tabulations) to exclude estimated amounts allocable to military agencies
as distinct from Government-owned defense plants.
The methodology underlying the 1942-45 estimates of equipment purchases
by Federal civilian agencies and by State and local governments is described
below under "Commodity Flow Estimates, 1948-52."
Chiefly because of weaknesses centering in the estimates of Federal equipment outlays for Government defense plants, the government-purchase
deductions for the years 1942-45 may be somewhat too large, and the breakdowns by commodity groups are clearly subject to considerable error. This
limitation of the government-purchase totals, however, may be viewed in
conjunction with the probable overstatement of the manufacturers' sales
estimates (as noted above) because of inadequate elimination of parts and
other unfinished products. The two biases, though not at all necessarily of
similar magnitude, would tend to offset.
For the years 1940 and 1941, the principal source of data for estimation of
government purchases of producers' durable equipment was the CensusCPA report on Facilities Expansion, July 1940-June 1945. From this report,
ratios of non-Federally financed equipment put in place to total equipment
put in place were calculated for industries considered to be the main purchasers of the commodoties in each major producer durable group. These
ratios were used to interpolate between ratios for 1939 and 1942 of private
domestic sales to total manufacturers' sales (before deduction of government
purchases and exports), and the ratios so derived for 1940 and 1941 were
applied to total manufacturers' sales for those years. In this process, explicit
allowances for State and local government purchases were made for some
commodity groups in order to avoid evident bias.
The facility equipment data were not very appropriate for this type of use.
The matching of equipment values classified by industry of purchaser with
producers' durable equipment purchased was necessarily quite rough.

Subtraction of government purchases

Other adjustments

For the period under review (except 1940 and 1941), estimates of government purchases of producers' durable equipment were made for four broad

As in the census-year benchmark estimates, the export adjustments for
1942-46 and the import adjustments for 1940-46 were based on detailed.




NATIONAL INCOME, 195 4 EDITION
commodity tabulations of the Bureau of the Census. To them were applied
the same producer-total ratios as used in deriving manufacturers' sales.
For 1940 and 1941, exports were covered in the step eliminating government
purchases.
For each of the producer durable groups, transportation charges, wholesalers' markups, and retailers' markups were estimated for the years 1940—46
by applying percentages obtained by straight-line interpolation of the
percentages which each of these several elements constituted of manufacturers' sales of finished producers' durables in the benchmark years 1939 and
1947.

Commodity Flow Estimates, 1948-52
For the commodity flow portion of the producers' durable series, the method
for 1948—52 entailed the following series of estimates: (1) derivation of
manufacturers' sales (prior to the government deduction) for the major
commodity groups, (2) addition of Federal manufacturers' excise taxes, (3)
deduction of government purchases, (4) addition of transportation charges,
(5) deduction of exports, (6) addition of imports, (7) subtraction of changes
in wholesalers' and retailers' inventories, and (8) addition of wholesalers' and
retailers' markups.

Derivation of manufacturers' sales
The Census Bureau's annual Survey of Manufactures provided the basis of
the estimates of manufacturers' sales (before deduction of government
purchases) for the years 1950—52. Covering all large establishments and a
representative sample of smaller establishments, the survey collected data on
manufacturers' shipments (sales) which, with respect to producers' durable
equipment, were available in satisfactory detail and showed generally small
sampling variation. In the task of classifying the census product data, the
assignments and allocations used were largely carried over from the 1947
benchmark study.
For the various producer durable groups, the commodity sales estimates
for 1948 and 1949 largely represent interpolations of the 1947 and 1950
benchmarks on the basis of industry sales data from the manufac urers' sales
series of the Office of Business Economics—derived from a sample of reporting
companies which currently account for about 45 percent of total manufacturing sales. Exceptions to this general procedure were afforded by construction machinery and aircraft, for which groups Census current reports on
manufacturers' commodity sales were available. In addition, the 1948-49
interpolations for cutlery and hand tools and railroad equipment were made
directly in terms of final cost to purchasers, not manufacturers' sales. Sales of
hardware stores from the Census Bureau's monthly retail trade sample were
used for the cutlery and hand tools group; Interstate Commerce Commission data on equipment purchases of Class I railroads, for the railroad equipment group.
Use of industry data on which to base the movement of commodity
sales has drawbacks. The data include indeterminate amounts of unfinished
and consumer products; and they are also inappropriate for any particular
group to the extent of including secondary products classified in other producer durable groups and of omitting products in that group included in the
sales of manufacturing industries other than the one used for an index.

Deduction of government purchases
In the derivation of Federal military agency purchases of producers'
durable equipment for the years 1948-51, estimates for 1947 and for the
latter half of 1951 were used as benchmarks. The 1947 estimates, as already
described, were based on detailed expenditure data reported by the Army
(including Air Force) and Navy. The second-half 1951 figures were made
from quarterly data reported by the National Production Authority on the
basis of tabulations by the Bureau of the Census from a sample of approximately 8,000 of the largest plants engaged in fabricating or assembling metal
products. The plants included in the sample had produced approximately
90 percent of the products of these industries in 1947.
For the reporting plants, the NPA data for the third and fourth quarters of
1951 showed for each commodity (in the same detail as the Census Annual




133

Survey) both total sales and those having military rating symbols. Only total
sales, however, were available for the first and second quarters. To obtain
estimated purchases by Federal agencies for the third and fourth quarters,
military-rated sales were raised to full coverage by the relationship of total
sales for the year reported in the Census Annual Survey to those shown in the
NPA reports. This procedure probably erred somewhat in the direction of
overstating Government purchases for items which could be used either as a
facility or as a component, since some purchases of this type by private firms
were accorded military ratings. In line with the usual procedure, the share of
total purchases allocable to the producers' durable category was estimated
from proportions applied in the allocation of sales at the manufacturing level.
For the period between 1947 and the third quarter of 1951, estimates of
Federal military agency purchases were obtained through interpolation by a
series based on Navy contracts awarded and Army obligations incurred.
These data, available in considerable commodity detail, were moved forward six months so as to make them reflect more nearly the period of actual
deliveries.
Purchases by Federal military agencies during 1952 also were based very
largely on the NPA data. However, for one important group—metalworking
machinery—the military rating treatment by NPA differed from that for
other commodities. It included many purchases by private firms and, therefore, could not be used to derive purchases by Federal military agencies.
For this commodity group, military purchases during 1952 were estimated at
the same proportion of manufacturers' total sales as obtained for the last
quarter of 1951.
For purchases of producers' durable equipment by Federal civilian agencies, benchmark estimates were prepared for 1951 from information collected
by the Bureau of Labor Statistics from the individual agencies. These estimates closely paralleled the ones that were made for the 1947 benchmark
study. Figures for 1948-50 were obtained by interpolation on the basis of the
National Income Division series (appropriately adjusted) on Federal purchases of goods and services for nonwar purposes. The same series—available
only in total and not in a commodity breakdown—was used to obtain estimates for 1942-46 and 1952 through extrapolation of the 1947 and 1951
benchmarks.
With respect to the State and local government series, the 1947 benchmark
figures were extrapolated to other years of the 1942—52 period by estimates of
equipment outlays by types of governmental unit and function, based on
data from governmental finance reports of the Bureau of the Census. To
derive these estimates, reported data on total capital outlays for equipment
for each type of unit were broken down by broad functional categories
according to Census data on total capital outlays (covering land and bi'ilrlings as well as equipment).

Other adjustments
Following the deduction of government purchases to arrive at manufacturers' sales of finished durable equipment by commodity groups, the first
adjustment made in the 1948-52 estimates consisted of the addition of Federal
manufacturers' excise taxes. For this relatively small item, concentrated
largely in the truck series, data were readily available from Internal Revenue
reports. Transportation charges were next added to the commodity group
estimates of manufacturers' sales. For this element, resort was mainly to the
application of 1947 percentages, although for some groups the figures based
on these percentages were modified after a summary check into the annual
changes indicated by ICC data on freight revenue.
The export and import adjustments of the various producer durable
groups for 1948-52 were made on the same basis as for the prior period.
Detailed product tabulations of the Bureau of the Census were grouped into
producer durable categories and then allocated on the basis of proportions
utilized in the classification of manufacturers' sales.
Adjustments for changes in wholesalers' and retailers' inventories—generally of negligible importance—were based mainly on the inventory series for
trade maintained by the Office of Business Economics. Inventory changes
computed for the benchmark year 1947 for major commodity groups were
projected by the OBE relevant line-of-trade figures.
The addition of wholesalers' and retailers' markups to manufacturers'
sales after 1947 was made through application of 1947 relationships.

134

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

coaches, on the basis of data of the Automobile Manufacturers Association
on the number of units sold and average factory value. The Census
Bureau reported data on factory sales of truck trailers. These series were
Business motor vehicles
converted to final purchase values by adding the same percentage allowances
The passenger car estimates have been discussed in the section on Personal
for transportation costs and dealers' markups as used for 1941, deducting
consumption expenditures for commodities. They consist of the business share
estimated government purchases, and adding 10 percent (again largely
of the combined consumer-business total of new car purchases and gross
arbitrary) of the final value of new trucks and busses for dealers' margins on
margins on used car sales. The percentage allocation between consumers and
sales of used vehicles.
business is based on limited information for the 1934-37 period and has been
The next step was the computation of a series to interpolate between 1947
held constant except for modifications in the World War II period to take
and the third quarter of 1951. Beginning with that quarter, the availability of
account of the probable effects of rationing.
NPA Government-purchase data (as noted above) again made the appliThe estimation of producers' purchases of trucks, truck trailers, and busses
cation of the commodity flow method feasible.
has varied during the period. For the years 1929—41, the method was essenThis interpolating series consisted of (1) the value of trucks and busses,
tially one of multiplying number of units by an estimated average price. For
derived as the product of estimated average price and the sum of new truck
1947, the second half of 1951, and 1952, the commodity flow method was
registrations and domestic shipments of motor coaches adjusted to exclude
used. Estimates for other years of the period after 1941 represent interpolations
civilian government purchases and (2) the value of truck trailer shipments,
on the basis of a series obtained by multiplying number of units by average
based on current Census reports. For the period 1947—49, average prices
price.
used for the truck and bus series were computed from Automobile ManufacFor the 1929-41 period, the number of trucks and busses purchased by
turers Association data on the total factory value of sales and number of
producers at retail was derived by deducting estimated government purunits sold, with adjustment to exclude exports as reported by the Bureau of
chases from comprehensive totals reported annually by the Automobile
the Census. To forestall the possibility that shipments of military vehicles
Manufacturers Association in Automobile Facts and Figures. The number of
might again make the price series based on Automobile Manufacturers
units purchased by producers directly from manufacturers was taken as 20
Association data inapplicable to private purchases, a different method of
percent of the number purchased at retail. This was a rough estimate based on
estimating average price was adopted for 1950 and 1951. The 1949 average
data in the 1929, 1935, and 1939 Census of the Distribution of Manufacturers'
price of trucks and busses was extended to the third quarter of 1951 by a
Sales covering sales of trucks, busses, passenger cars, bodies, parts, and
quarterly series computed in somewhat indirect manner. A constant dollar
accessories.
series was first obtained by multiplying new registrations for each of six
The number of trucks and busses purchased directly from manufacturers
size classes (from the R. L. Polk Co.) by a rough estimate of average price for
was multiplied by an annual average factory price as computed from data
1948. The quarterly totals of the six classes were converted to current dollars
in Automobile Facts and Figures and raised to cover transportation charges. The by means of the Bureau of Labor Statistics index of wholesale prices of trucks,
latter were estimated on the basis of information provided by the Office of
and the current dollar totals were then divided by total new registrations.
Price Administration. Producers' purchases at retail were multiplied by this
average price series after it had been further increased to allow for dealers'
markups, computed from data on passenger cars published in Automobile Railroad and transit equipment

Other Methods

Facts and Figures.

The estimated total value of trucks and busses purchased by final users
both directly and at retail was adjusted to include bodies sold separately from
chassis.
Purchases of truck trailers were estimated separately and added. The value
of truck trailers produced, benchmarked on Census of Manufactures data for
1935, 1937, and 1939, was adjusted to exclude exports and government purchases and to include transportation charges. A markup allowance was added
to the estimated portion of the total sold through dealers.
The methods employed for making the prewar estimates were not suitable
for the years 1942-45. The price series used in the 1929-41 estimates were
greatly affected during the war by the large volume of high priced military
vehicles. Also, the methods of estimating government purchases would not
have been satisfactory for the war period, when such purchases were of substantial magnitude.
Data on the numbers of heavy, medium, and light trucks and busses shipped
to civilian domestic users—the approximate measure desired—were supplied for the years 1942—45 by the Office of Defense Transportation. Separately for the three size classes of vehicles, numbers shipped were multiplied
by price series representing special estimates for 1939 extrapolated to 1942
by indexes of the Bureau of Labor Statistics and to 1945 by unpublished
data of the Agricultural Marketing Service. The weights used were for 1939
in order to exclude the effects of war-period military purchases.
The Office of Defense Transportation also provided figures on the number
of truck trailers shipped for civilian domestic use during the years 1942-45.
These totals were multiplied by an average 1939 factory value extrapolated
to 1942 and then to 1945 by the series for light trucks (noted above) used in
estimating new truck purchases.
Dealers' gross margins on sales of used vehicles, assumed to be reflected in
the average price data used for the 1929—41 estimates, were explicitly added
to the series beginning with the 1942 estimates. For the years 1942-45,
such margins were taken as 15 percent of estimated purchases of new trucks
and busses.
For the interpolating series, the 1946 and 1947 estimates involved, first,
derivation of the total factory value of domestic sales of trucks and motor




The most important component of this group, equipment expenditures of
class I railroads, was obtained from the Bureau of Railway Economics of the
Association of American Railroads for the prewar and war periods and from
the Interstate Commerce Commission for 1946. This series was raised (about
6 percent) to allow for the estimated expenditures of class II and class III
railroads. Other components of the group total include (1) equipment
expenditures of transit corporations for electric cars and trolley coaches, data
for which have been published annually by the American Transit Association
in its Transit Fact Book; (2) expenditures for Pullman Corporation cars, as
published in Statistics of Railways by the Interstate Commerce Commission:
and (3) the value of tank car purchases (not included in the railroad equipment figures), estimated by multiplying the number of tank cars as reported
by the American Railway Car Institute by the average value of tank cars as
published annually in Statistics of Railways.

As covered earlier, this series was transferred to the commodity flow category of producers' durable equipment beginning with 1947.

Ships and boats
For the years 1929-38, business purchases of ships and boats were derived
by the commodity flow method. The estimates since 1939, which are subject
to significant limitations, have been prepared as the sum of separate series for
subsidized ships, ships completed under private contract, and boats.
The Maritime Adiminstration has provided data on sales of subsidized merchant vessels of 2,000 tons or more. Valuation is at cost to the purchaser,
exclusive of the Government subsidy. In order to match the timing of entries
in the government account, no attempt has been made to convert sales to
value of work in place.
The Maritime Administration also has provided lists of merchant vessels
completed under private contract. These have been valued at figures given by
the same source for comparable vessels, either subsidized or Government
purchased, and for recent years with the aid of specific price data furnished
by the National Federation of American Shipping. In addition, for the postWorld War II period allowance has been made for the capitalized cost of
repairing ships and converting them to peacetime use. This was done largely

NATIONAL INCOME, 1 9 5 4 EDITION
on the basis of information reported in the Marine Engineering and Shipping
Review.
The estimates for smaller craft—covering private business purchases of
motor vessels and barges, lighters, and other unrigged boats—represented
about three-fifths of the ships and boats total (excluding conversion cost) in
1947. The series is benchmarked on Census of Manufactures data for 1939
(value of work done) and for 1947 (value of deliveries), adjusted to exclude
exports and Government expenditures on vessels built in private ship yards.
The figures for 1940-46 and 1948-52 are interpolations and extrapolations
based on a series representing annual tonnages of private vessels under 2,000
tons documented for commercial use (after elimination of the estimated
tonnage of used vessels documented) as published by the Bureau of Customs,
adjusted for price changes by the Maritime Administration index of construction costs.

Nature of Recent Revisions
As mentioned in the introduction to this section, the series on producers'
purchases of durable equipment shown in table 32 of Part V represent a
revision of previous estimates for the entire period since 1929. Important data
not previously available were incorporated into the estimates, a number of
changes in procedure were made, and the Standard Industrial Classification
(1945) was adopted as the basis for the producer durable commodity groupings. As a general result, the new estimates are believed to constitute a
significant improvement over those published in previous editions of the

135

Incorporation of this new information, it may be noted, has permitted
publication of a commodity breakdown of producers' durable equipment for
1946 and subsequent years, as well as for the prior period. Until this revision
work, the basic statistical data were not suitable for extension of the commodity groups on a satisfactory basis after 1945.
Exhibit 5 presents a comparison of the new and previous estimates of
producers' durable equipment for the years 1929- 52. It will be seen that for
a number of years the revisions were appreciable. In very large degree they
occurred in the segment of the series comprised of products customarily
charged by business to current expense, instead of amortized through depreciation allowances. As indicated earlier, this "fringe" segment of producers' durables is difficult to measure.
For the large bulk of the producers' durable total comprised of items
normally capitalized by business, the statistical revisions were, on the whole,
rather small. They reflected, of course, a tendency for changes in the underlying data to be offsetting. For example, the estimates of manufacturers'
sales of finished producers' durables for 1942—45 were revised downward
considerably, but this change was substantially counterbalanced in the final
estimates by the marked reduction that was made in the subtraction for
government purchases. Also, although the revisions (in the capitalized segment) are not relatively large for the total, they represent varying and
partially compensating changes in the commodity groups.

NATIONAL INCOME supplement.

Exhibit 5.—Comparison of New and Superseded Series for Private Purchases
of Producers' Durable Equipment, 1929-52
[Billions of dollars]

New series
Year
Total

Superseded series

Capital
Capital
outlays Excluding
outlays Excluding
charges
charges
charged to current Total' charged to current
to current expense
to current expense
expense
expense

1929

5.8

0.3

5.6

6.4

0.6

5.8

1930
1931
1932
1933
1934

4.5
2.8
1.6
1.6
2.3

.2
.2
.1
.1
.2

4.2
2.7
1.5
1.5
2.1

4.9
3.2
1.8
1.8
2.5

.5
.4
.3
.3
.3

4.4
2.8
1.5
1.5
2.2

__

3.1
4.2
5.1
3.6
4.2

.2
.2
.2
.2
.2

2.9
4.0
4.9
3.5
4.0

3.4
4.5
5.4
4.0
4.6

.4
.5
.5
.4
.5

3.0
4.0
4.9
3.5
4.0

-

5.5
6.9
4.3
4.0
5.4

.3
.4
.3
.4
.5

5.2
6.6
4.1
3.7
5.0

6.1
7.7
4.9
4.1
5.7

.7
.8
.6
.5
.6

5.4
6.8
4.3
3.6
5.2

. . . -.

7.7
10.7
16.7
19.1
17.8

.5
.7
.8
.9
.8

7.2
10.0
15.8
18.2
17.1

7.5
12.3
17.1
19.9
18.7

.7
1.4
1.6
1.9
1.7

6.8
10.9
15.5
18.0
17.0

21.1
23.2
23.3

.9
1.1
1.0

20.2
22.3
24 fi
22 1
22.3 ! 25.4

2.1
2.6
2.8

20.2
22.1
22.6

1935
1936
1937
1938
1939

.
.

._

1940
1941
1942
1943
1944

..

-

1945
1946
1947
1948
1949
1950
1951
1952

.

--

-- --

.

--

-

1. As published in the National Income Number (table 2) of the July 1953 SUBVEY OF
OUEEENT BUSINESS.

Statistical sources from which new data were obtained for the producers'
durable series consisted mainly of the 1947 Census of Manufactures, the 1948
Census of Business, the Census Bureau's annual (sample) surveys of manufacturers for 1950—52, tabulations of Government purchases of producers'
equipment for military purposes during World War II, and reports of the
National Production Authority containing Government purchase data for
1951—52. In addition, it was possible through the availability of specific
commodity studies and the cooperation of specialists, both Government and
private, to improve the assignments and allocations of data on manufac-




11. CHANGE IN BUSINESS INVENTORIES
This section deals with the change in nonfarm business inventories and the
inventory valuation adjustment. Farm inventories are covered in the section
on the Income of unincorporated enterprises.
The basic sources of the nonfarm inventory estimates are reported accounting data on the value of inventories. These data have a high degree of coverage. Internal Revenue Service tabulations from annual corporation tax returns alone account for about four-fifths of the estimated value of nonfarm
inventories. Periodic data on the value of noncorporate inventories are available from Internal Revenue Service and Census Bureau compilations.
The adequacy of the inventory estimates included in national income
statistics is, however, less than might be suggested by the coverage of the
book value data and the reliability of the basic sources from which they are
drawn. The extension of reported values to full coverage introduces seme
uncertainty into the estimates, but their main source of error stems from the
fact that the accounting methods underlying the reported data are divergent
and inappropriate for national income purposes. Inventory calculation at the
level of the individual business firm is a complex problem, and existing accounting methods vary widely both with respect to the scope of the cost elements (especially overhead costs) included in the inventory account and with
respect to the costing procedures used to charge goods to cost of sales and to
inventories, respectively.
The scope-of-cost limitation of the basic data reported by business is accepted in estimating the inventory components of national income and
product. However, a uniform and appropriate costing procedure is substituted for the divergent procedures used by business firms (such as first-in,
first-out and last-in, first-out). This results in a measure, for inclusion in the
gross national product, of the physical volume change in inventories valued
at average prices during the period. The excess of this measure over the bookvalue change in inventories represents the "inventory valuation adjustment."
This adjustment is added to the business-income components of national income and secures measures of earnings from current production consistent
with the treatment of inventories in the gross national product.
This adjustment of the reported book value data is a quite difficult procedure, involving the revaluation of the entire volume of nonfarm business
inventories given only limited knowledge of the prices reflected in them.
Also important is that the estimates of inventory change included in national income and product are calculated as the difference between large and
usually volatile inventory totals at two points of time. Even small errors in the
data on total inventories can lead to large relative errors in the estimates of

136

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Finally, it should be noted that the comprehensive accounting data on in*
ventories become available only after a lag of several years. Current estimates
must be based upon less satisfactory sources.

The following comments deal with those steps of the estimating procedure
which need amplification.

Step 1: The book value aggregates
General Estimating Procedure
The procedure for deriving the current value of the physical change in
nonfarm business inventories and the associated inventory valuation adjustment is carried out separately by detailed industry groups. The limitations
already noted of the all-industry estimates of the inventory valuation adjustment attach to the individual industry estimates to an even greater degree.
The general procedure of estimation involves six principal steps, as summarized below.
(1) Reported book values of year-end inventories are raised to complete
coverage.
(2) Estimates of the portion of total book value that is reported on a last-in,
first-out (LIFO) basis are deducted from the totals and separately processed.
This step is necessary because the valuation procedure underlying LIFO
inventories requires an adjustment procedure which differs from that applicable to the remainder of business inventories.
(3) The estimates of book value of non-LIFO inventories are converted to
a constant price basis by means of price deflation procedures.
(4) The change in these inventories at constant prices is obtained by subtracting beginning from ending inventories at constant prices.
(5) The current value of the physical change in inventories is obtained by
multiplying the change in inventories at constant prices by the ratio of current prices to the constant price base.
(6) The inventory valuation adjustment is obtained by subtracting the
change in the book value of inventories from the current value of the physical change in inventories.
Step (5) yields the inventory component (other than LIFO) of gross national product. The result of step (6) constitutes the corresponding adjustment to corporate and noncorporate enterprise incomes, which are calculated initially on the basis of the inventory accounting methods underlying
reported inventory data.

Sources and methods of estimating the book values differ for the corporate
and noncorporate sectors, and as between past periods and the recent years
for which final information is not yet available.
The final source of data on the book value of corporate inventories is
Statistics of Income—Part 2, the annual compilation of corporate income tax
returns published by the Internal Revenue Service. The reported totals are
raised on the basis of cost of goods sold (by about 1 percent) to take account
of corporations not reporting balance sheet data. Since corporations account
for about four-fifths of nonfarm inventory holdings, a substantial portion of
the estimates of the total book value of nonfarm inventories rests upon a
source considered to have a high degree of reliability.

Exhibit S.—Estimated Book Value of Total Corporate and Noncorporate
Inventories and Industrial Breakdown of Noncorporate Inventories, End of
1950
Item

Millions
of dollars

Percent

Total, all industries

69,343

100.0

Corporate.-.

55,101

79.5

Noncorporate

14,242

20.5

159
630
1,238
11,801

1.8
17.0

2,189
9,612

3.2
13.9

Mining
Contract construction
Manufacturing
Wholesale and retail trade
Wholesale trade
Retail trade
Services

.2
.9

414

As can be seen from Exhibit 2, the bulk of estimated noncorporate invento-

Exhibit 1.—Derivation of Change in Business Inventories and Inventory
ries is in wholesale and retail trade. Estimates are also made for mining, conValuation Adjustment for the Corporate Sector of the Apparel Manutract construction, manufacturing, and services. Data on noncorporate infacturing Industry, 1950
ventories in finance, insurance, and real estate; transportation; and comValue data in millions
of dollars

Item

1949
1. Book value of year-end inventories.

- -

1,012
95.1
95.2

3. (a) Index of cost valuation (1947=100)
(b) Index of market valuation (1947=100).
(e) Year-end inventories in 1947 dollars (line 1 divided by line 3a
or 3b, whichever is lower)

1,004

1950
1,501
110.2
114.4
1,362

4. Change in inventories in 1947 dollars (from line 3c)

298

5. (a) Index of current valuation (1947=100)
(b) Change in inventories in current dollars (line 4 times line 5a) _

100.2
299

6. (a) Change in book value of inventories (from line 1)
(b) Inventory valuation adjustment (line 5b minus line 6a)

.

489
-190

The estimating procedure is illustrated in Exhibit 1 by actual data for a
single industry. Calculations of a similar nature are made for mining, construction, transportation, communications and public utilities, wholesale
trade, retail trade, and for about 20 industries in manufacturing. In the
alcoholic beverages and tobacco manufacturing industries, however, direct
quantity data instead of deflation procedures are used to estimate changes in
inventories.
To simplify the exhibit, an industry was selected in which LIFO inventories
are negligible, and consequently step (2) is omitted. Also, the exhibit is confined to the corporate sector of the industry. Similar calculations are made
for the noncorporate sector of each industry listed above for which book
value estimates are available. As will be seen, the derivation of book values
differs for the two sectors. However, the data and methods used in the revaluation of these book values are the same within each industry.



munications and public utilities are lacking. The amounts involved must be
insignificant and are not included in the estimates.
The main sources of the noncorporate inventory estimates for trade were the
Censuses of Wholesale and Retail Trade for 1929,1933,1935,1939, and 1948
and Internal Revenue Service tabulations for 1939 from the income tax returns
of sole proprietorships and partnerships. The procedures for estimating noncorporate inventories in trade vary considerably according to the nature of
the available information. For some years census data on total inventories
are available which can be accepted as benchmarks. For non-Census years
prior to 1939 the procedure involves the multiplication of noncorporate sales
series derived in the estimation of noncorporate business income by inventorysales ratios. These ratios are benchmarked on tax return and industrial
census data for unincorporated business. Corresponding corporate inventorysales ratios are used widely for interpolation and extrapolation of the noncorporate ratios. Since 1939 the interpolation and extrapolation of census
benchmark estimates of noncorporate trade inventories for 1939 and 1948
have been based on the movement of the noncorporate components implicit
in the published wholesale and retail inventory estimates of the Office of
Business Economics.
A diversity of procedures is followed for estimating noncorporate inventories
in industries other than trade. As can be seen from Exhibit 2, the amounts
involved are very small.
As already noted, the final sources for estimating the book value of inventories, both corporate and noncorporate, become available only with a
considerable lag. Prior to the receipt of this information, inventory book
values are extrapolated on the basis of interim data. The extrapolation for
the large manufacturing sector—accounting for 50 percent of the nonfarm
total at the end of 1950—is based upon the Industry Survey of the Office of
Business Economics. Reports of inventory holdings tabulated in connection

NATIONAL INCOME, 1954 EDITION
with this survey cover more than one-half of the estimated total and are
weighted by industry group and asset size class.
The extrapolation of wholesale inventories (15 percent of the nonfarm
total) is on the basis of the Office of Business Economics series on wholesale
inventories. This is derived mainly from a sample of inventories of merchant
wholesalers reporting to the Bureau of the Census, together with Department
of Agriculture data on warehouse-stocks of selected farm products. These two
sources cover about one-tenth of total stocks, and the data tabulated are
weighted by kind of business.
The extrapolation of retail trade inventories (one-fourth of the 1950 nonfarm total) is on the basis of the Office of Business Economics series on retail
inventories. This in turn rests upon Census Bureau estimates of total year-end
inventories in retail trade built up from a complete count of large multi-unit
organizations and geographic area samples of all other stores.
Other nonfarm inventories are extrapolated into the current period mainly
by data collected by the Securities and Exchange Commission for its reports

137

Step 3: Conversion of non-LIFO inventories to constant
prices
Separate composite price indexes are constructed to deflate the book values
of total non-LIFO inventories in each industry. For this purpose it is necessary, first, to select commodity price indexes that are representative of the
commodities included in inventories; second, to weight these price indexes in
accordance with the relative importance in the book value of inventories of
the commodities which they represent; and, third, to determine the periods
to which the unit prices reflected in the book value data pertain, so that the
price indexes to be used for deflating the book values can be related to these
same periods.

Construction of composite price indexes

The selection and weighting of the price series used in the construction of
on Working Capital of United States Corporations.
the composite price indexes was for the most part an interrelated operation,
based upon the estimated commodity composition of inventories in each
industry.
Step 2: Adjustment for UFO inventories
The estimates of the commodity composition of inventories in manufacturing were built up mainly from the 1939 and 1947 Censuses of ManufacEstimates of LIFO inventories are deducted from total book values and
tures tabulations of inventories held by industry. Similar estimates for trade
separately processed because the revaluation procedure for non-LIFO invenwere derived from the 1939 and 1948 Censuses of Wholesale and Retail
tories is not applicable to them.
Trade, which showed the distribution of inventories by kind of business. In
The estimates of LIFO inventories in manufacturing are based primarily
most instances, the type of inventory commodity involved could be identified
on two special questionnaires, the first for 1947 and the second for 1951,
on the basis of the census designation of the industry or kind of business
which were submitted to manufacturing corporations in connection with the
Industry Survey of the Office of Business Economics. (For a summary of the holding the inventory. The full industry and kind-of-business detail given in
the censuses, together with the further breakdown of manufacturing invenresults of the 1951 LIFO survey, see "LIFO Inventories and National Income
tories into finished products and materials, supplies, etc., was utilized in
Accounting", SURVEY OF CURRENT BUSINESS, May 1953.) Additional informaestimating the commodity composition of inventories. More summary
tion on the relative importance of LIFO inventories in 1947 was found in a
study of the LIFO method by J. Keith Butters and Powell Niland. (Effects methods were used in instances in which price information was lacking to
match the full detail of the commodity breakdown.
of Taxation—Inventory Accounting and Policies, Graduate School of Business
Other sources consulted for this purpose included the Census of ManufacAdministration, Harvard University, 1949.) Ratios of LIFO to total inventures tabulations of materials consumed in selected industries; materialstories derived from these inquiries were applied to total corporate inventories
consumed data published by private industrial research groups; material
by industry and asset size class to estimate total LIFO inventories in manufacturing. Information from Moody's Manual of Industrial Securities is used to requirements studies of the War Production Board; financial reports of
certain large corporations in the iron and steel industry giving information
extrapolate these ratios and also to derive similar ratios for department
on the commodity composition of their inventories; Internal Revenue
stores, the only other industrial sector in which LIFO inventories appear to
Service data on industrial inventory holdings included in the "Source Book"
be important. Noncorporate LIFO inventories are neglected throughout.
underlying Statistics of Income—Part 2; and sales and value-of-product data,
The concentration of the LIFO method among larger firms indicates that
when more pertinent information was not available.
the omission is insignificant.
With few exceptions the monthly price series used in the construction of
The Internal Revenue Service has recently completed a special tabulation
the composite price indexes were taken from the wholesale price index of the
for 1950 wherein reported year-end corporate inventories in each major
Bureau of Labor Statistics. The BLS index, prior to its revision in 1952, conindustry are classified into seven separate categories according to the invensisted of approximately 900 separately coded commodity price series comtory accounting method used by the reporting firm. The returns can be
bined into 49 subgroups, 10 major groups, and 5 economic groups. The
readily regrouped into three categories, viz., firms using only the LIFO
revised wholesale price index includes over twice as many individual commethod, firms using LIFO as well as other valuation methods, andfirmsnot
modity price series as the earlier index as well as a greater number of major
using the LIFO method. Since the inventory data for firms using LIFO as
group and subgroup composites. In addition, over 200 product class indexes
well as other methods are not broken down into separate LIFO and nonhave been developed to represent groups of related commodities with generLIFO shares, it is possible to obtain only upper and lower limits within
ally similar price movements. The new index, on a 1947-49 = 100 base, is
which the correct LIFO ratio falls. The final tabulation is planned for release
available in full detail for the period beginning January 1947, and was
in 1954; preliminary results are generally consistent with the sample-based
incorporated into the estimation of net inventory change and inventory
LIFO ratios used in estimating net inventory change.
valuation adjustment beginning in 1948.
The estimated value of LIFO inventories currently represents roughly oneFor several reasons, BLS subgroup and product class indexes were used
tenth of the total book value of nonfarm inventories. In some individual
most frequently in the construction of the composite price indexes. It was
industries the proportion is much higher.
not possible, in general, to estimate the commodity composition of inventories
To the extent that the physical volume of inventories increases, changes
in a detail sufficient to calculate weights for individual price series. Also, in
in the book values of LIFO inventories already reflect changes in the physical
the cases in which the available price information did not match the estivolume expressed at current prices and no adjustment of these book values
mated commodity composition of inventories it was thought that the use of
to national income definitions is needed. To the extent that physical volumes
group rather than individual price series would tend to give the better
decrease, changes in LIFO book values reflect these decreases in prior-period
representation of inventory price movements.
prices and a conversion to current prices is made. The price data used in
Two other considerations favored the use of subgroup and product class
each industry are the same as those described below. To date, price adjustindexes. Because of specification changes, the elimination of existing series,
ments for declines in LIFO inventories have been negligible.
and the introduction of new ones, it is frequently difficult to maintain the
continuity of the individual commodity price series. Also, in many instances
Gaps in the basic information may cause sizable errors in the estimates of
the portions of the inventories which could be related to individual comthe change in LIFO inventories. However, in view of the fact that LIFO
modity price series were too fragmentary, from the point of view of the total
inventories are a small fraction of the total, the possible error introduced
industry group, to warrant separate treatment.
into the overall figures is much less significant.



138

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

Further limitations of the composite price indexes may be noted. The price
series underlying them do not take full account of quality change and of
divergences between quoted prices and prices actually charged in market
transactions. In addition to these shortcomings generally encountered in
price deflation, two others appear in connection with the deflation of inventories because the price series (1) do not consistently measure the prices of
purchased inventories at the transaction stage at which they are acquired
by the inventory ho'der and (2) do not measure directly the costs that are
reflected in the valuation of goods-in-process and finished product inventories.
With respect to the weighting system employed, it should be noted that
it is not strictly appropriate for the purpose at hand. Prior to 1948, fixed
weights based on the relative commodity composition of inventories in 1939
are used to combine price index series selected from the unrevised Bureau of
Labor Statistics wholesale price index into a composite inventory deflator for
each industry. Starting at year end 1947 a new set of fixed weights, based on
the distribution of year-end inventories in the 1947 Census of Manufactures
and of beginning inventories in the 1948 Censuses of Wholesale and Retail
Trade, is used to combine the price index series selected from the revised
BLS wholesale price index. Ideally, a system of shifting weights reflecting
the changing annual commodity composition of inventories should have been
used in constructing the inventory deflators, but this could not be done for
lack of information. However, inventory deflators for the various industries
calculated on the basis of 1939 and of 1947 weights were in general very
similar. This suggests that the error introduced by using fixed weights over a
long period is not likely to be significant. Revisions in the estimates of the
inventory valuation adjustment for years since 1948 are mainly due to the
adoption of more comprehensive price data from the revised wholesale
price index rather than to weight changes.

allow for the fact that the use of a single turnover rate for each industry,
instead of separate rates for inventories having a different turnover, overweights inventories that have a relatively quick turnover. These three factors
cannot be accurately measured, but they were believed to be of sufficient
importance to warrant increasing by 50 percent the turnover periods as initially calculated.
The calculation of inventory turnover periods for trade was based upon the
1939 and 1948 Censuses of Wholesale and Retail Trade. Separate calculations were made by kind of business. Census data were used because they
covered noncorporate establishments, which are of particular importance in
trade, as well as corporations. Since the cost of goods sold was not reported
in these censuses, sales had to be used as the numerator of the turnover rates.
This tends to understate the turnover period, since sales include a gross profit
margin whereas inventories are valued at cost. To offset this bias, as well as
the three factors already mentioned, the turnover periods initially calculated
were increased by 75 percent.
The cost valuation indexes applicable to year-end inventories were obtained by averaging the composite price indexes for a number of months prior
to the year-end equal to the estimated turnover periods.
In addition to the uncertainties that have already become apparent, it
may be noted that the calculations involve the assumptions that for the years
for which estimates of turnover periods were made the size of year-end
inventories was an approximation to the average of the inventory holdings
during the year, and also that the turnover periods estimated on the basis of
1939 and 1947-48 data are applicable to the entire periods 1929-47 and 1948
to date, respectively. With respect to the latter assumption, experimentation
with alternative turnover periods has indicated that even considerable modification in their assumed length does not in general greatly affect the statistical result.

Synchronization of composite price indexes with year-end
book values

Step 5: Conversion of deflated change to current prices

To take account of the widespread practice of valuing inventories at the
lower of "cost" or "market," separate deflating indexes are constructed
to represent each of these valuations at the year-end, and the lower of the
two is used to deflate year-end inventories. Indexes of market valuation
are derived by averaging the composite price indexes for December and
the following January in order to approximate year-end prices. Market
indexes are not calculated for the transportation, communications, and
public utilities industries, where inventories consist largely of purchased
materials customarily valued on a straight cost basis.
The construction of indexes of cost valuation is difficult. The particular accounting method used by business firms to charge goods out of inventory to
cost of sales determines the period whose prices are reflected in the year-end
inventories valued on a cost basis. The construction of the cost valuation
indexes applicable to year-end inventories is adapted in the first instance to
the first-in, first-out (FIFO) method, which is the basis of valuation of the
bulk of non-LIFO inventories. Owing to the lack of information, no separate
procedure is developed for other methods of inventory accounting except
LIFO—such as the specific identification and average cost methods—which
underlie some of the book value data. It is believed that the implicit allowance for them described below gives generally satisfactory results.
The prices reflected in year-end inventories valued on a FIFO-cost basis
are the prices of a period immediately preceding the year end, the length of
which depends on the rate of turnover of inventories. The period for which
the composite monthly price indexes for each industry had to be averaged
in order to reflect the cost prices implicit in year-end inventories was based
on the estimated inventory "turnover period" in the industry. Initial calculations of turnover periods for industries other than trade through the year
1947 were made by dividing year-end inventories in each industry into the
total cost of goods sold, as reported in Statistics of Income—Part 2 for 1939, and
then dividing the resulting turnover rate into 12, the number of calendar
months. Starting with the deflation of year-end 1947 inventories, these turnover period calculations were based upon 1947 Statistics of Income.
The turnover periods so calculated were then lengthened for three reasons:
first, to take account of non-FIFO inventories (other than LIFO), whose
general effect is to extend the length of the prior periods whose prices are reflected in year-end inventories; second, to give effect to the lapse of time between the purchase and delivery dates of inventory goods; and third, to



The indexes used to convert the deflated value change of inventories to
current prices represent the annual average of the monthly composite price
indexes already described. Thus, to recapitulate, the same composite priceindex series are used for computing the "market," "cost," and "current"
indexes. The three differ only with respect to the time period to which they
refer—that is, to the span over which the monthly price indexes are averaged.

Characteristics of Revisions
The annual inventory estimates published each July are based upon preliminary data for at least the two most recent years. Under the present schedule, the Internal Revenue Service corporate income tax return tabulations,
on which the book value of corporate inventories is based, are not available
for the two latest years. Revisions may affect earlier years also, mainly because the noncorporate book value estimates are based on less regular benchmarks. Introduction of new benchmarks, such as periodic industrial census
tabulations, may change prior year estimates through the substitution of interpolation for previous extrapolation procedures.
Revisions of the inventory change estimates for recent years occasioned by
the incorporation of comprehensive data are sometimes quite sizable. As
noted earlier, even small percentage revisions in the book value aggregates
can cause substantial revisions in the increments computed from them.
Revisions are usually much larger in the "change in business inventories"
item than in the "inventory valuation adjustment." Revisions in the latter
typically are not large. The bulk of revisions in the underlying book value
estimates is generally reflected in the change in business inventories rather
than in the inventory valuation adjustment. Also, the price information used
to revalue business inventories becomes available fairly promptly, and subsequent routine revisions of it are usually minor.
The smallness of the revisions in the inventory valuation adjustment should
not be taken to mean that this item, or the underlying revaluation of book
values, is firmly based. Quite to the contrary, as already has been emphasized, this is the most difficult step in the estimating procedure. The absence
of major revisions in these estimates signifies only that the final sources of
price information do not differ significantly from the preliminary information
on which the estimates are initially based.

NATIONAL INCOME,

12. NET FOREIGN INVESTMENT
The net foreign investment component of gross national product measures
the net export balance on goods and services (less the net outflow of gifts),
which is necessarily financed by international investment. It is taken from
the official balance of payments of the United States, where it appears as
the "Balance on goods and services and unilateral transfers."' Accordingly,
the statistical sources and methods discussed below are those of the Balance
of Payments Division of the Office of Business Economics. The balance of
payments of the United States is published quarterly in the Survey of Current
Business, and in occasional bulletins, the latest of which is The Balance of
Payments of the United States, 1949-51, a supplement to the Survey of Current

Business (Washington, United States Government Printing Office, 1952). A
more detailed description of the methods of estimating the international
transactions of the United States appears in that volume.
These methods also produce an alternative estimate of net foreign investment in terms of the change in international assets and liabilities. Usually
there is a statistical discrepancy (labeled "errors and omissions" in the
balance-of-payments statement) between the two estimates. As has been
noted, it is the former estimate which is included as a component of gross
national product, although for most years it cannot be assumed to be more
exact than the one based on changes in assets and liabilities.
Official estimates of the United States balance of payments have been prepared for all years beginning with 1919. Over this period, the sources and
methods underlying the estimates have changed significantly. This section
is devoted primarily to a description of the current methodology.
A substantial increase in the accuracy of the estimates, especially in the
last decade, has stemmed from the inauguration of new techniques and data
sources discussed below. In spite of the improvement of the data, it should
be recognized that net foreign investment is subject to possibly large percentage errors of estimation because it is calculated as the difference between
gross outflows and gross inflows which are usually large in relation to the
net balance.
The following discussion deals in turn with net factor payments to the
United States and with net purchases from the United States (including
gifts), which together compose the "Balance on goods and services and unilateral transfers". The balance is obtained as gross receipts of the United
States less gross payments by the United States. In the following discussion,
the gross receipts and payments are described separately, with no further
reference to the fact that they are differenced to obtain the net entries for
net foreign investment.
The balance on goods and services and unilateral transfers reflects all current transactions of the United States with the rest of the world. Hence, a
large number of the flows differenced to obtain net foreign investment also
become explicit elements (with appropriate sign) of various other components
of national income and national product. Thus, net factor incomes received
from abroad become elements of the wages and salaries, interest, and corporate profit components of national income and personal income, and net
purchases made abroad (including gifts) become elements of personal consumption expenditures for commodities and services and of government
purchases of goods and services. The estimation of these elements is described
below. However, only in the sections dealing with the national income and
product components in which they are included are they explicitly identified
as constituting such elements. For instance, net interest received from abroad
is described in the present section under the heading Net payments of
factor income, but the fact that this item is included in the interest component
of national income and personal income is stated explicitly only in the section
on Interest.
1. An adjustment is made for geographical coverage, since United States territories and
possessions are considered part of the United States for balance-of-payments purposes but
are included with the rest of the world for national income purposes. (Statistically, the
adjustment was confined to 1941-46 when its quantitative importance called for a rough
order-of-magnitude calculation in spite of the absence of solid data.) Itshould be noted also
that in the official balance-of-payments statement account is taken of unilateral transfers
in kind, which are ignored in table 11, Part V of this report. Exclusion of these transfers
affects debits and credits equally, and does not alter the net balance of transactions. Other
differences between table 11 and the balance-of-payments statement are matters of classification and will come to note as the derivation of the entries in table 11 is described.

291692°—54- - 1 0



195 4

EDITION

139

Net Payments of Factor Income to the United States
Factor income transactions consist of wage and salary receipts and the
international flows of property income. The latter represent interest, dividends, and branch profits. Undistributed profits of foreign subsidiaries are
not included. All property incomes are measured net of taxes levied by the
paying country.

Wages and salaries
This item (included in "private miscellaneous services" in the official
balance-of-payments classification) represents wages and salaries received
by United States residents in this country from (1) foreign governments and
(2) international organizations. Component (1) is estimated from fragmentary
information supplied by certain foreign missions. Component (2) is reported
by the international organizations.
It will be noted that, from the standpoint of the definitional framework
set out in Part II of this report, this series is incomplete because it does not
segregate the net wages and salaries of "border workers." (It does not include
the wages and salaries of United States residents whose place of work is
abroad, and it does not deduct payments by United States employers to
foreign residents working in the United States.) Statistical information for
making this segregation is lacking.

Interest
United States receipts of interest from abroad are estimated in four parts.
(1) Interest from foreign branches and affiliates (including subsidiaries)
of American corporations is estimated from a 1950 benchmark provided by
the Commerce Department's census of American direct investments abroad
(Foreign Investments of the United States, a supplement to the Survey of Current

Business, Washington, United States Government Printing Office, 1953).
This benchmark figure is extrapolated and linked to a previous benchmark
based on the Treasury Department's 1942 Census of American-Owned
Assets in Foreign Countries (Form TFR 500) on the basis of quarterly sample
company reports, supplemented by information taken from income tax returns, annual reports to the Securities and Exchange Commission (Form
10K), and annual corporate reports to stockholders. Currently the quarterly
sample reports cover about 350 companies accounting for around 75 percent
of total direct investment; about 90 percent of total investment income (interest, dividends, and branch profits) earned from foreign branches and affiliates
is actually reported in one or more of the sources used for extrapolation of the
1950 benchmark.
(2) Interest received by the United States Government is reported by the
recipient agencies.
(3) Interest on foreign dollar bonds is estimated chiefly from regular
questionnaire surveys of the debtors' fiscal and paying agents in the United
States, corrected for basic coverage by reference to holdings disclosed in the
1942 Treasury Census.
(4) Other foreign interest received by the United States is calculated by
applying average bond yields, as reported from the principal issuing countries,
to holdings series benchmarked on the 1942 Treasury Census and extrapolated
by reference to foreign bond purchases and sales reported on Foreign Exchange Forms S—1/3. These statistics of purchases and sales, it should be
noted, do not distinguish the bonds by country of debtor or by currency in
which payable. The extrapolation, therefore, is a relatively rough process.
However, the series appears reasonable when checked against official estimates for Canada, which accounts for most of the receipts in this category.
United States payments of interest are estimated as the sum of (1) interest
subject to withholding tax, and (2) interest not subject to withholding tax.
(1) Interest subject to withholding tax paid by the United States to
foreigners is reported to the Internal Revenue Service on withholding tax
return forms 1012 and 1013, for corporate and Federal bonds, and Form 1042,
for other obligations. Special tabulations from these returns are made available to the Office of Business Economics. Net income payments by fiduciaries
are allocated somewhat arbitrarily between interest and dividends. The tax
return data are not available in time for use in the estimate for the latest year,
which is therefore made by extrapolation with reference to reported changes
in foreign holdings.

140

A SUPPLEMENT TO THE SURVEY OF CURRENT BUSINESS

(2) Interest not subject to withholding tax—about 60 percent of the total
in 1950—represents largely payments on foreign government holdings of
United States Federal Government long- and short-term issues, and to a
much smaller extent payments on foreign holdings of State and municipal
government bonds and payments to certain countries which are exempt from
withholding taxes by international treaties. It is estimated in the main by
applying average-yield series to amounts of foreign holdings. Foreign government holdings of long-term Federal issues are benc