Investigation of concentration of economic power; monograph no. 1[-43]

^^3d SessKm ^^} SENATE COMMITTEE PRINT

INVESTIGATION OF CONCENTRATION
OF ECONOMIC POWER

TEMPORARY NATIONAL ECONOMIC
COMMITTEE

A STUDY MADE UNDER THE AUSPICES OF THE DEPART-
MENT OF COMMERCE FOR THE TEMPORARY NATIONAL
ECONOMIC COMMITTEE, SEVENTY-SIXTH CONGRESS,
THIRD SESSION, PURSUANT TO PUBLIC RESOLUTION
NO. 113 (SEVENTY-FIFTH CONGRESS), AUTHORIZING
AND DIRECTING A SELECT COMMITTEE TO MAKE A
FULL AND COMPLETE STUDY AND INVESTIGATION
WITH RESPECT TO THE CONCENTRATION OF ECONOMIC
POWER IN, AND FINANCIAL CONTROL OVER,
PRODUCTION AND DISTRIBUTION
OF GOODS AND SERVICES

MONOGRAPH No. ^-6
CONCENTRATION AND COMPOSITION OF
INDIVIDUAL INCOMES, 1918-1937

Printed for the use of the
Temporary National Economic Committee

UNITED STATES

GOVERNMENT PRINTING OFFICE

WASHINGTON : 1940

f^ORTHEASTERN UNIVERSfTY SCHOOlof LftWDBRRRl

TEMPORARY NATIONAL ECONOMIC COMITTEE

(Created pursuant to Public Res. 113, 75th Cong.^

JOSEPH C. O'MAHONEY, Senator from Wyoming, Chairman

HATTON W. SUMNERS, Representative from Texas, Vice Chairman

WILLIAM H. KING, Senator from Utah

WALLACE H. WHITE, Jr., Senator from Maine

CLYDE WILLIAMS, Representative from Missouri

B. CARROLL REECE, Representative from Tennessee

THURMAN W. ARNOLD, Assistant Attorney General

•WENDELL BERGE, Special Assistant to the Attorney General

Representing the Department of Justice

JEROME N. FRANK, Chairman

•SUMNER T. PIKE, Commissioner

Representing the Securities and Exchange Commission

GARLAND S. FERGUSON, Commissioner

•EWIN L. DAVIS, Chairman
Representing the Federal Trade Commission
ISADOR LUBIN, Commissioner of Labor Statistics
•A. FORD HINRICHS, Chief Economist, Bureau of Labor Statistics
Representing the Department of Labor
JOSEPH J. O'CONNELL, Jr., Special Assistant to the General Counsel lr^r» 1

•CHARLES L. KADES, Special Assistant to the General Counsel lS^> 1'

Representing the Department of the Treasury IO"l^ ''

Representing the Department of Commerce

* * * k.-

CO r*

LEON HENDERSON, Economic Coordinator

DEWEY ANDERSON, Executive Secretary C3

THEODORE J. KREPS, Economic Adviser
'Alternates

^

Monograph No. 4

CONCENTRATION AND COMPOSITION OF INDIVIDUAL INCOMES,

1918-1937

BY

ADOLPH J. QOLDENTHAL

REPRINTED
BY

WILLIAM S HEIN & CO , INC

BUFFALO. N. Y.
1968

ACKNOWLEDGMENT

This monograph was written

by
ADOLPH J. GOLDENTHAL

Economic Analyst, National Income Division, Department oj Commerce
Under the general supervision of

ROBERT R. NATHAN

Chief, Xational Income Division, Department of Commerce

The Temporary National Economic Committee is greatly indebted
to the author for this contribution to the literature of the subject
under review.

The status of the materials in this volume is precisely the same as that of
other carefully prepared testimony when given by individual witnesses; it is
information submitted for Committee deliberation. No ntatter what the
official capacity of the witness or author may be, the publication of his
testimony, report, or monograph by the Committee in no way signifies nor
implies assent to, or approval of, any of the facts, opinions or recommenda-
tions, nor acceptance thereof in whole or in part by the members of the
Temporary National Economic Committee, individually or 'collectively.
Sole and undivided responsibility for every statement in such testimony,
reports, or monographs rests entirely upon the respective authors.

(Signed) Joseph C. O'Mahoney,
Chairman, Temporary National Economic Committee.

in

TABLE OF CONTENTS

Page

Letter of transmittal. ix

Summary of statistical findings xi

CHAPTER I

Introduction 1

I. Distribution of income and problems before the Temporary

National Economic Committee 1

II. Purpose of present inquiry 6

III. General characteristics of data 6

CHAPTER II

The concentration of income, 1918-37_- 9

I . Definitions 9

1. "Earning Power" 9

2. "Purchasing Power" II

3. "Real Income" ^ 11

4. Income Recipient 12

II. Statistics of income concentration, 1918-37 13

1. The highest 1 percent of income recipients 14

2. Selected proportions of income recipients 20

CHAPTER III

The composition of income, 1918-37 35

I. Introductory 35

II. Composition of income: All income recipients and the highest

] percent ._ 37

III. Shifts in the composition of income of the highest 1 percent of

income recipients 41

IV. Concentration of types of income 44

V. Composition of income by income classes 47

VI. Composition of income and concentration of income 49

CHAPTER IV

The concentration of "purchasing power" and the effects of relief and

vetei'ans' adjusted-service payments on income concentration 55

I. The concentration of "purchasing power," 1918-37 55

1. The transition from "earning power" to "purchasing
power" 55

2. Statistics of "purchasing power" concentration 59

II. Relief and veterans' adjusted-service payments 65

1. Statistics of income concentration, excluding work-relief

wages, 1931-37 , 67

2. Statistics of income concentration, including direct and

work relief and veterans' adjusted-service pavments,

1934-37 1 69

APPENDIX A

Notes to tables 71

Note A- h (tables 1-5) 71

Note A-2 (table 6) 79

Note A-3 (compensation and rumber of employees of State and local

governments) ' 81

V

TI TABLE OF CONTENTS

Notes to tables — Continued. Pagfe

Note A-4 (table 10) 83

Note A-5 (table 12) . 89

Note A-6 (table 15) __.. 96

Note A-7 (tables 16 and 18) 97

APPENDIX B

Some aspects of the statistics on high incomes 99

Note B-1 Relation of Statutory Net Income to Economic Income 99

Note B-2 Effect on Measures of Income Concentration of Income

Concept in Use During the Years 1934-37, 101

Note B-3 Characteristics of Income-Tax Data and Method of Measur-
ing Income Concentration 104

Index _ 109

SCHEDULE OF TABLES AND CHARTS

TABLES

Pag*

1. Shares of total individual income received by the highest 1 percent of

income receipients, 1918-37 ^ 16

2. Shares of total individual income received by selected proportions of

income recipients, 1918-37 22

3. Shares of total individual income received by selected proportions of

income recipients, 1926-37 23

4. Minimum statutory net incomes of selected proportions of income

recipients, 1918-37 26

5. Minimum economic incomes of selected proportions of income re-

cipients, 1934-37.-_. - '. . - 26

6. Number oif individuals in the selected proportions of income recipients,

1918-37 27

7. Average statutory net incomes of selected proportions of income re-

cipients, 1918-37 ■ 28

8. Average economic incomes of selected proportions of income recipients,

1926-37 :..__.- 29

9. Percentage increases in shares of total individual income from year of

lowest to year of highest income concentration 32

10. Composition of total individual income, 1918-37 37

11. Percentage distribution of total individual income, by type of receipt,

1918-37 1 ---. 38

12. Composition of incomes of the highest 1 percent of income recipients,

1918-36 39

13. Percentage distribution of incomes oi the highest 1 percent of income

recipients, by type of receipt, 1918-36 — 40

14. Percentages 'of each type of income received by the highest 1 percent

of income recipients, 1918-36 45

16. Composition of incomes by income classes, 1926, 1929, 1932, 1935 48

16. Shares of total "purchasing power" available to the highest 1 percent

of income recipients, 1918-37 60

17. Effect of adjustments for "purchasing power" on the income shares

of the highest 1 percent Of income recipients, 1918-37 61

18. Shares of "purchasing power" available to selected proportions of

income recipients and effect of adjustments for purchasing power,
1926and 1936 62

19. Proportions of economic incomes of highest 1 percent of income re-

cipients paid in Federal income taxes, 1918-37 63

20. Direct and work relief and veterans' adjusted service compensation

payments 1929-37 --- 67

21. Shares of total individual income, excluding relief and veterans'

adjusted service payments, received by selected proportions of income
recipients, 1934-37_-_ - - - - 68

22. Shares of total individual income, including relief and veterans'

adjusted service payments, received by selected proportions of
income recipients, 1934-37 70

CHARTS

I. Concentration of income and total income per recipient, 1918-37^ 18

II. Shares of total individual income received by selected proportions of

income recipients, 1918-37 30

III. Composition of incomes of the highest 1 percent of income recipients,

1918-36 - -- 42

vn

VIII SCHEDULE OF TABLES AND CHARTS

APPENDIX TABLES

Page

A-1. Aggregate statutory net income of selected proportions of income

recipients, 1918-37 78

A-2. Aggregate economic income of selected proportions of income re-
cipients, 1926-37 78

A-3. Total individual income excluding compensation of State and local

governmental employees, 1918-37 79

A-4. Number of persons with gainful occupations. 1918-37 80

A-5. Compensation and number of employees of State and local govern-
ments, 1918-37 83

A-6. Federal income taxes paid by all income recipients and the highest

1 percent, 1918-37 - , 98

A-7. Federal income taxes paid by selected proportions of income re-
cipients, 1926 and 1936 98

B-1. Ratios of statutory net income to economic income — Highest 1

percent of income recipients, 1918-37 99

B-2. Relative mean deviations of the* two measures of income concentra-
tion, 1926-37 -< - 100

B-3. Ratios of statutory net income to economic income — Selected pro-
portions of income recipients, 1926-37 101

B-4. Nurober of individuals among the various proportions of income
recipients reporting a statutory capital loss of $2,000 and over,
1934-36 104

LETTER OF TRANSMITTAL

This report concerns itself with the problems of how the distribution
of individuals' income has changed since 1918, and how it varies from
year to year. Attention is centered particularly on the degree of
concentration in the hands of the largest income receivers. The
record is based upon income-tax material, with various adjustments
making the data as comparable as possible, despite the many changes
in tax laws and regulations. Further analysis breaks these incomes
down according to source, in order to find explanations for the
shifts. Finally, corrections are made for relief and veterans' bonus
payments, and estimates are made after taxes, in order to picture the
concentration of purchasing power.

It should not be necessary to argue that such information is of
great significance to any understanding of the functioning of the
economic system. It is basic to the problem of capital accumulation
and the underlying adjustment of consumption, savings, and invest-
ment, so necessary to attaining full activity and employment. How-
ever, it has a more direct connection to the subject of monopoly.
Large fortunes have most often had their basis in situations in which
strong elements of monopoly appear. And the possession of financial
strength is helpful, to say the least, in obtaining or maintaining in-
dustrial controls which take various monopolistic forms and which
perpetuate or enhance the concentration of income. Of course,
incomes would be by no means equal in a perfectly competitive
economic system. However, the degree of income concentration can
be taken as one indicator of the degree to which monopolistic elements
are present in our economy.

This report has been prepared in the National Income Division of
the Bureau of Foreign and Domestic Commerce, under the direct
supervision of Robert R. Nathan.

WiLL.\RD L. Thorp,
Adviser on Economic Studies, Department of Commerce.

Washington, D. C, March 1, lO^O.

SUMMARY OF STATISTICAL FINDINGS

This summary is confined to the statistics presented in this study.
With the exception of the introduction, each chapter contains, in
addition to statistical data, material on the meaning of income con-
centration and cortiposition which, it is believed, will be of assistance
in properly interpreting the statistical findings. Chapter II presents,
for the past two decades, measures of the concentration of income
received by individuals in the United States in return for personal
services and the ownership of property. The concentration of income
is measured by the shares of the income of all individuals received by
the highest 2 percent and smaller proportions of the Nation's income
recipients. The statistics reveal that the shares of total individual
income received by the higher income recipients have generally
increased during periods of business expansion and declined during
periods of business contraction. Subject to certain exceptions, the
larger the average income of all income recipients the greater has been
the degree of income ..concentration.

For the period from IWS through 1937 the degree of income con-
centration was lowest in the depression years 192D and 1932 and
highest in the prosperous years 1928 and 1929. There has been no
significant trend over the period as a whole and the degree of income
concentration during the recent years 1934 through 1937 has been at
approximately the same level as during the years 1918 through 1924.
For these two periods the average income, corrected for price changes,
was also approximately the same. The years 1925 tluoiigh 1933 wit-
nessed wide fluctuations in income concentration. In contrast, from
1918 through 1924 and 1934 through 1937 the degree of income con-
centration fluctuated within relatively narrow limits. Apart from
minor variations associated with the short cyclical movements in
business activity, the concentration of income increased during the
period of rising income and expanding business activity from 1922
through 1928. Most of this increase took place during the years 1925
through 1928. After 1929 the iricome shares of the higher income
recipients declined sharply along with general business activity,
reaching a minimum in 1932 and 1933. These shares then increased
from 1934 until 1937. The sharp reversal in the upward trend of
business in the middle of 1937 was accompanied by a decline for that
year in the income shares of the higher income recipients.

Measiu'es of income concentration are presented for five proportions
of income recipients, varying in size from the highest 2 percent to the
highest one one-hundredth of 1 percent. The changes in the income
shares received by these various proportions of income recipients in the
higher brackets followed for the most part the same general pattern,
increasing in years of business prosperity and declining in years of
business depression. However, the smaller the group of income re-
cipients, the larger, relatively, were the year-to-year changes in the size
of the income shares. The income shares of the smaller groups, that

XII SUMMARY OF STATISTICAL FINDINGS

is, the groups with the higher incomes, dedined more during periods
of business depression and increased more during periods of business
prosperity than the shares received by the more inclusive groups. This
difference in the behavior of the income shares received by the various
proportions of income recipients was larger as the size of the group
varied from the highest 2 percent to the highest one one-hundredth of
1 percent. The increases in the shares received by the higher income
groups were particularly large during the period of marked business
prosperity ft-om 1925 through 1928. For example, the income share
of the highest 2 percent of income recipients rose by 35 percent from
1924 through 1928 while the income share received by the highest
one-tenth of 1 percent rose by as much as 75 percent. There are some
further differences in the nlovement of the shares of income received
by these various proportions of income recipients which are described
in chapter II.

Tables are also presented in chapter II showing the number of
persons included in the various proportions of income recipients, the
income levels which separate these income groups, and the average
incomes. In the year 1936, for example, there were approximately
50,363,000 income recipients, excluding employees of State and local
governments. In this year the highest 2 percent of income recipients
included slightly over 1,000,000 persons with incomes of $4,390 and
over.' The average income of this group was $11,955 as compared
with an average of $1,065 for the other 98 percent and $1,275 for all
income recipients. The smallest group for which income shares are
measured is the highest one one-hundredth of 1 percent of income
recipients which in 1936 .included 5,036 persons with incomes above
$116,430. The average income of this group was approximately
$250,000.

During the period studied, the minimum income levels of the various
proportions of income recipients have been subject to wide fluctuations.
In 1928, for example, persons with incomes above $10,140 were
included within the highest 1 percent of income recipients whereas in
1934 it would have required an income of but $5,375 to be included
with the same group. In 1932 and 1933 the minimum incomes for
the highest 1 percent of income recipients were even lower than in
1934, but for reasons indicated at a later point no data are presented
for these years. The income levels which separated the various
smaller proportions of income recipients have also varied widely dur-
ing relatively brief periods. In 1934 anyone with an income in excess
of $80,775 would have been included with the highest one one-hun-
dredth of 1 percent of income recipients, while in* 1936 an income at
least 44 percent "larger, $116,430, was necessary for inclusion in this
category.

Chapter III presents statistics showing how the various types of
inconie, such as employee compensation, dividends, and interest were
combined each year to produce the incomes of all income recipients
and of the various proportions of higher income recipients whose
income shares were given in the preceding chapter. In addition, the
concentration of each type of income among the higher income
recipients is shown. These statistics have a twofold purpose: First,
they indicate how the different types of income were distributed

' Thi? income level rpprcserts an understatement as the income tax statistics which were used as the ,
basis for the estimate were not adjusted for nonreporting or underreporting tf incomes. See ch. II. n. 15.

SUMMARY OF STATISTICAL FINDINGS XIII

among the various groups of income receivers and how this distribu-
tion changes with increases and declines of the total income and of
each type of income. Second, they provide the basis for an analysis
of the shifts in income concentration in terms of the composition of
income.

In the year 1922 when the degree of income concentration was
approximately equal to the average for the past two decades, the
highest 1 percent of income recipients received 29 percent of their
income from salaries and wages, 19 percent from entrepreneurial net
income (net profits from unincorporated businesses), 25 percent from
dividends, 14 percent from interest, 8 percent from profits on the sale
of property, and 5 percent from net rents and royalties. In this year,
63 percent of the income of all individuals was derived from salaries
and wages, 19 percent from entrepreneurial net income, 5 percent
from dividends and 6 percent from interest, about 2 percent from net
profits on the sale of property, and 6 percent from net rents and royal-
ties. Comparison of these two sets of percentages reveals that divi-
dends, interest, and net profits from the sale of property were inuch
more important sources of income for the highest 1 percent of income
recipients than for the other income recipients. If employee com-
pensation and entrepreneurial net income are classified as income
primarily from personal service and the other income sources as income
primarily from property, the former accounted, in 1922, for 82 percent
of the income of all income recipients and 48 percent of the income of
the highest 1 percent.

For all income recipients and for the highest 1 percent the relative
importance of the various income sources has varied with changes in
business conditions. Salaries and wages were generally a larger
proportion of income in times of business depression and dividends
and net profits from the sale of property were generally larger pro-
portions of income in times of business prosperity. There has been
during the period studied a general decline in the importance of net
rents and royalties and entrepreneurial net income. The latter source
contributed a considerably smaller proportion to the income of the
highest 1 percent during recent years than during earlier years of a
comparable degree of income concentration. This decline has been
accompanied by a marked increase in the importance of employee
compensation and in recent years these two sources together consti-
tuted a slightly larger share of the income of the highest 1 percent
than in previous years characterized by a similar degree Of income
concentration.

The type of income contributing the largest share to the income of
the highest 1 percent of income recipients has changed several times
since 1918. In 11 years during the 20-year period from 1918 through
1937, the two sources classified as income primarily from personal
service (employee compensation and entrepreneurial net income)
contributed a greater share to the income of the highest 1 percent of
income recipients than did the sources classified as income primarily
from property. Employee compensation was the largest single income
component in 12 years, dividends in 6, net profits from the sale of
property in 1, and entrepreneurial net income in 1.

Data are presented which show the proportion of each type of
income received by the highest 1 percent of income recipients. The
extent of concentration varied from a small proportion of salaries

XIV SUMMARY OF STATISTICAL FINDINGS

and wages — 6 to 7 percent — to the major portion of dividends and
net profits from the sale of property. The proportion of total divi-
dends received by this group in different years ranged from 58 to 70
percent.

On the basi^ of data on the composition of income by income classes,
it is shown that the importance of employee compensation, entre-
preneurial net income, and net rents and royalties fell as the size of
income increased and t' e importance of dividends and net profits on
the sale of property rose sharply as the size of income increased.
Interest constituted an increasingly important source for the larger
incomes until the very high incomes were reached, in most years
$100,000 and over, after which it dropped in importance. Above this
amount incomes were derived chiefly from two sources — dividends
and net profits on the sale of property with profits being of greater
importance in years of high business activity. In 1932 virtually all
income classes incurred losses from the sale of property.

The data on the composition of individual incomes, and on the
concentration of the various types of income, throw considerable light
on the "causes" of the changes in income concentration. Shifts in
income concentration are largely explainable in terms pf year-to-year
differences both in the relative importance of different income sources
and in the concentration of these sources. Due to both the high
concentration of dividends and net profits from the sale of property
and the relatively large variation in their volume from year to year,
shifts in income concentration were usually associated with fluctua-
tions in these two sources of income. Increases in their volume which
were greater than the increases in total individual income accounted
in large part for .the sharp rise in income concentration during the
years 1925 through 1928. The decline in income concentration after
1929 may be attributed in large part to net losses from the sale of
property, chiefly in the form of securities. The reduction in the
amount of dividends after 1930 was also an important 'factor in the
lessened concentration of income during the early thirties.

In contrast to the large changes in the concentration of income
during the years 1925 through 1933, the degree of income concentra-
tion fluctuated within fairly narrow limits from 1918 through 1924.
During this period dividends as a share of individual income did not
vary greatly and the volume of net profits from the sale of property
was relatively small. The different behavior of these two income
sources during the two periods, 1918 through 1924 and 1925 through
1933, suggests that in the absence of marked fluctuations in dividends
and profits and losses for the sale of property, the concentration of
income w6uld not vary much from year to year.

The fluctuations in the volume of dividends and net profits from
the sale of property were responsible in large part for the greater
variability in the income shares received by the highest income
groups, as the higher the income of the group, the more important
were these two sources. In a few years the changes in the income
shares of the various groups of income recipients departed from the
usual pattern of increased variability for the higher income groups.
These divergent movements are also explainable in terms of the
changes in the composition of individual incomes which took place.

In chapter IV statistics are presented which . measure to a con-
siderable extent the shifts in the concentration of ''purchasing power,"

SUMMARY OF STATISTICAL FINDINGS XV

that is, in the portion of current income which is available to indi-
viduals for spending or saving. In order to obtain the distribution of
purchasing power, the distribution of income received in return for
personal service and for the ownership of property should be adjusted
to take account of income taxes and transfers of income such as gifts,
contributions, and direct relief. The most striking difference between
the trends in the concentration of income received in return for per-
sonal or capital services and the concentration of purchasing power
is the greater increase in the concentration of purchasing power that
occurred during the twenties. The share of total income received by
the highest 1 percent of income recipients increased by 51 percent
from 1918 through 1928 while the increase in the share of income
available for saving and expenditure by this proportion of income
recipients was 60 percent. This larger increase in the concentration
of purchasing power was due to the decline during this period in the
rates of Federal income taxation, particularly after 1924. In 1918
Federal income taxes absorbed 13.5 percent of the income of the
highest 1 percent of income recipients and in 1928 only 7.4 percent.

In contrast to this sharper increase in the concentration of pur-
chasing power than in the concentration of income for the years
1918 through 1928, the rise in the concentration of purchasing power
after 1934 was not as great as the rise in the concentration of income.
In 1936, when the surtax rates on incomes above $50,000 were raised,
the share of income received by the highest 1 percent of income
receivers rose by 8 percent whUe the increase in the share of pur-
chasing power available to this proportion of the Nation's income
recipients was only 3 percent. For the highest one-tenth of 1 per-
cent of income recipients there was, in fact, no increase in the share
of purchasing power available from 1935 to 1936, although the pro-
portion of total income received by this group rosa by as much as
10 percent.

In the latter part of chapter IV the effect of relief payments and
the veterans' bonus on the concentration of income is determined.
The measures of income concentration presented in chapter II included
work-relief earnings in total individual income, but excluded direct-
rehef payments and the veterans' bonus. If work-relief payments are
also excluded, the income shares of the higher-income recipients are,
of course, slightly larger. The differences, however, are slight. The
largest difference was in 1936 when the highest 1 percent of income
recipients received 14.53 percent of the total income with work relief
earnings included and 15.09 percent when this income is excluded.
The effect of including work-relief wages in income is to diminish
shghtly the fluctuations in income concentration from 1934 through
1937.

The inclusion of direct-relief and veterans' payments as well as
work-relief payments also diminishes the magnitude of the fluctuations
in mcome concentration. From 1934 to 1936 the income share of the
highest 1 percent of income recipients rose by 9.5 percent when
relief and bonus payments are included in the total income of indi-
viduals as compared with 12.4 percent when these items are excluded.

CHAPTER I
INTRODUCTION

I. DISTRIBUTION OF INCOME AND PROBLEMS BEFORE THE TEMPORARY
NATIONAL ECONOMIC COMMITTEE

The distribution of the Nation's total income among individuals
and the changes in this distribution are of such fundamental- impor-
tance that knowledge concerning them is relevant to a consideration
of many of the problems before the Temporary National Economic
Committee. A brief statement of the relation of the distribution of
income to some of these problems will be useful in evaluating the
significance of Che statistical findings of this study.

Of foremost interest is the intimate relation of the distribution
of income to monopoly and the effectiveness of competition in general.
The distribution of income in recent times reflects the existing
monopolistic elements in the economy and, through the transmission
of wealth derived from earlier monopolistic situations, the elements
of industrial monopoly that have prevailed in the past. Historically,
public indignation concerning the large fortunes and hence large
incomes which had as their source either a strategic control of an
industry, a particular product, or a natural resource, provided much
of the impetus for the enactment and enforcement of laws designed to
curb monopolistic practices. From what is known of the rise of great
fortunes and incomes, very many, possibly the majority, resulted from
the exploitation of circumstances in which strong elements of monop-
oly were present.

The dominant source of these fortunes has varied with the economic
development of the Nation. Rising land values; poorly organized
markets; control of natural resources such as oil, copper, aluminum,
and timber or of important stages in their fabrication ; railroad devel-
opment; strategic positions in the Nation's financial markets; pro-
motions of consolidations in industry; monopolistic practices and
conditicTns in industries such as steel and tobacco; and the rapid
expansion of new industries such as agricultural machinery, electrical
appliances, motion pictures, chemicals, and radio — all these situations
provided the basis for many of the large fortunes and in all of them
strong elements of monopoly are found. The incomes derived from
a monopolistic situation often have provided the financial power
which has been used to further the concentration of industrial control.
Such increases in the power of monopoly have in turn made it possible
to sustain a level of profits sufficiently high to perpetuate and augment
these incomes. In this way the relationship between the concentra-
tion of income and industrial monopoly has tended to be interacting
and cumulative. The inheritance of these great fortunes operates
to continue the concentration of income. The perpetuation of large
fortunes, however derived, is a potent factor in diminishing equality
of opportunity and, therefore, results in a greater degree of inequality
in the distribution of income than would otherwise be the case.

1

256149— 40— No. 4 2

2 CMDNCENTRATION OF ECONOMIC POWER

The large incomes derived from monopolistic practices illustrate in
a striking manner'the relation between the concentration of income
and monopoly. This relationship, however, is of more general im-
portance than is indicated by confining attention to the very large
incomes. There are, of course, many more incomes of lesser size that
are a result of a wide variety of practices which may be broadly
described as monopolistic. With the exception of incomes derived
from inherited wealth, main reliance is placed upon competitive forces
to determine the distribution of the Nation's income among individuals.
To the extent that competitive situations are lacking in our economy,
incomes larger than competitive conditions wo-uld permit are present.
These smaller incomes also play an important role in the concentration
of income.

It may be confidently stated that were it not for past and present
monopoly in one forrrl or other, the prevailing distribution of income
would be considerably more equal. Or, to put the same thought in
other words, the eradication of the monopolistic elements in our
economy and the repeal of the privileges which persist as a result of
monopoly existing in earlier periods of the Nation's history w^ould
bring about an income distribution significantly different from that
which now obtains. This is not to say that increasing the effectiveness
of competition would yield equal incomes for all. A considerable
degree of income inequality seems to be consistent with a high degree
of competition and is probably an essential characteristic of a com-
petitive economy. The elimination of a wide variety of monopolistic
controls and practices would, however, have the effect of reducing
the degree of income concentration. The task of eradicating the
monopolistic elements in the economy may be expressed positively
as the fulfillment of the conditions necessary for the maintenance of
competition in the production and marketing of goods and services.
Furthermore, it would involve the correction of the abuses in cor-
porate finance made possible by strategic positions of individuals or
groups of individuals. In areas where the maintenance or extension
of competition as a regulative force is not feasible it would be necessary
to maintain effective controls limiting the incomes to reasonable,
competitive amounts. The accomplishment of these tasks would
diminish the concentration of incomes and, by eliminating the source
of the type of large incomes which most men consider unjust, would
remove to a considerable extent the doubts which are entertained as
to the equity of the present pattern of income distribution.

The importance of competition as a factor in bringing about a more
equal distribution of income may be further illustrated by the unem-
ployment problem and the concentration of dividend receipts. Inso-
far as the present large volume of unemployment is attributable to
causes which may be ultimately traceable to the enervation of the
competitive forces in the economy, strengthening of these forces will
result in the reemployment and the consequent increase in the in-
comes of those dependent on public and private relief. Reemploy-
ment attained through this method will bring about both an increase
in the national income and a change in the distribution of income in
favor of the unemployed. The importance of controlling corporate
profits through effective competition is readily appreciated by ex-
amining the relation of dividend receipts and the concentration of
income. Inasmuch as the ownership of American corporations is

CONCENTRATION OF ECONOMIC POWER ' 3

highly concentrated among individuals, as evidenced by the concen-
tration of dividend receipts, a decline in the effectiveness of industrial
competition, accompanied by no other significant changes, is likely
to be reflected in an increase in the concentration of income.*

The relation between the distribution of income and industrial
organization and control may be approached from another angle;
namely, the supply of capital for enterprises of various sizes. The
availability of funds for relatively small-scale enterprises is probably
closely related to the character of the distribution of income. Indi-
viduals either in small businesses or organizing them typically obtain
their equity capital from their own savings or from the savings of
friends and relatives. The relative importance of moderately sized
incomes is, therefore, a basic factor in determining the importance of
small-scale enterprises.

While the concentration of income is riot a direct measure of the
concentration of economic power, as the term is generally understood,
the two are closely related. The concentration of economic power
and control arises chiefly from devices of corporate organization,
business practices, strategic positions in business and finance, the
control of organized groups, and ownership. The economic power
exercised by specific individuals through these methods is not neces-
sarily commensurate with the size of their incomes. Largely be-
cause of the separation of ownership and control in corporations,
receivers of large incomes from dividends may not themselves exer-
cise much control. In addition, incomes derived from interest pay-
ments do not usually carry with them significant amounts of control.
On the other hand, even if large dividend recipients do not them-
selves exercise directly the power which goes with their incomes,
others exercise this power for them and the result, as far as concen-
tration of power is concerned, is somewhat the same. Often the
opposite situation takes place when a large dividend recipient owning
a large block of stock exercises a greater amount of control than many
individuals who own, in the aggregate, a larger portion of the cor-
poration. Furthermore, concentration of income among individuals
may not reveal the true extent of the concentration of income among
members of the same family and they may jointly exercise power
greater than their individual incomes would indicate.^ It is readily
appreciated that while there is a significant relationship between the
concentration of individual income and the concentration of economic
power, this relationship is neither simple nor direct.

As profits from the sale ol securities were a major factor in produc-
ing the increase in the degree of income concentration during the
twenties, the problem of preventing a repetition of the stock-market
situation of those years assumes added importance. In this connec-
tion, appropriate banking, credit, and fiscal policies would do much
to prevent a recurrence of the rapid increase in the concentration of
income such as took place after 1924. In addition, the rise in secu-

' See ch. Ill for data on the concentration of dividend payments.

' Family is used here in broader sense than a household. An illustration of this aspect of concentration
of power was provided bj' the testimony of Robert H. Jackson before the Senate Committee on Finance.
Basine his statement on unpublished incomp-tax statistics, Mr. Jackson revealed that of the f.8 taxpayers
with eross income (defined as total taxable and nontaxable income before deductions for realized capital
losses, taxes paid, etc.) of •'51,000,000 and over in 1032, 38 were accounted for by membership in 14 families.
See Hearings Before Committee on Finance on H. R. 8974, 71th Cone., 1st scss., p. 176.

4 CONCENTRATION OF ECONOMIC POWER

rity prices during the twenties was also related to the merger or com-
bination movement in industry and to the rapid increase in the cor-
porate profits which characterized these years. Both of these devel-
opments are a part of the problem of mahitaining effective competition
in industry.

In recent years there has been a growing interest in the influence
of the volume of consumption, saving, and capital formation on
economic stability and the full utilization of resources. As the
volume of individual saving is determined, for the most part, not
only by the size of the total income but also by the distribution of
that hicome, knowledge of the changes in income distribution is im-
portant in this connection. The description presented in this report
of the trends in the concentration of income over the past 2 decades
may, therefore, be of assistance in furthering the understanding of the
relationship of income distribution to the functioning of the economic
system as it is influenced by changes in the ratio of saving to con-
sumption. With reference to this problem, information on the dis-
tribution of income and the changes therein is basic to both an eval-
uation of the effect of existing tax policies on the volume of saving
and, if such measures are considered desirable, to the formulation of
fiscal measures designed to promote economic stability and progress
by influencing the volume of saving.

Furthermore, the distribution of income is a factor of cardinal im-
portance in determining the demand for specific goods and services
and the allocation of the Nation's resources among various uses. In
particular, the rapidity of capital formation is dependent, in large
part, on the nature of the income distribution. With the increase
in the importance of durable consumers' goods and luxury and semi-
luxury goods, the problems of marketing and producing these goods
are closely related to the size and distribution of income. Shifts in
the direction of consumers' demand for such goods have introduced
elements of uncertainty in many industries, and information concern-
ing these shifts may be of assistance in coping with the problems of
producing and marketing consumers' goods.

The distribution oi income is, of course, the basic factor in deter-
mining the distribution of welfare among individuals and families ^
and knowledge of the trends in the concentration of income is, there-
fore, fundamental to an appraisaj or an evaluation of the functioning
of our economy as it is reflected in the welfare of the various groups
in the population. Are the rich becoming richer and the poor, poorer?
Is the middle class being wiped out? Few questions of a social or
economic nature seek answers more insistently. Indeed, in the light
of the basic philosophy of democracy, the statistics on the national
income and the volume of physical production, when used to indicate
the economic progress of the Nation, are, meaningful only when the
distribution among individuals of the ability to claim the products of
the Nation's output is known. Moreover, historical comparison

3 The relationship between the distribution of income and welfare, however, is not a direct one nor one
upon which there is general agreement. For example, under certain assumptions regarding this relation-
sbip income inerjuality may decline during a depression year, but welfare inequality may rise because of
thc'fall in the level of incomes. For a brief discussion of the viewpoints concerning the relation of income
and welfare sor" the article by Simon Kuznets on the National Income in the Encyclopedia of Social
Sciences, vol. XI, pp. 220-22J. The above example is presented in this article.

CONCENTRATION OF ECONOMIC POWER 5

of these indexes of economic well-being is valid only insofar as the
distribution of inconrie is unchanged.^

While it is beyond the scope of this introductory note to consider the
influence of the various factors which determine the distribution of
income, the fact that many of these are within the commonly accepted
sphere of governmental activity makes a brief mention desirable.
Public policy has generally attempted to eradicate various monop-
olistic abuses and practices. The basic requirement is that competi-
tive situations in the Nation's markets for commodities and individual
services be maintained, and, where they do not exist, that competi-
tive situations be brought about or, if this is not feasible, that other
methods of control be instituted. The importance of this task for the
distribution of income has already been noted.

Of far-reaching importance is the relation of taxation and public
expenditures to the distribution of income. The more obvious changes
in the distribution of income are brought about by the imposition of
personal income taxes, estate taxes, and inheritance taxes. The fact
that the concentration of income differs in various phases of the busi-
ness cycle might lead to a consideration of the desirability of a flexible
system of income taxation — that is, a system whereby personal
income-tax rates would be raised during periods of prosperity and
lowered during periods of depression. By diminishing the fluctuations
in income concentration a flexible-rate structure may, if accompanied
by other essential measures, reduce the intensity of the cyclical
movements in business activity.

Taxation and public expenditure, however, affect the distribution of
income in many other ways. It is commonplace knowledge that,
as the value of the benefits a specific individual receives from public
expenditures is rarely equal to the direct and indirect taxes he pays,
the act of taxation and the provision of governmental services in-
evitably involves a change in the distribution of income. The tax
system not only directly influences the distribution of income but also,
through its influence on business activity, the size of the national in-
come. The size of the national income in turn has widespread effects
on the distribution of income and welfare. In addition, the purposes
to which governmental expenditures are devoted influence both the
current distribution of income and the future distribution of income.
One of the ways by which governmental expenditures affect the future
distribution of income is through their influence on the efficiency and
mobility of labor. These expenditures may take the form of providing
greater equality of opportunity to secure education and specialized
training, improving the labor market by employment exchanges, and
various other measures for increasing the efficiency and mobility of
labor. If the system of education including vocational guidance and
training brings about a greater equality of opportunity, a more equal
distribution of individual productive capacity, and an increase in the
efficiency and mobility of labor, these would, in themselves, result in a
more equal distribution of income. However, this result w luld be
realized only if a competitive market for services exists, particularly

* The national income, for example, is the net value of the poods and services produced in a given year.
As the typej of goods and services produced in a given year are influenced by the size distribution of income,
the interpretation of the national income totals of different years, after taking account of price changes,
requires information on the income distributions in the respective years.

Q CONCENTRATION OF ECONOMIC POWER

with regard to freedom of entry into occupations and flexibility of the
rates of pay. Devoting tax revenues to the improvement of the
efficiency and mobility of labor would also make possible a larger
national income. The^ state resorts to more direct methods of in-
■fluencing the distribution of income when it sets wages and prices,
subsidizes specific groups, and supports the monopoly power of others.
These methods of direct governmental intervention in the distributive
process have been steadily increasing.

II. PURPOSE OF PRESENT. INQUIRY

In the light of the foregoing discussion, the highly restricted purpose
and scope of this investigation make it a modest contribution, indeed,
to an understanding of these problems. The purpose of the present
inquiry is to fill a, gap in the body of information upon which judg-
ments concerning these problems must be based. It is largely a
statistical examination of the trends in the concentration and com-
position Jof income over the past two decades. In the immediately
following cliapter, measures of the concentration of income from 1918
through 1937 are presented and the indicated trends are related to the
broad economic movements of the period. In the third chapter an
explanation of the trends in the concentration of income is sought by
an analysis of the changes in the composition of incoro.e; that is, the
shifts in the relative importance of wages, interest, dividends, etc.
In the last chapter the effect of certain types of income and taxes on
the concentration of income is shown and measures of the concentra-
tion of "purchasing power" for the years 1918 through,! 93 7 are pre-
sented.

III. GENERAL CHARACTERISTICS OF DATA

It might be well to indicate at the outset the general characteristics
of the basic data utilized in this study. The available information
falls far short of that required to answer all the important and far-
reaching questions concerning the distribution of income. In view of
the significance of the subject, considerable effort has been expended
to make the fullest use of the available information and, so far as
practicable, to correct the deficiencies of the basic data.

Information on the incomes in the higher brackets was obtained
from tabulations of the Federal income-tax data. The official income-
tax statistics have been published annually in volumes entitled the
Statistics of Income. Unfortunately, these data are not easily
a,dapted for use in economic and social analysis. The most important
limitation of the Federal income-tax information is the fact that it
covers with reasonable completeness but a small percentage of the Na-
tion's income recipients. Furthermore, the data on these relatively
high incomes are merely a byproduct of the administration of the Fed-
eral income-tax law. As such they are subject to important limitations
for use in economic analysis resulting chiefly from the special methods
of defining income embodied in the effective income-tax laws, the
cornplexity and changing character of these laws, and the practices
which have developed in connection with the law. In addition, the
methods adopted in tabulating such a complex mass of data have
created problems, especially when one is interested in comparable in-
formation on total income and specific sources over a fairly long

CONCENTRATION OF ECONOMIC POWER 7

period. Considerable effort has been devoted to analyzing and cor-
recting to some extent the deficiencies of these data, but as is obvious
to those familiar with the income-tax laws and the derived data, a
complete analysis would probably fill several volumes. Examination
of the income-tax legislation, court decisions, the voluminous regula-
tions and rulings of the Commissioner of Internal Revenue, and the
publications of various commercial services which seek to interpret the
laws, provides ample evidence of this fact. However, the vast ma-
jority of changes in the definition of income are not of sufficient im-
portance to affect seriously the basic data. It is hoped that the
important changes have been noted and taken into account so that
the final measures of income concentration are reasonably correct.
The other principal data utilized are annual estimates of the total
income of all individuals and of the number of income recipients.

In order not to divert the attention of the reader from the basic
objectives of the study, the detailed discussion of the underlying
statistics, the methods of estimation and adjustment, and the more
technical aspects of the study have been placed in an appendix. The
concepts of the study, the sources of the data and their more important
limitations, and the extent to which the basic statistics are adjusted,
are all indicated in the text. Reference is made in the various chap-
ters to relevant appendix notes.

CHAPTER II
THE CONCENTRATION OF INCOME: 1918-37

I. DEFINITIONS

Each year individuals receive shares of varying sizes of the Nation's
total income. Two time-honored questions concerning this process
continue to occup}^ a prominent position in the thoughts of those
interested in understanding the operation of, our business economy
and in evaluating the various aspects of social and economic life.
How is the total income of individuals distributed among the Nation's
income recipients in a given year? Has this distribution changed
from 3^ear to year and what has been the trend? It is to the second of
those questions that this chapter is primarily devoted. An attempt
to answer these questions cannot be undertaken until the meaning of
income and income recipient is given. The definition of these terms
will vary with the objectives of the inquiry. Depending upon the
objective, different stages in the process of income circulation may be
selected at which to measure the distribution of income. Without
entering upon a detailed discussion of the complex question of the
definitions of income, three stages may be distinguished. These may
be conveniently described as the distribution of "earning power,"
"purchasing power," and "real income." The following paragraphs
will indicate briefly the purposes which each of the distributions is
designed to serve and the income items included in each. The defi-
nition of the term "income recipient" will also be presented.

1. ^^ Earning Power."

One of the purposes of the present study is to develop for the past
two decades measures of the concentration of income among individual
income recipients which would indicate their relative success in
obtaining primary distributive shares of the Nation's total income.
The primary distributive shares are those received as a direct result of
participation in the conduct of the business economy. In other words,
one of the aims is to learn what have been the changes in the distribu-
tion of income as received directly in return for personal services and
for the ownership of property before income is redistributed either by
public policy or individual choice. This statement should not be
interpreted to mean that public policy does not play an important role
in the primary distribution of incorrie and primary distribution might
well be qualified by the phrase, within the framework of existing legal
and economic institutions. ^ Defining the objective in this manner, no
account would be taken of transfers of income from one individual to
another or from the State and private organizations to individuals in
cases where no services are performed. Nor would there be any
deduction of taxes from income.

' Some of the ways in which public policy influences the primary distribution of income are briefly ind'
catcd in ch. I.

IQ CONCENTRATION OF ECONOMIC POWER

Studies of the distribution of income using this concept of individual
income would be of value to persons interested in the distribution of
"earning power" or acquired income. The term "earning power""
is not altogether satisfactory and no ethical significance should be
attributed to its use. Quantitative measures embodying this concept
would also provide the factual background for an analysis of the basic
economic factors influencing the degree of income concentration, re-
gardless of which concept of income is considered most relevant. This
follows frorn the fact that the distribution of acquired income is funda-
mental in determining the secondary distributions of income. To
summarize the above discussion in terms of specific income items:
Income as a measure of individual earning power should be taken
before deduction for direct or indirect taxes, gifts, and contributions to
charitable organizations. Likewise, gifts, inheritances, and income
from free public services should be excluded from the income of recipi-
ents. In recent years direct-relief payments should also be excluded
from income although as a transfer item it possesses certain unique
characteristics wliich will be discussed at a later point. ^ The Federal
income-tax data adjusted for certain deductions and exclusions
fulfill to a considerable extent the need for income statistics based on
such a concept of individual income. Estimates of the total income
of all individuals can be developed which will embody this income
concept.

Two concepts of personal income are utilized to measure the con-
centration of earning power — economic income and statutory net
income. The former is more inclusive and closely approximates what
is commonly regarded as- the total income of an individual less strictly
business expenses. Included are such sources of income as wages,
salaries, and fees (including wages and salaries paid in kind) , pensions,
net incomes of independent businessmen, net rents and royalties,
profits and losses from the sale of property, dividends, and interest.^

2 See pp. 66-7 of ch. IV. Work-relief wages are included in the total income of individuals when the
concentration of earning power is measured. See discussion in ch. IV, pp. 65-6.

3 There is no necessary identity of the total of individual incomes with the national income. All of the
above-mentioned income sources, with the exception of the income from the sale of property, are usually
included in the national income. Except insofar as such profits are received by dealers in assets or insofar
as they reflect the indirect receipt of undistribi^ted earnings of corporations, this type of gain or loss does not
result from an addition to the net value of the goods and services produced— the national income. These
gains or losses, however, do influence the shares of the Nation's output of goods and services which indi-
viduals may claim and constitute a source from which many individuals may be said to acquire additions
to their other income. To an individual, net profits from the sale of property are very much like other
current income and may be, spent for consumption goods or saved. Therefore, when primary interest lies
in the distribution of individual incomes received in return for personal services or made possible through
the ownership of property, the inclusion of this type of income seems to be proper. Some part of the income
from the sale of property included in the basic statistics represents nominal income occasioned by general
changes in prices. However, the inclusion of this type of gain or loss does not appreciably aSect the findings
of this study.

It should be noted that for purposes other than those of this study, a different treatment of this type of
Income may be desirable. Thus, if interest lies in the determination of the individual savings available
for new investment, thB inclusion of realized gains or losses in the income total may yield misleading indi-
cations as to the volume of funds (exclusive of funds made possible through bank credit) available for this
purpose; and, therefore, the aggregate of total individual income, as' given in this study, would have to be
adjusted for duplications. From some points of view, it would be desirable to include as individual income
unrealized profits or losses due to changes in the value of property. This treatment follows from Robert
Murray Haig's definition of income which has as its purpose the measurement of the "money value of the
net accretion to economic power between two points of time." If this definition is accepted, the treatment
of changes in the value of property used in this study represents a compromise forced by practical diffi-
culties and possible constitutional limitations. A suggested alternative involves the averaging of the
realized gain or loss in some way over a period of years determined either by the period the asset was held
or by some arbitrary number of years. On the other hand, because of the variability of income from the
sale of property, it may be of interest to know the distribution of income excluding this form of income as
well as including it.

It is generally appreciated that because of the divergent purposes which income statistics serve, there is
no method of treating capital gains or losses that will be satisfactory for all purposes. For a more compre-
hensive treatment of these matters see Facing the Tax Problem, particularly pp. 484-491 (Twentieth Century
Fund, New York, 1937); ch. VII of Personal Income Taxation by Henry C. Simons (Chicago, 1938); and
various discussions in the volumes published by the Conference on Research in National Income and
Wealth entitled. Studies" in Income and Wealth, particularly those by Clark Warburton in vol. I, pp. 97-
101, and by Roy Blough and W. W. Hewett in vol. II, pp. 191-268.

CONCENTRATION OF ECONOMIC POWER H

The concept of statutory n6t income is used in parts of this study
largely as a substitute for economic income where the latter cannot
be derived satisfactorily. It is the product of the income-tax law
and as such is not entirely suitable for economic analysis. Statutory
net income is defined as the income of an individual after deductions
for interest paid, certain types of direct taxes paid, noninsured losses
by theft, fire, etc., losses incurred from bad debts, contributions, and
a number of miscellaneous deductions. Also excluded from statutory
net income are wholly tax-exempt interest on Government obligations
and the compensation of State and local government employees which
was exempt from, income taxation during the period covered in this
study. All these items with the exception of the compensation
received by State and local government employees are included in
economic income. It would have been desirable to utilize the concept
of economic income throughout the study but the nature of the basic
source material made it necessary to use statutory net income for cer-
tain years. Wherever possible the data have been adjusted to secure
what has been defined as economic income. The statistics on income
concentration presented in this chapter are intended to measure the
changes in the concentration of earning power.

2. "Purchasing Power."

Persons interested in the effects of changes in the concentration of
income on the expenditures of individuals would use the income con-
cept of "purchasing power" or disposable income. The distribution
of purchasing power would show the distribution of current income
among those who ultimately dispose of it by consumption or saving.
Precise definition of this concept is difficult but some of the major
adjustments to the distribution of earning power or acquired income
may be indicated. First, most of the direct personal taxes paid out
of acquired income should be deducted. Second, the distribution of
acquired income should be adjusted, by deductions from the incomes
of the donors and additions to the incomes of the recipients, for the
transfer of certain types of income among individuals either directly
or through charitable organizations. These transfer income items
include certain gifts among individuals and contributions to, and
payments from, charitable organizations. It may be mentioned that
some have argued for the inclusion of gifts as income to the recipient
and for no deduction of gifts from the income of the donor, considering
gifts as a form of consumption. If gifts were deducted from the in-
comes of the donors, it would be necessary to define the term carefully.
In addition, the distribution of purchasing power is influenced by the
receipt of inheritances and insurance benefits, neither of which are
included in the measures of acquired income or earning power. Third,
direct-relief payments and the veterans' adjusted-service compensa-
tion (the soldiers' bonus) should be added to the incomes of the recipi-
ents. Chapter IV contains data on the trends in the concentration
of purchasing power and a more detailed discussion of the necessary
adjustments to income.

3. "Real Income."'

Although no data are presented embodying the concept of "real
income," a brief discussion of its meaning may be of assistance in

* The term "real income" is employed in a different sense than is usually attached to it in economic and
statistical literature. In conventional usage money income after correction^for price changes is defined as
real income. The term, however, seemed to be also appropriate for use in the present connection.

22 CONCENTRATION OF ECONOMIC POWER

interpreting the data on the distribution of earning power and pur-
chasing power. For those interested in the distribution of income
in the form of money, goods, and services which is at the disposal of
individuals after income has been redistributed either by public
policy or individual choice, it will be necessary to go beyond the
distribution of purchasing powtr and take into account, first, the
incidence by income classes of all taxes, direct and indirect, and
second, the receipt by income classes of the benefits of free services
furnished by public and private agencies. This would involve net
additions to the incomes ol some groups in the population and net
deductions from the incomes of other groups. Studies based on
this income concept would presumably constitute a step in determining
the distribution of welfare. Strictly interpreted, the practical applica-
tion of a concept of income designed to show the distribution of "real
income" is apparently beyond the limits of the available information.
Before an investigation of the distribution of income from this point
of view is feasible, it will be necessary to resolve the theoretical and
statistical difficulties involved in determining the incidence of the
Nation's tax system on the various income groups and in allocating
the benefits of free public and private services among these groups.
There are, however, possibilities for fruitful research along these
lines. Aside from a few cursory remarks in chapter IV, (p. 58)
concerning the incidence of the tax system, this concept of income is
not treated in this study.

4. Income Recipient.

The unit adopted in this report for distributing income is the
individual income recipieiit. The use of this unit follows from the
primary purpose of this study which is to measure the concentration
of earning power. It is the individual who comes in direct contact
with the distributive system and receives an income in return for
personal services and for the ownership of property. After receiving
their incomes individuals spend them either on an individual, family,
or household basis and, therefore, the income-consuming unit may,
and usually does, differ from the income-receiving unit. The dis-
tribution of purchasing power may be studied at two stages: First, the
distribution among those who acquired an income after the indicated
adjustments, and, second, the distribution among the consuming
units by which the income is spent. The data in'chapter IV are based
upon the former distribution. The latter distribution is of interest
primarily in expenditure or budget studies.

For the purposes of this study the tenn "income recipient" includes
all persons receiving an income as defined above and all employable
persons attempting to earn one. During periods of prolonged unem-
ployment the latter group has been numerically important. They
support themselves and their dependents by disbursing savings and
incurring debts or else through gifts and public or private relief. It
will be recalled that these items are not included as income when
studying the distribution of earning power. The adoption of this
rather broad definition of income recipient is necessitated by the
objective of the analysis of income concentration; namely, to develop
measures of the concentration of income among persons either feceiving
or usually receiving an income. It is believed that by using this
lefinition the measures of income concentration would possess greater

CONCENTRATION OF ECONOMIC POWER 23

significance. Were the study confined only to individuals actually
receiving income, the shifts in the concentration of income which may
be indicated from years of prosperity to years of depression would be
of limited value. Thus it would be possible for income to become
more concentrated among individuals normally receiving an income,
yet measures of the concentration of income among those individuals
actually in receipt of an income might well lead to the conclusion
that incomes became less concentrated.

The need for a quantitative measure of the number of income re-
cipients is met reasonably well by the enumeration of persons with
gainful occupations in the decennial censuses. Persons classified as
gainful workers include all who pursue an occupation though not
necessarily employed at the time of enumeration. A person was not
classified as gainfully occupied if he usually spent less than 1 day a
week at an occupation. Persons retired or incapacitated were ex-
cluded. It was necessary to adopt a generally accepted adjustment
of 325,000 persons in the 1930 census for an estimated undercount
of young people who were omitted from the census enumeration be-
cause of a lack of previous work experience. The most serious defect
of these data for the purpose of this study is that there undoubtedly
are persons receiving income from property or from pensions who are
not classified as gainfully occupied. As there seems to be no reason
to believe that the proportion of such nongainfully occupied income
recipients to the total number of persons with gainful occupations has
changed materially during the period under consideration, no estimate,
necessarily arbitrary, was hazarded of the number of such incom-e
recipients. The trends in income concentration would not be affected
if the absolute totals for each year are not precisely accurate. It may
be mentioned in this connection, first, that rnany income recipients
without an occupation nevertheless report one to the census enumera-
tor and, second, that the exclusion of these nongainfully occupied
income recipients is offset, in part at least, by the inclusion as gainful
workers of many part-time workers and children on the borderline of
being classified as income recipients. Among these may be mentioned
the 1,202,000 gainful workers from 10 to 24 years of age attending
schools and 469,000 children on farms under 15 years who were
classified in 1930 as gainful workers. Annual estimates of the number
of gainful workers have been made by several organizations. The
series selected takes into account the changing age composition of the
population and immigration and emigration. A further discussion of
the data on income recipients particularly with reference to the
income tax data will be found in appendix note A-2. (p. 79).

II. STATISTICS OF INCOME CONCENTRATION, 1918-37

The income concentration statistics in this section are intended to
measure the concentration of earning power. The method adopted
for measuring the concentration of earning power involves determining
the proportions or shares of the total income of all individuals received
by fixed percentages of the Nation's income recipients.^ In view of
the relatively high exemptioTns of the Federal income-tax law, it is
possible to measure at the most the snare of the total income received

s For some comments on methods of measuring income "inequality" and concentration, see appendix
Dote B-3, p. 104.

J4 OONCENTRATIO^J OF ECONOMIC POWER

by the highest 2 percent of income recipients. Much interest is
generally attached to evidence of changes in the shares of the total
individual income received by the various proportions of the higher
income recipients. As a relatively small proportion of income
recipients in the high-income brackets largely determine the volume
of personal or noncorporate saving, this is particularly true in con-
nection with questions involving the ratio of saving to consumption,
discussions of which have occupied a prominent position in the analysis
of the factors determining the size of the national mcome.

1. The Highest 1 Percent of Income Recipients.

Table 1 shows the shares of the total individual income received
by the highest 1 percent of income recipients from 1918 through 1937.
The study begins with the year 1918 because this was the first year
the inco'me-tax statistics were considered comparable with those for
the following years.^ It ends in 1937 as statistics on the higher in-
comes are not yet available for 1938 and 1939.

Income is economic income as defined. The income tax data,
which are the source of information on the higher incomes, were ad-
justed in various ways and several items were estimated in order to
arrive at a comparable and complete concept of income. The detailed
procedures are described in appendix note A-1, The legal definition
of income has been changed several times during, this period. For the
years 1918 through 1931 a id the year 1934, the measures of income
concentration are based o i a comparable income concept. An im-
portant change in the defiiJtion of income occurred in 1932 and again
in 1934. These changes in the law concern the treatment of realized
capital gains and losses. For 1932 and 1933, losses from the sale or
exchange of stocks and bonds held 2 years or less were limited to gains
from such transactions. As it was not possible to adjust the data so as
to make them comparable with the preceding or following years, no
data are presented in table 1 or subsequent tables for these 2 years. ^

« The Federal individua] incometax law has been in effect cdfntinuously since Mar. 1, 1913. In view of
thje questionable reliability of the statistics for the years prior to 1918, the comparatively small number of
persons filing income-tax returns, the changes in the definition of income and in the manner in which the
statistics were tabulated, no measures of income concentration are presented for these years.

The questionable reliability of the data is largely due to the initial difficulties of enforcement. At least
several years were required before the meaning of such a comprehensive and complex law was developed
and the enforcement machinery organized. The data for these early years are examined in considerable
detail in chs. 22 and 30 of Income in the United States, vol. II, published by the National Bureau of Eco-
nomic Research. In this book both O. W. Knauth and F. R. Macaulay conclude that the income-tax
data are considerably more accurate for 1918 than for the preceding years. The practice of requiring em-
ployers to report the amounts over $1,000 paid to employees was instituted in 1918 and this probably led to
a considerable improvement in the accuracy of the statistics. Prior to 1917 the exemptions were quite high,
especially when the prevailing level of incomes is taken into account, and consequently the number of re-
turns filed was comparatively small. The exemptions wore $4,000 for married persons and $3,000 for single
persons. In 1917 the exemptions were lowered to $2,000 and $1,000, respecti-<'ely, for married and single per-
sons. As the level of income was higher beginning in 1917, the coverage of the total distribution of income
was even larger than the difference in exemptions sugpests.

From 1913 thi'ough 1915 losses sustained in transactions entered into for profit but not connected with
business or trade were not deductible. In 1916 and 1917 such losses were deductible to the extent of the
aggregate income from such transactions. In 1918 and subsequent years until 1932 such losses were com-
pletely taken into account in cne form or another. In addition to these specific items there were numerous
refinements of the definition of income during this early period of experimentation. Prior to 1916 little
detail was published. However, in 191G the data were tabulated in a way as to render them incomparable
with preceding'or succeeding years. The net income of returns filed separately by husband and wife was
combined and the total was classified as one return. Failure to note this unique met'hod of tabulation
has resulted in questionable conclusions concerning the pre-war concentration of income. Contributions
to charitable and scientific organizations, etc., not exceeding l.l percent of the net income before deducting
the contributions with certain exceptions, were deductible in 1917 and subsequem, years. However, in
1917 contributions were tabulated with net income in contrast to later years when contributions were
excluded from net income.

' See, however, footnote 11, p. 16, which presents data for 1932.

CONCENTRATION OF ECONOMIC POWER 15

Beginning in 1934 varying proportions of gains and losses were in--
cluded in net income depending on the length of time the asset was
held and losses were limited to $2,000 in excess of gains. The defi-
nition of capital asset was also broadened slightly. With the use of a
special tabulation of actual realized capital gains and losses for 1934
made available by the Treasury Department, it was possible to
adjust the basic data on the higher incomes for 1934 so as to make
them comparable to the preceding years. The data for 1935, 1936,
and 1937 could not be corrected. The second measure of income
concentration for 1934 and the measures for 1935 to 1937 are there-
fore, not strictly comparable with the measures for the preceding years.
As is readilj' seen by comparing the two figures for 1934, the measures
of income concentration for the period 1934-37 slightly overstate the
degree of income concentration.^

In interpreting the data it is important to note that the level of
the percentages represents an understatement for each year of the
proportion of total individual income received by the highest 1 percent
of income recipients. This consistent understatement does not dimin-
ish the value of the data for the present purpose, which is to determine
the year-to-year changes in degree of income concentration. The
understatement arises largely from the fact that the data on the higher
incomes have been used without correction for nonreporting or under-
reporting of income to the income-tax authorities. It is generally be-
lieved that underreporting and nonreporting of income does exist and,
therefore, the given percentages do not purport to show the actual
degree of concentration for any year. It was not thought feasible
to attempt to correct the data as any adjustment would necessarily be
largely arbitrary. This understatement was offset in small part by
the manner in which the salaries and wages of State and local govern-
ment employees were treated. Their salaries and wages were ex-
empted from Federal income taxation during the period covered and
hence from the basic data on the higher incomes. It was found expe-
dient for various reasons to exclude the employees of State and local
governments and their compensation fram the other data upon which
the measures of income concentration are based. The exclusion of
this group does not significantly affect the movement of the indexes
of concentration and they may be taken as indicative of changes in
the concentration of income among all income recipients.

For those interested in an estimate of the actual degree of concen-
tration of income for a given year, the following data may be cited from
two studies which had this as their purpose. In a study for 1929 by
the Brookings Institution it was estimated that the highest 1 percent
of income recipients received about 24 percent of total individual in-
come.^ The comparable percentage for 1929 in table 1 is 18.5. For
1918 it was estimated by the National Bureau of Economic Research
that the highest 1 percent of income recipients received 13.7 percent of

8 In addition the extent of overstatement varies slightly from year to year in the following order of im-
portance: 193? or 1937, 1935, and 1936. Thp characteristics of the data for these years are described in some
detail in appendix note B-2. • The directions of the changes in income concentration from 1934 through 1937,
however, are correctly shown by the data in tabk 1. See appendix note B-2 and text below for interpretation
of changes in income concentration during these years.

» M. Leven, H. O. Moulton. and C. Warburton. America's Capacity To Consume, Washington, D. C.
1934, p. 207. The estimate of the distribution of income by size is credited to Maurice Leven.

16

CONCENTRATION OF ECONOMIC POWER

the total individual income. The comparable percentage in table 1 is

12.8.'°

Table 1. — Shares of total individual income received by the highest 1 -percent of
income recipients, 1918-37 '

Year

Percent 8f
total in-
come re-
ceived by
highest 1
percent of
income re-
cipients

Index

(1918=
100)

Minimum

incomes of

highest 1

percent

$6, 385
7,910
8,010
6,845
7,445
7, 505
8,040
9,380
9,655
9,590

Year

Percent of
total in-
come re-
ceived by
highest 1
percent of
income re-
cipients

Index

(1918 =
100)

Minimum

incomes of

hiehest 1

percent

1918

12.79
13.35
12.42
13.57
14.24
12.95
14.17
16.39
16.21
17.18

100.00
104. 38
97.11
106. 10
111.34
101.25
110.79
128. 15
126. 74
134. 32

1928

19.26
18.47
14.63
13.72

12.66

13.03
13.41
14.53
13.29

1.50. 59
144. 41
114.39
107. 27

98.98

101.88
104.85
113.60
103.91

$10. 140

1919

1920

1921

1922

1923

1924

1925

1926...

1927

1929

1930

193L......

1934

1934

1935

19.36.

1937

9, 975
8,080
6,595

5. ,175

5. 375
5,800
6.880
6,940

1 Economic income. The 1934 figure of 12,66 percent is comparable to percentages for the preceding years.
The second percentage for 1934 and the perceutagos for 1935-37 are slightly overstated relative tr) the pre-
ceding yews due to a change in the income concept. F.or comparability of the data for 1934-37 see text
and appendix note B-2. Total individual income for 1934-37 includes work-relief wages.

Source: See appendix note A-1.

It should be noted that the statistics presented in this study show
the concentration of income each year. No information is presented
concerning" the year-to-year stability of the incomes of identical indi-*
viduals. That is, the extent to which individuals maintain their
relative position in the income distributions of different years. This
aspect of the concentration of income is obviously important, but it
has npt received the attention it deserves largely because of the lack
of adequate information.

Changes in income concentration. — For the years covered by the
data the degree of income concentration, as measured by the income
shares received by the highest 1 percent of income recipients, was
lowest in 1920 and 1934 and highest in 1928.^' There has been no
significant trend over the period as a whole. The level and the ex-
tent of variation in the degree of income concentration was about the
same during recent years, 1934 through 1937, as during the years 1918

10 W. C. Mitchell, W. I. King, F. R. Macaulay, and O. W. Knauth, Income in the United States, vol. I.
New York, 1921, p. 133. The individual distribution is the work of F. R. Macaulay. It should be noted
that each investigator applied different correction factors to the income-tax data. The differences between
the corrected and uncorrected percentages for the 2 years, therefore, represent in part differences in judg-
ment as to the degree to which the higher incomes are understated in the income-tax data of each year.

The two figures for 1918 are based on data including the military personnel. Dr. Macaulay has esti-
mated that there were 2, .500,000 soldiers, sailors, and marines who had little, if any, income beyond the
pay, food, and clothing provided by the Goverrunent which he valued at .$700 per man (op. cit., p. 131).
If these men and their incomes are excluded from the datt. the figure corresponding to that in table 1 of
this study is 12.9 percent .ind to thfjt from the Xational Bureau study is 13.8.

U N'o measures of income concentration were presented in table 1 for the years 1932 and 1933 due to the
effect on the income tax statistics of the change in the law limiting losses from the sale or exchange of stocks
and bonds held 2 years or less to the profits from such sales. However, it is virtually certain that the con-
centration of income continued to decline in 1932. Based on the concept of income embodied in the income
tax law effective in 1932, the share of the total individual income received by the highest 1 percent of income
recipients, in terms of statutory net income, was 10.11 percent. If realized losses from the sale of property
were fully included as income, the share of total individual income received by the highest 1 percent of
income would doubtless be less than this figure. As the percentage share of 10.11 for 1932 is expressed in
terms of statutory net income, it should be compared with the percentages in table 2, where it is shown
that the percentage of total income received by the highest 1 percent of income recipients was 11.27 in 1931
and 10.20 in 1934. The degree of income concentration, therefore, probably reached a minimum in 1932.
It seems that the income share of the highest 1 percent of income recipients in 1932 was about the same .
size as in 1920 when the income share bas?d on statutory net income was 10.05 percent.

CONOENTRATION OF BOONOMIG POWER 17

through 1924. The average share of total income received by the high-
est 1 percent of income recipients was 13.36 percent for the years 1918
through 1924 and 13.56 percent for the years 1934 through 1937.^^
The maximum and minimum income shares for the early period were
14.24 percent in 1922 and 12.42 percent in 1920 and for the latter
period 14.53 percent in 1936 and 12.66 percent in 1934. In contrast
to comparative stability of the income shares during these two periods,
the period 1925 4,hrough 1932 witnessed wide fluctuations in the size
of the shares of income received by the highest 1 percent. The income
share received by the highest 1 percent of incorrie recipients in 1928
was more than one-third larger than in 1924 and the share received iii
1932 was more than one-third less than in 1928. Apart from minor
Variations associated with short cyclical movements in business activ-
ity, the concentration of income increased from 1921 through 1928
with most of the increase taking place after 1924, After 1929 income
concentration declined sharply, reaching a minimum in 1932 and 1933.
The concentration of income then increased from 1934 through 1936
and declined in 1937.

Table 1 also shows the income level above which the highest 1
percent of income recipients were located each year. The income
levels are also understated because no account was taken of under-
reporting and nonreporting of incomes to the Bureau of Internal
Revenue. In the estimate previously cited for 1918 the income level
for the highest 1 percent was $7,700 as compared with $6,385 in table
1. In the Brookings estimate for 1929 the income level was $12,850
and in table 1, $9,975. During the period covered by the data the
minimum income of the highest 1 percent of income recipients has
been subject to marked fluctuations. In 1934 anyone with an income
in excess of $5,375 was among the highest 1 percent of income recip-
ients, whereas in 1928 an income of $10,140 or more was necessary
for inclusion in this category. These variations reflect, in large part,
the movements in the level of prices and incomes as well as the shifts
in the degree of income concentration. Evidence of changes in the
minimum incomes of the highest 1 percent of income recipients reveals
the fallacy of taking the proportions of total income received in each
year by individuals with incomes above a fixed amount as indicative
of changes in the degree of income concentration. During periods of
changing price and income levels, this procedure, which has some-
times been adopted, will yield misleading results.

Business activity and income concentration. — While it is not the pres-
ent purpose to undertake an analysis of factors responsible for the
indicated changes in the degree of income concentration, the more
obvious trends and their relationships to broad economic movements
will be outlined. Chart I presents in graphic form the data in table 1,
the size of the total income per income recipient (deflated roughly to
eliminate the influence of price changes), and the turning points of the
short business cycle, as defined by the National Bureau of Economic
Research. This chart serves the purpose of relating the changes in
the concentration of income to the size of the total income and to
general business conditions. The changes in the magnitude of the

" Because of the limitations on the deductibility of realized capital losses the percentages for the latter
period are slightly overstated. Therefore, the two averages are probably more nearly equal than is indicated
by the figures.

256149-40 -No. 4 -3

18

CONCENTRATION OF ECONOMIC POWER

total income per recipient may also be taken as indicative of the
amplitude or lq tensity of the cyclical movements.

The chart reveals in striking fashion that the changes in degree of
income concentration as measured by the proportion of totaJ income
received by the highest 1 percent of income recipients have conformed
to a cyclical pattern. Income concentration increased during periods
of business expansion and declined during periods of business con-
traction. The only exception was the slight contraction in business
activity during 1927 which was not severe enough to produce a decline
in total income or in the index of income concentration. As the
income accounting period is the calendar year, the relationship between

C14ART 1

Concentration of income and total income per recipient. 1918-37

P£fiC£NT

,^ \c\ I rrn

DOLLARS

SMAses OF roTAi iNcoMe aeceiveo
Br MiOHesr of/e ptuctNT of income

MEOPIENTS

scAie ON leFT

2,000
1,600

'3S "36 -37

Source: Income shares from Table 1. Income per recipient was obtained by deflating Total Income, as
given in Table 10, by the cost-of-living index of the National Industrial Conference Board and dividing
the deflated Total Income by the number of income recipients, as given in Appendix Table A-4. Data
on turning points in the business cycle, except in 1937, from' Bulletin 69 of the National Bureau of
Economic Research.

Note. — The income shares understate for each year the proportion of total income received. See text
and footnotes to Tables 1 and 10.

income concentration and business conditions is somewhat obscured.
If the income accounting period coincided with the phases of the
business cycle, the relationship between business activity and income
concentration would be more impressive and the fluctuations in
income concentration would probably be somewhat larger.

There has been a marked tendency for the degree of income con-
centration to vary with the size of total income per income recipient.
High income concentration has been associated with a large income
per recipient and low income concentration with a small income per
recipient. This is clearly shown by the chart. Measured on a
calendar-year basis the degree of income concentration has been more
sensitive to changes in business activity than to changes in the size

CONCENTRATION OF ECONOMIC POWER IQ

of total income. A rise or decline in business activity during the
closing months of a year was generally accompanied by a rise or decline
in income concentration. As will be shown in chapter III this close
correlation is traceable in large part to the highly variable source of
income, profits and losses from the sale of property. The absolute
amounts of annual total income were not so susceptible to these
changes in business activity though the rises or declines in business
activity during the closing months of a year did influence, of course,
the rate of increase or decrease in annual total income. Thus, income
concentration rose but total income declined in 2 years — 1921 and
1924, In both of these years there was an expansion of business
activity during the latter part of the year. Income concentration
declined" and total income rose in 4 years — 1923, 1926, 1929, and 1937.
In each of these years a decline in business activity set in during the
latter part of the year. Were the income accounting periods deter-
mined by the rises and declines in volume of income, measm^ed on a
monthly basis, the above-noted lack of correspondence between
increases and decreases in total income and in the degree of income
concentration would largely disappear. The analysis presented in
chapter III of the relation between the concentration of income and
the composition of incomes will throw further light on some of the
factors responsible for the shifts in income concentration.

The relationships between changes in income concentration, total
income, and business activity will, perhaps, be brought out more
clearly by a brief year-by-year summary of the changes in income
concentration. In 1919, the year of the short-lived post-war pros-
perity, total individual income rose as did the degree of income con-
centration. During 1920 business conditions experienced a severe
and rapid contraction. Total income declined and the share received
by the highest 1 percent was less than in the previous year. In 1921
a decline in total income from the 1920 level was associated with an
increase in the income share of the highest 1 percent of income recip-
ients. However, business conditions, which were depressed in 1920
and most of 1921, began to improve during the closing months of
the latter year. In 1922 there was an increase in total income and
in the concentration of income. The year 1923 was one during which
the total income rose sharply, but the share of the total income received
by the highest 1 percent of income recipients declined. Though both
the national income and total individual income showed substantial
increases in 1923, the latter half of the year was characterized by a
contraction in general business activity. In 1925, a year of rising
income, the increase in income concentration was the largest, both in
percentage and absolute terms, of any of the years covered. The
slight contraction in business activity during 1927 was not sufficiently
severe to produce a decline in either the size of the total income or in
the index of income concentration. During the years of high business
prosperity, 1928 and 1929, total individual income and the degree of
income concentration reached the highest level of any of the years
which the data include. The share of the total income received by
the highest 1 percent of income recipients in 1928 was larger than in
1929. Again it may be noted that while the total individual income
and the national income were larger in 1929 than in 1928, the latter
part of 1929 was a period of declining business activity. The period

20 CO^'CENTRATION OP ECONOMIC POWER

from 1922 through 1929 may be considered as the expansion phase of
a long business cycle and the concentration of income increased, apart
from minor variations, with the rising amplitude of the cycle.

During the depression beginning in 1929 income concentration
experienced a sharp decline from the peaks of 1928 and 1929 and reached
a minimum in 1932 and 1933. This drop was paralleled by a diminish-
ing total income. During 1934, 1935, and 1936 the concentration of
income increased moderately, with the increase in concentration during
1936 being of largest proportions. While the total income of all
indi\aduals in 1937 was greater than in 1936, business conditions
declined sharply in the latter half of the year. As was the case in
previous years characterized by a sharp reversal in an upward move-
ment of business activity, the income share of the highest 1 percent
of income recipients was reduced. In consequence of the income
concept used during the years 1934 through 1937, the actual increase
in income concentration from 1934 through 1936 was sornewhat
greater than indicated in table 1 and the deconcentration of income
during 1937 was also greater.^^ As noted above, the average income
share of the highest 1 percent of income recipients was approximately
the same during the years 1918 through 1924 and 1934 through 1937.
During these two periods the deflated average income of all income
recipients was also approximately the same, $1,466 for the years
1918 through 1924 and $1,419 for the years 1934 through 1937.^"

2. Selected Proportions of Income Recipients.

The measures of income concentration presented thus far have been
confined to the highest 1 percent of income recipients. Before the
data at our disposal are fully exploited, it will be possible to present
evidence on the changes in the shares of the total income received by
various other proportions of income recipients. Considerable impor-
tance is attached to changes in the degree of income concentration with-
in the highest 1 percent and such changes could very well be obscured
by considering only the highest 1 percent of the recipients. Further-
more, in order to interpret properly the measures of income concentra-
tion for the highest 1 percent, it is necessary to know the distribution
of income within this group. The available information enables us to
cover at most the highest 2 percent of income recipients and even for
this group the estimates are probably not so reliable as those for the
smaller proportions of income recipients. ^^ The proportions of income

" The effect of the peculiar definition of income in use during 1934 through 1937 on the measures of income
concentration is analyzed in appendix, note B-2.

» Both averages in terms of 1923 prices. The average for the latter period is slightly overstated due to
the incopic concept used for these years. See chart I.

" The most important reason why the income measures for the highest 2 percent of Income recipients are
probably less reliable than those for the smaller proportions is the fact that the lower the minimum income
level for the croup of income recipients, the more important is nonreporting and underreporting of income
to the Bureau of Internal Revenue. In no year, however, was it necessary to go below the net income for
which a return was required by law. In all years except two, tlie lower limit of the highest 2 percent of in-
come recipients was at least .several thoasand dollars above the net income lor which a return was required
by a married person. In 1931, the minimum income level of the highest 2 percent of income recipients wa^
$3,9«j0, while tne net income for which a return was required was $3,500 for a married person and $1,500 for a
single person or a married person not living willi husband or wife. In 1931 the minimum net income level
was $.1,275 while the net incomes requiring a return were $2,"W and $1,000, respectively.

Another reason of lesser importance is the fact that the data for net incomes under $5,000 were climated
on the basis of a sample for the years 1918 through 1927 and for 1929; and partly estimated and partly tabu-
lated for 1928, 1930 and subsequent years. For discussion of validity of method used prior to 192S see Statis-
tics of Income for 1928, pp. 19-24. For the minimum net income levels of the highest 2 percent of inconn
recipients see table 4, p. 26.

CONCENTRATION OF ECONOMIC POWER 21

recipients in terms of which income concentration is measured were
selected in order to present a reasonably complete picture of the shifts
in the income structure of the Nation within the limits of the available
information. With regard to the comparative reliability of the meas-
ures of income concentration for the various proportions of income re-
cipients, the measures for the highest 1 percent of income recipients
are believed to be the best set of data. This superiority largely arises
from limitations of certain necessary adjustments of the income-tax
data and certain practices which have developed in connection with
the operation of the income-tax law.^^

Character oj the statistics. — For the years 1918 through 1925 the in-
come concept used to measure the incomes of these selected groups
of income recipients is statutory net income. While it was possible
for the years 1918 through 1926 to convert the basic data for the
highest 1 percent to economic income with assurance of substantial
accuracy, it was not feasible to do so for the other proportions of
income recipients. ^'^ Table 2 contains for the years 1918 through
1931 and 1934 through 1937 the shares of total economic income
represented by the statutory net incomes of the selected proportions
of income recipients. For the period 1926 through 1931 and 1934
through 1937, estimates are presented in table 3 for the various pro-
portions of income recipients in terms of economic income. The
statutory net incomes of the various proportions of income recipients
are less than their economic incomes as the former concept does not
include the various types of deductions and tax-exempt interest.
Therefore, the use of statutory net income results in smaller shares of
total income. For example, in 1929 the statutory net income of the
highest 1 percent was 16.03 percent of the total economic income
(table 2) while the economic income of this group was 18.47 percent
of the total economic income (table 3).

" See appendix note A-1, pp. 71-2, 75-7, and appendix note B-3, p. 106. In addition to the factors men-
tioned in these two notes, the measures for the highest 1 percent are more reliable than those for the highest
2 percent for reasons indicated in the precedine footnote.

'• The principal reason for the inability to present estimates for all proportions of income recipients and all
years in terms of economic income was the method followed in the Statistics of Income in tabulating the var-
ious deductions from total income. For the years prior to 1926, realized capital losses which should be de-
ducted from "total income" to secure economic income were tabulated with other deductions which should
be added to net income to obtain economic income. While it was practicable to separate realized capital
losses from the other deductions for the highest 1 percent of income recipients, such a separation for the other
proportions of income recipients would be subject to considerable error and, as will.be shown, is not essential
for the present purpose.

22 OONriENTRATiaN OP ECONOMIC POWER

Table 2. — Shares of total individual income received by selected proportions of
income recipients, 1918-37 *

Group of income recipients

Year

Highest
2 percent

Highest
1 percent

Highest

n of 1 per

cent

Highest

Mo of 1 per

cent

Highest
Hoo of 1
percent

Percentages

Ibis -- -

14.12
14.49
13.37
14.47
15.46
14.37
15.32
17.86
17.62
18.65
20.68
19.84
16.14
15.33

13.85

14.31
14.77
16.16
14.71

10.82
11.21
10.05
10.80
11.83
10.76
12.02
14.12
13. 91
14.82
16.81
16.03
12.35
11.27

10.20

10.61
11.15
12.42
11.15

8.43

8.03

7.54

8.06

9.02

8.20

9.20

11.04

10.81

11.67

13.60

12.94

9.43

8.33

7.61

7.95
8.39
9.49
8.42

4.55
4.47
3.62
3.84
4.56
4.06
4.57
5.88
5.84
6.46
8.05
7.78
4.99
4.06

3.78

3.90
4.18
4.75
4.16

1.65

1919

1.63

1920 - -

1.12

1921

1.16

1922 -

1.57

1923

1.37

1924

1.54

1926

2.27

1926 . . i

2.32

1927

2.60

1928

3.60

1929

3.65

1930

1.97

1931

1.44

1934

1.31

1934

1.30

1935

1.41

1936

1.57

1937.

1.37

Indexes (1918=100)

100.00
102. 62
94.69
102. 48
109. 49
101. 77
108. 50
126.49
1?4. 79
132. 08
146. 46
140.51
114.31
108, 57

98.09

101. 35
104.60
111.45
104. 18

100.00
103.60
92.98
99. 82
109. 33
99.45
111.09
130.50
128.56
136. 97
155. 36
148.15
114.14
104. 16

94.27

98.06
103. 05
114.79
103. 05

100.00
102. 37

89.44

95.61
107.00

97.27
109.13
130. 96
128.23
138.43
161.33
153.50
111.86

98.81

90.27

94.31
99.53
112. 57
99.88

100.00
98.24
79.56
84.40
100.22
89.23
100.44
129.23
128.35
141.98
176. 92
170. 99
109. 67
89.23

83.08

85.71
91.87
104. 40
91.21

100.00

1919

92.73

1920 -

67.88

1921 . .

70.50

1922 -.

95.15

1923

83.03

93.33

1925 -

137.58

1926

140.61

1927

157.58

1923

212. 12

1929

215. 15

193C .

119.39

1931

1934..

87.27
79.39

1934

78.79

1935

85.46

1936 .

95. 15

83.03

• statutory net income of the selected proportions of income recipients. The first group of percentages
for 1934 are comparable to those for the precedinR years. Except for the highest Moo of 1 percent in 1934
through 1937 and the highest Mo of 1 percent in 1935 through 1937, the second group of percentages for 1934
and the percentages for 1935-37 are slightly overstated relative to the percentages for the earlier years due
to a change in the income concept. The percentages for the highest Moo of 1 percent in 1934-37 and the
highest Mo of 1 percent in 1935-37 are slightly understated relative to the figures for the preceding years.
For comparability of data for 1934-37 see text and appendix note B-2. Total individual income for 1934-37
includes work relief wages.

Source: See appendix note A-1.

CX)NCENTRATION OF ECONOMIC POWER

23

Table 3. — Shares of total individual income received by selected proportions of
income recipients, 1926-37 '

Group of income recipients

Year

Highest
2 percent

Highest
1 percent

Highest
\i of 1 per-
cent

Highest
Ho of 1
percent

Highest
Moo of 1
percent

Percentages

1926

20.52
21.54
23.65
22.84
18.99
18.51

16.95

17.41
17.81
18.83
17.37

16.21
17.18
19.26
18.47
14.63
13.72

12.66

13.03
13.41
14.53
13.29

12.60
13.51
15.56
14.89
11.22
10.19

9.57

9.86
10.19
11.17
10.14

6.81
7.50
9.18
8.'90
6.00
5.10

4.94

5.03
5.22
5.74
5.15

2.70

1927

3 02

1928...

3.94

1929 - _.

4.00

1930 . .

2 38

1931 -

1.88

1934

1. 7>

1934 . .

1.75

1935... -

1.81

1936

1.97

1937

1.77

Indexes (1926=100)

1926

100.00
104. 97
115.25
111.31
92.54
90.20

82.60

84.84
86.79
91.76
84.65

100.00
105.98
118.82
113.94
90.25
84.64

78.10

80.38
82.73
89.64
81.99

100.00
107.22
123.49
118. 17
89.05
80.87

75.95

78.25
80.87
88.65
80.48

100.00
110. 13
134.80
130.69
88.11
74.88

72.64

73.86
76.65
84.29
75.62

100.00

1927

111.86

1928

145.93

1929

148.16

1930

88.16

1931 - . ....

69.63

1934 .

66. 19

1934.

64.81

1935

67.04

1936

72.96

1937

66.66

1 Economic income of the selected proportions of income recipients. The first group of percentages for
1934 are comparable to those for the preceding years. Except for the highest Moo of 1 pmrcent in 1934 through
1937 and the highest Mo of 1 percent in 1935 through 1937, the second group of percentages for 1934 and the
percentages for 1935-37 are slightly overstated relative to the percentages for the earlier years due to a change
in the income concept. The percentages for the highest Moo of 1 percent in 1934-37 and the highest Ho of 1
percent in 1935-37 are slightly understated relative to the figures for the preceding years. For compara-
bility of data for 1934-37 see text and appendix note B-2. Total individual income for 1934-37 includes work
relief wages.

Source: See appendix note A-1.

24 CX)NCENTRATION OF ECONOMIC POWER

Aside from the additional degree of understatement, the principal
defect of the use of statutory net income for the present purpose is
that the statutory net income shares are somewhat more variable than
the economic income shares. This difference is readily appreciated
by a study of chart II which is drawn on a logarithmic or ratio scale,
or by comparison of the two indexes of the percentages of total
income received by the highest 1 percent — economic income in table
1 and statutory net income in table 2. The index in table 1 has a
minimum of 97.11 in 1920 and a maximum of 150.59 in 1928, while
the index of the highest 1 percent in table 2 has a minimum of 92.88
and a maximum of 155.36.^® The increased variabiUty arises from
the fact that the statutory net income is generally a larger percent-
age of economic income in years of high income concentration than
in years of low income concentration. As indicated by a comparison
of the measures for the years 1926 through 1937 in tables 2 and 3,
the difference between the variabiUty of indexes of income concen-
tration based on statutory net income and economic income increases
shghtly as the proportion of income recipients becomes smaller.
Appendix note B-1 examines in som^ detail the relationship of statu-
tory net income to economic income for different years and for dif-
ferent proportions of income recipients. The conclusions of this note
may be summarized by stating that apart from the differences in
the degree of variability both for different years and for the different
proportions of income recipients, the data on income concentration
in terms of statutory net income will be indicative of the shifts in
the concentration of economic income.^® However, due to the chang-
ing ratio of statutory net income to economic income, comparison of
the levels of iucome concentration in 2 years will yield, in several
instances, different results according to whether the statutory net in-
come measures or the economic income measures are used. Thus,
the statutory net income data of a given year sometimes show a
lower degree of incor- e concentration than in an earlier year, whereas
the economic income data show a higher degree of income concentra-
tion. Comparison of the two measures for the highest 1 percent in
1918 and 1921 provides an illustration of this limitation of the statutory
net income data.

A further qualification with regard to the data in tables 2 and 3
should be noted; namely, that for the years 1924 through 1931, the
percentages of total income received by the three smaller proportions

1' The relative mean deviation of the percentages for the years 1918 through 1937 in table 1 is 0.1.1 and in
table 2 (highest 1 percent of income recipients), 0.13. The relative mean deviation is a measure of the rela-
tive variability of a scries of numbers and is defined as the arithmetic mean of the deviations (disregarding
signs) from tne arithmetic mean of the original data divided by the arithmetic mean of the original data.
Thus, for the data in table 1, the arithmetic mean of the percentages is 14.54; the arithmetic mean of the
differences (disregarding signs) between the various percentages and 14.54 is 1.57; the relative mean deviation

is, therefore,-^-— or 0.11.

" A factor which may result in a greater variability than is actually the ease in the economic income shares
received by smaller proportions of income recipients is mentioned in appendix note A-1, pp. 75-7.

CONCENTRATION OF ECONOMIC POV/ER 25

of income recipients slightly understate the concentration of income
relative to the percentages for earlier years. This understatement is
due to the effect on the method of tabulating the income-tax data of an
alternative method of treating reahzed capital losses on assets held
over 2 years as a tax credit introduced in the income-tax law in 1924.
In the income-tax data individuals were classified according to size
of their "statutory net incomes" which did not take account of the
losses reported for tax credit. In the adjusted data on the higher
incomes used in this study, realized capital losses are fully deducted
but some individuals are left in higher income classes than their
income permits after deduction of the reahzed capital losses on assets
held over 2 years reported for a tax credit. The understatement of
the incomes of the various proportions of income recipients resulting
from this procedure is relatively larger for the smaller groups of income
recipients. It is negligible for the 1 and 2^ percent groups. The
understatement is more important in years of large reahzed capital
losses. For a more detailed treatment of this matter see appendix
note A-1 (pp. 71-2).

The comments made in connection with the data in table 1 for the
highest 1 percent regarding the understatement in the level of the
measures of income concentration and the comparability of the data
for the years 1934 through 1937 are also applicaole to the data in the
tables for the selected proportions of income recipients. In addition,
the degree of overstatement which characterizes the measures of in-
come concentration for 1934 through 1937 declines in importance for
the smaller proportions of income recipients. This is readily seen by
comparing the two sets of figures for 1934. There is actually a very
slight understatement in the second 1934 figure for the highest
one one-hundredth of 1 percent. In 1935, 1936, and 1937 the data for
this group -and probably for the highest one-tenth of 1 percent group
as well are also subject to some understatement.^*^ It is necessary,
therefore, to exercise considerable care in interpreting the data for the
years 1934 through 1937, particularly for the smaller proportions of
income recipients. Thus, while the share of total income received
by the highest one one-hundredth of 1 percent of income recipients
shown in table 2 is the same in 1922 and 1936, the figure for 1936 is
known to be somewhat understated relative to that for 1922. The
share of total income received by this group in 1936 probably lies
between the 1922 figure of 1.57 percent and the 1918 figure of 1.65
percent. On the other hand the 1936 percentage shown for the highest
1 percent of income recipients is known to be somewhat overstated
and is less than the share received by this group in 1930 and approxi-
mately equal to the share of 12.02 percent received in 1924 by this
group.

'"See appendix note B-2, p. 101, which analyses the data for these years.

26

CONCENTRATION OF ECONOMIC POWER

Table 4. — Minimum statutory net incomes of selected proportions of income recipients,

1918-37 1

Group of income recipients

Year

Highest
2 percent

Highest
1 percent

Highest
Hofl
percent

Highest
Hoof 1
percent

Highest
Hoo of 1
percent

1918 . --

$3,980
4,510
4,615
4,160
4,320
4,380
4,605
4,840
5.030
5,060
5,390
5,380
4,605
3,960

3,275

3,275
3,460
3,900
4,100

$5, 510
6,760
6,910
5,800
6,235
6,560
7,045
8,130
8,205
8,310
8,820
8,680
6,980
5,660

4,610

4,610
5,005
6,040
6,075

$8,555
10.605
10, 876
8.990
9.895
10. 570
11,480

13, 495
13, 330
13,600

14. 690
14, 030
10. 705

8,415

6,840

6,840
7.550
9.575
9, 4 15

$26, 160
31,455
29, 945
24,790

28, 8«)

29, 825
34,230
40,005
39, 365
41,710
48, 510
44,750
29, 520
21,415

18,250

18, 250
20, 700
27, 700
26, 220

$106, 500

1919 - --

118,000

1920 — -

96, 440

1921 -

79, 280

1922 -- -

98, 930

1923 . -

99,800

1924

118,400

1925

158, 100

1928 -

157, 450

1927 .

182,000

1928 -.-

239, 750

1929 -

231,000

1930 -

119,500

1931 -

1934

79, 950
63, 300

1934 l

63,300

1935 -

73,100

1936

96,800

1937

90,150

' Due to the change in the definition of income after 1934 the income levels for the years 1934 through 1937
are slightly higher than they would be were the data comparable throughout the whole period. In addition
the income levels for the 3 smaller proportions of income recipients are somewhat too high for the years 1924
through 1931 owing to the tax-credit method of reporting capital losses on the sale of assets held more than 2
years. See text and appendix note'A-1, pp. 71-2.

Source: See appendix note A-1.

Table 5. — Minimum economic incomes of selected proportions of income recipients,

1934-87

Group of income recipients

Year

Highest 2
percent

Highest 1
percent

Highest

ViOf

1 percent

Highest

Mo

of 1 percent

Highest

Moo of

1 percent

1934

$3,775
3,990
4.390
4,640

$5, 375
5,800
6,880
6,940

$8, 105
8,835
10,965
10,825

$22, 210
24, 735
32,090
30. 235

$80, 755

1935

91,685

1936

116 430

1937

106,650

Source: See appendix note A-1.

Table 4 shows the statutory net income levels which separate the
various proportions of income recipients and table 5 shows these levels
for the years 1934 through 1937 in terms of economic income. The
minimum economic income levels of the highest 1 percent of income
recipients are presented in table 1 . The income levels which separated
the various proportions of income recipients vary significantly during
relatively brief periods. Table 5 indicates that in 1934 an individual
with a dollar income of $80,775 would have been included with the
highest one one-hundredth of 1 percent of income recipients, while in
1936 it would have required an income 44 percent larger, $116,430, to
be included within the same group. This shift from 1934 to 1936 in
the minimum income of this grpup is a result of the general increase
in the level of. incomes and the rise in income concentration. Table 6

CONCENTRATION OF ECONOMIC POWER

27=

presents the number of income recipients included within each group of
income recipients each year and table 7 contains the average incomes
in terms of statutory net income of the various proportions of income
recipients. The average economic incomes for the years 192G through
1937 are presented in table 8.

Table 6. — Numbet of individuals in the selected proportions of income recipients

1918-37

Group of income recipients

Year

All income
recipients '

Highest 2
I)ercent

Highest 1
percent

Highest M
of 1 per-
percent

Highest
Mo of 1
percent

Highest
Moo of 1
percent

1918

40,461,000
40, 306. 000
39, 975, 000
40, 768, 000
41,300,000
42, 064, 000
42, 928, 000
43,576,000
44,209,000
44, 852, 000
45,506,000
46,169,000
46, 845, 000
47, 438, 000

49, 260, 000
49,848,000

50, 363, 000
50, 924, 000

809,220
806, 120
799,500
815,360
826,000
841, 280
858, 560
871. 620
884,180
897,040
910,120
923, 380
936,900
948,760

985, 200

996,960

1, 007, 260

1, 018, 480

404, 610
403,060
399, 750
407,680
413,000
420, 640
429,280
435, 760
442,090
448.520
455. 060
461, 690
468.450
474, 380

492,600
498. 480
503. 630
509,240

202, 305
201, 530
199, 875
203,840
206, .^,00
210, 320
214. 640
217, 880
221, 045
224.260
227, 530
230, 845
234, 225
237,190

246,300
249, 240
251,815
254,620

40,461
40,306
,■'9, 975
40,768
41, 300
42, 064
42,928
43, 570
44,209
44,852
45,506
46, 169
46, 845
47,438

49,260
49,848
50,363
50,924

4,046

1919

4,031
3,998
4,077

1920

1921

1922

4,130

1923

4,206

1924

4,293

1925 ...

4,3.58

1926

4,421

1927

4, 485

1928_... -

4,551

1929

4,617

1930

4,685

1931

4,744

1934

4,926

1925

4,985

1936

5,036

1937

5,092

' Excludes employees of State and local governments.
Source: See appendix note A-2.

In interpreting the data on income concentration expressed in terms
of percentages of total income it is essential to take account of the
structure of the distribution underlying the measure of concentration.
It will be readily appreciated that identical degrees of income con-
centration may result from widely different distributions of income.
Thus, one distribution with a relatively small income range, for exam-
ple, up to $50,000, and another distribution with incomes running into
the millions may yield the same measure of income concentration and
even the same minimum income level for a group such as the highest
1 percent; yet the significance attached to the identical measures of
income concentration derived from these two distributions will obvi-
ously be quite different. A further contrast will be provided when
the minimum incomes of the two groups of income recipients are not
the same. The share of income received during recent years by the
highest 1 percent of income recipients will be given a considerably dif-
ferent significance according to whether the minimum income is closer
to $8,000 or $25,000. The preceding tables presenting minimum in-
come levels, number of income recipients, and average incomes, serve
the purpose of disclosing important characteristics of the income dis-
tributions. As noted above, the tables on minimum incomes are par-
ticularly interesting in this connection in that they indicate in fairly
compact fashion the income range and the general distribution within
that range. This aspect of the income structure is generally consid-
ered along with the degree of concentration as fundamental to an eval-

28

CONCENTRATION OF ECONOMIC POWER

uation of the manner in which incomes are distributed. The varying
relationships in different years between the minimum incomes and
average incomes of the selected proportions of income recipients or,
in other words, the changes in the structure of the distribution, reflect
the shifts in income concentration which will now be considered in
some detail.

Table 7. — Average statutory net incomes of selected proportions of income recipients,

1918-37 1

Group of income recipients

Year

Highest
2 percent

Highest
1 percent

Highest
^ofl
percent

Highest
Hoofl
percent

Highest
Moo of 1
percent

1918 - ---

$9,880
11,610
11,345

9,430
10, 775
11,390
12,030
14,840
14, 845
15,520
17,995
17,415
11,785

8,770

6,940

7,270
8,085
10,255
10, 025

$15, 140
17, 965
17,060
14. 080
16, 500
17, 060
18, 880
23, 455

23, 435

24, 670
29,270
28,135
18,040
12,900

10, 235

10, 780
12,205
15, 765
15,190

$23,605
27, 665
?5, 605
21,000
25, 165
25, 985
28,890
36, 670
36, 430
38,840
47,360
45, 440
27,545
19,060

15, 275

16,140
18, 355
24, 085
22, 935

$63, 715
71, 700
61, 390
50,090
63. 610
64,285
71.700
97, 785
98, 485
107, 485
140,090
136, 605
72,900
46,460

37,900

39,645
45, 780
60,360
56, 615

$231, 340

1919

245, 595

1920

190. 095

1921 ---

150,845

1922 --.

218, 400

1923

217, 545

1924

241, 325

1925 - -

376, 770

1926 -

390,635

1927

433, 445

1928

608,440

1929 - ,

623,130

1930 — - - -

287, 510

1931

164, 420

1934 -

131,345

1934

132, 360

1935 --.

153,860

1936

199, 565

1937 •....

186,565

Indexes (1918=100)

1918 ....

100.00
117.51
114.83

95.45
109.06
115.28
121.76
150.20
150.25
157. 09
182. 14
176. 27
119.28

88.77

70.24

73.58
81.83
103. 80
101.47

100.00
118.66
112.68

93.00
108. 98
112.68
124. 70
154. 92
154. 79
162. 95
193. 33
185. 83
119. 15

85.20

67.60

71.20
80.61
104. 13
lOo. 33

100.00
117.20
108. 47

88.96
106. 61
110.08
122. 39
155. 35
154. 33
164. 54
200. 64
192.50
116.69

80.75

64.71

68.38
77.76
102. 03
97.16

100.00
112.53
96.35
78.62
99.85
100. 89
112.53

153. 47

154. 57
168.70
219. 87
214. 40
114.42

72.92

59.48

62.22
71.85
94.73
88.86

100.00

1919

106.16

1920

82.17

1921

65.20

1922 . .

94 41

1923

94.04

104 32

1925

162 86

168 86

1927

187 36

1928

263.00

1929

269 36

1930

124.28

1931

1934

71.07
56 78

1934

57.21

1935...

66 51

1936

86 26

1937

80 65

1 The second sot of averages for 1934 and those for 1935-37 are slightly overstated relative to the averages
for the preceding years. See text and appendix note B-2.

Source: Averages calculated from data in table 6 and appendix table A-1.

CONCENTRATION OF ECONOMIC POWER

29

Table 8. — Average economic incomes of selected proportions of income recipients,

1926-87 1

Year

Group of income recipients

All income

Highest 2

Highest 1

Highest H

Highest Mo

Highest
Moo of 1
percent

recipients '

percent

percent

of 1 percent

of 1 percent

$1, 670

$17, 285

$27, 315

$42,445

$114,820

$454, 650

1,650

17,925

28,590

44, 960

124.855

502, 785

1.725

20, 585

33. 535

54, 175

159, 735

686, 660

1,740

20, 050

32,420

52,285

156, 295

702, 620

1,460

13,865

21, 360

32, 765

87, 650

347,280

1,160

10. 595

15, 705

23,315

58, 415

215, 430

1,015

8,500

12, 700

19,200

49, 535

176, 210

1,030

8,840

13. 235

20,035

51,115

177,225

1,105

9,750

14, 675

22,300

57, 175

197, 595

1,275

11,955

18,420

28, 355

72, 870

249,800

1,365

11, 830

18, 105

27, 630

70, 105

241, 165

Indexes (1926=100)

1926
1927
1928
1929
1930
1931

1934.

1934
1935
1936
1937

1926
1927
1928
1929
1930
1931

1934

1934
1935
1936
1937.

100.00

100.00

100.00

100. 00

100.00

98.80

103. 70

104. 68

105. 93

108. 74

103. 17

119.08

122, 78

127.65

139. 12

104. 25

115.99

118.71

12.3. 19

136. 12

87.43

80.20

78.19

77.19

76. 34

69.42

61.28

57.49

54.93

50.88

60.80

49.16

46.50

45.24

43.14

^1.52

51.14

48.46

47.21

44.52

66.13

56.40

53.73

52. .54

49.80

76.18

69.16

67.45

66.80

63.46

81.74

68.44

66.28

65.10

61.06

100.00
110.59
151.03
154. 54
76.38
47.38

38.76

38.98
43.46
54.94
53.04

' The second set of avoraee= for 1934 and those for 1935-37 are slightly overstated relative to the averages
for the preceding years. See text and appendix note B-2.
2 For these averages employees of State and local governments are included in all income recipients.

Source: Averages of selected proportions of income recipients are calculated from table 6 and appendix
table A-2. Averages for all income recipients are calculated from total individual income as given in table
10 and data in appendix table A-4.

Changes in income concentration. — Tables 2 and 3 and chart II
reveal that the changes in the degree of income concentration among
the various proportions of income recipients follow for the most part
the same general pattern as the highest 1 percent of income recipients.
However, significant differences in both the relative importance of the
changes in the degree of income concentration and in the direction of
year-to-year changes are indicated among the various income groups.
As the vertical scale of chart II is logarithmic, the chart shows not only
the direction of the year-to-year changes but also the relative import-
ance of these changes for the various income groups. The indexes in
tables 2 and 3 are also useful in comparing the various proportions of
income recipients with respect to the relative importance of the
indicated changes in the income shares.

30

CONCENTRATION OF ECONOMIC POWER

CHART I I

Shares of Total Individual Income Received by Selected Proportions
OF Income recipients. 1918-37

(Logarithmic scale)

PERCENT

3 0

2 6

20

/ 5

-

y

^
\

\

fGMES
'6HES

-

-

>\

/

r

X

y

^

>;

T 2 PEA

•*

^

/

y

•N

/

r

X

y

t

i

\1

T / AEffC£Nk

1 \iVl^».i^

/ n

>

V

y

r

V

Nw

k 1 ^./ >x

9

a

7

-

/

\

T! K. I IT;* 1 ^1 -

--

-\

y

/"^

\/

r

^

^

\1

-

V

^

t

r

^'

\

1

-

^

/

\

t

6
5

3

~

J

/

\\

HiST,

—

-

i

r

-•

V

f

^\

S~

—

■>

J

^

V

J

\

V

^

^/oOf

/PI

■/f^E

VT-
\

-

\.

y

r

1

1

-

•
>

'J

\\

2

f

y

\^

\

»EST}

N

A

\

i

r

V

)

1

\

V

7

>

SACt

■NT

\.

J

FCO/voM/c //vco/ne

STATUTORY NET INCOME

'34

f9/8 '/9 '20 7/ 2^ '£3 74 75 76 77 78 79 '30 '31

'34 '35 '36 '37

OD- 40-/70

Source: Tables 2 and 3.

Note— The vertical scale is logarithmic; therefore, equal vertical distances represent equal percentage
changes. The continuous lines from 1918 through 1931 represent the shares that the statutory net incomes
of the selected propcrtions of income constitute of total individual income and the broken lines from
1926 through 1931 and from 1934 to 1937 represent the shares that the economic incomes of the selected
proportions of income recipients constitute of total individual income. See footnotes to Tables 2 and 3.

CONCENTRATION OF ECONOMIC POWER 31

Focusing attention first upon the broad movements, it is readily
seen that the share of the total income received by the three larger
proportions of income recipients was higher in 1919, 1922, and 1924
than in 1918. In contrast, the shares of the two smaller groups did
not exceed the 1918 level until 1925. Thus, it is interesting to note
that the war year of 1918 was characterized by a higher degree of
income concentration among the two highest income groups (those
with statutory net incomes above $26,160) than the immediately
following years. On the other hand, the highest 2 percent group
received a smaller income share in 1918 than in the immediately
following years, with the exception of 1920.

After 1924 the shares of the smaller proportions of income recipients
increased relatively more than those of the larger proportions. The
share of total income received by the highest one one-hundredth of 1
percent in 1928 and 1929 was more than twice as large, as in 1924
whereas the share received by the highest 2 percent of income recipients
was only slightly more than one-third larger than in 1924. The year
1925 witnessed the greatest percentage increases in the shares received
by all proportions of income recipients. The increases varied from 17
percent for the highest 2 percent of income recipients to 47 percent for
the highest one one-hundredth of 1 percent.

The shares of total individual income received by the various
proportions of income recipients declined sharply after 1929. The
smaller the proportion of income recipients, the larger was the rela-
tive decline in its share of total income. After receiving a share
of the total individual income in 1928 and 1929 which was twice that
of 1918 and three times as large as the previous low of 1920, the share
of total income received by the highest one one-hundredth of 1 percent
was reduced in 1934 to 80 percent of the 1918 share. The decreases
in the shares of total income received by the highest one one-hundredth
of 1 percent and the highest one-tenth of 1 percent of income recipients
were especially large in 1931. In 1934 all proportions of income recip-
ients had shares of the total income less than in 1918 but larger than
1920, the year when the degree of income concentration was least.
During 1935 the shares of income received bv the various proportions
of income recipients experienced a fairly uniform increase. In 1936 a
more substantial increase in income concentration took place. The
indicated increases in the income shares varied from 10 percent for the
highest one-tenth of 1 percent of income recipients to 6 percent for the
highest 2 percent of income recipients. During 1937 a decline in in-
come concentration occurred and the various proportions of income re-
cipients received income shares of about the same size as in 1934.
The actual increases in income concentration during 1935 and 1936
and the decrease in income concentration during 1937 were somewhat
more substantial than indicated in table 3 because of tl^e differences in
the extent to which the data for each year overstate the degree of
income concentration.

It is evident from the above description of the broad movements
in income concentration that there are significant differences among
the various proportions of income recipients in the shifts in income
concentration over the period covered. These differences are more
marked between the most inclusive and the least inclusive groups of
income recipients, but the differences become more important as the
groups of income recipients vary in size. Probably the most impor-

32

OONCENTRATION OF ECONOMIC POWER

tant diflference between the various proportions of income recipients
lies in the variabiHty of the shares of income received by them.
The smaller the proportion of income recipients, the more variable
has been the share of income. ^^ As is shown in appendix note B-1,
only a small part of this difference in variability is accounted for
by the use of statutory net income as the income concept. The
following table shows the percentage increases in the shares of total
income from the year of lowest to the year of highest concentration.

Table 9. — Percentage increases in shares of total individual income from year of
lowest to year of highest income concentration

Group of income recipients

Highest 2 percent

Highest 1 percent , ., —

Highest one-half of 1 percent

Highest one-tentli of 1 percent

Highest one one-hundredth of 1 percent

Year of
lowest con-
centration

1920
1920
1920
1920
1920

Year of
highest con-
centration

1928
1928
1928
1928
1929

Percentage
increases

54.7
67.3
80.4
122.4
217.0

Source: Percentages computed from data in table 2, p. 22.

The variability of income shares received by the highest 2 percent
of income recipients is due in large part to the fact that this group
inclttdes the smaller proportions of income recipients whose income
shares fluctuate markedly. Thus, if we are interested in the changes
in the income shares received by the "upper middle class," it would be
necessary to calculate the share of income received each year by those
falling between the lower limit of the highest 2 percent and the lower
limit of the highest one-half of 1 percent. This group has been
located between the statutory net incomes of $3,275 and $6,840 in a
low-income year such as 1934 and $5,390 and $14,690 in the high-
income year, 1928. In contrast to an increase of over one-half in
the income share of the highest 2 percent from 1920 to 1928 (table 9),
the share of the "upper middle class" rose by but slightly more than
one-fifth during this period (calculated from data in table 3). The
decline from 1928 to 1934 was also much less, being slightly over one-
ninth as compared with a decline of one-thrrd in the share received
by the highest 2 percent of income recipients. It is readily evident
that if the income groups are compared in this fashion — that is, exclu-
sive of the groups above them — the contrast in the relative variabihty
of the income shares received would be much greater than shown
in preceding income concentration tables.

Another difference between the various proportions of income
recipients lies in the direction of the year-to-year changes in the degree
of income concentration. These differences in direction are not as
characteristic as the similarity in the year-to-year movements, but
divergent shifts in income concentration have taken place. In 1919
the three largest proportions of income recipients received larger

2' The relative mean deviations of the income shares are-
Highest 2 percent of income recipients __ 0. 11

Highest 1 percent of income recipients I. ."Ill" I I'l'l'll'lll .13

Highest one-half of I percent of income recipientsfJ.'''"'!""'"""!"" II lllVllllllllll .IS
Highest one-tenth of 1 percent of income recipif^nts - . 20
Highest one one-hundredth of 1 percent of income recipients :.. 31

Source: Computed from data in table 2, p. 22.

CONCENTRATION OF ECONOMIC POWER 32

shares of the total mcome than in 1918, whereas tlie highest one-tenth
of 1 percent and the highest one one-hundredth of 1 percent of income
recipients received smaller shares than in the previous year. The
share of the highest one one-hundredth of 1 percent was 7 percent
less while the share of the highest 1 percent was about 4 percent
larger. The changes in income concentration during 1921 departed
from the general pattern of increased variability for the less inclusive
proportions of income recipients, the smaller groups not experiencing
increases in their shares of total income as great as the larger pro-
portions of income recipients. In 1926 the shares of all proportions
except the highest one one-hundredth of 1 percent declined. The
latter group received a share which was slightly larger than the
previous year. In 1929 the income shares of all proportions except
the smallest declined. While the shares of total income received by
the four largest proportions of income recipients declined twice during
the period 1924 through 1929, the share received by the highest
one one-hundredth of 1 percent of income recipients increased in each
of these years. After 1929 the income shares of all groups dropped
sharply, the groups of income recipients with the largest incomes
experiencing the most severe reduction in their shares. The decrease
in the income shares was larger for 1930 than for 1931.

The change in the income shares of the various proportions of
income recipients from 1930 to 1931 is of special interest. The share
of total income received by the highest 2 percent of income recipients
was almost the same in the 2 years (18.99 percent in 1930 and 18.51
in 1931, see table 3) yet the shares received by the four smaller
proportions of income recipients declined considerably.^^ The de-
creases in the income shares of other proportions of income recipients
were, beginning with the smallest group: 21, 15, 9 and 6 percent.
The shift in income concentration in 1931 serves to illustrate the
danger of assuming that changes in the degree of income "inequality""
among the higher-income recipients are always indicative of changes in
the entire distribution. In this instance the highest 2 percent of
income recipients received shares of approximately the same size in
2 successive years yet the distribution of income within this group was
considerably different in each of the 2 years.

22 The actual decline? were somewhat less than shown as the shares of the three smaller proportions of
income reoipients are understated more in 1931 than in 1930. The shares of the highest 1 and 2 percent of
income recipients are probably correct for each year as given. See appendix note A-1, p. 71-2.

256149— 40— No. 4-

CHAPTER III
THE COMPOSITION OF INCOME: 1918-37

I. INTRODUCTORY

In the preceding chapter, attention was directed to the shares of the
Nation's total income received by selected proportions of income
recipients. In order to increase our understanding of the process of
income distribution, we may probe beneath this distribution of the
total individual income by shares and inquire as to how the various
types of income such as salaries, dividends, and interest were combined
to produce these income shares. This approach from the viewpoint of
the sources of income immediately suggests a series of interesting
questions. What has been the difference between the composition of
the incomes of all income recipients and of the proportions whose
shares have been measured? Has this difference been stable or has it
been subject to variation as business conditions change? Has there
been a shift over the past two decades relative to the importance of the
various types of income received by the higher income groups? How
have the increases and decreases in the various income streams been
reflected in the income composition of the different groups of income
recipients? What has been the degree of concentration among
individuals of each type of income and how has this concentration
varied from year to year?

Turning from a descriptive to an analytical approach, the problem
arises of relating these questions to the degree of income concentration
and the fluctuations therein. With this shift in interest the questions
may be rephrased as follows: What has been the effect of changes in
the magnitude of the various income streams on the degree of income
concentration? For example, what changes in income concentration
have taken place during the years when an income source which is
highly concentrated or, in other words, unequally distributed, increased
more than total income? Which types of income have contributed
most to the concentration of income and to the changes in income
concentration? When the answers to these questions have been
discovered, the first steps will have been taken toward understanding
the factors responsible for the shifts in income concentration outlined
in the preceding chapter.

The questions on the composition of personal incomes have been
framed largely with reference to one aspect of these statistics; namely,
their relation to the subject of income concentration. The statistical
information presented in this chapter is interesting on its. own account
and in connection with broader questions of business cycle analysis
dealing with the flow of income through the economic system. Stu-
dents of income taxation may be interested in the information in
connection with variable rates of taxation for different types of income.
Some types of income are regarded as being socially more beneficial
than others and the taxation of certain sources has less of a depressing
■effect on the volume of business activity. Also, as will be indicated,
some types of income are associated with an increase in income con-
centration. Income taxation, as well as inheritance taxation, may
take- account of these considerations. The information presented on

35

3g CONCENTRATION OF ECONOMIC POWER

the composition of income at various income levels, the concentration
of types of income, and the changes in the composition of income from
periods of business prosperity to periods of business depression is
obviously important in this connection. In addition, it may be noted
that the data on the distribution of the Nation's total income by type
of income take on added significance when it is known how the types
of income are distributed with respect to the size of individual incomes.

With the purposes of our inquiry set by these questions, we may now
turn to a consideration of the information on the composition of
incomes. Tlie annual issues of the Statistics of Income provide data
on the aggregate amount of each type of income received by all the
individuals in a given income class. On the basis of these statistics,
with necessary adjustments for completeness and comparability, it is
possible to determine the composition of the incomes of those indi-
viduals included within the various groups of income recipients for the
years 1918-31 and 1934-36. It is important to note that when the
income source data are presented in terms of the aggregate amounts
for each income group, there is no direct evidence as to how the
various distributive shares were combined to yield the incomes for
specific individuals; that is, how many individuals in the various
income classes received their incomes from one, two, or all the sources
and how important were these various sources? ^ An examination of
the tables on the composition of the incomes of those filing income-tax
returns available since 1934 and 1935 indicates that for the higher
incomes the total income of specific individuals is to a large extent
derived from several sources.^ The data as given in subsequent tables
of this study show the aggregate amount of income from each source
received by the individuals included within the highest 1 percent of
income recipients. Some data on the proportion of the individuals in
the highest 1 percent of income recipients receiving each type of income
are also presented.

In the interest of emphasizing the significant facts concerning income
composition with relation to income concentration, data on the com-
position of incomes of only the highest 1 percent of income recipients
will be presented. Changes in the income shares of the various

' It may be well to indicate briefly the extent of our knowledge of the composition of the higher incomes
by outlining the manner in which the basic statistics are tabulated. In addition to showing the aggregate
amount of each type of income received by all individuals in a given income class, tables were presented in
the Statistics of Income beginning in 1927 showing the frequency distribution of each of the principal sources
of Income; that is, the number of individuals receiving a specified amount of income from a jiiven source.
In 1935 and 1936, more detail was published; the size of the specific source was cross classified with the net
Income so that it is possible to determine, for example, how many individuals with a net income of $40,000
to $50,000 received an income from dividends of $20,000 to $25,000. Beginnine in 1934, the Statistics of In-
come show the number of individuals in a given net income class reporting each type of income. Thus it is
possible to determine the number of individuals with a net income of .$5,000 to $10,000 who reported an
income from salaries and wages or the number receiving dividends. However, these statistics do not tell
us how many individuals reported an income from one or several sources, nor do we know the relative im-
portance of these sources for individual income recipients. This type of data is available for Wisconsin.
See Wisconsin Individual Income Tax Statistics, 193fi. vols. IVA and IVB, also vols. I for 1929 and 1935.
Special tabulations of the HOr. Federal income-tax returns, now being completed by the Treasury Depart-
ment, will also provide tnis type of information.

2 For example, the text table first published in the Statistics of Income for 1935 (pp. 13-23) which cross
classified the size of net income by the size of specific sources. Statistics of Income table 7, which first con-
tained the number of individuals by income classes reporting each source of income in 1934, is also useful in
this connection.

CONCENTRATION OF ECONOMIC POWER

37

groups of income recipients exhibited the same general pattern and the
differences between the composition of all incomes and the incomes of
the highest 1 percent and the smaller proportions of income recipients
are largely a matter of degree. The relative importance of the income
sources for the various groups of income recipients may be determined
with the use of the data which will be presented for 4 years showing the
composition of income by income classes.

II. COMPOSITION OF

income: all income
highest 1 percent

recipients and THE

Tables 10 through 13 show the composition of income in both
aggregate and percentage form for all income recipients and for the
highest 1 percent. There are minor differences in the classification
of income sources for all income recipients and for the highest 1 per-
cent which are noted in the respective appendix notes. When the
differences are important they are discussed in the text. In the main
the classifications are similar.

Table 10. — Composiiion of total individual income, 191R-S7 ^
[Millions of dollars]

Year

Com- Entre-
pensa- | prc-
tion of I ncurial
cmpoy- net

ees 2 income

1918-
1919.
1920.
1921.
1922.
1923.
1924.
1925-
1926.
1927.
1928
1929.
1930.
1931.

1934.

1934.
1935.
1930-

34, 628, 1 8, 841

34.628 I 8,841

37, 417 I 10, 103

42.911 I 11,678

1937 1 I 47,568 , 12,550

445
823
565
545
753
986
937
055
796
933
400
238
693
991

17, 266
19, 210
14,807
10. 307
11,519

12, 957
13,305
14, 076
13,661

13, 042
13, 530
13, 527
10, 404

7,436

Income

Real-

pri-

ized

marily

capital

from

pains

personal

and

service

losses

50, 711

220

57, 033

880

59, 372

870

45, 852

210

49, 272

870

56, 943

980

57.242

1. 330

60, 131

2,600

62,457

2,210

01,975

2,610

! 63,930

4,020

1 66,765

2,890

! 59,097

-1,410

1 48,427

-2, 770

43, 469

-620

43, 469

10

47. 520

390

54, 589

920

60, 118

170

Net
rents
and
royal-
ties

2,258
2,783
2,950
3,052
3,504
3,641
3,809
.3,811
3,589
3, 357
3,458
3,419
2, 763
2,083

1,690

Divi-
dends

3,452
2,895
3,215
2,932
3,006
3,823

3, 763

4, 362
4,736
5,036
5,362
5,978
5,801
4,335

2,775

1,690 2,775

1,917 I 3,038

2,275 1 4,807

2,525 5,424

Income

Divi-

pri-

Interest

dends
and in-

marily
from

terest 3

prop-
erty

2,504

5,930

8,408

3,012

5, 907

9,570

3,377

6,592

10,412

3,511

6,443

9,705

3,640

6,646

11,020

3.884

7,707

12, 328

4,115

7,878

13,017

4, 375

8,737

15, 238

4,541

9,277

15, 076

4,816

9,852

15, 819

5,123

10, 485

18, 563

5,356

11,518

17, 827

5, 575

11,600

12, 953

5,522

10, 158

9,471

5,109

7,993

9,063

5,109

7,993

9,693

4,927

8,005

10, 312

4,847

9,631

12, 826

4,878

10, 225

12, 920

Total
income

59, 119
66. 603
69, 784
55, 557
60,292

69, 271

70, 259
75, 369
77,533
77, 794
82, 493
84,592
72. 050
57, 898

52,532

53,162
57,832
67.415
73,038

1 Because of a change in the definition of realized capital gains and losses the table is divided into 2 parts:
1918-34 and 1Q34-37. In the latter period varying proportions of gains and losses are included depending on
the length of time the asset was held and lossesof an individual are limited to $2,000 in excess of gains. The
estimates of realized capital gains and losses are understated for each year. See appendix note A-4.

2 Includes work-relief wages but excludes direct-relief payments and adjusted-service certificate payments
to veterans (veterans' bonus).

' Includes also net balance of international flow of property incomes.

* The 1929-37 data are based on estimates appearing in the June 1939 issue of the Survey if Current
Business. After the present report was completed, estimates including 1939 were published in the Survey
of June 1540. The revisions for earlier years were of a minor character except for the preliiniiary 1937
dividend estimate which was revised downward in accordance with data not previously available from
corporate income-tax returns. This change resulted in an absolute decline from 1936 to 1937 of thn '3-tenths
of 1 percent in the percentage share of dividends rather than the increase of the same size shown in ible 11.
The concentration measures are but slightly affected by the revision, there being a decline from 1936 to
1937 of 8 percent in the income share of the highest 1 percent group instead of 8.5 percent as calculated from
table 1.

Source: Based, with exception of estimates of realized capital gains and losses, on income estirnates of the
Department of Commerce, for 1929-37, extrapolated on basis of estimates of Simon Kuznets in National
Jncome and ranital Fi/rination, for 1919-29, and estimates of Willford I. King in National Income and Its
Purchasing Power, for 1918. See appendix note A-1 for detailed explanation of methods.

38

CONCENTRATION OF ECONOMIC POWER

Table 11. — Perceniagfi distribution of total individual income, by type of receipt,.

1918-37 '

Year

"1

Total
income

;

fcom-
•pensa-
t ion of
employ-
ees'

Entre-
pre-
neurial
net in-
come

Income
prima-
rily
from
person-
al serv-
ice

Rea-
lized
capital
gains
and
losses

Net
rents
and
royal-
ties

Divi-
dends

Interest

Divi-
dends
and
inter-
est'

Income
prima-
rily
from-
prop-
erty

1918

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

100.0
100.0
100.0
100.0

66.6
56.8
63.9
64.0
62.6
63.5
62.6
61.1
63.0
62.9
61.1
62.9-
67.6
70.8

65.9

65.2
64.7
63.6
65.1

29.2

28.8
21.2
18.6
19.1
18.7
18.9
18.7
17.6
16.8
16.4
16.0
14.4
12.8

16.8

16.6
17.5
17.3
17.2

85.8
85.6
85.1
82.5
81.7
82.2
81.5
79.8
80.6
79.7
77.5
78.9
82.0
83.6

82.7

81.8
82.2
80.9
82.3

0.4
1.3

1.2

.4

1.6

1.4

1.9

3.5

2.8

3.3

5.6

3.4

-1.9

-4.7

-1.1

(<)
.7
1.4
.2

3.8
4.2
4.2
5.5
5.8
5.3
5.4
5.1
4.6
4.3
4.2
4.1
3.8
3.6

3.2

3.2
3.3
3.4
3.5

5.8
4.4
4.6
5.3
6.0
5.5
5.4
5.8
6.1
6.6
6.5
7.1
8.1
7.5

5.3

5.3
6.3

7.1
7.4

4.2

4.6
4.8
6.3
6.0
5.6
6.9
5.8
6.9
6.2
6.2
6.4
7.8
9.6

9.8

9.6
8.5
7.2
6.7

10.0
8.9
9.6
11.6
11.0
11.1
11.2
11.6
12.0
12.7
12.7
13.6
16.1
17.5

15.2

16.0
13.8
14.3
14.0

14.2

1919

14.4

1920. ---

14.9

1921

17.5

1922

18.3

1923

17.8

1924 .-.

18.5

1925

20.2

1926

19.4

1927

20.3

1928

22.5

1929....

21.1

1930 . .

18.0

1931

16.4

1934

17.3

1934

18. ?

1935

17.8

1936...

19.1

1937 «_^

17.7

' Because of a change in the definition of realized capital gains and losses the table is divided into 2 parts.
1918-34 and 1934-37. In the latter period varying proportionsof gains and losses are included depending on
the length of time the asset was held and losses of an individual are limited to $2,000 in excess of gains. The
estimates of realized capital gains and losses are understated for each year. See appendix note A-4.

' Includes work-relief wages but excludes direct-relief payments and adjusted-service certificate pay-
mentsto veterans (veterans' bcnus).

' Includesalso net balanceofinternationalflow of property incomes.

< Less than .1

» See footnote 4 of table 10.

Source: Percentagesderivedfrom data in table 10.

As was to be expected, the relative importance of the income
sources of the highest 1 percent of income recipients differs in many
important respects from the composition of the income of all indi-
viduals. The year 1922 may be taken as illustrative since in that
year the income share of the highest 1 percent of income recipients
was 14.24 percent, onlv slightly below the average of 14.54 for the
period 1918-34 and 1934-37. In 1922 the highest 1 percent of
income recipients, received 29 percent of their income from salaries,
wages, fees, and pensions, whereas 63 percent of the income of all
individuals was derived from this source. The relative importance
of this source for the highest 1 percent of income recipients has
varied from a minimum of 20 percent in 1928 to a maximum of 40
percent in 1934. For the total income of all individuals the variations
were not so large, the minimum being 57 percent in 1918 and the
maximum, 71 percent in 1931. Entrepreneurial net income, that is,.
the net profit of independent businessmen, constituted about one-fifth
of the income received by all recipients and by the highest 1 percent.
There has been a sustained decline in the share of entrepreneurial
net income in the incomes of all income recipients and of the highest
1 percent. From a maximum of 29 percent of total individual income
in 1918 the contribution of this source fell to a miminum of 13 percent
in 1931. In recent years this income category constituted about 17
percent of total income. Paralleling this movement, entrepreneurial
net income dropped from a high of 27 percent of the income of the

CONCENTRATION OF ECONOMIC POWER

39

highest 1 percent in 1919 to a low of 14 percent in 1930' and 1931 and
during recent years it rose to contribute about 16 percent to the
income of this group. These percentages probably understate som.e-
what the importance of entrepreneurial net income for the highest
1 percent owing to the underreporting of this type of income to the
income-tax authorities. Except, perhaps, for realized capital gains-
this source is probably subject to more underreporting than the
other income sources. A small part of the understatement of entre-
preneurial net income is due to the exclusion from taxable income of
the value of farm products raised and consumed on the farm.

Realized capital gains were largely received by those with higher
incomes and hence constituted a more important income source for
the highest 1 percent than for all income recipients. In 1922, of the
total income of the highest 1 percent of recipients, 7.6 percent was
derived from this source as compared with 1.4 percent for all income
recipients. Both of these percentages understate somewhat the
proportion of income derived from realized capital gains as this type
of income is probably subject to more underreporting to the income-
tax authorities than are the other types with the possible exception
of entrepreneurial net income.^ This source of income was the most
volatile of all. In some years it was a negative item and in other
years very large. In 1928, realized gains rose to constitute 27 percent
of the income of the highest 1 percent of all income recipients.

Table 12. — Composition of incovies of the highest 1 percent of income recipients,

1918-36 i

[Millions of dollars]

Year

1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931

1934

1934
1935
1936.

Com-
pen-
sation
of em-
ployees

1,963
2,311
2,548
2,461
2,373
2,539
2,647
2,792
2,946
3.020
3,098
3,167
3,141
2,908

2.540

2,540
2,707
2,923

Entre-
preneur'
ial net
income

1,683
2,314
1,874
1,345
1,578
1,586
1,780
2,137
2,072
2,064
2,245
2,185
1,414
1,061

1,043

1,043
1,171

Income
primarily

from
personal

service

3,646
4,625
4,422
3,806
3,951
4, 125
4,427
4,929
5,018
5,084
5,343
5,352
4,555
3,969

3,583

3,583
3,878

Realized

capital

gains

and

losses

149

507

391

182

621

627

958

2,274

1,826

2,250

4,151

3,556

85

-615

-204

66
331
711

Net
rents
and
royal-
ties

355
354
352
377
412
414
445
495
492
451
426
437
374
277

180

180
198
235

Divi-
dends

1,996
2,005
2,201
T,847
2,066
2,329
2,495
2,842
3,297
3, 472
3,671
3,869
3, 468
2,511

1,609

1,609
1.814

Interest
and mis-
cella-
neous
prop-
erty

income

1,097
1,133
1,059
999
1,149
1,196
1,228
1,324
1,443
1,566
1,669
1,755
1,522
1,306

1,088

1,088
1,097

Income
primarily
from
prop-
erty

3,597
3,999
4,003
3,405
4,248
4,566
5,126
6.935
7,058
7,739
9,917
9,617
5,449
3,479

2,673

2,940
3,440

Total
income-

7, 243:

8,624:

8, 425.

7,211

8,199-

8,691

9,553.

11,864

12,07&

12, 823^

15, 260

14, 969'

10, 004

7,448

6, 256

6,52&
7,318
9.290-

' Definition of realized capital gains and losses was changed beginning in 1934 with result that gains for
1934-36 are overstated. (See appendi.x note B-2, pp. 101-2). Estimate of loss of 204 millions for 1934, how-
ever, is based on same definition that prevailed during 1918-31. The other sources may be compared from
year to year without appreciable error. For 1936, only sources for which data are presented are comparable-
to those for the preceding years. Data for 1937 were not available. See text and appendix note A-5 for
definitions and year-to-year comparability of individual sources.

Source: Adjusted data from the Statistics of Income for the respective years. See appendix note A-6.

3 Unlike the estimate."; for the other sources of total individual income, those for total realized capital-
gains are based on data from individual income-tax returns. See appendix note A-4, pp. 84-9.

40 OONCENTRATION OF ECONOMIC POWER

Table 13. — Percentage distribution of incomes of the highest 1 percent of income
recipients, by type of receipt, 1918-36 ^

Year

Total
income

Com-
pensa-
tion of

em-
ployees

Entre-
preneu-
rial net
income

Income

pri-
marily

from
p ersonal
service

•Real-
ized

capital

gains
and

losses

Net
rents
and
royal-
ties

Divi-
dends

Interest

and
miscel-
laneous
prop-
erty
income

Income

pri-
marily
from
prop-
erty

1918. -,

1919 -.

100.0
100.0
100.0
100.0
100.0

100. n

100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0

100.0

100.0
100.0
100.0

27.1
26.8
30.2
34.1
28.9
29.2
27.7
23.6
2' A
23.6
20.3
21.2
31.4
39.1

40.6

38.9
37.0
31.5

23.2

26.8
22.3
18.7
10.3
18.3
18.6
18.0
17.2
16.1
14.7
14.6
14.1
14.2

16.7

16.0
16.0

50.3
53.6
52.5
52.8
48.2
47.5
46.3
41. .5
41.6
39.7
35.0
35.8
45. 5
53.3

57.3

54.9
53.0

2.1

5.9

4.6

2.5

7.6

7.2

10.0

19.2

1.5.1

17.5

27.2

23.8

.9

—8.2

-3.3

1.0
4.5

7.7

4.9
4.1
4.2
5.2
5.0
4.8
4.7
4.2
4.1
3.5
2.8
2.9
3.7
3.7

2.9

2.8
2.7
2.5

27.6
23.3
26.1
25.6
25.2
26.8
26.1
23.9
27.3
27.1
24.1
25.8
34.7
33.7

25.7

24.6
24.8

15.1
13.1
12.6
1.3.9
14.0
13.7
12.9
11.2
U.9
12.2
10.9
11.7
15.2
17.5

17.4

16.7
15.0

49.7
46.4

1920- --

47.5

1921..

1922

47.2
51.8

1923

52.5

1924 --

53.7

1925

58.5

1926..

58.4

^927 -.--

60.3

1928

65.0

1929

64.2

1930..

1931

54.5
46.7

1934

42.7

1934

45.1

'935 -

47.0

1936

' Percentages for 1918-31 and first set for 1934 may be compared from year to year without appreciable
«rror. Definition of realized capital gains and losses was changed beginning in 1934 with result that in
second set of percentages for 1934 and in sets for 1935 and 1936 the proportion"; of income derived from realized
capital gains are overstated and the proportions derived from the other sources are slightly understated. See
text and appendix note A-5 for definitions and year-to-year comparability of individual sources.

Source: Percentages calculated from data in table 12.

Net rents and royalties were about an equally important source of
income for both groups amounting in 1922 to about 5 percent of total
income. For the entire period under consideration, this source was
a slightly larger proportion of income for all income recipients than
for the higher-income group. The rise in the relative importance of
rent as an income source during the early twenties and the decline
after 1929 reflect the trends of the level of rents.

Similar to realized capital gains the receipt of dividends is highly
concentrated among individuals with high incomes. While dividends
constituted but 6 percent of the total income of individuals in 1922,
this source accounted for 25 percent of the income of the highest 1
percent of income recipients. The shifts in the importance of divi-
dends in the total income of individuals and in the income of the
highest 1 percent were generally in the same direction. However,
during the years of large realized capital gains, 1927 through 1929,
the importance of dividends declined for the highest 1 percent of in-
come recipients in contrast to the increase of dividends as a share in
the total income of all individuals.

In 1922 interest constituted 14 percent of the total income of the
highest 1 percent of income recipients and only 6 percent of the total
income of all individuals. As indicated by the heading in table 13,
the interest source for the highest 1 percent of income recipients in-
cludes some miscellaneous property income, such as part of the in-
come received through fiduciaries, a small amount of dividends, and
income not classified elsewhere.* The inclusion of these items offsets

< See appendix note A-5, p. 95. Dividends received through fiduciaries were classified with the divi-
dend source.

CK)NCENTRATION OF ECONOMIC POWER . 41

to some degree the understatement wliich characterizes the interest
source proper. This understatement is due to the indirect manner
in which a considerable portion of interest payments are received by
or accrue to individuals through life-insurance companies, savings
banks, building and loan associations, and other institutions for col-
lective saving and investment. A large part of tliis interest is prob-
ably unrecorded in the income-tax statistics. Another source of
understatement of interest in the incomes of the highest 1 percent of
income recipients is the fact that tax-exempt interest received by
these individuals is probably not fully included in the statistics col-
lected in connection with the Federal income-tax iaw.^ The impor-
tance of interest for the total income of all individuals in-
creased quite uniformly each year from 1920 through 1934. In con-
trast, interest payments as a share of the income of the highest 1
percent slowly declined in importance from 1923 until 1929. The
contribution of interest to the incomes of all' income recipients and
of the highest 1 percent rose sharply dm'ing the depression begimiing
in 1929 owing to the relative stability of this as compared with the
other sources of income.

III. SHIFTS IN THE COMPOSITION OF INCOME OF THE HIGHEST 1 PERCENT
OF INCOME RECIPIENTS

Chart III shows in graphic form the percentage composition of the
income of the highest 1 percent of income receivers during the years
1918-36. The general impression conveyed by the chart is that the
relative importance of the income sources from which those included
each year in the highest 1 percent derived their incomes has been
subject to considerable variation over the past two decades. If the
various sources are broadly classified into two groups, income pri-
marily from personal service and income primarily from property,
the data indicate that generally during years of depressed business
conditions, income primarily from personal service was responsible
for the major portion of the incomes of the highest 1 percent. In
1934 these sources accounted for 57 percent of their incomes. Con-
versely, during years of business prosperity income primarily from
property provided the bulk of the incomes of this group. In 1928
the percentage of income from this source rose to 65 percent. In
9 out of the 16 years covered by the statistics, income primarily
from property represented a larger share of the income of the highest
1 percent than the sources classified as income from personal service.
However, if the 4 years for which data are omitted, 1932, 1933, 1936^
and 1937, are taken into account, income primarily from personal
service undoubtedly was the largest of these two main sources in 11
years during this 20-year period. According to the more detailed
classification of income sources, the compensation of employees was
the largest single income component in 8 years, dividends in 6, net
realized capital gains in 1, and entrepreneurial net income in 1. Dur-
ing the 4 years for which data are omitted or not complete, employee
compensation doubtless constituted the largest income source.

Confining attention to the broad movements, it is readily seen that
changes in the relative importance of the various sources were influ-
enced to a marked degree by the fluctuations of the highly variable

• See appendix note A-5, pp. 95 and 96.

42

CONCENTRATION OF ECONOMIC POWER

Chart III

•Composition of incomes of the highest one percent of income recipi-

ENTS. 1918-36

PERCENT OF INCOME

60
50

4^0

30

>

r'^^

^

/A/COME PRIMARILY FROM
^^ PERSONAL SERVICE*

r

y

"v

*-,

^

■^

>

v^

Ma_—

4

/

\

\^

J

50

40

30

20

J

A

C

OMR
£M

ENS
PLC

ATIC
tYEt

->N C

-s

-)F

//

V

>.»

N

\

*T

^
N^>

D

fVIQ

ENL

^7
OS

=V

^

•""

1

''J

f

-

-

20

/ 0

0

»»

«<».„

w^

^""

<

NTL

■RES

T

>>*•.

.>

X^

^""

"~

«,

'•h

-

^^M

^

i«

A/£T RENTS /

WD ROYALTIES-^

\ 1 1

~"

■—

^30
*20

^/ 0

0

-/ 0

y

^\

N

em

7?£/
VET

^RE/VEO
INCOM

>

r

■\

^^

•»•"

J

r

^

J^

•»«,

\

f"

-

,*

—

—

y

/^

^

A

^

'EAL
A/N

IZEl
S Ai

VL> L

P/Ta
OSS

^L
"S

I

y

y

y

\

\.

y

*

'9/8 '19 'ZO '2/ '22 23 24 25 26 27 28 29 '30 '31

'34

•34 '35 '36

OD--W-I74

■Source: Table 13.

Note.— Percentages for 1918-31 and first set for 1934 may be compared from year to year without appre-
ciable error. The second set for 1934 and the sets for 1935 and 1936 are slightly understated with exception
of percentages for realized capital gains which are overstated. See footnote to Table 13.

• Total of employee compensation and entrepreneurial net income.

CONCENTRATION OF ECONOMIC POWER 43

income, realized capital gains and losses. The fluctuations in the
importance of this source were closely related to general business
conditions, rising in years of prosperity and declining in years of
depression. Reflecting the violent fluctuations in business conditions
that characterized this period, the share of net profits from the sale
of property varied from a negative contribution to the income of the
highest 1 percent of income recipients in 1931 to a larger share than
any other single source in 1928.

Year-to-year shifts in importance of the compensation of employees
as a source of income for the highest 1 percent of income recipients
were inversely relaited to changes in profits from the sale of property.
In every year when the importance of such profits increased, employ-
ees' compensation declined and, conversely, when realized capital
gains declined, employees' compensation rose. In 1931 and 1934
wages and salaries accounted for as much as 40 percent of the income
of the highest 1 percent. In years of high business prosperity it
contributed but slightly more than 20 percent to the income of this
group. During recent years the share of employees' compensation
in the income of the highest 1 percent has been considerably larger
than in previous years of comparable income concentration. The
increased importance of this source has been obtained largely at the
expense of entrepreneurial net income which has experienced a marked
decline.

Entrepreneurial net income, after rising in importance in 1919 to
constitute the largest income source, declined sharply in 1920 and
1921. Except for sinall increases in 1922 and 1924 this source experi-
enced a slow decline until 1928. A small part of this decline in entre-
preneurial net income is probably due to the change in classification
which excluded from this source, for the years 1922 through 1931,
a portion of the realized capital gains received through partnerships
from the sale of assets held over 2 years. ^ Entrepreneurial net income
contributed about the same proportion to the incomes of the highest
1 percent from 1929 through 1931. When the slight difference in
comparability of this source beginning in 1934 is taken into account,
the share of the income of this higher income group represented by
entrepreneurial net income was approximately the same for the years
1934 and 1935 as for the years 1929 through 1931.^ Over the period
from 1918 through 1935, a significant decline in the importance of
entrepreneurial net income has taken place both for the highest 1
percent and all income recipients. These declines do not appear to
be associated with the cyclical movements in business activity. The
reduction in the share of total income represented by entrepreneurial
income is apparently due to the declining importance of agriculture
where the noncorporate form predominates. To some extent, the
decline in the importance of entrepreneurial income for the income of
the highest 1 percent may be also attributable to the tendency on the
part of businessmen to shift to the corporate form of economic
organization or to the supplanting of noncorporate business by con-
cerns organized under the corporate form.^ Entrepreneurial net

« See appendix note A-5. pp. 90-1.

' Idem.

' A thorough analysis of the reasons for these declines in entrepreneurial net income would carry us far
jafleld. The above comments indicate some of the more probable and general causes. These and other
tactors deserve further investigation. Improved data on entrepreneurial net income would doubtless be
■of much value for this purpose.

44 CONCENTRATION OF ECONOMIC POWER

income is a mixed income category composed of wages and salaries^
rent on land, interest on invested capital, and entrepreneurial profits.
The shift toward incorporating businesses, therefore, has tended to
resolve this source into its constituent elements. Inasmuch as the
bulk of entrepreneurial net income may be considered as a return for
personal services rather than a return on capital, this source has been
combined with wages and salaries in the third column of tables 10
through 13 to yield the income derived primarily from personal service.
During recent years these two sources contributed a slightly larger
proportion of the income of the highest 1 percent of income recipients
than during earlier years characterized by a comparable degree of
income concentration or business activity.

For all years, net rents and royalties constituted a minor share of
the incomes of the highest 1 percent of income recipients. Generally
over the period covered by the statistics, the importance of net rents
and royalties declined. After rising slightly in 1921 to 5.2 percent of
income as a combined result of the rapid rise in rents during this year
and the decline in the other types of property income, there was a
steady decline until 1929 when this source accounted for but 2.9
percent of the incomes of the highest 1 percent of income recipients.
In 1930 and 1931 thissource of income became slightly more important
for the highest 1 percent of income recipients. However, it declined
again during the years from 1934 through 1936, largely as a conse-
quence of the greater increases in the other types of property income.

Dividends, as a share of the income of the highest 1 percent, ex-
hibited a slight upward trend from 1919 through 1927. Except for
profits from the sale of property, dividends were the only source that
increased in importance during these years. Despite this general
increase along with realized capital gains, the year-to-year variations
in the percentages of income from dividends were inversely related tO'
the shifts in percentages of income from realized capital gains. Divi-
dends rose when realized capital gains declined and dividends declined
when realized capital gains rose. After 1929 when profits from the
sale of property declined sharply in relative importance, dividends
became a more hnportant source for the higher incomes.

The total amount of interest income received by individuals is
fairly stable from year to year. Largely because of this stability,
this source declined m importance for the highest 1 percent during
years of business prosperity when other sources of income increased,
and rose in importance during periods of depression when the amounts
of other income were reduced. In prosperous years interest and mis-
cellaneous property income accounted for slightly more than 10 per-
cent of the income of the highest 1 percent whereas in years of business
depression this source was responsible for more than 17 percent of the
income of this group. As noted above these percentages represent
somewhat of an understatement of the interest received directly and
indirectly, by this group.

IV. CONCENTRATION OF TYPES OF INCOME

Table 14 indicates the extent to which the different sources of in-
corne were concentrated among the highest 1 percent of income
recipients. The degree of concentration varies from a small pro-
portion of employees' compensation to a major proportion of dividends

CONCENTRATION OF ECONOMIC POWER

45

and realized capital gains. It should be noted that the measures of
concentration in table 14 do not show the extent of concentration of
each type of income among those actually receiving such income.
The data show the concentration among those included within the
highest 1 percent of income recipients and not all of these received
each type of income. For example, in 1927 table 14 shows that 69
percent of all dividends paid to individuals were received by those
individuals falling within the highest 1 percent. If we limit ourselves
to those receiving dividends in 1927 and classify these dividend re-
cipients according to the size of dividend income, a different result
will be obtained. A table in the Statistics of Income for 1927 (p. 10)
makes possible such a calculation. Using this table, a number of
persons equal to but one-half of 1 percent of all income recipients re-
ceived about 71 percent of all dividends. Similar calculations could
be carried out for some of the other sources for the years 1927-1936.^

Table 14. — Percentages of each type of income received by the highest 1 percent of

income recipients, 1918-36

Year

Com-
pensa-
tion of

em-
ployees

Entre-
pre-
neurial

net
income

Income

pri-
marily

from
personal
service

Realized

capital

gains

and

losses '

Net

rents

and

royalties

Divi-
dends

Interest'

Divi-
dends
and
interest

Income

pri-
marily
from
prop-
erty

1918

5.9
6.1
5.7
6.9
6.3
5.8
6.0
6.1
6.0
6.2
6.2
5.9
6.5
7.1

7.3

7.3
6.8
6.8

9.7
12.0
12.7
13.0
13.7
12.2
13.4
15.2
15.2
15.8
16.6
16.2
13.6
14.3

11.8

11.8
11.6

7.2
8.1
7.4
8.3
8.0
7.2
7.7
8.2
8.0
8.2
8.4
8.0
7.7
8.2

8.2

8.2
8.2

67.7
57.6
44.9
86.7
71.4
64.0
72.1
84.5
82.6
86.2
89.9

123.0

(')
< 22.2

<32.9

(3)

84.7
77.4

15.7
12.7
11.9
12.4
11.8
11.4
11.7
13.0
13.7
13.4
12.3
12.8
13.5
13.3

10.7

10.7
10.3
10.3

57.8
69.3
68.5
63.0
68.7
60.9
66.3
65.2
69.6
69.0
68.6
64.7
59.8
57.9

58.0

58.0
59.7

44.3
37.6
31.4
28.5
31.6
30.8
29.8
30.3
31.8
32.5
32.6
31.7
26.2
22.4

20.9

20.9
22.1

52.2
53.1
49.5
44.2
48.4
45.7
47.3
'47.7
51.1
51.1
50.9
48.8
43.0
37.6

33.7

33.7
36.4

42.8

1919

41.8

1920

38.4

1921

1922

1923

35.1
38.5
37.0

1924 .

39.4

1925 .

45.5

1926

46.8

1927

48.9

1928

53.4

1929.. ..

53.9

1930

42. 1

1931

36.7

1934-

29.5

1934

1935

1936

30.3
33.4

■ The definition of realized capital gains and losses was changed beginning in 1934. See footnotes to tables
10 and 12.

2 Source for highest 1 percent includes interest and miscellaneous property income as given in table 12.
Source for all income recipients includes interest and net balance of international flow of property incomes.

3 In 1930 total realized capitalloss was $1,410,000,000 and the highest 1 percent of income recipients received
capital gains of $85,000,000; in 1934 the total gain was $10,000,000 and the gain of the highest 1 percent of
income recipients was $66,000,000.

* Percentage of total loss received by highest 1 percent of income recipients.

Source: Calculated from data in tables 10 and 12.

The year-to-year differences in the degree of concentration of each
type of income suggest that in seeking to explain the shifts in the
degree of income concentration by means of changes in the composi-
tion of total individual income, it will be necessary to take into account
not only the change in the relative importance of each type of income

' It would be possible to use these tables showing the distribution of the size of specific sources of income
for an analysis of the "inequality" in the distribution of these income items by means of Lorenz or Pareto
curves. Such an analysis would be limited by the fact that the data are limited to income items reported
by individuals with net incomes of $."),000 and over and fiMns, on Treasury Form 1040, and by the manner
in which realized capital gains were reported for tax purposes See appendix note A-5, pp. 92-3.) It may
be stated that sources which are highly concentrated as shown in table 14 are also unequally distributed
among the individual income recipients covered in the Statistics of Income tables.

46 CONCENTKATIOx\ OF ECO^•OMIC I'OAVER

but also the changes in the degree of concentration of each source.
However, the degree of concentration of certain of the sources dis-
played considerable stability. The extreme variability in the degree
of concentration of realized capital gains from year to year is ac-
counted for by the extreme fluctuations in the total amount of this
income and by the fact that individuals with realized capital gains
are usually in the higher income brackets. The sustained decline in
the concentration of interest income probably reflects the methods of
financing the war in 1918 and the increased holdings of bonds by
institutional investors. However, interest as a share of the income
of the highest 1 percent displays no tendency to decline.

In connection with the data on the composition of the incomes of
the highest 1 percent and on the concentration of the types of income,
it is of interest to learn what proportion of those included in this
income group received an income from each source. Data enabling
one to determine this have been published in the annual volumes
of the Statistics of Income since 1934 (table 7). Thus in 1936, 62
percent of the individuals included in the highest 1 percent group^
received an income from salaries and wages. The category entrepre-
neurial net income is composed of the Statistics of Income classifica-
tions, business profit, partnership profit, business loss, and partner-
ship loss. Some individuals may report an income from two of these,
for example, both business and partnership profit, and, therefore, when
the percentages for each c these sources are totaled, the resulting:
figure of 35 percent overst- tes to some extent the proportion of indi-
viduals in the 1 percent group receiving entrepreneurial income.
Thirty-two percent of those in this group realized a capital gain and
12 percent a capital loss in 1936. Thus, 44 percent reported a profit
or loss from the sale of property. Net rents and royalties was an.
income source for 22 percent of the individuals in this group. Sixty-
nine percent reported an income from dividends in 1936. The classi-
fication "interest and micellaneous property income" is the total of
four separate items as reported in the Statistics of Income. Of these
four, the necessary data as to the number reporting are available for
fiduciary income and taxable interest, exclusive of the taxable interest
on Government obligations. Twelve percent of the highest 1 percent
reported an income from the former source and 49 percent from the
latter. Doubtless many individuals reported an income from both of
these sources.

CONCENTRATION OF ECONOMIC POWER 47

V. COMPOSITION OF INCOME BY INCOME CLASSES

The differences in the percentage composition of the incomes in the-
various income classes are shown in table 15. These data indicate the-
extent to which the composition of income varies within as small a
group as the highest 1 percent of income recipients and provide the-
basis for determining the reasons for the differences among the selected,
proportions of income recipients in the magnitude and direction of the
changes in income concentration that were outlined in the preceding
chapter.

Owing to certain peculiarities of the basic data some care should
be exercised in interpreting the data presented in this table. The
classification of individuals by income classes is by their statutory
net income whereas the economic income of the individuals in each
class is distributed by sources. Statutory net income is the legal defi-
nition which was in effect during each year ^° and provided the basis^
for the classification of individual incomes in the annual issues of the
Statistics of Income. It differs from economic income in that the
legal deductions from income and wholly tax-exempt interest are not
included in statutory net income. In addition, for the years 1926,
1929, and 1932 statutory net income excluded realized capital losses.
on assets held more than 2 years that were reported for a tax credit.
This method of reporting such losses was used only by indivi ualff
with net incomes of $30,000 and over." In order to obtain the data
shown in table 15 these losses wei'e deducted, but the individuals were
not transferred to their proper income classes after deduction of the
losses. This treatment explains the apparent inconsistencies in the
data for 1932 when for some classes the loss is greater than the total
income. It is necessary to take account of this distortion of the data
in determining the relative importance of the various sources in 1932.^^
In 1926 and 1929 it may be ignored as the losses reported for tax
credit were quite small. For 1935 the definition statutory net income
differs from the preceding years due to the changes in the law regard-
ing the definition of realized capital gains and losses and the limita-
tion on the deductibility of losses. ^^ As a result of the treatment ac-
corded gains and losses the unportance of gains is slightly overstated
except for the very high income classes over approximately $200,000.
For these high-income classes the fact that only various proportions
of gains are included more than offsets the limitation on the deducti-
bility of losses and, therefore, gains are slightly understated.

m Capital net gains from the sale of assets held over 2 years were taxed at a special rate in certain Income
classes during the years 1922 through 1933, but these gains were included in statutory net income in the-
data published in the Statistics of Income.

" See p. 93, footnote 31, for explanation.

« In addition, for 1932, losses from the sale or exchange of stocks and bonds held 2 years or less were
limited to gains from such transactions.

13 Varying proportions of gains and losses were included according to the length of time the asset was-
held and losses were limited to $2,000 in excess of gains. See appendix note B-2,. p. 101, for definition..

^g CONCENTRATION OF ECONOMIC POWER

Table 15. — Composition of incomes, hrj income classes, 1926, 1929, 1932, 1935

Com-

En-

In-
come

In-
terest
and

In-

pen-

tre-

pri-

Real-

Net

come

Percent

Net income class

Total
in-
come

sa-
tion

of
em-
ploy-
ees \'

pre-
neur-
ial
net
in-
come

ma-
rily
from
per-
sonal
serv-
ice

ized
capital
?aius
and
losses

rents
and
roy-
alties

Divi
dends

cel-
lane-
ous
prop
erty
in-
come

pri-
ma-
rily
from
prop-
erty

of all
income
recipi-
ents
in clas;:

Number
of income
recipients

Percentages, 1926

All income classes

100.0

62.9

17.6

80.5

2.9

4.6

6.1

5.9

19.5

100.00

46, 412, 000

Under $5 000

100.0
100.0

71.4
28.5

17.2
19.2

88.6
47.7

.4
12.9

4.7
4.5

1.8
23.4

4.5
11.4

11.4
52.3

98. 07
1.93

45, 517, 132

$5,000 and over

894, 868

$5,000 to $10,000_-

100.0

42.3

27.1

69.4

5.2

5.9

9.8

9.7

.30.6

1.21

560, 549

$10,000 to $25,000

100.0

33.2

21.4

54.6

9.1

5.1

19.3

12.0

45.4

.53

246, 730

$25,000 to $50,000

100.0

24.2

16.3

40.5

12.1

4.4

29.5

13.5

59.5

.12

57, 487

$50,000 to $100,000

100.0

17.8

14.2

32.0

15.0

3.4

36.2

13.5

68.0

.04

20,520

$100,000 to $500,000

100.0

11.3

10.7

22.0

24.4

2.3

39.8

11.5

78.0

.02

8,883

$500,000 to $1,000,000-. -

100.0

4.6

6.8

11.4

34.7

1.9

41.7

10.4

88.6

.001

468

$1,000,000 and over

100.0

2.9

2.2

5.1

50.1

.8

36.6

7.3

94.9

.0005

231

Percentages, 1929

All income classes

100.0

63.0

16.0

79.0

3.4

4.0

7.1

6.5

21.0

100.00

48, 555, 000

Under $5,000 -

100.0
100.0

73.5
27.0

15.7
16.8

89.2
43.8

-1.2
19.4

4.2
3.4

2.6
22.2

5.2
11.2

10.8
56.2

97.87
2.13

47, 522, 929

$5,000 and over

1, 032, 071

$5,000 to $10,000

100.0

46.8

24.5

71.3

4.3

5.0

9.9

9.5

28.7

1.36

658, 039

$10,000 to $25,000

100.0

33.7

19.5

53.2

10.0

4.4

20.0

12.5

46.8

.56

271, 454

$25,000 to $50,000

100.0

23.3

13.8

37.1

15.8

3.4

29.3

14.4

02.9

.13

63, 689

.$50,000 to $100,000

100.0

16.0

11.3

27.3

21.9

2.6

34.0

14.2

72.7

.05

24, 073

$100,000 to $500,000

100.0

9.1

11.4

20.5

36.6

1.4

31.3

10.3

79.5

.03

13, 327

$500,000 to $1,000,000...

100.0

3.8

10.1

13.9

49.5

.8

28.3

7.4

86.1

.002

976

$1,000,000 and over

100.0

1.9

6.6

8.5

59.4

.7

24.5

6.9

91.5

.001

513

Percentages, 1932

All income classes

100.0

70.6

1

n.5

82.1

-3.8

3.4

0.1

12.2

17.9

100. 00

50, 503, 000

Under .$5,000

100.0
100.0

73.8
43.0

11.4
12.4

85.2
55.4

-2.7
-13.1

3.4
3.4

3.0
32.2

11.1
22.2

14.8
44.6

99.29
.71

50, 146, 558

$5,000 and over

356, 442

$5,000 to $10,000

100.0

.55. 4

13.7

69.1

-3.9

3.5

15.5

15.8

30.9

..50

251,014

$10,000 to $25,000

100.0

38.8

12.1

50.9

-4.8

3.6

28.2

22.1

49.1

.16

79, 210

$25,000 to .$.50,000

100.0

33.8

12.8

40.6

-21.9

3.3

41.3

30.7

53.4

.04

18, 480

$50,000 to $100,000

100.0

30.0

11.5

41.5

-43.7

2.9

65.2

34.1

5.8.5

.01

5,902

$100,000 to .$500,000

100.0

19.6

7.5

27.1

-53. 9

2.3

90.2

34.2

72.9

.003

1,730

$500,000 to $1,000,000...

100.0

11.5

.9

12.4

-102.2

2.2

151.3

36.3

87.6

0)

86

$1,000,000 and over

100.0

1.7

-.1

1.6

-32.8

.03

111.1

20.2

98.4

(')

20

Percentages, 1935

100.00

All income classes

100.0

64.7

17.5

82.2

.7

3.3

5.3

8.5

17.8

i2, 327, 000

Under $5,000...

100. 0

68.7

17.7

86.4

.1

3.4

2.4

7.7

13.6

99.04

51,826,885

$5,000 and over

100.0

37.0

16.0

53.0

4.5

2.7

24.8

15.0

47.0

.96

500, 115

$5,000 to $10,000

100.0

51.9

19.6

71.5

2.2

3.3

11.5

11.5

28.6

.65

339, 842

$10,000 to $25,000

100.0

38.3

17.0

55.3

4.4

3.0

21.3

16.0

44.7

.24

123, 564

$25,000 to .$.50,000

100.0

26.9

13. 5

40.4

6.3

2.4

32.3

18.6

59.6

.05

26,029

.$50,000 to $100,000

100.0

20.2

12.2

32.4

7.6

1.8

40.1

18.1

67.6

.02

8,033

$100,000 to $500,000

100.0

11.9

9.5

21.4

7.1

1.0

52.8

17.6

78.6

.005

2,497

$500,000 to $1,000,000...

100.0

2.3

3.5

5.8

13.0

.4

65.5

15.4

94.2

(1)

109

$1,000,000 and over

100.0

.9

.6

1.5

9.0

.1

78.0

11.5

98.5

(')

41

1 Less than 0.0005.

Note.— See text and appendix note A-6 for limitations of data.

Source: Adjusted data from Statistics of Income and estimates of total individual income. See text and
appendix note A-6 for methods of deriving table.

CONCENTRATION OF ECONOMIC POWER 49

The table reveals in striking fashion how the relative importance
of the different income streams varies within a comparatively small
group as persons with incomes of $5,000 and over. Employee compen-
sation, dividends, and realized capital gains provide the best illus-
trations of this. The very high incomes were derived largely from
dividends and realized capital gains. In 1932 and 1935, however,
dividends provided by far the main source of income for these groups
with gains being of relatively minor importance. In 1932 there was
probably a net realized capital loss even in these high brackets.

In all years dividends generally rose in importance with an increase
in the size of income. In a year of high business prosperity such as
1929, however, this source declined somewhat for the incomes above
$100,000. For these very high incomes, profits from the sale of
property exceeded dividends in importance. In 1926 and 1929 as
the size of income increased the proportion of income derived from
realized capital gains rose sharply accounting in these years for more
than one-half of the incomes of those in the $1,000,000 and over class.
In all years the importance of interest and miscellaneous property
income generally increased until about the $100,000 income group and
then declined for the very high incomes.

VI. COMPOSITION OF INCOME AND CONCENTRATION OF INCOME

Thus far data have been presented which have made possible an
analysis of the composition of incomes of the higher income recipients
in relation to the composition of all incomes. The changes in the
composition of income of the highest 1 percent of income recipients
have been examined and the degree of concentration of particular
income sources has been noted. All this evidence suggests that
during periods when types of income which are highly concentrated
increase more or decline less than the total income of individuals, an
increase in the income share of the higher income recipients would
take place. The converse proposition expressed in terms of larger
increases and smaller declines of sources which are not highly con-
centrated and less income concentration follows. In addition to
changes in the relative importance of the various sources in the total
income, changes in the degree of concentration of these sources also
influences the degree of income concentration. Examination of the
data in the foregoing tables in conjunction with the data on income
concentration yields the conclusion that all these factors played . a
part in determining the degree of income concentration in different
years. Shifts in the degree of income concentration were associated
with changes in the relative importance of the various sources of
income and changes in the degree of concentration of these sources.
Analysis of the data on income composition also reveals to a large
extent how the various changes in the distribution of income within
the highest 2 percent of income recipients were brought about.

It is difficult to formulate generalizations concerning the relation
of income concentration to specific income sources that will be valid
for the whole period largely because of the shifting importance of the
sources of income. Furthermore, to ascribe, in terms of the com-
position of income, the "cause" of a given change in income con-

256149 — 40— No. 4-

50 CONCENTRATION OF ECONOMIC POWER

centrttion to a specific income source would be to oversimplify a
complex situation. A given shift in the degree of income concentration
may be viewed as the joint product of a combination of changes in
the volume of all income sources. In addition, changes in the volume
of a single source of income cannot be considered as being independent
of the rates of change of the other income sources. These statements,
however, should not be interpreted as denying that for some years
fluctuations in income concentration can be attributed in large part
to fluctuations in a particular source.

Of all income sources, changes in the volume of realized capital
gains and losses were most closely associated with changes in the
degree of income concentration during this period. In all years,
except 1921, an increase in income concentration was associated with
increases in realized capital gains which were greater than the increases
in total income or decreases which were less than decreases in total
income. During the years from 1918 through 1923 not much im-
portance can be attached to this correlation between realized capital
gains and income concentration because gains constituted but a small
proportion- of total individual income and of the income of the highest
one percent of income recipients. The changes in 1921 which pro-
vided the exception to the close relation of changes in realized capital
gains and losses and income concentration will serve to confirm this
observation. In this year an increase in income concentration was
accompanied by a decline in realized capital gains that was rela-
tively lai^er than the decline in total income. The shift in income
concentration in 1921 may be explained on the basis of the cornbined
effect of changes in the other income sources. There were declines of
20 percent and 30 percent in the aggregate amounts of employee
compensation and entrepreneurial net income, respectively. In con-
trast to these sizable declines, there was a fall of but 9 percent in
dividends and increases in rents and interest of about 4 percent.
Thus the rise in income concentration during 1921 may be ascribed
to the sharp declines in employee compensation and eintrepreneurial
net income which sources are largely received by the lower income
recipients, and the relative stability of dividends which flow for the
most part to the higher income recipients. The increases in volume
of rents and interest also contributed to the rise in income concen-
tration." The effect of the relatively larger decline in realized capital
gains than total income is reflected, however, in the differences among
the selected proportions of income recipients in the smaller importance
of the increase in income concentration. As was noted in chapter II
(p. 33) the higher the incomes of the group, the smaller, relatively,
was the increase in the income share received. This difference among
the various proportions may be attributed largely to the decline in
realized capital gains as this source generally accounts for a substantial
p^rt of the incomes in the verj'^ high brackets.

Doubtless a substantial portion of the rapid increase in concentra-
tion from 1924 through 1929 was due to the realization of capital
gains on a large scale. This increase in profits from the sale of prop-
erty, largely in form of securities, was related to some extent to the
increase in dividends. Until 1927 the realization of capital gains may
bo regarded to a considerable degree as resulting from the revaluation

'* For the effect of these changes on the compo.sition of the income of the highest 1 percent see table 13.

CONCENTRATION OF ECONOMIC POWER 51

of securities in anticipation of an increased volume of dividends. The
rise in dividend income during this period was also an important
factor in the increase in income concentration. The sharp decline in
income concentration after 1929 is, in large part, traceable to the wide-
spread realization of losses from the sale of property. However,
realized capital losses cannot be regarded as the sole factor in this
decline. After 1930 dividends declined much more than total income
and hence this source was also a contributing factor in the lessening
of income concentration that took place during these years.

In contrast to the violent fluctuations in the concentration of in-
come during the years from 1925 through 1934, the degree of income
concentration from 1918 through 1924 fluctuated within relatively
narrow limits. The highest degree of concentration as measured by
the income share of the highest 1 percent of income recipients was
14.2 percent in 1922 and the lowest was 12.4 percent in 1920. During
this period, dividends as a share of total individual income did not
vary greatly and the volume of realized capital gains was small. The
radically different behavior of these income sources during the two
periods, 1925 through 1934 as compared with the years 1918 through
1924, suggests that in the absence of fluctuations in dividends and
profits and losses on the sale of property which are much larger than the
fluctuations in total individual income, the concentration of income
would not vary greatly from year to year. This statement assumes,
of course, that no important independent factor is introduced to
influence the degree of income concentration.

Table 15 reveals that the very high incomes are derived to a con-
siderable extent from realized capital gains. The greater variabiUty
of the income shares received by the smaller groups of income re-
cipients noted in the preceding chapter is largely due to the wide
fluctuations in the volume of this income source. The increases and
declines in dividend income which were generally larger than the
changes in the total income of all individuals also contributed to the
increased variability of these income shares.

The changes in income concentration in 1919 were of special interest
as the three larger proportions of income recipients (the highest 2, 1,
and one-half of 1 percent) received greater shares of the total income
than in the previous year while the shares of the highest one-tenth
and one-hundredth of 1 percent of income recipients were smaller.
Examination of the precedm^ tables^ indicates the manner in which the
changes in the various sources of income combined to bring about this
result. Realized capital gains increased, but in 1919 this type of
income constituted a small proportion of the total income of the
highest 1 percent of income recipients (5.8 percent). While dividends
declined, the proportion of total dividends received by the highest
1 percent jumped from 58 to 69 percent. The income source that
apparently played the dominant role in influencing the shifts in
income concentration in 1919 was entrepreneurial income. In 1919
the increase in this source apparently resulted in a larger than pro-
portionate increase in moderately high incomes — incomes between
$5,000 and $30,000. Table 15 indicates that this type of income is
more important for this income group than for the others and table 12
shows that in 1919 the entrepreneurial net income was the largest
income source for the highest 1 percent of income recipients.

,52 OONCENTriATION OF ECONOMIC POWER

As noted in the preceding chapter, the highest 2 percent of income
recipients received about the same share of the total income in 1931
as in 1930 while the shares of the smaller proportions declined con-
siderably. In terms of the changes in the composition of income, this
shift in the income structure was apparently due to the large decline
in property income caused by the diminished volume of dividend
payments and the increase in realized capital losses. Total losses
were nearly twice as large in 1931 as in 1930 and dividends declined
by one-fourth. The decline in the compensation of employees was
not so severe, being less than one-sixth. This change in the com-
position of total income, while it did not materially affect the division
of total income between the highest 2 percent and the lower 98 percent
of the income recipients, did decrease the share of total income re-
ceived by individuals with very Jarge incomes which are derived
chiefly from dividends and realized capital gains.

Due to the interest in changes in income concentration during
recent years, the role of the various income sources in producing the
indicated shifts in income concentration will be briefly outlined.
Comparing 1934 with 1931 there was a decline in the share of total
individual income received by the highest 1 percent of income recipi-
ents from 13.7 to 12.7 percent. This decline in income concentration
was accompanied by a substantial decline in dividend payments which
is reflected in the reduction of the share of dividends in the total
income of the highest 1 percent, 34 percent in 1931 as compared with
26 percent for 1934. The percentage declines in the sources classified
as income primarily from property was about twice as large as the
decline in the two sources classified as income from personal service.
-The increase in the degree of income concentration in 1935 was due
jointly to a rise in realized capital gams and dividends, and to a sharp
increase in entrepreneurial income. The fairly uniform increase in
income concentration among the selected proportions of income
recipients in 1935 is attributable to the relatively large increases in
entrepreneurial net income and net rents and to the fact that though
realized capital gains increased they were relatively small in size for an
individual and constituted a small proportion of total income. The
compensation of employees increased nearly as much as total income,
the percentage increases being 8.1 and 8.8 percent, respectively. The
increase in concentration in 1936 is clearly a result of a relatively
greater increase in the income sources derived chiefly from the owner-
ship of property than in those derived from personal service. The
rise in dividends and profits from the sale of property were especially
important in bringing about an increase in concentration. The decline
in concentration in 1937 was accompanied by increases in the com-
pensation of employees and entrepreneurial net income which were
larger than the increases in the total income of all individuals. Losses
from the sales of securities also contributed to the decline in con-
centration.

The analysis of the relation of the changes in concentration and the
composition of income could doubtless be profitably continued. Tie
general lines of further investigation would be to integrate a more
detailed study of the changes in composition and concentration of
income with an examination of the factors determining the composition

CONCENTRATION OF ECONOMIC POWER 53

of income. Part of this more detailed study would take the form of
analyzing the extent to which incomes of specific individuals are
derived from several sources. As indicated in the introductory portion
of this chapter, knowledge of this aspect of the composition of incomes
is of considerable importance for an analysis of the composition and
concentration of income. From the viewpoint of income sqprces, a
given degree of income concentration may be said to be the result of
three conditions: First, the relative importance of various income
sources in the total income of individuals; second, the extent to which
the receipt of each source is concentrated; and third, the extent to
which individuals receive incomes from more than one source. The
first two have been analyzed in this study and the tliird has been
treated only indirectly. When more comprehensive data become
available, the tliird factor may be investigated more thoroughly.
For the present purpose the analysis has been carried far enough to
disclose the intimate relationship that exists between shifts in the
composition of income and shifts in the concentration of income. On
the basis of the indicated relationsliips it is possible to determine with
considerable accuracy the effect on the degree of income concentration
of changes in the volume of the distributive shares and to single
out the types of income chiefly responsible for shifts in the degree of
income concentration. Such information raises the level of our
knowledge of the process of income distribution and lays the founda-
tion for an evaluation of the various methods for influencing the
degree of income concentration embodied in Federal and State law.

CHAPTER IV

THE CONCENTRATION OF "PURCHASING POWER" AND THE
EFFECTS OF RELIEF AND VETERANS' ADJUSTED-SERVICE
PAYMENTS ON INCOME CONCENTRATION

I. THE CONCENTRATION OF "PURCHASING POWEr": 191^-37

As indicated in the introductory part of chapter II the distribution
of income may be measured at different stages in the process of income
circulation. In chapter II income was measured when it was acquired
in return for personal or capital services. Much interest is attached to
another stage in the circulation of income — the distribution of what may
be termed, in the interest of convenience, "purchasing power."

1. The Transition From "Earning Power" to "Purchasing Power."

Purchasing power may be briefly defined for the present purpose as
that portion of current income which is available to individuals either
for spending or saving; that is, the receipts of individuals over which
they have control of disposition. This part of income may be spent
for consumer goods or saved. To arrive at this concept of income it is
necessary to make a number of adjustments to the distribution of
acquired income. First, most of the direct personal taxes paid out of
acquired income should be deducted. Second, the distribution of ac-
quired income should be adjusted, by deductions from the incomes of the
donors and additions to the incomes of the recipients, for the trans-
fer of certain types of income among individuals and during recent
years for the disbursement of direct-relief payments and the veterans'
adjusted-service compensation. In addition, the distribution of
purchasing power is influenced by the receipt of inheritances and in-
surance benefits, neither of which are included in acquired income.
The distribution of purchasing power is not so uniquely defined as the
distribution of earning power. The purchasing power exercised by
individuals or families may be determined not only by their current
income but also by their assets and ability to secure credit. The data
in this study, however, are confined to the distribution of purchasing
power derived from current income. In addition, some persons are
interested in the purchasing power of individuals after certain ex-
penditures which are regarded as having a prior claim on income.
Therefore, other adjustments might be made, such as deduction of
insurance premiums or the inclusion of consumer credit when extended
and the deduction of repayments on consumer debts. The statistics
showing the shifts in the concentration of purchasing power presented
in this section, due to the absence of essential information, show only
partially the effect of taxes and transfers of income on the concentra-
tion of acquired income.

55

56

CONCENTRATION OF BC50NOMIC POWER

Personal taxes. — There is some difficulty in determining the specific
direct personal taxes which should be deducted from income. Since
the purpose of this study is to segregate the income available for
expenditure, the various types of indirect taxes should not be deducted
from income inasmuch as these taxes are included, for the most part,
in the prices of goods and services. The personal-property tax on
owner-occupied homes and the motor-vehicle license fees for the
personal use of automobiles are taxes which are paid directly by con-
sumers out of their incomes, but these taxes may be considered ^s
indirect taxes included in the expenditures for housing and for the
use of an automobile. The difference between these two taxes and
indirect taxes seems to be superficial and in studying the expenditures
of individuals it seems proper to treat them in the same fashion as the
other types of indirect taxes included in the price of goods and services.
Federal and State personal-income taxes, the poll tax, and probably
the major portion of the personal-property taxes on intangible and
tangible property other than real estate, are paid directly from income
and should be deducted in arriving at the amount available for ex-
penditure. Of these taxes only the Federal individual income tax is
deducted in this study. The other taxes cannot be deducted because
of the lack of information. The total amounts of some of the taxes
are not known and information on the distribution by income classes
of all these taxes, except the Federal income taxes, is unavailable.^

The inability to deduct these taxes from income is not believed to
be a serious limitation of the data on the concentration of purchasing
power. The Federal income tax constitutes the bulk of the taxes
which should be deducted, accounting in recent years for about 65 to

' The poll tax is relatively unimportant for the present purpose. The total amount collected in the fiscal
year ending in 1936 is estimated at $17,000,000. (See Tax Re«!earch Foundation, Tax Systems of the World,
7th ed., 1938, p. 391.) The deductions from income for State income taxes are considerably larger but com-
plete data are not available either as to the aggregate amounts, or the distribution by income classes. The
following information partly estimated and partly from State reports is available as to the amount of tax
collections for the State income taxes for recent years. The figures giv«n in the following table are estimated
collections for the fiscal years ending during the given year and, therefore, apply for the most part to incomes
during the previous year.

Estimated collections from Slate individual income taxes in fiscal years ending in 1930-36, 1938
[Millions of dollars]

1930

1931

1932

1933

1934

1935

1936

1938

136

83

67

64

79

100

144

249

Source: 1930-35: Collections from Selected State-Imposed Taxes, 1930-36; table 4 (U. S. Treasury
Department, Division of Research and Statistics. 1936). 1936; Facing the Tax Problem, table J (p.
530) (Twentieth Century Fimd, New York, 1937) 1938; Bulletin of the Treasury Department,
August 1939. p. 4.

It is much more difficult to obtain an estimate of pesonal-property taxes on intangible and tangible prop-
erty (other than real estate). For 1936 the total tax on nonbusiness personalty was estimated at .fnO.OOO.OOO.
(See Twentieth Century Fund, Studies in Current Tax Problems, New York, 1937, p. 18, footnote 14.)
This figure serves merely as an approximation of the tangible and intangible property taxes which should
be deducted from income for the present purpose.

CONCENTRATION OF ECONOMIC POWER

57

.75 percent of the total collections from these taxes; ^ arid, as will be
shown, the Federal income tax has varied considerably in importance
relative to the incomes of the highest 1 percent of income recipients
and to the total income of individuals. In contrast, the other taxes
constitute a small part of the total and it is probable that the effect of
these taxes on the concentration of income is relatively stable for the
period covered by the statistics. There are some specific data on State
personal income taxes which is informative in tms connection. As-
suming that the State individual income taxes are paid wholly by the
highest 1 percent of income recipients the share of total income
received by this group in 1935 is reduced from 13.2 (table 22) to 13.0.
While there have been changes since 1918 in the number of States
with personal income tax laws, two of the three States which collected
in recent years more than four-fifths of the total State income taxes
have had personal income tax laws for the whole period and the third
introduced the income tax in 1919.^ On th'e basis of available in-
formation it seems likely that the increase in the number of States
with income taxes, as well as some increase in rates, since 1929, has
made the State income tax relatively more important since that date
than prior to it. Because of the comparatively small amounts
collected, however, changes in the number of States with income taxes
and changes in the rates of State income taxation undoubtedly have
had little effect on the trends in the concentration of purchasing power.
In view of the foregoing discussion it seems reasonable to conclude
that the effect of direct personal taxes in producing shifts in the con-
centration of purchasing power wiU be largely taken into account by
the deduction of Federal income taxes from income.

' The following table presents data on the specific taxes:

Collections from selected taxes in fiscal years ending in 1930, 1934, 1936, and 1938.
[Millions of dollars]

Type of tax

1930

1934

1936

1938

Federal individual income tax

1,147

135

170

17

420
79
170

17

674

144

170

17

1,286

State personal income tax

249

Nonbusiness personalty tax -- . . .-

170

Poll tax- . . - - -

17

Total

1,469

78

686
61

1,005
67

1,722

Federal individual income tax as percent of above total

75

Source: Federal individual income-tax collection from Annual Report of Secretary of Treasury for
1939 (p. 375).

State personal income-tax collections from sourcos cited in preceding footnote.

Nonbusiness personalty taxes and poll-tax source cited in preceding footnote. Amounts kept
constant owing to a lack of data. Nonbusiness personalty taxes should be somewhat higher in 1930
and 1938 and lower in 1934.

' These States are Massachusetts, New York, and Wisconsin. The New York law was first effective in
1919. From 1923 to 1929 there was no increase in the number of States having personal income taxes. After
1929, 18 States enacted income-tax laws. (See National Industrial Conference Board, State Income Taxes,
New York, 1930. vol. I, and for list of States imposing a personal income tax on Jan. 1, 1937, Twentieth
Century Fund, Facing the Tax Problem, New York, 1937. p. 15.) For data showing that these 3 States
regularly collected from 1930 through 1935 more than four-fifths of the State personal income taxes see the
U. S. Treasury Department publication Collections from Selected State-Imposed Taxes, 1930-36, table II.
Table 4 of this publication shows the number of States with personal income-tax laws in recent years.

58 CONCENTRATION OF ECONOMIC POWER

The definition of purchasing power adopted in this report measures
the income that individuals are free to utilize as they wish either for
current consumption or as an addition to their savings before this
income is disbursed or saved. A further step would involve taking
account of the various indirect taxes persons pay when they expend
this income on goods and services. Some taxes, however, such as
the corporate net income and capital stock taxes, have the effect of
reducing individual incomes before they are received and, tLsrefore,
are not included in the measures of earning power or purchasing power.
As pointed out in the beginning of chapter II, the determinati6n of
the incidence of all types of taxes constitutes a step in measuring the
distribution of what has been termed "real income." However, the
term "purchasing power" is sometimes defined to include this type of
adjustment.^

It should, perhaps, be noted that the taxes which have been singled
out for deduction from income in measuring purchasing power are
for the most part progressive in incidence; that is, they constitute a
larger proportion of individual income as the size of income increases.
These direct taxes are but a small portion of all taxes that are paid
directly and indirectly and obviously the incidence of these taxes would
not be representative of the incidence of total tax collections by all
governmental units which in 1938 amounted to $14,811,000,000.^
Collections during this fiscal year from the Federal and vState personal
income taxes, the poll tax and the miscellaneous personal-property
taxes, exclusive of real-estate tax, probably did not exceed
$1,800,000,000. Indirect taxes, such as sales, excise, and other taxes
shifted for the most part to the consumer, in contrast to the income tax,
are largely regressive in incidence.^ Collections from these taxes are
at least several times as large as the revenue from the three types of
taxes which should be deducted from income for the purpose of
measuring purchasing power, as defined in this report.

Income transfers. — In addition to the deduction of direct personal
taxes, the distribution of purchasing power is influenced by gifts
among individuals, contributions to charitable organizations, inheri-
tances, insurance benefits, and in recent years the receipt of direct

* In order to show the effect of this adjustment in a single year, it would be necessary to determine, by
income classes, the incidence of all taxes. When it is desired to show the shifts in the concentration of
purchasing power, according to this alternative definition, over a period of years, it would possiblv he simpler
and more satisfactory to take account, for the most part, of the changing incidence of all taxes Hy securing
data— first, on the changes in the prices of the goods and services purchased by various income classes;
second, on the changes in the allocation of income between various types of goods and services; and, third,
on the changes in the ratio of consumption to saving. In this manner the effect of indirect taxes would be
largely reflected in the price data. Unfortunately the magnitude of this task precluded the possibility of
undertaking it in the present study. An easjer though probably less satisfactory method would be to
investigate the changes in percentages of total tax collections from the various types of taxes and taking
account of then incidence. See chapter 3 of Facing the Tax Problem (Twentieth Century Fund, New York
1937) for a description of the changes since 1913 in the relative importance of the various taxes.

» I-iuUetin of the Treasury Department, August, 1939. p. 4.

' For an analysis of the incidence of the total tax "burden" on different income classes by the use of hypo-
thetical examples see Twentieth Century Fund, Facing the Tr.x Problem, New York, 1937, ch. 16, and for a
more detailed analysis of the same data see Newcomer, Mabel; "Estimate of the Tax Burden on Different
Income Classes" in Studies in Current Tax Problems; The Twentieth Century Fund, 1937. It appears on
the basis of the examples in this study that for the income classes not subject to income and death taxes the
taxation system in effect during 1936, is repressive. For the income classes subject to income and death
taxes. Miss Newcomer concludes that the tax system is progressive in incidence. Because of the changes
from year to year in the relative importance of progressive and regressive taxes, the incidence of the tax
burden in 1936 cannot be taken as representative of all years. For example in 1930 the Federal Government
derived OS percent of its revenue from taxes which are largely progressive in incidence (income taxes, corpo-
rate and individual, estate and gift taxes). In 1934 only 34 percent of the revenue was obtained from these
sources. In 1937, however, 49 percent of Federal revenue was obtained from these sources. (The 1934 and
1937 percentages include also capital-stock and excess-profits taxes which were not in effect in 1930. Data
from Annual Reports of the Secretary of the Treasury.)

CONCENTRATION OF ECONOMIC POWER 59

relief and the veterans adjusted-service compensation.^ Little is
known concerning the magnitude and distribution by income classes
of most of these items ^ In this study it was practicable to take into
account only the last two — direct relief and the veterans' adjusted-
service compensation. It may be indicated that the effect on the
distribution of pm-chasing power of the gifts, inheritances, insurance
benefits, and contributions is likely to be fairly stable over short
periods so that the failure to take them into account may not appre-
ciably affect the trends in the concentration of purchasing power which
are shown in this section.® The shares of purchasing power at the
disposal of the highest 1 percent of income recipients would probably
be only slightly overstated.'" The absence of information on the
receipt of inheritances, gifts, insurance benefits by income classes
makes it impossible to arrive at a more positive conclusion. Probably
only the portions of these types of receipts that are used for current
expenditures should be taken into account for the present purpose.

2. Statistics qf " Purchasing Po'mer" Concentration.

In view of the limited availabihty of the necessary basic informa-
tion, the data on the concentration of purchasing power presented in
this section take into account only the effect on the concentration of
earning power of Federal income taxes, relief payments, and the
veterans' adjusted-service compensation. As indicated by the fore-
going discussion it is beheved that the trends in the concentration of
purchasing power will be revealed when these adjustments are made.
Due to the inabihty to carry through all of the indicated corrections,
the level of the purchasing power shares shown in table 16 is probably

' For a discussion of the characteristics of relief and veterans' adjusted-service payments as a form of in-
come see below, pp. 65-7.

* In a recent publication of the National Resources Committee entitled Consumer Expenditures in the
United States it was estimated that gifts and contributions by families and single individuals to indi-
viduals and organizations amounted to $2,178,000,000 or 3.7 percent of total consumer income (p. 46). These
figures relate to a 12-month period during 1935-36. The amount for gifts to individuals includes gifts such
as Christmas and birthday presents. It is doubtful whether these should be included as transfer income
items when the distribution of purchasing power is under consideration. As was mentioned at an earlier
point the gifts included for the present r urpose would need to be limited both as to type and relationship
of donor and recipient. The break-down of gifts between gifts to individuals and gifts or contributions to
organizations is available in this publication only for families. Of total gifts of $1,324,000,000 made by
families (exclusive of single individuals) $723,000,000, or 55 percent, were given to individuals. Data on
the contributions to charitable organizations reported by individuals filing income tax returns are avail-
able in the annual issues of the Statistics of Income since 1922; but data on total contributions made to these
orgamzations are difficult to obtain. Estimates are presented for the years 1909, 1914, and for alternate years
from 1919 to 1931 in W. H. Lough's High Level Consumption (p. 245). For 1929 an estimate of $1,712,000,000
is given. An estimate for 1935-36 of $980,000,000 may be derived from the National Resources Committee
ttudy by assuming that the distribution of gifts between those to individuals and to organizations is the
same for single individuals as for families.

Ucing the Statistics of Income (table 7) to obtain data on the contributions of the highest 1 percent of in-
come recipients, the share of total contributions to charitable organizations made by them was 18 percent for
1929 and 17 percent for 1935-36. These percentages may be compared with the share of total income received
by the highest 1 percent of income recipients— 18.5 percent in 1929 (table 1) and 13.2 percent in 1935 and 14
percent in 1936 (table 22). Net much significance can be attached to these percentages, however, because of
the slender basis for the estimates of total contributions and the probable underreporting of contributions to
the Bureau of Internal Revenue. In addition, the contributions deductible under the income-tax law may
not exceed 15 percent of net income before deduction of the contributions.

» It may be noted that contributions as a percentage of the income of the highest 1 iwrcent of income
recipients display considerable stability, rising somewhat, however, in years of business depression and
declining in years of business prosperitv. For the period 1922-31 and 1934 these percentages varied from 2.1
percent in 1925, 28 and 29 to 2.6 percent in 1931. Durirg th^ years 1934-37 when, due to the change in the
definition of income, the percentages would slightly understate the proportion, the highest figure was 2.3
percent in 1937 and the lowest, 2.0 percent in 1935 and 1936.

"> On the basis of the data on the distribution of expenditures for gifts and contributions by size of family
incomes (Consumer Expenditures, table 27A, p. 87) and the use of extreme assumptions, it seems likely
that the effect of the deductions for contributions and the transfers involved in gifts (accepting the broad
definition of gift;, would reduce the share of total incom received by the highest 1 percent of income recip-
ients by less than 1 percent (absolute amount). For example, in 1936 the share of acquired income received
by the highest 1 percent of income recipients was 14.5 percent and it seems probable that this group's sharp
of income after taking into liocount the change in the distribution of income resulting from gifts and con-
tributions ^ould be not less than 13.5 percent. The changes in the income shares of the smaller proportions
of income recipients due to these two items are probably of greater importance.

60

CONCENTRATION OF ECONOMIC POWER

high by an absolute amount of about 1 percent. It should be noted,
however, that all the data on the concentration of purchasing power
as well as on the concentration of earning power understate the actual
degree of concentration because of the understatement which charac-
terizes the basic data on the high incomes."

Table 16. — Shares of total ^'purchasing power" available to the highest 1 percent of
income recipients, 1918-37 '

Year

Percent of
total pur-
chasing
power
available

Index
(1918 = 100)

Year

Percent of
total pur-
chasing
power
available

Index
(1918=100)

1918

11.30
11.87
11.32
12.59
13.14
12.23
13.35
15.56
15.42
16.28
IS. 10

100.00
105. 04
100.18
111.42
116.28
108.23
118. 14
137. 70
136. 46
144.07
160. 18

1929

17.45
14.03
13.07

11.64

12.02
12.21
12.58
11.80

154. 42

1919

1930 _

124. 16

1920

1931

115. 66

1921.

1934

1922

103.01

1923

1934

1924

106. 37

1925

1935

108. 05

1926 .. .

1936 --

111.33

1927

1937

104.42

1928

1 See text for limitations of data. The second percentage for 1934 and those for 1935 through 1937 are over-
stated for reasons indicated on p. 15 of ch. II and in appendix note B-2

Source: See appendix note A-7.

Table 16 shows the shares of the total purchasing power available
to the highest 1 percent of income recipients when, first, the Federal
income taxes are deducted from the incomes of all income recipients
and the highest 1 percent, and, second, work- and direct-relief payments
and the veterans' adjusted-service compensation are included in the
total income of individuals. The influence of year-to-year changes
in prices is largely eliminated by comparing the concentration of pur-
chasing power in one year with the concentration in another year.
However, insofar as there were differences in the movements of the
prices of goods and services purchased by various income groups, these
differences would have to be taken into consideration in interpreting
the data.^^

Effect of adjustments for "purchasing power" on the concentration of
"earning power." — Table 17 shows the percentage reductions in the in-
come shares of the highest 1 percent of income recipients resulting from
these adjustments. These percentage reductions vary from a mini-
mum of 4.1 percent in 1930 to a maximum of 13.4 percent in 1936. In
1930 the highest 1 percent of income recipients received 14.63 percent
of the total incorne received for personal and capital services (table 1,
p. 16) and, as indicated in table 16, the share of this group after deduc-
tion for income taxes was 14.03. In 1936 when the reduction in the
share of the highest 1 percent of income recipients was greatest, the
share of total purchasing power was 12.58 percent and the share of total
acquired income was 14.53 percent. Part of the difference between the
two shares in 1936 is due to the inclusion of direct relief and the soldiers'

" This is a result of nonreporting and underreporting of incomes to the Bureau of Internal Revenue.
See ch. II, pp. 15.

1! Some statistics, based on scanty data, showing price movements for different income groups were pub-
lished by Willford I. King in The National Income and Its Purchasing Power, New York 1930, pp. 68-69.
The available data on this subject are extremely inadequate.

CXDNCENTRATION OF ECONOMIC POWER

61

bonus in the total estimate of purchasing power and not entirely to
the deduction of income taxes. If these two items are excluded and
the effect of the deduction for income taxes is alone considered, the
reduction in the income share of the highest 1 percent of income
recipients is somewhat less. Thus, the deduction of Federal income
taxes reduced the share of the highest 1 percent from 14.53 to
13.00 percent. This is a reduction of 10.5 percent in the income
share of this group which may be compared with the reduction of 13.4
percent when the combined effect of income taxes and direct relief and
the veterans' bonus is measured. In other years direct relief and the
veterans' bonus exercised much less influence as these two items con-
stituted a smaller proportion of total income. ^^ The reductions in
the share of income shown by the data in table 17 are, therefore, almost
wholly attributable to the deduction of income taxes. Considering
the effect of income taxes alone, the reduction in the income share of
the highest 1 percent of income recipients in 1936 was also slightly
larger than in any other year with the exception of 1918 and 1919.

Table 17. — Effect of adjustments for "purchasing power" on the income shares of the
highest 1 percent of income recipients, 1918-37

Year

Percentage

reductions

in income

shares

Year

Percentage

reductions

in income

shares

1

Year

Percentage

reductions

in income

shares

1918

11.7
11.1
8.9
7.2
7.7
5.6
5.8

1925 _.

5.1
4.8
5.2
6.0
5.5
4.1
4.7
1

1934

8 1

1919

1920

1926

1927

1928

1934

7 8

1921..

1935 .. .

8 9

1922

1929

1930

1936

13 4

1923

1937

11 2

1924

1931

Note.— The data show the percentage by which the share of total income is reduced when Federal income
taxes are deducted and direct relief and the veterans' adjusted-service compensation are included in the total
income.

Source: Calculated from data in tables 1 and 16.

Because of the progressive rates of income taxation, the deduction
of income taxes results in a relati\ ely larger reduction in the income
as the size of income increases. ^"^ The reductions in the income shares
of the proportions of income recipients less than 1 percent are, there-
fore, more substantial than those shown for the highest 1 percent of
income recipients. Examples for the years 1926 and 1936 will be
presented as illustrative of the reductions which take place when the
income shares of higher income groups are adjusted to obtain measures
of purchasing power. The rates of taxation were low in 1926 and
high ui 1936.1^ Table 18 contams for 1926 and 1936 the shares of pur-
chasing power received by selected proportions of income recipients
and the percentage reductions in these shares due to the adjustments
for purchasing power. The efl'ect of progressive rates of taxation in
each year and of higher rates of taxation in the latter year are clearly
shown by this table. In 1926 the income share of the highest 1

'3 See table 2U, p. 67.

'* This is true, of course, only if the surtax rates continue to increase with the size of income. The surtax
rates have generally remained constant after a given income is reached. To take the extreme instances,
during the years 1925 through 1931, the surtax rates were constant above $100,000. In the years 1936
through 1938 the surtax rates rose for incomes up to $5,000,000.

'• See table 19, p. 63, for data showing the proportions of the incomes of the highest 1 percent paid in
income taxes during the years 1918-37.

62

CONCENTRATION OF ECONOMIC POWER

percent of income recipients was reduced by 4.8 percent and the share
of the highest one one-hundredth of 1 percent by 13.3 percent. In
1936, when the rates of income taxation were considerably higher, the
reductions for these two groups were 13.4 percent and 40.6 percent.
As indicative of the differences between these 2 years in the effect of
the adjustments for purchasing power it may be noted that percentage
reduction in the share of the highest one one-hundredth of 1 percent of
income recipients in 1926 was the same as the reduction in the share
of the highest 1 percent of income recipients in 1936. The proportions
of income paid in income taxes were also about the same for these two
groups — 14.1 for the highest one one-hundredth of 1 percent of income
recipients in 1926 and 13.4 percent for the highest 1 percent of income
recipients in 1936. Due to differences in the income-tax-rate structures
of the 2 years, the percentage reductions in the income shares resulting
from the indicated adjustments increased more with the size of income
in 1936 than in 1926. Thus, in 1926 there was a reduction of 10.0
percent in the income share of the highest one-tenth of 1 percent of
income recipients and of 13.3 percent in the income share of the highest
one one-hundredth of 1 percent. In 1936 there was a decline of 24.9

Table 18. — Shares of "purchasing power" available to selected proportions of income
recipients and effect of adjustments for purchasing power, 1926 and 1936 '

Year

Group of Income recipients

Highest
^^of 1
percent

Highest
Moof 1
percent

Highest
Moo of 1
percent

Share of purchasing power

Proportion of income paid in Federal income taxes.
Percentage reduction in income share

1936

Share of purchasing power

Proportion of income paid in Federal income taxes
Percentage reduction in income share

Percent
11.79
7.29
0.43

9.35
15.13
16.29

Percent
6.13
10.97
10.00

4,31
23.81
24.91

Percent
2.34
14.09
13.33

1.17
39.90
40.61

' Shares for 1936 are slightly overstated. Proportion of income paid in income taxes is slightly under,
stated for 1936. See appendix note B-2.

Source: See appendix note A-7.

percent in the income share received by the former group and a
decline of 40.6 percent in the share received by the latter group.

The effect of income taxes in reducing the incomes of the highest 1
percent of income recipients is also shown by table 19. The percent-
ages of economic income paid to the Federal Government in the form
of income taxes given in this table are not directly comparable to
Statistics of Income data on effective tax rates, as the latter are based
on statutory net incomes while the percentages in table 19 are based on
economic incomes.'^ The changes in the percentage of income paid
for income taxes reflect changes in the rates of taxation ; changes in the
treatment for taxation of certain income, such as dividends, realized
capital gains and losses, and "earned" income; changes in personal
exemptions and credits allou'^d for dependents; and the changing levels
of income with constant rates for specific amounts of income. This

'• See ch. II, pp. 10-11, for definitions of economic income and statutory net income..

CONCENTRATION OF ECONOMIC POWER

63

last factor explains the sharp decline in the percentages from 1928
through 1931 when rates were constant" and treatment of certain
types of income was unchanged.

Table 19.^ — Proportions of economic income of highest 1 percent of income recipients
paid in Federal income taxes, 1918-37 '

Year

Percent

Year

Percent

Year

Percent

1918

13.41

12.81
10.30
8.51
9.10
6.51
6.74

1925..

5.98
6 85
6.28
7.43
6,62
4.60
3.18

1934...

7.69

1919

1926

1934

1920

1927

7.38

1921

102S

1935

8.43

1922 ..

1929

1936....

12.23

1923

1930 .

1937

11.39

1924

1931

1 The second 1934 percentage and those for 1935 through to 1937 understate slightly the proportion of
Income paid for Federai-income taxes. See text and appendix note B-2.

Source: Appendix tables A-1 and A-7.

Income-tax payments as a proportion of the income of the highest
1 percent of income recipients were highest in 1918, 1919, and 1936.
The 1036 percentage is slightly understated when compared to the
percentages for the years prior to 1934 so that the income-tax payments
in 1936 were about as large a proportion of the income of this group as
in 1919.^^ While the surtax rates were higher in 1936, the "normal"
tax rates were higher in 1918 and 1919. This difference in rate struc-
tures yielded higher total tax rates, surtax and normal tax, for incomes
of less than $25,000 in 1918 and 1919 than in 1936.

Year-to-year changes in the concentration of "purchasing power." — The
general movements in the concentration of purchasing power are
similar to those already described in chapter II for the concentration
of earning power. Apart from minor variations, the degree of
concentration rose after 1921 and reached a maximum in 1928.
After 1928 the concentration of purchasing power declined sharply
and then increased somewhat from 1934 through 1936. In 1937 the
share of total purchasing power at the disposal of the highest 1 per-
cent of income recipients declined slightly. ^^

Examination of table 17 is probably the simplest method for isolat-
ing the differences between the magnitudes of the year-to-year varia-
tions in the concentration of purchasing power and acquired income or
earning power. After 1919 the differences between the shares of pur-
chasing power and earning power declined. The decline continued,
interrupted only by a few minor increases, until 1934. After 1934 the
difference became greater.

'' Rates were constant for this period except for a modification in 1929. In 1929 the rates of the normal
tax on individuals were reduced from V/i, 3, and 5 percent to H, 2, and 4 percent by joint resolution of Con-
gress. The first percentage is the normal rate on the first $4,000 of net mcome, the second on the second
$4,000, and the third on the balance over $8,000. (See Statistics of Income for 1929, p. 1.)

" The understatement arises from the change in definition of income beginning in 1934. Relative to the
earlier years income of the highest 1 percent is overstated during 1934 through 1937. For the years 1934
through 1937 taxes paid would, therefore, be a somewhat larger percentage of income than shown in table 19
if there were no change in the definition. See ch. II, p. 15 where the definition of income for these years is
explained and appendix note B-2.

" To some extent income taxes may be considered in part a.s a deduction from the income of the following
year rather than of the year the income on which the tax is paid was received. This is so because the tax is
not paid during the year the income was received but in the subsequent year. Only those with high incomes
set up special reserves for income taxes, but most taxpayers probably make some provisions for taxes out of
their current income. In years when the rates of income taxation were increased or income declined sharply,
income taxes probably reduce to some extent the income of the following year. The opposite effect results
when rates decline and incomes increase. This aspect of the statistics on purchasing power concentration
should be taken into account when using them for certain purposes.

54 OONCENTRATION OF ECONOMIC POWER

Thus, the most striking differences between the trend in the con-
centration of earning power and purchasing power are clearly revealed.
The first is the greater increase in the degree of concentration of
purchasing power that occurred during the twenties. As shown in
table 1 of the second chapter (p. 16) the share of the acquired income
or earning power received by the highest 1 percent of income re-
cipients increased from 1918 to 1928 by 51 percent. The comparable
increase in the share of total income available for expenditure and saving
by this proportion of the Nation's income receivers was 60 percent.
This sharper rise in the concentration of purchasing power is, of course,
due to the decline, particularly after 1924, in the rates of Federal
income taxation. ^° In 1918 Federal income taxes absorbed 13.48
percent of the income of the highest 1 percent of income recipients
and in 1928 only 7.43 percent. The percentage declines after 1928
in the shares of purchasing power and earning power received by the
highest 1 percent of income recipients were approximately the same.

The changes in the shares of purchasing power received by the
groups of income recipients of less than 1 percent were also in the
same direction as the changes in the shares of earning power received
by these groups. As a result of the lower surtax rates in effect from
1924 through 1928, the differences between the increases in the shares
of earning power and purchasing power were greater for these smaller
proportions of income recipients than for the highest 1 percent. As
shown in chf pter II, the increases in the shares of earning power
received by t.ie smaller proportions of income recipients were greater
than those received by the larger proportions. Therefore, the
increases in the shares of purchasing power received by the groups
with less than 1 percent of the income recipients were considerably
larger than the increases in the shares received by the 1 and 2 percent
groups.

In contrast to this greater increase in the concentration of pur-
chasing power, the rise in the concentration of purchasing power after
1934 was not so great as the increase in the concentration of earning
power. This is particularly true of 1936 when the surtax rates on
incomes above $50,000 were raised. In 1936 the share of acquired
income received by the highest 1 percent of income recipients rose
by 8.4 percent (table 1, p. 16) while the increase in th^ share of pur-
chasing power available to this proportion of the Nation's income
receivers was only 3.0 percent. A further contrast between the
changes in concentration in these two periods is provided by the
changes in the shares of purchasing power received by the smaller
proportions of income during the years 1934 through 1936. During
1936 there was an increase of 10 percent in the share of earning power
received by the highest one-tenth of 1 percent of income recipients
(table 3, p. 23). On the other hand, the share of purchasing power
of this group was about the same.^^ There was, in fact, a slight
decline in the share of purchasing power received by the highest
one one-hundreth of 1 percent of income recipients in contrast to an
increase of over 9 percent in the share of earning power received by this

'" Part of the decline in the rates of taxation is attributable to the more favorable treatment accorded
realized capital gains or assets held over 2 years.

" 4.25 percent in 19.34, 4.30 pcrcont in 193.5, 4.31 percent in 1930, and 4.00 percent in 1937. Federal income
taxes paid by this grou)) were .$383,000,000 in 1934, .$490,000,000 in 193.'i, $874,000,000 in 1936. and $799,000,000
in 1937. See appendi.x note A-7 for method of securing tliese data. For the purpose of comparing changes
in shares of earning and purchasing power, t he fact that actual changes for these years differ somewhat from
those shown by the data is not particularly important as both sets of data are biased to same degree. In
1936 the rise in concentration is slightly greater than shown. See appendix note B-2.

CONCENTRATION OF ECONOMIO POWER g5

group. ^^ The share of purchasing power available to this group in 1936
was less than in 1934 while the share of earning power was more
than 13 percent larger. This divergence between the movements in
the concentration of earning power and purchasing power is largely a
consequence of the mcrease in surtax rates in 1936. The general rise
in incomes also played a part, as this factor can account for an increase
in the tax rate for these groups of income recipients with no change in
the rate structure.

II. RELIEF AND THE VETERANS' ADJUSTED SERVICE PAYMENTS

In recent years the receipt of direct and work relief and the veterans'
adjusted service compensation has introduced new elements into the
income structure of the Nation. These sources of income possess
characteristics which make it desirable to prepare additional measures
of income concentration. Statistics of the concentration of income
will be presented in this section which include and exclude these three
types of income. Individual income is taken as received; that is,
prior to the adjustments for purchasing power. Before the statistics
are given, the characteristics of these types of income will be analyzed
briefly in the light of the discussion in chapter II on the objectives
of measuring the concentration of income. While exception may be
taken by some to parts of the following analysis, it is believed that the
measures of concentration according to the several income totals are
of interest in that^ they enable the individual to select the total best
suited to his particular purpose.

In chapter II, work-relief wages were included as income when
the concentration of earning power or acquired income was under
consideration. In doing so, it was recognized that tliis type of wage
income, though similar in some respects to the same type of income
derived from private and public employment at "regular" govern-
mental functions, possesses certain distinguishing features. The
most important for the present purpose is that the wage payment to
an individual, wlule subject to variation according to the class of work
and to the region, was largely determined by criteria related to need.
For most of the period, the rates of pay for the various classes of
work were equal to the rates prevaiUng in private industry, but the
total wage payment to an individual was limited through control of
hours of work. The pei*sonnel on work relief projects were largely
drawn from famiUes certified as being in need of reUef and the type of
project was to a considerable extent determined by the certified per-
sonnel. It seems proper to conclude, therefore, that while work-rehef
wages were paid for personal services, neither the size of the payment
nor the services were determined by forces witliin the exchange
economy which, for the most part, directly determined them in private
employment and to a considerable extent, though indirectly, in regular
public employment.^^

" Shares of purchasing power available to the highest Moo of 1 percent of income recipients were 1.27 per-
cent in 1934, 1.26 percent in 1935, 1.17 percent in 1936, and 1.10 percent in 1937. Federal income taxes paid
by this group in 1934 were $235,000,000 in 1935, $295,000,000 in 1936, $502,000,000; and in 1937, $464,000,000.
Seeappendix note A-7for method of securing these data. For data on concentration of earning power see
table 3, p. 23.

" The problem which work-relief earnings present for this study should be distinguished from the problem
these earnings present for estimates of the national income. For estimates of the national income the
answer to the question as to what portion of work-relief earnings should be included in the national income
is to be found in an evaluation of the goods and services produced by work-relief projects. The problem
work-relief earnings raises for this study is whether in measuring the concentration of earning power it is
proper to include these earnings in the total income of all individuals. As indicated above, a different
set of considerations is relevant for the latter problem.

256129— 40~No. 4 6

^g CONCENTRATION OF ECONOMIC POWER

Largely because of the fact that this wage income was paid in return
for personal services and because of its importance both directly as a
source of income and indirectly through its influence on the income
structure, it was beheved that the measures of the distribution of
earning power would possess greater significance if work-rehef earn-
ings were included in the total income of all individuals. In view of
the distinctive character of this income, measures of income concen-
tration will be presented in this chapter excluding work-relief wages.
However, the data indicate that the inclusion or exclusion of work-
rehef wages results in but a small difference in the degree of income con-
centration and in the magnitude of the year-to-year changes. Because
of the direct and indirect effects of work-relief wages on the whole
income structure, care must be exercised in interpreting the measures
of income concentration, excluding work-relief wages. In particular,
it cannot be assumed that the concentration sho\v n by the data would
have prevailed had there been no made-work program. When work-
relief wages are included in total income and direct relief excluded, as
in chapter II, the shifts in the program of the Federal Government
with respect to the relative importance of work relief and direct rehef
as a means of deahng with the unemployment problem will influence
the resulting trends in concentration of earning power as shown by
the figures.

It seems evident that direct relief should be treated as a transfer
item and excluded from income when studying the distribution of
earning power. While the other transfer items involve for the most
part ^* deductions from the incomes of some individuals and additions
to the income of others, the method adopted in the financing of direct
relief by borrowing involves the creation of a transfer income item
without a commensurate deduction from the incomes of either indi-
viduals or business enterprises.^^ If the funds for direct relief were
obtaiTied by taxation it would be necessary to deduct the amounts
from the incomes of the taxpayers. For some types of income analysis
the distinction between transfer incomes financed by taxation and by
borrowing may be unimportant. However, for the present purpose
this distinction is crucial. Although the money expended for direct
relief was obtained from the savings of the purchasers of the bonds,
for the present purposes this type of transfer cannot be considered as
a deduction from their incomes. The incomes received in the given
year by the investors were not diminished by the fact that they chose
to buy a Government bond. On the other hand, these funds provided
a source of income for the recipients of relief. In the case of a gift
from one individual to another the donor's income may be said to be
reduced by the amount of the gift, as he has no claim to the money
spent for the gift. However, when a transfer income is created by
the Government borrowing and disbursing funds, the individuals
providing the funds receive a claim for the return of the funds in the

" The other exceptions are gifts paid from the accumulated savings of the donor or free public services
made available to individuals which are financed by the taxation of business enterprises. Direct relief
payments of a given year may be considered as the disbursement of the past and current savings of those
•who purchased the Government bonds.

21 The assumption that direct relief payments were financed by borrowing requires, perhaps, some ex-
planation. In general, there is no earmarking of Federal Government receipts for particular purposes.
Inasmuch as relief expenditures were classified as emergency expenditures and at the same time as direct
relief was disbursed, additional funds were obtained by borrowing, it may be presumed that the direct
relief payments of the Federal Government were financed by borrowing. In the case of State and local
governments, relief disbursements were financed to a considerable extent by general tax revenues with
emphasis on sales taxes. Therefore, to the extent that funds raised in this manner are included in direct
relief disbursements there is a duplication of incomes.

CONCENTRATION OF ECONOMIC POWER

67

form of a bond which is, of course," similar in this respect to other
bonds. In this restricted sense total individual income may be said to
be increased by the disbursement of direct relief. The method of
financing relief by borrowing may result in a different distribution of
income in future years than would otherwise prevail, but this aspect
is not relevant for the present purpose. In addition, the expenditure
of direct relief funds has an indirect effect on the general level and
volume of income.

The veteran's adjusted-service compensation also is excluded from
acquired income but on different grounds than direct relief. These
payments may be considered as part of the compensation for services
rendered at the time of the war. As such they represent a payment
which was separated by many years from the performance of the
service. While other items included in acquired income, such as
pensions, also possess to some degree this characteristic, the unusually
large amounts involved and the nonrecurrent and noneconomic
nature of this disbursement point to the desirability of excluding this
income in studying the concentration of earning power. Measures of
income concentration will be presented in this section, excluding w-.rk
relief, direct relief, and the veterans' bonus, and including these income
items. Table 20 shows the amounts of these tlu-ee types of income
and the percentages they constituted of total income.

Table 20. — Direct and work relief and veterans' adjusted-service compensation

payments, 1929-37 »

Type of income

1929

1930

1931

1934 <

1935

1936

1937

Inmillions of dollars

Direct relief -.

48

84

204

696

1,430

24

948

1,339

24

660
2,383
1.428

876

Work relief- _

1.739

Adjusted-service compensation '

912

132

Total

48

84

1,116

2,150

2,311

4,471

2,747

As percentages of total income

Direct relief

0.1

0.1

0.4

1.4
2.8

1.7
2.4

1.0
3.7
2. 2_

1.2

Work relief

2.4

Adjusted-service compensation '

1.6

.2

Total

.1

.1

2.0

4.2

4.1

6.9

3.8

I In 1932, direct relief payments were $444,000,000 and the veterans' adjusted service payments $144,000,000.
In 193"? direct-relief payments were $552,000,000, work-relief wage $640,000,000, and veterans' adjusted-service
payments $60,000,000,

' Veterans' adjusted service compensation is included in the year during which loans were made on the
certifieates and the final payments on the balance were made.

' I^e.ss than 0.1 percent.

< The various types of income constitute the same percentages of the 1934 total income which is comparable
to the total income figures for the preceding years.

Source: Data on various types of relief and bonus payments from the National Income Division, Depart
ment of Commerce. Total mdividual income is taken from table 10, less work-relief earnings.

1. Statistics of Income Concentration, Excluding Work-Relief Wages,
1934-37.

Table 21 contains for the years 1934 through 1937 the shares of
total individual income received by the selected proportions of income
recipients as these were defined in chapter II with the exception that
work-relief wages are excluded from total income. As the exclusion

68

CONCENTRATION OF ECONOMIC POWER

of work-relief wages affects only the total income of all individuals,
and not the incomes of the selected proportions of income recipients,
the differences among the various proportions in the magnitude
of the year-to-year changes in income shares are the same as shown in
table 3 and discussed in connection with that table. The income
shares for each group are, of course, somewhat larger than when work
relief is included in total income. The greatest difference is in 1936
whet) work-relief payments constituted a larger proportion of total
individual income than in any of the other years from 1934 through
1937. In 1936 the highest 1 percent of income recipients received
14.53 percent of the total individual income when work-relief earnings,
are included (table 3) and 15.09 percent when this income is excluded
(table 21).

Table 21. — Shares of total individual income, excluding relief and veterans' adjusted-
service vayments, received by selected proportions of income recipients, 1934-37 '

Group of income recipients

Year

Highest
2 percent

Highest
1 percent

Highest
H of 1 per-
cent

Highest
1/10 of 1
percent

Highest
1/100 of 1
percent

Percentages

1934 ,

17.92
18.26
19.50
17.81

13.42
13.74
15.09
13.63

10.15
10.44
11.60
10.40

5.18
5.35
5.96
5.28

1.80

1935

1.85

1936 -

2.01

1937

1.82

Indexes (1934=100)

\034

100.00
101.90
109. 15
99.39

100. 00 100. 00
102. 38 102. 86
112.44 114.29

101. 56 102. 46

100.00
103. 28
115.06
101. 93

100.00

1936

102. 7S

1936

113. 33

1937

101. 10

1 Economic incomes with the statutory realized capital gains and losses. Total individual income excludes
work relief, direct relief, and veterans' adjusted-service compensation. For limitations of data resulting
from inclusion of statutory realized capital gains and losses rather than actual realized capital gains and
losses see appendix note B-2.

Source: Incomes of selected proportions of income recipients from appendix table A-2. Total individual
income, excluding relief and veterans' adjusted-service payments, was obtained by deducting from data in
appendix table A-3 the figures on work -relief wages given in table 20.

Work-relief wages as a percentage of total income varied from 3.7
percent in 1936 to 2.4 percent in 1935 and 1937. Because of this
varying importance, there are small differences in the magnitude of
the year-to-year shifts in income concentration depending on whether
work-relief wages are included in total income, as in chapter II, or
excluded from total income, as in table 21. Work -relief earnings were
a slightly larger proportion of total income in 1934 than in 1935 and
consequently the income shares of the higher income groups are
increased more in 1934 than in 1935 by the exclusion of work-relief
wages. The rise in income concentration from 1934 to 1935 is, there-
fore, reduced somewhat when work-relief earnings are excluded from
total income. In contrast to 1935. the increase in income concen-
tration in 1936 is greater when work-relief wages are excluded. This
follows from the fact that the exclusion of work-relief earnings results

CONCENTRATION OF ECONOMIC POWER 69

in a larger increase for 1936 than for 1935 in the size of the income
shares received by the higher income groups. For the 2-year period
of rising income concentration from 1934 through 1936, the exclusion
of work-relief wages results in a slightly larger increase in income con-
centration. There is an indicated rise of 12.4 percent in the income
share of the highest 1 percent, excluding work relief from total income,
and a rise of 11.5 percent, including work relief as income. .In
1937, the decline in income concentration is greater with work-relief
payments excluded from the total individual income. Largely
because the year of highest income concentration, 1936, was also the
year in which work-relief earnings were the largest percentage of
total income, the fluctuations in income concentration during this
period are slightly smaller when work-relief payments are included
in income than when these payments are excluded.

^. Statistics of Income Concentration, Including Direct and Work Relief
and Veterans' Adjusted-Service Compensation Payments, 1934-37.

Table 22 shows the size of the income shares of the selected pro-
portions of income recipients when work relief, direct relief, and the
soldiers' bonus are included in the total of individual incomes.-® As
indicated at an earlier point, tlie iuclusion of these items involves no
double counting and for this reason separate measures of income
concentration may be presented with no deductions for the transfer
of incomes. The inclusion of direct relief and, from some points of
view, work relief and the veterans' bonus involves a departure from
the concept of earning power or acquired income. The data m this
section are of interest in showing the effect of these payments on the
concentration of income among income recipients. The use of income
recipients as the units for distributing income when direct relief
payments are mcluded is not altogether satisfactory. It will be
recalled that when the term, "income recipient," was defined,'''^
persons receiving direct relief were not necessarily included as income
recipients. This term was restricted to persons usually in receipt of
income from personal services or from the ownership of property.
To the extent that direct relief was received by totally dependent
individuals or families without an income recipient, the concentration
of income shown in table 22 is understated.

2« It was assumed that the entire bonus was paid to the lower 98 percent of income recipients. Actually, a
small part of the bonus was received by the highest 2 percent but as the bonus was exempt from income
taxation, the amounts received are not included as income for the higher income recipients. The amount
of bonus received by the higher proportions of income recipients undoubtedly constituted a negligible addi-
tion to their incomes and no attempt was made to correct for its exclusion.

" See ch. II, pp. 12-13.

70

CONCENTRATION OF ECONOMIC POWER

Table 22. — Shares of total individual income, including relief and veterans' adjusted
service payments, received by selected proportions of income recipients, 19S4-S7 *

Oroup of income recipients

Year

Highest
2 percent

Highest
1 percent

Highest
H of 1 per-
cent

Highest
Hoofl
percent

Highest
Moo of 1
percent

Percentages

1934

17.16
17.50
18v24
17.12

12.85
13.17
14.02
13.10

9.72
10.00
10.81

9.99

4.96
5.13
5.56
5.07

1.72

1936

1.77

1936

1.91

1937 ..--

1.74

Indexes (1934=100)

1934

100.00
101. 98
106. 29
99.77

100.00
102. 49
109. 49
101.95

100.00
102.88
111.21
102. 78

100.00
103. 43
112. 10
102. 22

100.00'

1935 . . .

102. 91

1936 .

111.05.

1937

101. 16

' Economic incomes with the statutory realized capital gains and losses. Total individual income includes
work relief, direct relief, and veterans' adjusted-service compensation. For limitations of data resulting from
Inclusion of statutory realized capital gains and losses rather than actual realized capital gains and losses see-
appendix note B-2.

Source: Incomes of selected proportions of income recipients from appendix table A-2. Total individual
Income, including relief and veterans' bonus payments, was obtained by adding to data in appendix table-
A-3 the figures on direct relief and veterans' adjusted-service payments given in table 20.

Prior to 1934 the inclusion of relief and bonus payments does not
significantly afi'ect the measures of income concentration. In 192&
and 1930, the shares received by the highest 1 percent of income
recipients are virtually unchanged by the additional types of incomes.
In 1931 the income share of the highest 1 percent of income recipients
is reduced from 13.72 to 13.44 percent, when these income items are
included in the total income. This reduction is largely the result of
the loans of $912,000,000 to veterans on their adjusted-service cer-
tificates. In 1934 the income share of the highest 1 percent of income
recipients is 12.85 percent as compared with 13.42 when the relief
and bonus payments are excluded. The year of the largest difference
is 1936 when these forms of income accounted for 6.9 percent of total
individual income. Table 21 shows that in 1936 the highest 1 percent
of income recipients received 15.09 percent of the total individual
income and table 20 shows that when these types of income are
included their share declines to 14.02 percent. If work relief earnings
are included in the total income, the share of this group is 14.53
percent.

The inclusion of the relief and bonus payments diminishes to a
small extent the magnitude of the fluctuations in the degree of income
concentration. Thus, as shown in table 21, from 1934 through 1936
the income share of the highest 1 percent of income recipients rose by
12.4 percent and according to the data in table 22 by only 9.5 percent
when the relief and bonus payments are included in the total income
of all individuals. The decline in the degree of income concentration
during 1937 is slightly smaller when the total individual income
includes relief payments and the veterans' bonus than when these
items are excluded.

APPENDIX A
NOTES TO TABLES

Note A-1

This note describes the methods used in deriving the data presented
in tables 1, 2, 3, 4, and 5 of chapter II. The limitations of the data
and of the adjustments are also indicated.

INCOME DISTRIBUTIONS FOR THE HIGHER INCOME RECIPIENTS

The basic data on the higher incomes were obtained from the
tabulations of the Federal income-tax statistics which have been pub-
lished annually by the Treasury Department in volumes entitled
Statistics of Income. Basic tables 3 and 7 on individual income-
tax returns of the annual issues of the Statistics of Income provided,,
for the most part, the required information on the number of indi-
viduals in each income class and on the amount of their incomes.'"
Table 7 has the following title in recent issues of the Statistics of
Income: "Individual returns, 1936, by net income classes — sources-
of income and deductions and net income * * *." This table
presents the data by 34 net-income classes above $5,000 and, except
for the years 1918-26, in one income class from zero to $5,000. For
the years 1918 through 1926 the data below $5,000 are presented in
thousand-dollar income intervals. Table 3 of the Statistics of Income
presents for all years the number of returns and net income by thou-
sand-dollar income classes below $5,000 as well as by the 34 income
classes above $5,000. The data published in the Statistics of Income
are based on unaudited returns and include data from amended returns
showing net income of $100,000 and over; but not from amended
returns with net income under $100,000 or from tentative returns.
For net incomes of $5,000 and over the data are tabulated from each
return and for net income of less than $5,000 the data are estimates
based on a sample.^

/. Statutory Net Incomes of the Higher Income Recipients.

For the years 1918-23 and 1934-37 the statutory net incomes were
taken directly from table 3 of the Statistics of Income of each year.^
For the years 1924 through 1931 it was necessary to adjust the data

1 The income distributions presented in the Statistics of Income include the incomes of trusts and estates.
In 1935 data on trusts and estates were first tabulated separately. (See Statistics of Income for 1935, tabte
6.) With the statistics for 1935 and 1936 it can be easily shown that the inclusion of these data in the income
distribution has but a slight effect on the income shares received by the various proportions of income
recipients and that the eflect on the indicated changes in income concentration may be ignored. In 1935,
for example, statutory net income of the highest 1 percent of income recipients including estates and trusts,
constituted 11.1 percent of total individual income and, if the net incomesof estates and trusts are excluded,
the percentage is 11.0 percent. The incomes of estates and trusts are included in the total of individual
incomes. See appendix note B-3, p. 106, for further discussion.

1 Individual returns with net income under $5,000 filed on Form 1040, which display income charac-
teristics similar to those usually found in returns with net income of $5,000 and over, are also ta: ulated.
The data for net incomes under $5,000 were estimated on the basis of a sample for 1918 through . "^"7 and
for 1929, and partly estimated and partly tabulated for 1928, 1930, and subsequent years. For discu; sion of
validity of method used prior to 1928, see Statistics of Income for 1928, pp. 19-24. Except for some of the
earlier years, the annual issues of the Statistics of Income contain a description of the methods used of
tabulation and estimation. For the earlier years, see "Income forecasting by the use of statistics of income
data," Review of Economic Statistics, vol. XII, No. 2 (May 1930), by J. F. Ebersole, S. S. Burr, and O. M.
Peterson.

• The statistics for 1923 were revised subsequent to the publication of the Statistics o( Income for 1923 and
the revised figures were used. (See Statistics of Income for 1925, pp. 28-29.)

71

72 CX)NCENTRATION OF ECONOMIC POWER

for the realized capital losses on assets held over 2 years that were
reported for a tax credit and not deducted from other income. The
amounts of these losses were obtained by multiplying by eight the
tax credit, 12)^ percent of the loss, given in table 2 of the annual
issues of the Statistics of Income. These losses were deducted from
the aggregate net income of each income class. No attempt was
made to transfer individuals to lower income classes because of the
deduction of these losses. The deduction of these capital losses
would obviously result in some redistribution of the individuals re-
porting such losses among the various income classes. It should be
noted that when income concentration is measured by taking the
aggregate income of various proportions of income recipients the dis-
tribution of income within each proportion need not be accurate.
However, when the adjustments to income result in the transfer of
individuals from one proportion to another, the measures of income
concentration are affected. The law so operated that only those
with statutory net incomes of over approximately $30,000 (over
$25,000 in 1924) reported their capital losses on assets held over 2
years by the tax-credit method. With the available data ^ and with
the use of certain assumptions it can be shown that the inability to
transfer these income recipients to their proper income class results
in only a slight understatement of the aggregate incomes of the
various proportions of income recipients and, hence, of the percentages
of total income received by them during the years 1924 through 1931.
The understatement is negligible for the highest 1 percent or 2 per-
cent but becomes somewhat more important for the smaller propor-
tions of income recipients. More important for the purpose of
measuring changes in income concentration is the fact that the
understatement is greater in years of large realized capital losses.
This tends to exaggerate the increases and declines in income concen-
tration. However, as indicated above, there is evidence that the
understatement is so small that rVar(;7C£? in the degree of understate-
ment may be disregarded.* As a further consequence of the method

* The available data show the total amount of these capital losses (Statistics of Income, table 2); the dis-
tribution of the amount of these losses by net income classes (table 2); and the distribution of individuals
reporting a loss for tax credit by the size of the capital loss (available only for the years 1929-31, see, for
example, the text table in Statistics of Income for 1929, p. 12).

» To cite two of the examples which have been worked out to determine the importance of this limitation
tor the Indicated changes in income concentration: In 1929 the minimum statutory net income of the highest
1 percent of income recipients was $8,680. If we make the extreme assumption that all those reporting a
capital loss for tax credit had a statutory net income of $30,000, then, according to the text table on p. 12 of
the Statistics of Income for 1929, 439 individuals reported a capital loss for tax credit larger than $21,320.
With a capital loss of $21,320 these individuals would fall below the minimum income level for the highest
1 percent of income recipients. The above assumption is the most extreme one that could be adopted as
most of those reporting a loss for tax credit had incomes considerably larger than $30,000, in fact, 66 percent
•of the losses for tax credit were reported by individuals with net incomes of $100,000 and over. We can
then make another far-fetched assumption that the deduction of these losses would give all these 439 indi-
viduals incomes of zero; that is, the capital loss was as large as their other income. If this were so, the in-
come of the highest 1 percent would be understated by $3,811,000. This figure is derived by substituting
for the assumed incomes of the 439 individuals reporting a capital loss for tax credit the same number of
incomes at the minimum income level .'or this group, $8,680. With these extreme assumptions the propor-
tion of total income received by the highest 1 percent of income recipients would be five-thousandths of 1
percent larger. The data as given to two decimal places in table 1 would not be affected. When the same
assumptions are m:Kle for 1931, the percentage of total income received by the highest 1 percent of income
recipients woultl be increased by two-hundredths of 1 percent.

The comparison between these 2 years is one of the most extreme that could be made and when the above
data are compared with the indicated changes ih income concentration it is readily seen that this limitation
may be ignored. Similar calculations may be carried out for the other proportions but the combination
of assumptions becomes too complicated for presentation. In 1929, 3,111 individuals reported a capital loss
for tax credit, in 1930, 4,318, and in 1931, 5,593. The number of individuals reporting such losses for the
years 1924 through 1928 was, of course, less than in 1930 and 1931, but no data are available for these years.
The amounts of losses reported for tax credit on the sale of assets held over 2 years are given below in millions
of dollars for the years 1924 through 1931. The data are taken from table 2 of the annual issues of the Sta-
tistics of Income:

fin million of dollars]

1924 72 1927_ 48 1930 81

192.5 61 1928 41 1931 193

1926 35 1929 . 43

CONCENTRATION OF ECONOMIC POWER 73,

of treating the incomes of individuals reporting a capital loss for tax
credit during the years 1924 through 1931, the minimum income
levels for these years (table 4) are subject to some overstatement and
the average incomes (tables 7 and 8) to some understatement.

Due to a change in the -Revenue Act of 1934 with respect to the
treatment of realized capital gaijis and losses it was necessary to
adjust the data for that year in order to make the income concept
comparable with the preceding years. ^ This adjustment was made
possible by a special tabulation of the 1934 income-tax data, made
available by the Treasury Department, giving the full amount of
realized capital gains and losses. With this information and the
data on statutory capital gains and losses it was possible to add and
subtract, by income classes, amounts so that realized capital gains
and losses would be fully taken into account.^

2. Economic Incomes of the Higher Income Recipients.

To statutory net income the following items were added by income
classes in order to obtain economic income for the years 1926-31 and
1934, and for the years 1934 tlirough 1937: Contributions, taxes paid,
interest paid, tax-exempt interest on governmental securities, and
"other deductions" consisting largely of losses due to bad debts, un-
insured losses due to fire and theft, losses from worthless securities,
and various other deductions. The items, exclusive of tax-exempt
interest, consist of legal deductions from total income to arrive at
statutory net income and were taken from the annual issues of the
Statistics of Income (table 7).^ Tax-exempt interest on govern-
mental securities for the years after 1923 was secured from a table
in the Statistics of Income showing the interest received from tax-
exempt obligations.^ The reporting of this income was for informa-

• For an explanation of the change in the law, see appendix note B-2, pp. 101-2.

' As was the case with the deductions for reahzed capital losses reported for tax credit this adjustment
results in a slight understatement of the incomes of the various proportions of income recipients as it was
not possible to transfer individuals from one income group to another. That the understatement is smair
may be determined with the use of the special tabulation of actual capital gains and losses showing by-
income classes the number of individuals with a capital loss of $2,000 and over and the amount of actual
capital losses. The number of individuals with losses of $2,000 and over in each of the selected proportions
of income recipients is given in appendix table B-4, p. 104. As a further consequence of the limitations of
this adjustment, the first group of minimum incomes for 1934 in table 4 are overstated and the first group
of averages for 1934 in tables 7 and 8 are understated.

8 As the minimum statutory net incomes of the highest 2 percent of the income recipients fell below
$5,000 during the years 1930, 1931, and 1934 through 1937 it was necessarv to estimate theamountof deduct-
ions for the $4,000 to $5,000 class and for the $3,000 to $4,000 class (1931, 1934, and 1935 only). This was done
by extrapolating the ratios of the sum of the four deductions and tax-exempt interest to net income into-
these income classes.

Business and partnership losses were not tabulated separately for the years prior to 1930 but were included
with "other deductions." As net losses incurred by individuals from unincorporated businesses and
partnerships should be deducted in order to secure economic income, it was necessary to prei)are estimates
of these losses in order to exclude them from "other deductions" for the years 1926 through 1929. The
estimates were made in the following manner: The ratios of business and partnership lo.sses to "all other
deductions" were computed for the years 1930 through 1936 by net income classes. In this case taxes paid
and interest paid were added to "all other deductions" sis those deductions were not tabulated separately
until 1928. On the basis of the year-to-year movement of these ratios and on information relating to the
volume of business and partnership losses, such as the volume of corporate deficits and business savings of
entrepreneurs, years from 1930 through 1936 were selected which were thought to be comparable, with
respect to the volume and distribution of business and partnership losses, to each of the years from 1926
through 1929. Thus the ratios for 1936 were applied to 1929 and 1927 and the ratios for 1935 to 1928 and
1926. Because of the relatively small amounts of business and partnership losses any error in the estimate
would have an almost insignificant effect on the incomes of the various proportions of income recipients.
In 1931, for example, business and partnership losses for net incomes of $5,000 and over amounted to $72,-
000,000, or less than nine-tenths of 1 percent of the economic income of this group.

In order to secure the economic incomes of the highest 1 percent of income recipients for the years 1918
through 1925, it was necessary to estimate the business and partnership losses of this group. This was done
in somewhat the same manner as for the years 1926 tlirough 1929 except that the losses were estimated for
the $5,000 and over group as a whole, and then separately for several of the thousand-dollar income classes
above $5,000.

» In the Statistics of Income for 1927, for example, this table appeared as table 8, pp. 85-86. The column
headed "Total Interest Received" was used. As this total includes the taxable interest on partially tax-

74 C50NCENTRATION OF ECONOMIC POWER

tional purposes only and the amount of tax-exempt interest shown in
these tables is understated because the schedule frequently is not
completely filled out.'°

In order to obtain the economic incomes of the highest 1 percent
of income recipients for the whole period it was necessary to estimate
realized capital losses for the years 1918-25 and tax-exempt interest for
the years 1918-19, and 1921-23 as this information was not available
for these years." The estimate of realized capital losses for the years
1918 through 1925 were needed since during these years they were
tabulated in one classification with interest paid, taxes paid, con-
tributions, and "other deductions." (See appendix note A-5, p. 93,
for method of estunate.) Following is a description of the method for
estimating tax-exempt interest:

(a) Estimate of tax-exempt interest received jrom wholly tax-exempt
securities by those with net incomes of $5,000 and over, 1918, 1919,
1921-23. — Ratios were calculated for each of the years 1920 and
1924 through 1936 of interest received by this group from wholly
tax-exempt securities to the amount of such securities out-
standing.*^ On the basis of a study of these ratios in relation to
changes in surtax rates and the number of returns and net income
above $5,000 it was decided to apply the ratio for 1920 to the
amounts of securities outstanding during the years 1918 thi-ough
1919 in order to obtain an estimate of interest received from
wholly tax-exempt securities by the $5,000-and-o^4er group. The
amount of interest for the year 1920 was kept constant for 1921.
The ratios for the years 1922 and 1923 were obtained by inter-
polating between the ratios for 1921 and 1924.

(6) Estimate oj taxable and nontaxable interest received from
partially tax-exempt securities by those with net incomes of $5,000
and over, 1918, 1919, 1921-23}^ — These estimates were prepared
by a method similar to that above. The ratios of this interest
received to the amount of partially tax-exempt securities out-
standing was obtained for the years 1924 through 1936. On the
basis of the behavior of these ratios, it was decided to use the
1924 ratio in each of the preceding years.

The above method for estimating interest received from gov-
ernmental securities is obviously arbitrary, but the results seemed
reasonable both in absolute amounts and as a ratio to the net
income of those in the $5,000-and-over net-income class. The

exempt securities, as given in Statistics of Income, table 7, the amount given in table 7 was excluded from
■economic income. In the 1936 volume this information was contained in an unnumbered table on p. 30.
These data are given only for those with net incomes of $5,000 and over.

10 See Statistics of Income for 1936, p. 28.

" The statistics on tax-exempt interest for 1920 were published in the Annual Reporter the Secretary of
Treasury, for 1923, p. 383. These data are somewhat defective for the present purpose in that presumably
they include tax-exempt compensation received; that is, salaries of employees of State and local governmental
•units and judges of Federal courts. However, the aggregate amount of this income received in 1920 by indi-
viduals whose statutory net income was in the income brackets with which we are concerned is relatively
insignificant. In view of this and of the under reportinc of tax-exempt interest, the inclusion of this com-
pensation is an unimportant limitation of the data. These statistics for 1920 do not include the taxable and
nontaxable interest on partially tax-exempt securities and it was therefore necessary to prepare an estimate
for It.

>' Net outstanding wholly tax-exempt securitie.« minus short- term securities as of December 31 of each
year. These totals, therefore, exclude such securities in State and Federal sinking funds and short-term
secuntic" which are largely held by flnancinl institutions. Data from Annual Reports of the Secretary of
Treasury.

" The taxable interest received from these securities Is included as part of statutory net income and is
tabulated separately in Statistics of Income, table 7. It was not thought feasible to attempt to estimate the
nontaxable interest received from partially tax-exempt securities separately and, therefore, the total amount
of interest received from partially tax-exemjit securities was estimated. The taxable interest shown in table
7 was then excluded from economic income.

CONCENTRATION OF ECONOMIC POWER 75

estimates of interest received from wholly tax-exempt and
partially tax-exempt securities by individuals with net incomes of
$5,000 and over for the years 1918 through 1923 are as follows:

[Millions of dollarsj

1918 15811920 18811922 210

1919 191 I 1921 18611923 223

In order to secure the amounts of interest received by the high-
est 1 percent of income recipients, it was necessary to know the
distribution of this interest for several of the thousand-dollar-net-
income classes above $5,000.'* The percentage distribution of
this interest in 1920'^ was applied to the 1918, 1919, and 1921
totals and the percentage distribution in 1924 was apphed to the
1922 and 1923 totals.'*

It should be noted that the passage from statutory net income to
economic income is not completely satisfactory as the addition of the
four types of deductions and tax-exempt interest to statutory net in-
come would result in some change in the ranking of individual incomes;
that is, the relative position of individuals when classified according
to the size of their economic incomes would be somewhat different
from the position they occupy when the classification is on the basis
of statutory net income. What the statistics on income concentration
actually show, for example, is the proportion of total economic income
received by the 1 percent of individuals with the highest statutory net
incomes. In the analysis of the data it was assumed that the distri-
bution of individuals by their economic incomes is identical with the
distribution of individuals by their statutory net incomes. While there
is doubtless a very high degree of correlation between the distribution
of individuals by statutory net income and by economic income, the
proportion of total economic income received by the individuals with
highest statutory net incomes would obviously be somewhat less than
the proportion of total economic income received by those with the
highest economic incomes. As this study is concerned more with
changes in uicome concentration than with the actual degree of con-
centration in any one year, this small understatement is not important,
provided it is approximately the same from year to year. Unfortu-
nately sufficient information is not available with which to determine
statistically the extent of the understatement nor of the changes in
the degree of understatement from one period to another.''

There is reason to believe, however, that the understatement is
slight and that, therefore, any year-to-year variation in the degree of
understatement is relatively slight. The items involved in trans-
forming statutory net income to economic income are each compara-
tively small and are of such a nature that the size of the items is
closely related to the size of statutory net income. In addition, with

X For the $6,000 to $6,000 net-income classes in 1918 and 1921 and for the $5,000 to $6,000 and the $6,000 to
$7,000 net-income classes in tho other years.

n Wholly exempt interest from Annual Report of the Secretary of Treasury for 1923 and taxable interest
from partially tax-exempt securities from Statistics of Income for 1920 (table 7).

'« As the amounts of this interest received in $5,000 to $6,000 and the $6,000 to $7,000 net-income classes is
relatively small, less than $6,000,000 at the most, any error in the estimated distribution may be ignored
for the present purpose.

" Ideally the necessary data would show for several years in different phases of the business cycle the dis-
tribution of individuals according to their economic incomes; that is, their statutory net incomes including
actual, as distinct from statutory realized capital gains and losses plus the 4 types of deductions and tax-
exempt interest. As will be made clear by the following discussion, It would be highly desirable to secure
Income data in this form.

76 CONCENTRATION OF ECONOMIC POWER

the exception of tax-exempt interest, the major portion of individuals-
in each income class reported each type of deduction.'^

It may be noted again that when income concentration is measured
by taking the aggregate income of the various proportions of income
recipients, the distribution of income within each proportion need
not be accurate. However, when the adjustments to income result
in the transfer of individuals from one proportion to another, the
measures of income concentration are affected. In determining
the degree of understatement in the measures of income concentra-
tion the amounts to be taken into consideration are the differences
between the economic incomes of those included in the given propor-
tion of income recipients and those who would be included if the in-
comes were classified according to economic income rather than
statutory net income. It is believed that only a relatively small
proportion of income recipients would be shifted from one group to
another and that the net difference in the aggregate income is small.

Some knowledge as to the year-to-year constancy or variability in
the degree of understatement may be obtained by examining the
amounts by which economic income exceeds statutory net income in
different years. Such information is presented in appendix note B-1.
These data show that the amounts to be added to statutory net in-
come in order to secure economic income were relatively larger in
years of diminished incomes than in years of increased incomes. As
indicated in appendix note B-1 this seems to be due to the fact that
from year to year the deductions and receipt of tax-exempt interest
were more stable than economic income and, therefore, when income
declmed the deductions constituted a larger percentage of income. It
may be that the larger the difference between statutory net income
and economic income the smaller is the correlation between the dis-
tributions of statutory net income and of economic income. If this
were so, the degree of understatement would be greater in years of
low income. In terms of the measures of income concentration this
would mean that the declines in income concentration would be over-
estimated. Several considerations indicate that, if this is the case,
the extent to which the changes in income concentration are over-
estimated is probably quite small.

First, it should be pointed out that with a general decline in in-
comes the difference between statutory net income and economic
income could increase with no change in the correlation between the
two distributions. This increase would be due to the greater stability,
as compared with economic income, of the statutory deductions from
income. It seems reasonable to expect that a substantial part of this
increase is due to a change of this sort.

Second, the year-to-year differences in the ratios of the four types
of deductions and tax-exempt interest to statutory net income are
relatively small for the larger groups of income recipients. However,
as the group of income recipients becoro.es smaller these differences
increase. The data in tables B-1 and B-3 of appendix note B-1
(pp. 99 and 101) are not in the ro.ost appropriate form for showing

" Data showing the number of individuals reporting interest paid, contributions, and taxes paid are
available by income classes for individuals with net incomes of $5,000 and over for the years 1934 through
1936. In 1936, for example, the proportion of individuals in each income class reporting interest paid as a
deduction varied from 50 percent for the $5,000 to $6,000 net income class to 80 nercent for the $1,000,000 and
over class. For taxes paid the corresponding percentages are 81 percent to 97 percent and for contributions
the percentages are 74 percent and 85 percent. See tabic 7 of Statistics of Income for 1936.

CONCENTRATION OF ECONOMIC POWER 77

the year-to-year differences. These tables present the percentages
that the aggregate statutory net income received by the various
proportions of income recipients constituted of their economic income.
However, the difference between the distributions according to
econom.ic income and statutory net income is important insofar as
individuals who are included, within the highest 1 percent of income
recipients, for example, when the classification is according to statu-
tory net incom.e would not be included when the basis for the classi-
fication is economic income. This redistribution would take place,
for the most part, around the lower incom.e limit of the highest 1
percent of income recipients and therefore the difference between
statutory and economic income should be measured at this point.
According to the data on minimum income levels as given in table 1
for economic income and those given in table 4 for statutory net
income, the statutory net incomes at the lower income limit of the
highest 1 percent of income recipients were increased by 15 percent in
1928, the year of greatest income concentration, and by 16 percent
in 1920 and 16.6 percent in 1934, the years of lowest income concen-
tration. It is not believed that differences of this magnitude are
indicative of appreciable differences among the various years in the
extent to which a redistribution of individuals would take place about
the lower income limit of this group. These differences are larger
for the smaller proportions of income recipients, and hence any year-
to-year differences in the degree of understatement that exist are
greater for these groups.

STATUTORY NET INCOMES AND ECONOMIC INCOMES OF SELECTED
PROPORTIONS OF INCOME RECIPIENTS

Thus far, the method of obtaining the distribution by income classes
of statutory net income and economic income has been described.
The next step was to secure the amounts of income received by the
selected proportions of income recipients. Using the number of
individuals in the various proportions of income recipients, shown in
table 6, the minimum statutory net incomes and the aggregate amounts
of economic and statutory net incomes of these groups were obtained
by interpolation.^^

The aggregate statutory net income received by the selected propor-
tions of income recipients for the years 1918 through 1931 and 1934
through 1937 are given in appendix table A-1 and the aggregate
economic income for the years 1926, 1931, and 1934 through 1937
are given in appendix table A-2. The amount of economic income
of the highest 1 percent of income recipients is given in table 12 for the
years 1918 through 1925 as well as for later years. ^^ The minimum
statutory net incomes of the selected groups of income recipients are
presented in table 4. The minimum economic income levels for the
highest 1 percent of income recipients, presented in table 1, were

i« The method of interpolation generally used was first to draw, with the aid of a flexible rule, a cumulative
curve of the number of returns. This was done for six or seven income classes adjacent to the class in which
the minimum income of the given group of income recipients was located. Second, the minimum income
ievel of the group of income recipients was "read" from this curve by taking the level above which the
number of income recipients in the given income group was located. These income levels are presented in
table 4. Third, cumulative income curves were drawn in the same manner as the curves for the number
of returns. Fourth, the aggregate income received by each of the income groups was obtained by "reading"
from the curve, the amount of income above the given income levels.

" The amount shown in table 12 for 1923 is overstated by $61,000,000, See appendix note A-5, p. 89.

78

OONCENTRATION OF ECONOMIC POWER

obtained by determining the difference between statutory net income
and economic income at the given statutory net income levels and
adding this difference to the minimum statutory net income levels.
The minimum economic incomes for the years 1934 through 1937^
shown in table 5, were obtained in the same manner.

Table A--1. — Aggregate statutory net income of selected proportions of income

recipients, 1918-37 »

[Millions of dollars]

Group of income recipients

Year

Highest
2 percent

Highest
1 percent

Highest
^ofl
percent

Highest
Mo of 1
percent

Highest
Moo of 1
percent

1918 -

7,995

9.360

9,070

7,690

8,900

9,580

10, 330

12,936

13, 125

13,920

16,380

16,080

11, 040

8,320

6,840

7,160
8,060
10, 330
10, 210

6,125

7,240

6,820

5,740

6,815

7,175

8, 105

10,220

10, 360

11,065

13,320

12,990

8,450

6,120

6,040

5,310
6,085
7,940
7,735

4,775
5,575
5,118
4,280
5,197
5,465
6,200
7,990
8,053
8,710
10, 776
10, 485
6,452
4,520

3,762

3,975
4,575
6,065
6,840

2,678
2,890
2,455
2,042
2,627
2,705
3,078
4,260
4,355
4,820
6,375
6,307
3,415
2,205

1,867

1,953
2,282
3,040
2,883

936

1919 ---

990

1920 -

760

J921

615

1922

902

1923 --

915

1924 --

1,036

1925 ---

1,642

1926

1,727

1927

1,944

1928

2,769

1929

2,877

1930 ----

1,347

1931 --

780

1934

647

1934

652

1935

767

1936 -

1937

1,005
950

1 Due to a change in the definition of realized capital gains and losses the table is divided into 2 parts —
1918-34 and 1934-37. In the latter period varying proportions of gains and losses are included depending
upon the length of time the asset was held and losses were limited for each individual to $2,000 in excess of
gains. The first set of figures for 19^4 are comparable to those for the preceding years.

Table A-2. — Aggregate economic income of selected proportions of income recipients,

1926-37 1

[Millions of dollars]

Group of income recipients

Year

Highest
2 percent

Highest
1 percent

Highest
Mof 1
percent

Highest
Moof 1
percent

Highest
Moo of 1
percent

1926

15,285
16,080
18, 735
18, 515
12,990
10,050

8,373

8,710
9,720
12, 040
12,050

12,075
12, 825
m, 260
14, 970
10, 005
7,450

6,255

6,525
7,320
9,290
9,220

9,382
10,083
12,327
12, 070
7,675
6,530

4,730

4,935
5,558
7,140
7,035

5,075
5,600
7.269
7, 215
4,105
2,770

2,440

2,518
2,850
3,670
3,670

2,010

1927

2, 2.55

1928..

3,125

1929

3,245

1930

1,627

1931

1,02?
S6&

1934

1934

873

1935

985

19.36

1,26*

1937

1,228

' Due to a change in the definition of realized capital gains and losses, the tabic is divided into 2 parts—
1926-34 and 1934-37. In the latter period varying proportions of gains and losses are included depending
on the length of time the asset was held and losses were limited for each individual to $2,000 in excess of
gains. The first group of figures for 1934 are comparable to those for the preceding years.

CONCENTRATION OF ECONOMIC POWER

79

PERCENTAGES OF TOTAL INDIVIDUAL INCOME RECEIVED BY THE SELECTED
PROPORTIONS OF INCOME RECIPIENTS

These percentages, presented in tables 1, 2, and 3, were derived
from the data in the preceding two tables and from estimates of total
individual income. For reasons indicated in appendix note A-3, the
salaries and wages received by employees of State and local govern-
ments were excluded from total individual income. Estimates of total
individual income are given in table 10. The salaries and wages of
employees of State and local governments, given in appendix table
A-5, p. 83, were deducted from these totals. Appendix table A-3
contains the estimated total income of individuals excluding the
compensation of State and local governmental employees.

Table A-3. — Total individual income excluding compensation of State and local
governmental employees, 1918-37

[In millions of dollars]

Year

Amount

Year

Amount

Year

Amount

1918

$56, 621
64,612
67,850
53,129
57,585
66,657
67, 422

1925..

$72, 406
74, 484
74,654
79,224
81, 050
68,409
54,288

1934....

$49,400

1919

1926

1934

1920

1927

50,028

1921

1928

1935

54,561

1922

1929

1936 .

63,943

1923

1930

1937. .

69, 387

1924

1931

Source: Table 10 and appendix tables A-4.

Note A-2.^' Table 6 — Number of Individuals in the Selected
Proportions of Income Recipients

The total number of income recipients, except for the years 1918
and 1919, were taken from estimates of the number of persons with
gainful occupations by Mr. Daniel Carson of the National Research
Project, Work Projects Administration.^^ These are annual esti-
mates of the average number of persons with gainful occupations as
enumerated in the decennial census. The estimates account for the
changing age composition in the population, for immigration and
emigration, and farm-city migration. An adjustment was made in
the 1930 census figure for an estimated undercount of 325,000 young
people who had entered the labor supply but were omitted because
of lack of previous work experience. Mr. Carson has also made other
adjustments of the census data for an undercount of farm family
workers and farm hired laborers and for the seasonal labor supply,
but these adjustments were not used in the present study. The
figures for the years 1918 and 1919 were derived from estimates for
these years by Willford I. King presented in his volume, The
National Income and Its Purchasing Power (p. 47). From the total

>' In connection with this note see discussion in ch. II, pp. 12-13.

" These estimates are presented in a study of the National Research Project entitled Labor Supply and
Employment, Preliminary Statement of Estimates Prepared and Methods Used, (mimeograplied) by
Daniel Carson assisted by Henrietta Liebman. See table 46, p. 136. The total of the first 2 columns in
this table were used.

so

CONCENTRATION OF ECONOMIC POWER

number of persons with gainful occupations, the number of employees
of State and local governments, given in appendix table A-5, was de-
ducted. The resulting total is presented in the first column of table
6. Table A-4 contains the estimated number of persons with gainful
occupations for the years 1918 through 1937.

Table A-4. — Number of persons with gainful occupations, 1918-37

Year

Persons

1918 .-

42,193,000
42,092.000
41,818,000
42, 633, 000

43, 218, 000

44, 052, 000
45, 008, 000

1919

1920 .

1921 -

1922 -

1923

1924 . .

'' Year

Persons

1925

45, 722, 000
46,412,000
47, 126, 000

47, 845, 000

48, 555, 000
49, 327, 000
49, 931, 000

1926 ---

1927

1928

1929

1930

1931-

Year

1932
1933
1934.
1935
1936.
1937.

Persons

50, 503, 000
51,065,000

51, 687, 000

52, 329, 000
52, 937, 000

53, 561, 000

Source: See text.

It should be noted that the term "income recipient" can be applied
to the income tax data only with certain qualifications. Strictly,
what the data show are the number of returns in the various income
classes. Because of methods of classification and the legal definition
of an income receiving imit, an income tax return may not necessarily
correspond to an income recipient as defined in this report.

The inclusion of trusts and estates as individual returns has beeh
discussed at an earher pc nt (appendix note A-1, p. 71). In the
income tax data the incom s of husband and wife filing a joint return
appear as one income. To a considerable extent a joint return repre-
sents two income recipients. If interest lies in showing the distribu-
tion of income in a given year, it seems that it would be desirable to
make some adjustment for this treatment of incomes. In the two
studies of the distribution of income in a single year cited in chapter II,
however, no mention is made of any adjustment. For the present
study, this type of adjustment is not important for two reasons.
First, the determination of the exact concentration of income in any
given year is not the primary objective, and, second, it seems probable
that joint returns are not generally filed in the income classes for wliich
the income tax data have been used. If a husband and wife have two
incomes and the sum of these two incomes places the combined income
in the surtax brackets, a tax saving would result from the filing of
separate returns and, doubtless, in most cases, individuals take account
of this and file separate returns. In all the years covered in this study
the minimum statutory net income of the highest 1 percent was weU
above the minimum income for which surtax rates are applicable or
the minimum income at which the normal tax rate is increased. This
is also true of the minimum statutory net income of the highest 2
percent with the exception of the years 1918, 1931, 1934, 1935, and
1936 when the rates were increased on net incomes above $4,000 and
the minimum net incomes were, respectively, $3,980, $3,960, $3,275,
$3,460, and $3,900 (table 4, p. 26). It may be, therefore, that in
these years there is some very slight overstatement in the income
shares shown for the highest 2-percent group. It should, perhaps, be
mentioned that if one spouse incurs a loss, a joint return may be filed
and, therefore, a joint return may appear in the income tax data
instead of a separate return of somewhat larger size.

CONCENTRATION OF ECONOMIC POWER gj

A related question concerns the community property laws of eight
States. Under these laws the incomes, with certain exceptions, of
both husbands and wives are evenly divided for income-tax purposes.
The question that is of present concern is the effect of this treatment
on the indicated shifts in income concentration. A study has been
made taking into consideration the number of States having these
laws over the period, the amounts of income involved, and the methods
of tabulating these returns. This study disclosed that there would
be a very slight effect on the indicated measures of income concentra-
tion of the changes in the above three factors. The most important
was the change in the California law effective in 1928 and under
extreme assumptions it can be shown that any decrease in the aggre-
gate income of the highest 1 percent due to the change in the method
of reporting may be ignored for the present purpose.

Another problem is introduced by the separate returns of husbands
and wives. To some extent wives filing separate returns on income
from property would not be classified as persons with gainful occupa-
tions and would, therefore, be excluded from the census enumeration,
and hence from the basic data on the number of income recipients.
For the present study, the relevant question is whether there has been
a disproportionate increase in these income recipients excluded from
the basic data on income recipients. A second and somewhat di*fferent
c(uestion relates to the division of uicome between hubsand and wife
for the purpose of minimizing tax liability and the effect of this on the
indicated shifts in income concentration. In these connections the
following data may be cited: The percentage of incomes included
in the highest 1 percent represented by those of wives filing separate
returns has been quite small, varying from 3.2 percent in 1919 to
7.4 percent in 1928 and the average for the years 1934-36 was 5.3
percent, only slightly higher than the average of 4.7 for the years 1918
through 1924. The increase has been somewhat larger for the smaller
proportions of income recipients. Part of this rise is due to the increase
in the proportion of women with gainful occupations and, therefore,
the data on the number of income recipients would take account of
these income recipients. (According to the 1920 census, 23.6 percent
of the women between the ages 20 to 64 reported an occupation and
the 1930 figure is 26.2.) Another factor in the increase is the changing
status of women as this is reflected in the separate maintenance of
property either held before marriage or inherited after marriage.
Presumably, the census data would not take account of any increase
in this practice. For some further discussion of the effect of the
division of income between husbands and wives on f'^e measures of
income concentration see appendix note B-3 (p, 104).

Note A-3. Compens.^tign and Number of Employees of State
AND Local Governments

Appendix table A-5 contains the estimated number of employees of
State and local governments and their compensation. As indicated
in chapter II, the compensation of State and local governmental em-
ployees was exempt from income taxation during the period covered
by this study. Information was not available on the frequency dis-
tribution of the incomes received by this group during the years
covered by this studj^ and these income recipients were also excluded

256149 — 40— No. 4 7

32 CONCENTRATION OF ECONOMIC POWER

from the data on total income and the number of income recipients.
The measures, therefore, actually cover the degree of income concen-
tration among income recipients exclusive of employees of State and
local governments. On the basis of a study of the compensation of
State and local government employees, considering such factors as the
movement of their average incomes in relation to the minimum in-
comes of the highest 1 percent and the relation of this compensation
to total income, it can be shown that the measures of income concen-
tration presented in this study can be used for determining the changes
in the concentration of income among all income recipients. This
seems to be the case for comparisons between different years as well
as for year-to-year changes. While the fact that the incomes of these
employees are more stable than other incomes would tend to result in
a smaller year-to-year variations in income concentration than shown
in this study, the actual effect is so slight that it may be disregarded.
For example, including with the income tax statistics an estimated
distribution in the higher brackets of the compensation of these em-
ployees for both 1929 and 1930 (see p. 97 for source), the decline in
the income share of the highest 1 percent in 1930 was 22.6 percent as
compared with 23.0 percent given in table 2 of this report or 23.3 percent
when the alternative method outlined below is used. In view of the
fact that the compensation of these employees declined from 1929 to
1930, this comparison, wliich assumed no change in the distribution
in the higher income brackets, exaggerates somewhat the actual
difference.

The year-to-year differences in income concentration are approxi-
mately the same whether this group is included in the data on total
income and on the total number of income recipients or excluded as
in the statistics in chapter II. The difference between the two meth-
ods in the size of the income shares is small. If the former method is
used the income share of the highest 1 percent of income recipients in
1936 is 14.04 percent as compared with 14.53 percent given in table 3,
and the income share of the highest one one-hundredth of 1 percent is
1.91 percent as compared with 1.97 percent shown in table 3. It
may be noted that the method used has some advantage in that it
raises the minimum income level for the larger groups of income
recipients and so obviates the necessity of going further down in the
income scale where the income tax statistics are less reliable due to
nonreporting and underreporting of incomes. Tliis is particularly true
of 1931 and 1934 in which years it is necessary to approach closely the
minimum income for wliich it is required to file a return. In addition
to nonreporting or underreporting of income for the purpose of avoid-
ing taxation, reporting of incomes just above tliis minimum income is
believed to be incomplete for another reason. To some extent in-
dividuals fail to file returns on nontaxable incomes above the mini-
mum, even though returns are required by law. Various exemptions
raise the size of a nontaxable income above the minimum for which a
return is required. In these income ranges a difference of several
hundred dollars in the minimum income of a group of income recipients
may be important.

CONCENTRATION OF ECONOMIC POWER

83

Table A-5.— Compensation and number of employees of State and local governments

1918-37 1

Year

Compensation

Number of
employees

Year

Compensation

Number of
employees

1918

$1, 369, 000, 000
1, 656, 000, 000
1, 945, 000, 000
2, 214, 000, 000
2, 382, 000, 000
2. 521, 000. 000
2, 681, 000, 000
2, 834, 000, 000

2, 001, 000, 000

3, 221, 000, 000

1, 732, 000
1, 786, 000
1, 843, 000
1, 865, 000
1, 918, 000

1, 988. 000

2, 080, 000
2, 146, 000
2, 203, 000
2, 274, 000

1928

$3, 388, 0./0, 000
3,541,000,000
3, 641, 000, 000
3, 609, 000, 000
3, 540, 000, 000
3, 164, 000, 000
3, 132, 000, 000
3, 272. 000, 000
3, 471, 000, 000
3, 650, 000, 000

2, 339, 000
2, 386, 000
2, 482, 000
2, "493, 000
2,471,000
2, 404, OOO

1919

1929 ...

1920 _-

1930

1921

1931

1922

1932

1923

1933

1924

1934

2 427 OOO

1925

1935

2,481,000
2, 574, 000
2, 637, 000

1926

1930

1927

1937

1 1929-37 from National Income Division, Department of Commerce. The 1929 estimate was extrapolated
backward until 1919 on the basis of unpublished data furnished by the National Bureau of Economic Ke-
search. The estimates for 1918 were based on data in National Income and its Purchasing Power, by WiU-
fordl. King, (p. 361).

Note A-4. Table 10. — Composition of Total Individual Income,

1918-37

For the period 1929 tlii'oiigh 1937 the data are based primarily upon
the estimates by the Department of Commerce of income paid out.^*
Changes in these estimates have been made in order to secure a
measure of the total of individual incomes as distinguished from the
concept underlying the estimates of income paid out. These changes
are noted below. For 1919 through 1928 the data are based upon
estimates by Simon Kuznets in National Income and Capital Forma-
tion, 1919-35, and for 1918 upon estimates by Willford I. King in
the National Income and Its Purchasing Power. Both of these
volumes are publications of the National Bureau of Economic Re-
search. In general, the estimates for the years prior to 1929 were
obtained in the following fashion: First, the estimates for each type
of payment were adjusted so as to conform to the income concepts of
the Department of Commerce, Second, the 1929 estimates of the
Department of Commerce were extrapolated on the basis of the move-
ment of the estimates of the National Bureau of Economic Research.
Third, the estimates were adjusted slightly to secure a measure of the
total of individual incomes. For the most part, the extrapolation of
the Department of Commerce series was done by Charles L. Merwin,
Jr., of the National Income Division, Department of Commerce.
It should be noted the estimates of total individual income are baaed
on a broader statistical foundation for the more recent years than for
the earlier years. The extrapolation back to 1919 is superior, for tb"
present purpose, to that for 1918 as the estimates through 1919 are
based on concepts and classifications quite similar to those of the
Department of Commerce. The figures for net rents and royalties
are probably subject to the largest margin of error, followed, in order,
by extrepreneurial income, interest, employee compensation, and
dividends.

" For a discussion of the concept of income paid out see any of the bulletins of Department of Commerce
on national income. Themostrecent is entitled Income in the United States, 1929-37, by Robert R.Nathan.
The estimates for 1929-37 of income paid out used in this study are those published in the June 1939 issue
of the Survey of Current Business.

g4 CX)NCENTRATION OF ECONOMIC POWER

As neither the Department of Commerce nor the National Bureau of
Economic Research has prepared estimates of reahzed capital gains
and losses of individuals, it was necessary to prepare such estimates
for this study. The method and limitations of the estimates of
realized capital gains and losses presented in this table are described
below. Following is a brief description of the adjustments of the
Department of Commerce estimates of income paid out.

COMPENSATION OF EMPLOYEES

Veterans' pensions were added to the following income items which
are included as employees' compensation in income paid out: Sal-
aries, wages, fees, commissions, etc., in cash or in kind including
work-relief wages; workmen's compensation benefits; pensions paid
in private industry; employee and employer contributions under the
Social Security and Railroad Retirement Acts; employee and em-
ployer contributions to Federal, State, and local government em-
ployees' retirement plans.

Thus this source is composed of all payments to employees for
present and past services; and since 1934, accruals to the benefit of
employees under certain pension and unemployment insurance plans.

ENTREPRENEURIAL NET INCOME

This source is the algebraic sum of entrepreneurial withdiawals,
included in income paid out and the business savings of entrepreneurs.

REALIZED CAPITAL GAINS AND LOSSES

These estimates are based principally upon data on the realized
capital gains and losses reported by individuals to the Bureau of
Internal Revenue and published in table 7 of the annual issues of
the Statistics of Income. This is practically the only direct source
of information on the volume of this type of income. However, as
the bulk of this type of income is received by individuals filing income-
tax returns any errors in the estimated amounts received by those
not filing returns are not believed to seriously affect the total estimate.
Because of the probable nonreporting of some of this income to the
Bureau of Internal Revenue, particularly in the lower brackets, the
data underestimate the total amounts actually received. These
estimates, however, are not characterized by the same degree of
accuracy as those for the other sources.

The realized capital gains and losses for individuals with net in-
comes above $5,000 were taken largely from table 7 of the annual
issues of the Statistics of Income. The net realized capital gain or
loss is the algebraic sum of the following items as classified in Sta-
tistics of IncomQ: (a) "Profit from the sale of real estate, stocks,
bonds, etc., other than taxed as capital gain from assets held more
than 2 years" (table 7); (6) ''capital net gain from the sale of
assets held more than 2 years" (table 7). This separate classification
was in use from 1922 through 1933; (c) "Net loss from the sale of
real estate, stocks, bonds, etc." (table 7); (d) "loss reported for tax
credit on capital net loss from sale of assets held more than 2 years"
(table 2). It was necessary to multiply the data on tax credits by

CONCENTRATION OF ECONOMIC POWER 35

8 as 12J2 percent of the loss was taken as the tax credit. This separate
classification was in use from 1924 through 1933.

The amounts of gains and losses received by all individuals with
positive net incomes below $5,000 were estimated by the method
indicated below. Since 1927 the realized gains and losses of indi-
viduals filing returns with negative net incomes have been pubUshed
in a special table in the Statistics of Income. For earlier years the
amounts were estimated. In addition it was necessary to prepare an
estimate of realized capital losses received by individuals with net
incomes of $5,000 and over during the years 1918 through 1925 com-
parable to data published for later years. A description of this esti-
mate is presented below. For these years the losses were not tabulated
separately in the Statistics of Income but were included with other
deductions from total income.

The estimate of realized capital loss for 1934 comparable in defini-
tion to the estimates for 1918 through 1931 was made with the aid of
an unpublished special tabulation of actual realized capital gains and
losses in 1934 which was made available by the Treasury Depart-
ment. This tabulation presents data on actual capital gains and
losses as distinguished from the data on statutory gains . and
losses published in the 1934 issue of the Statistics of Income. Statu-
tory realized capital gains and losses include varying proportions of
realized capital gains and losses depending on the length of time the
asset was held and the capital loss for each individual was limited
to $2,000 in excess of gains.

The definition of capital assets is slightly more inclusive beginning
in 1934. Included as a capital asset before 1934 was all property not
connected with the taxpayer's business or trade. In 1934 the definition
was broadened to include all property held, whether or not connected
with trade or business, except stock in trade, property which would
be included in inventory, or property held for sale in ordinary course
of business. Consequently for the years 1934 through 1937 the defi-
nition of capital asset dift'ered from that of earlier years, in that prop-
erty connected with trade or business, exclusive of stock in trade, was
included within the definition. The basic data for the 1934 estimate
also differ slightly from the data for previous years due to the method
of classifj'^ing realized capital gains received through partnerships and
fiduciaries. Prior to 1934 that portion of capital gains and losses
from the sale of assets held over 2 years received through partnerships
and fiduciaries and classified by individuals filing returns as ''capital
net gains"' (1922-31) or as tax credit for loss (1924-31) Avas not in-
cluded with partnership or fiduciary income (as was true of all other
realized capital gains and losses) but with realized capital gains and
losses. Beginning in 1934 all realized capital gains and losses received
through partnerships and fiduciaries were classified with those sources.
There is no way of determining on the basis of avnilable information
the net effect of these two factors on the comparability of the estimate
of realized capital loss for 1934. To some extent these two changes
oft'set each other.

As has been already indicated the estimates of realized capital
gains and losses for the years 1934 through 1937 are not comparable
to the estimates for previous years. In addition to the two differences
mentioned directly above, varying proportions of realized capital gain?

86

CONCENTRATION OF ECONOMIC POWER

and losses depending on the length of time the assets were held, with a
limit of $2,000 on the size of net loss, were included in the income-tax data.
Thus 100 percent of the gain or loss on assets held 1 year or less is
taken into account, 80 percent on assets held 1 year but not over 2
years, 60 percent on assets held 2 years but not over 5, 40 percent on
assets held 5 years but not over 10, and 30 percent on assets held
over 10 years.

No attempt was made to correct the income tax data for under-
reporting and nonreporting of realized capital gains to the Bureau of
Internal Revenue. It is generally believed that such income is subject
to a considerable degree of underreporting and nonreporting and the
amounts of realized capital gains are consequently understated for
each year. As the chief interest of the present study lies in year-to-
year movements rather than in correct absolute totals for any year, the
understatement, provided it is of the same degree each year, is not a
serious defect.

The estimates include the realized capital gains and losses received
directly by individuals and do not include the realized capital gains and
losses of those institutions designated in national-income studies as
aggregates of individuals. (These include savings, building-and-loan
associations, mutual savings banks, mutual insurance companies,
and other associations for the collective savings and investment of
funds). As the other property income received by these institutions
is included in the estimates, the treatment accorded realized capital
gains and losses may seem inconsistent. However, there is reason to
believe that the relatively small amount of such income received by
these associations is treated in a different manner than other property
income. Presumably, according to conservative accomiting practice,
the gains would be segregated so as to offset losses of later years. Over
long periods the net gain or net loss may be received by the members
of these associations, but over short periods it is doubtful if the income
position of the members is much affected by such gains or losses. At
any rate, largely because of the type of assets held by these institu-
tions, the amount of gains and losses realized are relatively small. This
is shown by the following tabulation of realized capital gains and losses
of certain of these associations taken from a detailed industrial tabu-
lation of corporate income-tax returns similar to that published by
broad industrial groups in the annual issues of the Statistics of
Income. It should be noted that the years for which data are avail-
able were characterized by abnormally large realized capital gains or
losses. No data are available for mutual savings banks and savings,
building-and-loan associations.

[In millions of dollars]

1929

1930

1931

1934 1

1935'

19361

National and State banks

45 1

35 1

(2) '

15.0

-14.0

-.5

-93
-70
-2

-8

60

121

Life insurance companies.

other insurance companies

-1

-2

9

Total...

80

.5

-165

-9

58

130

» $2,000 limitation on loss.

' Losses not available, gains were 0.2.

CONCENTRATION OF ECONOMIC POWER §7

Methods of Estimate.

(a) Realized capital gains and losses of individuals with net incomes
under $5,000. — These estimates were obtained by extrapolating
separately, by thousand-dollar-income classes, the amounts of gains
and losses for the net-income classes below $5,000. The algebraic
sum of the estimates of realized capital gains and losses was taken as
the net realized gain or loss. Data from table 7 of the annual Statistics
of Income were used for the income classes above $5,000. The
extrapolation curves were drawn with the aid of a flexible rule and with
the following information wliich was utilized in improving the accuracy
of the extrapolation:

1. As the exemptions for the Wisconsin State income tax are
lower than those of the Federal income tax the behavior of gains
and losses in the lower-income classes was studied. The data are
published in Wisconsin Individual Income: 1936 Income,
volume 1 .

2. The unpublished Treasury tabulation for 1934 which pre-
sented actual gains and losses by $500 class intervals below $5,000.
This information was useful in view ol the fact that the statistics
by source of income are presented in the annual Statistics of
Income by income classes only for net incomes above $5,000
except for the years 1918 through 1926. These data are useful
onl}^ above the exemption limit wliich in 1934 was $2,500 for a
married person and are limited for the classes immediately above
by the nonreporting of incomes and the understatement of gains
which characterize these income classes.

3. The estimates obtained by separately extrapolating the
amounts of capital gains and losses were checked by extrapolating
into the lower-income classes the ratios of gains and losses to net
income. Estimates of the distribution of income below $5,000
by the National Bureau of Economic Research for 1918 and by the
Brookings Institution for 1929 were useful in determining the
approximate distribution of incomes in the classes from zero to
$5,000. Estimates of the average income of all income recipients
were used to vary the distribution from year to year. The
information derived from extrapolating the ratios was especially
helpful in guiding the extrapolation by amount of gain and loss
in the lowest income classes.

As is readily seen the estimates for each year of realized capital
gains or losses for income recipients with net incomes under $5,000 are
subject to understatement as the data used for extrapolation are under-
stated. Since the cliief interest of the present study is in year-to-year
movements rather than showing correct absolute totals, the under-
statement, provided it is of the same degree each year, is not a serious
defect.

(6) Realized capital losses: 1918-25. — For these years the realized
capital losses of individuals filing positive net income returns were not
shown separately in table 7 of the Statistics of Income but were
included with "Other deductions". Hence it was necessary to prepare
an estimate of these losses.

3g C50NCENTRATI0N OF ECONOMIC POWER

In general, the magnitude of realized capital losses is determined by
the movements of prices of several classes of goods, by the prices and
volume of purchases at various times in the past related to the prices
and volume of sales, and by the volume of transactions by individuals.
It V, as not found feasible to develop a function or formula which would
relate these factors to the value of capital losses for this period. The
reasons for the inability to do so may be briefly summarized as fol-
lows: Absence of adequate data on most of the above-mentioned
factors; the gains resulting from transactions in the various classes
of goods vary in relative importance from year to year; many diverse-
price combinations are possible and the distribution of losses from
assets sold by length of time they were held probably varies signifi-
cantly during these years; the proportion of the total volume of trans-
actions carried on by individuals filing income-tax returns as compared
with corporations and individuals not reporting to the Bureau for
Internal Revenue probably varies from year to year. In view of the
variation of these factors influencing the volume of capital losses and
the changes in the relative weights of these factors during this period
it was not found satisfactory to develop a mathematical formula of
estimating the losses. Instead estimates were prepared which con-
tain arbitrary aspects and these estimates were checked by another
method. It is believed that the error in the final estimates is rela-
tively small.

Briefly, the following were the main steps taken in the preparation
of the estimates:

The ratios of realized capital gains of individuals reporting gains to
the losses of individuals reporting losses were calculated for the year
1926 through 1931. On the basis of the movement of these ratios
in conjunction with the movement of series related to the magnitude
of capital losses estimates were prepared of ratios for the earlier years.
Data on the amounts of realized capital gains were taken from table 7
of the annual issues of the Statistics of Income for 1918-25. The
principal series which were used in estimating the size of the ratios
were the movement of comprehensive indexes of stock prices, data on
the volume of stock-market transactions, available mdicators of the
prices of land and real estate. It was necessary to attach varying
weights each year to the various indicators because of the reasons
mentioned in the general discussion above.

The estimates were checked by use of the classification in table 7
of the Statistics of Income headed "General deductions" which includes
realized capital losses, net business and partnership losses, interest
paid, taxes paid and other deductions. Business and partnership
losses were estimated by the method outlined in appendix note A-1,
page 73. The remaining deduction items exclusive of capital
losses are fairly stable from year to year and were estimated. Deduct-
ing these estimated items from the total of "General deductions,"
the residual is the amount of capital losses. The two estimates of
capital losses were then compared, discrepancies were investigated
and a final estimate arrived at.

As realized capital losses were not estimated by income classes but
for all individuals filing income-tax returns as a group, it was not possi-
ble to obtain the amounts for those not filing returns by extrapolation
as was done for the years after 1925. These realized capital losses

CONCENTRATION OF ECONOMIC POWER §9

were estimated on the basis of the experience derived from the extra-
polations of later years. Account was taken of changes in the size of
minimum income for which a return was required by law.

Note A-5. Table 12 — Composition of Income of the Highest
1 Percent of Income Recipients, 1918-36

The basic data for this table were derived for the most part from
table 7 of the annual issues of the Statistics of Income, which is
entitled for recent years, "Individual returns, 1936, by net income
classes: Sources of income and deductions and net income * * *."2i
A number of adjustments were made in order to improve the year-to-
year comparability of the individual sources of income. As some of
the sources presented in the Statistics of Income are of interest only
from the viewpoint of income taxation, they were combined with
other sources for the present purpose. Combining various sources
also makes for greater year-to-year comparability. It is believed
that the data for each of the sources of income as given in table 12,
except the second realized capital gain figure for 1934, and the capital
gain figures for 1935 and 1936, may be compared from year to year
without appreciable error. However, there are year-to-year differ-
ences of minor importance in the composition of some of the sources,
particularly entrepreneurial net income and realized capital gains
and losses. The definition of each source of income, the adjustments
made, and the qualifications with regard to comparability are presented
below. ^^

The Statistics of Income tables present the sources of income by
34 income classes for returns with a net income of $5,000 and over
and, with the exception of the years 1918-26, in 1 income class from
zero to $5,000. As the lowest net income included with the highest
1 percent of income recipients in 1934 was below $5,000, estimates of
the sources of income for the $4,000 to $5,000 class were prepared by
extrapolation of ratios of each source to net income. The amount of
each type of income received by the highest 1 percent of income
recipients was obtained by straight-line interpolation using the
minimum statutory net income levels for tliis income group. ^® The
method of obtaining the minimum income levels is described in
appendix note A-1. Except where specified otherwise, the comments
refer to the data for the years 1918-31.

compensation of employees

This item is taken directly from the Statistics of Income (table 7)
classification "Wages and salaries." In recent years the title of the

2< The statistics for 1923 were revised subsequent to the publication of the Statistics of Income for 1923.
(See pp. 2S-29 of the Statistics of Income for 1925.) The revi.sed data which affected the distribution below
$15,000 wore given only in terms of the number of returns and net income and no attempt was made to adjust
the various income sources. The aggregate income of the highest I percent of income recipients is, therefore,
overstated by ?61, 000,000 in table 12. The percentage distribution in table 13 is not significantly affected
by this revision.

25 The description of the various sources of income is based for the most part on notes in the annual issues
of the Statistics of Income, particularly the i.ssue for 19.3fi, and notes in "Income Forecasting by the Use of
Statistics of Income Data," Review of Economic Statistics, vol. Xll, No. 2 (May 1930) by J. F. Ebersole,
S. S. Burr, and G. M. Peterson. The Statistics of Income for the earlier years do not contain much in the
way of definition and explanation and the article in the Review of Economic Statistics was of assistance in
this connection. Examination of the income-tax forms and income-tax regulations for each year was also
helpful.

^ The difference between the aggregate economic income of the highest 1 percent of income recipients
obtained in this fashion and that obtained by graphical interpolation as indicated in appendix note A-1
was negligible.

90 CONCENTRATION OF ECONOMIC POWER

classification has been changed to "Salaries, wages, commissions, fees,
etc," wliich describes more fully the content. In general, all types
of compensation from outside sources for personal services are included.

The data do not include allowable expenses paid out ot the wages,
salaries, etc., as these were to be deducted in advance. Wages and
salaries paid by State and local governmental units are excluded as
these were tax-exempt for the period under study. Also excluded for
the same reason were the salaries of judges of Federal courts.

Prior to 1924 the wages and salaries of the individual, his wife,
or dependent minors derived from a business conducted by the
individual were partly included in this category. It seems that such
income was reported under either wages and salaries or business
income for this period at the choice of the income recipient. Beguming
in 1924 this type of income was definitely classified under business
profit. Prior to 1927 the taxpayer's "earned income" from a partner-
ship was also included in the classification "Wages and salaries," but
since that date, such income has been included with partnership
income. No. adjustment was made for either of these items. It is
believed that the amounts of income involved are relatively small.
The composition of this source was not changed for the years 1934-36.

ENTREPRENEURIAL NET INCOME

This source is the algebraic sum of the net profit and the net loss
from unincorporated businesses and the individual's share of the net
profit and net loss from partnerships, syndicates, pools, etc. The
data for the most part are taken directly from the corresponding items
as given in table 1 ."" Due to several changes in its composition,
which are noted below, this source is slightlv understated, relative to
the other years, for the period 1922 through^ 1930. In 1934 and 1935
it is somewhat more inclusive than m the preceding years.

Among different individuals there is the possibility of some variation
in the treatment of similar types of income. Thus, one person in the
real-estate business might classify his income as business income while
another might treat the same type of income under rents and royalties
or as profits from dealings in property and include his business expenses
in deductions from total income. While this variation in the treat-
ment of income probably has an effect on the distribution among
types of income in a given year, it seems doubtful that the year-to-year
comparability of the income sources is much influenced.

Certain types of income received through partnerships are included
in other sources:

(a) Profits received through partnerships from the sale of
assets held more than 2 years wliich were classified by the indi-
vidual for tax purposes as capital net gain from assets held more
than 2 years were included with capital net gain and not with
partnership income for the period 1922-31. The partnership loss
on the same type of assets reported for tax credit was included
with capital net loss for the period 1924-31. Thus, partnership
income is understated or overstated during the years 1924
through 1931 depending on the relative size of such gains or losses

" Prior to 1930 business and partnership losses were not presented separately in table 7 of the Statistics of
Income but were included with "all other deductions." An estimate of such losses was made for each year
from 1918 throu£li 1029. See appendix note A-], p. 73, for the method of securing the estimates.

CONCENTRATION OF ECONOMIC POWER 91

each year. For the years 1922 and 1923 this source is under-
stated relative to the preceding years as this special treatment
applied only to realized capital gains. As only a portion of the
total profits or losses from the sale of property was classified as
capital gain or loss from the sale of assets held more than 2 years,
the bulk of such income has been included with partnership
income. ^^ The effect of this treatment of part of the realized
capital gains and losses received through partnerships on the size
of entrepreneurial net income is quite small in view of the follow-
ing considerations: (1) In 1926, for example, less than 25 percent of
entrepreneurial income was received through partnerships in
the income brackets where gains and losses on assets held over 2
years were reported separately from other partnership income.
(2) Undoubtedly only a small portion of this partnership income
was accounted for by the gain or loss from the sale of assets
held over 2 years. It should be indicated that for each of the
years a greater proportion of the total gains or profits from the
sale of property was reported as capital net gain from the sale of
assets held more than 2 years than the proportion of losses
reported as capital net loss from the sale of assets held more than
2 years. In 1934, 1935, and 1936 realized capital net gains and
losses as defined by the effective income-tax laws were fully
included with partnership income.

(6) Dividends received through partnerships on the stock of
domestic corporations subject to taxation under title I of the
effective revenue acts and dividends on the stock of foreign cor-
porations deriving more than half their gross income from sources
in the United States were classified under dividends for the years
1918 through 1931. (See description of dividend source.) As
these constitute the bulk of dividends received, only a small
amount of dividends is included with partnership income during
these years. In 1934 and 1935 the treatment of dividends received
through partnerships was unchanged except for the fact that all
dividends received on the stock of foreign corporations were classi-
fied with partnership income. This change in classification is of
relatively minor importance as dividends received annually by all
individuals on the stock of foreign corporations probably amounted
to from $10,000,000 to $15,000,000 during recent years.^^ In
1936 all dividends received through partnerships were included
with partnership income. Primarily because of this change in
classification data on entrepreneurial net income are not presented
in tables 12 and 13 for 1936.

(c) Interest on tax-free covenant bonds received through part-
nerships was included with "other income" in 1920. (See below
"Interest and miscellaneous property income.") Taxable and
tax-exempt interest received • hrough partnerships has been classi-
fied with "Interest and miscellaneous property income."

The data for the years 1918 through 1921 were adjusted as follows;
For the period 1918 through 1921, income from fiduciaries and
income received through personal-service corporations were classified

'> See footnote 31 on p. 93 for explanation of cases where gains were classified as capital net gain from the
sale of assets held over 2 years and where losses were reported for tax credit.

2' This estimate was kindly furnished by Dr. J'aul Dickens of the Finance Division, Department of
Commerce.

92 CONCENTRATION OF ECONOMIC POWER

with partnership income. An estimate of fiduciary income received
by those with net income of $5,000 and over for the years 1918-21 was
made by applying, to the total of partnership, fiduciary, and personal-
service corporation income, ratios of fiduciary income to the total of
partnership and fiduciary income. These ratios were selected on the
basis of an examination of the ratios for the years 1922 through 1936
in relation to the ratios of total property income to entrepreneurial
net income and property income. In order to secure the amount of
fiduciary income received by the highest 1 percent of income recipients
it was also necessary to estimate the fiduciary income received by
those in the $5,000 to $6,000 net-income class in 1918 and 1921 and
the $5,000 to $6,000 and $6,000 to $7,000 income classes for 1919 and
1920. The estimates of partnership income and fiduciary income still
include the income from ])ersonal-service corporations for the years
1918 through 1921. There is reason to believe, however, that the
amount of this income is relatively small. A personal-service cor-
poration was defined in the Revenue Acts of 1918 and 1921 (sec. 200)
as "a corporation whose income is to be ascribed primarily to the
activities of the principal owners or stockholders who are themselves
regularly engaged in active conduct of the affairs of the corporation,
and in which capital (whether invested or borrowed) is not a material
income-producing factor. * * *" These corporations were per-
mitted for income-tax purposes to treat their income in the sanle
manner as partnership income during the years 1918 through 1921.
In 1918, 3,503 such corporations (all were required to file a return)
reported a net income of $50,000,000. When this amount is com-
pared with the $900,000,000 of net income reported by partnerships
in 1918 (p. 11, 1918 Statistics of Income), it is readily seen that the
overstatement of partnership income due to the inclusion of the income
from personjd-service corporations is slight. It should be noted that
only a portion of the above amounts are included with the incomes
of the highest 1 percent of income recipients as the amounts cited
are taken from the informational returns which partnerships and
personal-service corporations were required to file and not from the
individual income-tax returns.

Whatever overstatement during the years 1918-21 in partnership
income and honce in entrepreneurial net income results from the partial
inclusion with partnership income of income received through personal
service corporations tends to be oft'set by the exclusion from entre-
preneurial income of wages and salaries of the individual, his wife and
dependent minor derived from an unincorporated business and the
"earned income" received from a partnership. As noted under the
description of employee compensation the former type of income was
partly included with employee compensation for the years 1918 through
1924 and the latter was also included with that source for the years
through 1927. During the indicated ye&iTs after 1921 the exclusion
of these two types of income adds to the above-mentioned under-
statement of entrepreneurial income which is due to the exclusion of
partnership realized capital gains reported for a tax credit.

REALIZED CAPITAL GAINS AND LOSSES

This item is partly estimated and partly taken directly from tables
2 and 7 of the annual issues of Statistics of Income. For the years

CONCENTRATION OF ECONOMIC PO^^•ER 93

1926 through 1931, this source is the algebraic sum of the following
items taken from the Statistics of Income: (a) "Profit from the sale
of real estate, stocks, bonds, etc., other than taxed as capital gain from
assets held more than 2 years" (table 7); (6) "capital net gains from
the sale of assets held more than 2 years" (table 7); (c) "net loss from
the sale of real estate, stocks, bonds, etc." (table 7); (d) "loss reported
for tax credit on capital net loss from sale of assets held more than 2
years." (Table 2. It was necessary to multiply these data by 8 as
12^^ percent of the loss was taken as tax credit.)

For the period 1918 through 1925, it was necessary to estimate-
the sum of items (c) and (d) as the data were not presented •separatel}''
in the Statistics of Income, but were included in "other deductions.''
The tax-credit alternative method of reporting losses on assets held!
more than 2 years was initiated in 1924. The losses reported for
tax credit were given for 1924 and 1925; however, in view of the small
amount of losses reported by this method relative to total losses, the
total loss from the sale of property was estimated as one item for these
2 years and tiu^ preceding years. The estimate of realized ^capital
losses was prepared in a manner similar to that described in appendix
note A-4 (pp. 87-9), except that it was done for incomes of $5,000
and over rather than for all individuals filing a return. Then in
order to obtain the realized capital losses for the highest 1 percent
group estimates were also prepared for several of the income classes
above $5,000. Total profits from the sale of property for 1918-25
are presented in table 7 of the Statistics of Income. Prior to 1922 the
"capital net gains from the sale of property held more than 2 years'"
were included in the Statistics of Income with "profits from the sale
of property." In 1922 the alternative method of re])orting suck
income as "capital net gain from the sale of property held more than
2 years" was introduced and beginning in that year items (a) and (6)
were totaled to obtain the realized capital gain.

This source is co.m.parable throughout the period 1918 through
1931, with the exception of the inclusion beginning in 1922 of gains
treated as capital net gain from assets held more than 2 years re-
ceived through partnerships and fiduciaries, and, beginning in 1924,
the losses from the sale of assets held more than 2 years reported for
tax credit and received through partnerships and fiduciaries.^^ It
should be noted that not all partnership and fiduciary gains or losses
on sale of assets held more than 2 years are included but only those-
which were so classified by the individual taxpayer. Wliether tax-
payers classified their gains or losses on assets held over 2 years in
this manner depended on the tax rate applicable to their net income.^"

In 1934, 1935, and 1936 no part of reahzed capital gain or loss
received by partnerships or fiduciaries was included in this source.
Beginning with 1934, the definition of realized capital gains and.
losses was changed as indicated in appendix note B-2 (pp. 101-2) with,
the result th-at the gains are overstated relative to those for the pre-

30 The comparability of the realized-gain-and-loss item is also affected by the fact that beginning on No-
vember 22, 1921, losses from the sales or other dealings in property were not recognized if substantially the
same property was acrjuircd 30 days before or after the date of such sale, except for dealers in stocks or
securities. Such losses are commonly referred to as "wash sales."

31 If the taxpayer were liable to a lesser tax by paying a flat tax of I2],i percent of his capital net gain od
assets held over 2 years, he would so classify such gains. If the taxpayer wore liable to a lesser tax by de-
ducting his capital net loss on assets held over 2 years from his other income than by taking 12!'2 percent
of the loss as a tax credit, he was required to use the latter method, that is, the method which produced the-
greater tax. The capital-gain provision affected only those with ret incomes approximately above $30,000'
from 1922 through 1931 and the capital-loss provision affected those with net incomes approximately above
$25,000 in 1924 and above $30,000 from 1925 through 1931.

94 CONCENTRATION OF ECONOMIC POWER

ceding years. The estimate of capital loss in 1934, comparable to a
considerable extent to the data for the preceding years, was derived
by the method described in appendix note A-1 (p. 73).

NET RENTS AND ROYALTIES

This source was taken directly from table 7 of the annual issues of
the Statistics of Income. All rents and royalties received by indi-
viduals except as part of business, partnership, and fiduciary income
are included. The rental value of owner-occupied homes is not
included, as no account is taken of such imputed inco.me under the
Federal income tax law. This form of income is not, at present, in-
cluded in the Department of Commerce estimates of net rents and
royalties.

DIVIDENDS

This item is taken directly from table 7 of the annual Statistics of
Income. Included for the years 1918-31 are dividends on the stock
of domestic corporations subject to taxation under title I of the
effective revenue laws, and the dividends received on the stocli: of
foreign corporations deriving more than half of their gross income
from sources within the United States. These dividends are in-
cluded when received directly by individuals or through partnerships
and fiduciaries. Excluded are dividends received on the stock of
domestic corporations not > deject to taxation under title I of the
effective revenue acts. Thf e corporations include building-and-loan
associations; mutual savingb banks; certain types of mutual insurance
companies other tban life; farmers' associations operated on a coop-
erative basis; various charitable, etlupational, fraternal organizations,
and others. Also excluded are dividends received on the stock of
foreign corporations deriving less than one-half of their gross income
from sources within the United vStates. These excluded dividends
were classified with "Other income" (see the income source entitled,
^'Interest and Miscellaneous Property Income"), except insofar as
received through partnerships or fiduciaries in which case they were
included with those income sources.

The data on dividends are comparable for the period 1918 through
1931 with two exceptions cited below. These two influences act to
oft'set each other and the net effect on the size of the dividend source is
probably slight. First, for the years 1918 through 192], dividends
received by the personal-service corporations were included with divi-
dends received by individuals. The inclusion of these dividends is
probably more than offset, because of the nature of such corporations,
by the exclusion of divideuds received by individuals from personal-
service corporations. Therefore, the treatment accorded to personal-
service corporations by the income-tax laws of these years (see state-
ment under entrepreneurial income above) may be a factor in under-
stating the dividends received by individuals for these years. Second,
stock dividends received bv individuals were included with dividends
m 1918 and 1919. In 1920, the Supreme Court decided that stock
dividends were to be excluded from income for income-tax purposes. ^^
However, later decisions set forth certain principles by which some

" Eisner v. Macomber (252 U. S. 189).

CONCENTRATION OF ECONOMIC POWER 95

stock dividends were included as taxable income. ^^ It does appear
that the amounts of stock dividends included as income in 1918 and
1919 were larger, relative to the amounts for later years.

In 1934 and 1935 all dividends on the stock of foreign corporations,
except when received through partnership and fiduciaries, were
mcluded with "other income." As was the case for the period
1918-31, the dividend source for 1934 and 1935 excludes dividends
on the stock of corporations not subject to taxation under title I of
the effective revenue acts. Except for the exclusion of dividends
on the stock of foreign corporations deriving more than half their
gross income from sources within the United States during 1934 and
1935, the dividend source in 1934 and 1935 is comparable to the source
for the preceding years. As indicated above, this change in classifi-
cation of a portion of the dividends from foreign corporations is of
minor importance.

In 1936 all dividends on the stock of both domestic and foreign
corporations were included in the dividend source except dividends
received through partnerships and fiduciaries. There is some offset-
ing influence involved in this change whereby the exclusion of those
dividends received through partnerships and fiduciaries, formerly
included in this source, is balanced to some extent by the inclusion of
all dividends on the stock of foreign corporations and of dividends on
the stock of domestic corporations not subject to taxation under title I
of the effective revenue acts. However, as the amounts of these two
sources are not known, no data on dividends are presented in tables
12 and 13 for 1936.

INTEEEST AND MISCELLANEOUS PROPERTY INCOME

This income source is composed partly of several items taken from
table 7 of the annual issues of the Statistics of Income and partly of
items estimated. In addition to interest, this source includes those
dividends excluded from the di^'idend source, except insofar as included
in partnership income; fiduciary income which is largely composed of
dividends not classified with dividends proper, profit or loss from the
sale of property other than those classified as capital gains or loss from
sale of assets held 2 or more years, rents and royalties, and interest;
also included was income not classified elsewhere. It is not possible
to present interest received separately, as interest Was combined with
"other income" for the years 1918 through 1926 under the heading of
"Interest and investment income."

Included in this source are the following: (a) "Interest and invest-
ment incom.e" for 1918-26 and the corresponding iteim.s of "Other
taxable interest" and "Other income" for the period 1926-31; (6)
taxable and tax-exem.pt interest on Government obligations. The
tax-exempt interest on Government obligations was estimated for the
years 1918 through 1923. A description of the method of obtaining
these estim.ates is given in appendix note A-1 (pp. 74-5). For the
years 1924-31, 1934-36, the interest on Government obligations was
secured from the table in the Statistics of Income on tax-exem.pt obli-
gations. (For example, table 8, pp. 85-86, in the 1927 Statistics of

M Soe list of decisions and interpretation of them in Regulations 101 Relating to the Income Tax Under the
Revenue Act of 1938, U. S. Treasury Department, pp. 308-310.

9Q CONCENTRATION OF ECONOMIC POWER

Income.) The amounts reported are known to understate the total
of this interest received by the individuals in the incom.e groups
included as the schedule from which this table is prepared is a supple-
ro.entary inform.ational schedule and frequently is not com.pletely
filled out;^* (c) fiduciary income. This was taken directly from table
7 of the annual issues of the Statistics of Incoro.e. As fiduciary income
was included with partnership incom.e in the basic data for 1918
through 1921, a separate estim.ate of fiduciary income was made for
these years. (See appendix note A-5, pp. 91-2, for the m.ethod of
estimation.) The treatm.ent of profits and losses received tlirough
fiduciaries from the sale of assets held ro.ore than 2 years and classified
by the individuals as profit or loss from the sale of assets held more
than 2 years were treated in the same fashion as the same type of
income received through partnerships.^^ Therefore, fiduciary incom.e
is understated or overstated during the years 1924 through 1931
depending on the relative size of such gains and losses. For the years
1922 and 1923 this source is understated relative to the preceding
years as this special treatment of gain and losses applied only to
realized capital gains during these 2 years.

In 1934 and 1935, all dividends on the stock of foreign corporations,
except those received through partnerships, were included in "Other
incom.e" or fiduciary income and hence in this source. For these 2
years and the period 1918-31, fiduciary incom.e includes dividends
received through fiduciaries on stock of dom.estic corporations not
subject to taxation under title I and dividends from foreign corpora-
tions deriving less than one-half of their gross income from, sources,
within the United States. In 1936 no dividends were included m
"Other income" and fiduciary incom.e includes all dividends received
through fiduciaries in contrast to earlier years when the major portion
of dividends received through fiduciaries were included with dividends.
(Sec discussion under "Dividends.")

It is believed that this source may be com.pared without appreciable
error for the years 1918-31, and 1934 and 1935, Due to the changes
in 1936, no data are presented for that year in tables 12 and 13.

Note A-6. Table 15 — Composition of Incomes by Income Classes,

1926, 1929, 1932, 1935

For sources and adjustment of data on the composition of incomes
for class of $5,000 and above see appendix note A-5. For the data
on composition of incomes of all income classes see appendix note A-4.
The amounts shown for the class under $5,000 were obtained by de-
ducting the amounts of each type of income for the $5,000 and over
class (derived chiefly from the annual Statistics of Income) from the
estimated amounts of each type of income received by all income
recipients.

Due largely to the imperfection, for this purpose, of the basic data,
the amount shown for interest and miscellaneous income is probably
overstated for the income class below $5,000. In the estimates for
the total individual income the property income, exclusive of realized
capital gains and losses, includes the amounts received by individuals
and aggregates of individuals. The latter term is defined to includ

" See Statistics of Income for 1936, p. 28.

36 See discussion under entrepreneurial net income pp. 90-1.

CONCENTRATION OF ECONOMIC POWER 97

insurance companies, building and loan associations, savings banks,
and other organizations devoted to the collective saving and invest-
ment of individual savings. While the estimates of total interest re-
ceived closely approximate the amount of such income received or
accruing to individuals either singly or as an association, it is not
known to what extent such receipts or accruals are imreported for a
given year in the basic data for incomes over $5,000, nor does present
knowledge of the flow of income through aggregates of individuals
permit one to determine to what extent the funds flowing to aggre-
gates of individuals in a given year are actually received by or accruing
to individuals in the same year. Although the relative proportion of
dividends flowing through aggregates of individuals is much less than
in the case of interest, these same considerations apply.

The amounts shown for the compensation of employees in the classes
above $5,000 excludes tax-exempt wages and salaries paid to em-
ployees ojf State and local governments and to judges of United States
courts. However, the source for all income classes includes such com-
pensation and the under $5,000 class, being a residual item, includes
all such compensation, a smaU portion of which belongs in the classes
over $5,000. The amounts involved are relatively small so that the
effect on the percentages of the total income received as employee
compensation for the various income classes is slight. Data from a
special study of the Treasury Department of the compensation in 1937
of the emploj^ees of State and local governments ^^ and data from the
Federal Budget on the salaries of Federal judges show that $118,000,000
in tax-exempt compensation was received by individuals in the wage
and salary bracket above $5,000. Using this figure for 1935 it is
readily seen that the effect of the exclusion of this compensation in
the income classes above $5,000 produces only a slight distortion in
the percentage composition of incomes for the various income classes. •
Thus the exclusion of such income from the under $5,000 class results '
in absolute deduction of less than one-tenth of 1 percent in the per-
centage of total income received as employee compensation, the
percentage being 68.6 rather than the 68.7 presented in table 15. For
the $5,000 and over class the table understates the employee-com-
pensation percentage by less than 1 percent, the percentage being 37.9
rather than 37.0. The effect on the $5,000 to $10,000 iucome class is
somewhat larger, the percentage being 53.5 as compared with 51.9
given in the table.

The source for all income classes headed "Interest and miscellaneous
property income" includes the net balance of international flow of
property incomes.

Note A-7. This Note Describes the Methods Used in Deriving
THE Data Presented in Tables 16 and 18

The purchasing power available to the highest 1 percent of income
recipients was obtained by deducting from the economic incomes of
this group, presented in table 12, the Federal income taxes paid by
this group. The income taxes paid are given in table 3 of the annual
issues of the Statistics of Income. With the use of the minimum-
income levels of the highest 1 percent as given in table 4 (p. 26) of this

" Published in Hearings Before a Special Committee on the Taxation of OoTemmental Securities and
Salaries, U. S. Senate, 76th Con?., 1st sess., p. 724.

256149 — 40-

98

OONOENTRATION OF ECONOMIC POWER

study the taxes paid by this group were calculated. These taxes are
presented in the appendix table A-6 below. To the total individual
income, presented in appendix table A-3 (p. 76), direct relief and the
veterans' adjusted-service compensation were added and the amount
of Federal income taxes paid by all individuals was deducted. Data
on the first two items are given in table 20 (p. 67), and on the total
amount of Federal income taxes in the following table. The taxes
are those paid on incomes of the given years.

The purchasing power shares available to the selected proportions
of income recipients for 1926 and 1936, presented in table 18, were
obtained in the same fashion. The aggregate mcome of these groups
is presented in appendix table A-2. The Federal income taxes paid
are given below.

Table A-6.

-Federal income taxes 'paid by all income recipients and by the highest
1 percent, 1918-37

[In millions of dollars]

Year

All income
recipients

Highest 1
percent of

income
recipients

Year

All income
recipients

Highest 1
percent of

income
recipients

1918

1,128
1,270
1,075
719
861
662
70-1
735
732
831

971
1,105
868
614
/46
562
644
709
706
805

1928

1, 164

1,002

477

246

511

657

1,214

1,142

1,134

1919

1929

991

1920

1930

1931

460

1921

237

1922

1934

1935

1923

481

1924 . - ..

617

1925

1936

1,136

1926 -

1937 _

1,050

1927 .-- --

Source: Table 3 of annual issues of Statistics of Income; 1923 data adjusted in accordance with revisions
indicated on p. 29 of Statistics of Income for 1925. The taxes are those paid on incomes of given years.

Table A-7. — Federal income taxes paid by selected proportions of income recipients,

1926 and 1936

[In millions of dollars]

Group of income recipients

Year

Highest ).i
of 1 percent

Highest Ho
of 1 percent

Highest Hoc
of 1 percent

1«26.„ .

684
1,080

557
874

283

1936.1

502

Source: Table 3 of Statistics of Income for the respective years.
of the given y^rs.

The taxes arc those paid on the incomes

APPENDIX B

SOME ASPECTS OF THE STATISTICS ON HIGH
INCOMES

Note B-1. Relation of Statutory Net Income to Economic Income

The purpose of this note is to examine the relation between statu-^
tory net income and economic income in order to learn to what extent
the data on income concentration for the years 1918 through 1925 in
terms of the statutory net income of the higher-income recipients may
be taken as indicative of shifts in the concentration of economic
income. Brief comparisons presented in chapter II (p. 24) indicated
that the use of statutory net income results in a small increase in
both the year-to-year variability of the income shares and in the dif-
ferences among the various proportions of income recipients in the
variability of their income shares. Before conclusions based chiefly
on the experience of the years 1926 through 1937 may be safely ex-
tended to the data for the years 1918 through 1925 it will be neces-
sary to determine more precisely the extent of the differences and
the explanations for them.

Another way of expressing the fact that the statutory net income
shares are more variable than the economic income shares is to state
that there is a changing ratio of statutory net income to economic
income. As will be shown statutory net constitutes a smaller propor-
tion of economic income in years of low-income concentration than in
years of high-income concentration. Therefore, the changes in the
concentration of economic income are slightly exaggerated when statu-
tory net income is used as a substitute for economic income. As was
indicated in chapter II, the following items were added to statutory
net income ^ to secure economic income: Interest and taxes paid, con-
tributions, "other deductions," tax-exempt interest on governmental
securities. Table B-1 contains the ratios of the statutory net income
to the economic income of the highest 1 percent of income recipients
for the years 1918-31 and 1934-37. The ratios generally rise in
years of high-income concentration and decline in years of low-income
concentration. To put it somewhat differently the deductions from
economic income are a larger proportion of economic income in years
of diminished incomes than in years of increased incomes.

Table B-1.-

-Ratios of statutory net income to economic income — highest 1 percent
of income recipients, 1918-37

Year

Ratios

Year

Ratios

Year

Ratios

1918—

.846
.840
.809
.79«,
.831
.831
.848

1925

.881
.858
.863
.873
.868
.845

1931.

1934

1934

19.35

1936

1937

.822

1919

1926

.806

1920 .

1927

1921. .

1928

.814

1922

1929

.832

1923

1930...

.85.1

1924

.839

Source: Calculated from data in table 12 and appendix table A-1.

' After adjustmenl of statutory net income for comparability by deducting for the years 1924 through
1931 capital losses on assets held over 2 years which were reported for a tax credit.

99

100

CONCENTRATION OF ECONOMIC POWER

The behavior of these ratios is apparently explained by the fact
that the deductions from economic income to arrive at statutory net
income are more stable from year to year than economic income and
hence constitute a larger proportion of economic income in years when
income is reduced. There does not seem to be any evidence of a
change in the relation of statutory net income to economic income
over the whole period. If such a change took place, it would further
limit or qualify the value of the measures of income concentration in
terms of statutory net income. The variations in the ratios are ap-
parently fully explained by the greater stability of the deductions as
compared with economic income.

The difference between the variability of the income concentration
measures based on statutory net income and economic income in-
creases slightly as the proportion of income recipients becomes smaller.
With the use of the relative mean deviations of the two sets of meas-
ures for the years 1926-37, shown in table B-2, the extent of the
increased variability can be determined. The table shows the differ-
ences between the two sets of relative mean deviations, but as the
absolute differences cannot be directly compared they are related to
the averages of the two relative mean deviations. These percentages
are presented in the last row of figures in table B-2. Except for the
largest group of income recipients, the percentages indicate a slightly
larger increase for the smaller income groups in the variability of the
statutory net-income measures.

Table B-2. — Relative mean deviations of the 2 m.easures of income concentration,

1926-37

uroup of income recipients

Income concept

Highest 2
percent

Highest 1
percent

Highest
Hof 1
percent

Highest
Mo of 1
percent

Highest
Moo of 1
percent

Statutory net income.

0.116
.099
.017

15.8

0.148
.128
.020

14.5

0.175
.150
.025

15.4

0.243
.204
.039

17.4

0.345

Economic income

.286

Difference

.059

Difference as percentage of average of 2
relative mean deviations

18.7

Source: Calculated from data in table 2, p. 22, and table 3, p. 23.

The increased variability of the statutory net-income measures for
the smaller proportions of income recipients is due to the behavior
of the ratios of statutory net income to economic income. In years
of reduced income, statutory net income constitutes a lesser percentage
of economic income, the smaller the pi^oportion of income recipients.
On the other hand, in years of increased income, statutory net income
either constitutes a larger percentage of economic income the smaller
the proportion of income recipients or the decline in the percentages
is less than in years of reduced income. These different relationships
are clearly shown in table B-3 which presents the percentages that
statutory net income constitute of economic income for the years
1926-34 and 1934-37. Thus in 1928 and 1929, both years of high
income, statutory net income is a larger percentage of economic
income as the proportion of income recipients becomes smaller. Ex-
cept for these 2 years and 1926 and 1927 when- the percentages are
approximately the same for all proportions, the percentages decline

CONCENTRATION OF ECONOMIC POWER

101

with the decrease in the size of the income group. However, the
lower the income, or the lower the income concentration, the greater
generally is the decline in the percentages. The iact that the per-
centages increase for the smaller proportions in 1928 and 1929 seems
to be due to the unusually large incomes received during these years.
In 1929, for example, the ratio of statutory net income to economic
income fell until about the $70,000 income level and then rose quite
sharply for incomes above that amount.

Table B-3. — Ratios of statutory net income to econcmic income — selected propor-
tions of income recipients, 1926-37

Group of income recipients

Year

Highest
2 percent

Highest
1 percent

Highest
Mof 1
percent

Highest
Mo of 1
percent

Highest
Moo of 1
percent

1926 --- ---

.859
.866
.874
.868
.850
.828

.817

.822
.829
.858
.847

.858
.863
.873
.868
.845
.822

.806

.814
.832
.855
.839

.858
.864
.874
.869
.841
.818

.795

.805
.823
.849

.8;%

.858
.861
.877
.874
.832
.7%

.765

.776
.801
.828
.808

.859

1927 - ---

.862

1928

.886

1929

.887

igso -

.828

1931

.764

1934 .-

.746

1934

.747

1935

.779

1936

.799

1937. ,

.774

Source: Computed from data in appendix tables A-1 and A-2, p. 78.

In view of the explanations for the differences between the behavior
of two indexes of income concentration for the years 1926 through
1934, there seems to be no reason to believe that the relationship
between the two indexes would not be approximately the same for
the years 1918 through 1925 as for the years 1926 through 1934 when
both indexes were available. We may, therefore, conclude that
when economic income cannot be derived, statutory net income
provides a method of approximating the changes in the concentration
of economic income, if account is taken of the fact that statutory net
income is somewhat more variable both from year to year and for
different proportions of income recipients. The use of statutory net
income is further limited, however, in view of the fact that measures
of income concentration based on it will sometimes give incorrect
information on the level of income concentration in 2 years. Com-
parison of the data in tables 1, 2, and 3 will readily confirm this.

Note B-2. Effect on Measures of Income Concentration of
Income Concept in Use During the Years 1934 Through 1937

The definition of income used for this period possesses certain
peculiarities, arising from the treatment of realized capital gain^ and
losses in the income tax law, which need to be taken into consideration
in interpreting the measures of income concentration. Wlien account
is taken of the effects of the definition of income on the measures of
income concentration, it will be possible to use these measures as
indicative of shifts in income concentration, income being defined as
it was for the j^ears 1918 through 1934.

102 CONCENTRATION OF ECONOMIC POWER

For the years 1918 through 1931 profits and losses from the sale of
property were completely taken into account in one way or another.
Beginning in 1934, varying proportions of such profits and losses are
included as income, the proportions depending on the length of time
the asset was held, and losses, as defined, were limited to $2,000 in
excess of gains.^ To tliis concept of gains and losses we may apply
the term "statutory capital gains and losses." One hundred percent
of the gain or loss on assets held 1 year or less was included as income;
80 percent on assets held 1 year but not over 2 years; 60 percent op
assets held 2 years, but not over 5 ; 40 percent on assets held 5 years
but not over 10, and 30 percent on assets held over 10 years. The
result of these provisions is that the importance of gains and losses
derived from the sale of assets held for long periods is lessened con-
siderably and capital losses are understated. It may be suggested
that while the change in the treatment of realized capital gains and
losses introduces a serious break in the continuity of the income-tax
data, taking varying proportions of gains and losses in the manner
provided in the law has some merit for our purposes. It seems
probable that short-term gains and losses are more closely related to
other types of current income in the treatment accorded them by
individuals than are long-term gains and losses. Another related
consideration and one that is partly responsible for this attitude is
the fact that a gain realized from an asset held for a long period may
be considered as having been accruing over these years and not
properly included in full as income in the year in which the gain was
realized. The emphasis given to short-term gains would be in line
with this reasoning ; however, unrealized gains currently accruing were
not included as income.

In order to properly interpret the income measures based on the
income concept embodying statutory net capital gains and losses, it
will be necessary to examine the effect on them of the use of this con-
cept. For the most part the provisions with regard to taking varying
proportions of gains and losses will be accepted and attention will be
directed to the effect of the $2,000 limitation on the size of an indi-
vidual's loss. By limiting the objective in this fashion, the number of
variables to be taken into account in interpreting the final measures
will be reduced to a manageable number.

In general, it seems that for the years covered the use of statutory
net capital gains and losses results in a small amount of overstatement,
except for the two proportions with the highest incomes, of the degree
of income concentration relative to the degree of income concentration
that would be shown when realized capital gains and losses are in-

' In addition to this change in the treatment of realized capital gains and losses, the definition of capital
asset was made slightly more inclusive beginning in 1934. Included as a capital asset before 1934 was all
property not connected with the taxpayer's business or trade. In 1934 the definition was broadened to
include all property, whether or not connected with trade or business, except stock in trade, property
which would be included in inventory, or property held for sale in the ordinary course of business. Con-
sequently, for the years 1934 through 1937 the definition of capital asset was more inclusive than for earlier
years in that property connected with trade or business, e.xclusive of stock in trade, was included within
the definition.

It should also be noted that the method of classifying realized capital gains or losses received through
partnershii)s and fiduciaries was changed somewhat in 1934. Prior to 1934 that portion of realized capital
gains or losses received on the sale of assets held over 2 years and classified by individuals as "capital net
gain from tjie sale of assets held over 2 years" or as a tax credit for "loss froo) sale of assets held over 2 years"
were included with these sources and not with partnership and fiduciary income. Beginning in 1934 all
realized capital gains and losses received through partnerships and fiduciaries were classified with those
income sources. Thus the realized capital gain and loss source is less inclusive for the years 1934 through
1937 because of this change in classification. These changes in the definition and classification of realized
capital gains and losses act to some extent to oflfset each other.

CONCENTRATION OF ECONOMIC POWER 103

eluded completely. The overstatement arises from the relatively
greater reductions in the incomes of the selected proportions of income
recipients than in the incomes of all recipients when realized capital
gains and losses are fully included. This can be shown for 1934 by
comparing the two sets of data. And there is reason to believe that
it is also true for the years 1935 through 1937. In 1934 the degree of
overstatement is approximately the same for the three larger pro-
portions, somewhat smaller for the highest one-tenth of 1 percent,
and for the highest one one-hundredth of 1 percent there is actually
a very slight understatement.^ The dechne in the degree of over-
statement for the highest one-tenth of 1 percent and the actual under-
statement for the highest one one-hundredth of 1 percent are due to
the fact that for these two groups, the necessary deductions for the
full inclusion of realized capital gains and losses are less important
than for the other three groups of income recipients. This is due, of
course, to the fact that realized capital gains are a large source of
income in these classes and that including them fully diminishes the
importance of the full deduction of realized capital losses.

In addition to the bias for a given year, the degree of overstatement
or understatement will vary somewhat with different years depending
lar^-ely upon the magnitude of realized capital losses.* For the present
purpose it is fortunate that the years 1934 through 1937 were not
charactepzed by unusualh^ large amounts of realized capital losses.
However, as is readily evident from table B-4, the importance of real-
ized capital losses has declined during these years and hence, excluding,
for the moment, the two smaller proportions of income recipients, the
degree of overstatement is larger in 1934 than in 1935 and larger in

1935 than 1936. Data for 1937 comparable to that shown for the other
years in table B-4 are not yet available. On the basis of the available
indicators of the importance of realized capital losses in 1937, it is
certain that the overstatement for 1937 is greater than for 1935 and

1936 and perhaps less than in 1934. In terms of the various propor-
tions of income recipients it seems likely that the reduction in the
importance of realized capital losses would result in an understatement
of the shares received by both the highest one-tenth of 1 percent, and
the highest one one-hundredth of 1 percent of income recipients in 1935
and 1936 and probably in .1937.

3 The concentration measures for 1934 with actual capital gains and losses and with statutory capital gains
and losses, both taken from table 3. p. 23, are presented in the following table:

Percent of total

income received

Group of income recipients

Income with

actual realized

capital gains

and losses

Income with
statutory capi-
tal gains and
losses

Highest 2 porcrnf ... .

IP. 95
12.66
9.57
4.94
1.70

17.41

Highest 1 percent .

13.03

Highest 1-2 of 1 percent..

Highest Mo of 1 percent _." ' . .

9.86
5.03

Highest Moo of 1 percent . . .- ...

1 75

The i-lifTercnces are somewhat les.s than iniiicated by the above comparison as the percentage'? in the first
column lire .-^nhjoct to ;; small amount of undcrsfatem.ent. fSee appendix note A-1, p. 73, where the adjust-
ment of tilt 1934 data for full inclu.=;ion of realizei! gains and losses is discussed.)

* The olhiT factor is the difference between various yars. iu thp distribution of gains and losses by the
length of time the assets sold were held.

104

CONCENTRATION OP* ECONOMIC POWER

Table B-4. — Number of individuals among the various proportions of income recip-
ient reporting a statutory capital loss of $2,000 and over, 1934-36

Group of income recipients

Year

Highest 1
percent

Highest Yi
of 1 percent

Highest Ho
of 1 percent

Highest Hoo
of 1 percent

Number

1934

26,356
18, 198
13, 678

20,275
13, 719
9,809

7, 871 1, 125

1935

1936 ^ — .—

5. 084 751

3.085 1 513

Number as percentage of income recipients in group

1934

6.35
3.65
Z70

8.23
5.50
3.90

15.98
10.20
6.13

22.84

1935 . ; -

15.07

1936 ,..-.: ^

10.19

Source: 1934 from unpublished data made available by the Treasury Department. The number of
Individuals is slightly understated as these data include only those returns filed through May 1936. The
Statistics of Income for 1934 includes returns filed to Dec. 31, 1935. 1935 and 1936 from Statistics of Income
for each year, text tables on pp. 20 and 21, respectively.

'The total number of income recipients in each group Is given in table 6, p. 27.

As regards the manner in which the year-to-year changes in the
importance of reaUzed capital losses affect the various proportions of
income recipients, the changes in 1936 will be described. In this year
the shares received by the smaller two proportions are understated
somewhat more than in the previous year and for the three larger
shares are overstated less. Of course, on the basis of the available
data it is not possible to determine with certainty whether the share
of the highest one-half of 1 percent in 1936 is overstated or under-
stated. It may be that for this group, as a whole, the provisions of
the law result in its share being the same as if realized capital gains
and losses were fully taken into account. This was approximately
true of the share received by the highest one one-hundredth of 1 percent
in 1934.

The above discussion may be summarized briefly by stating that as
a result of the bias in the income-tax data, the increase in income con-
centi-ation from 1934 through 1936 was actually somewhat more than
indicated by the data in tables 1, 3, 21, and 22. This is true for all
the selected proportions of income recipients. The decline in income
concentration during 1937 was also somewhat greater than shown by
the data.

Note B-3. Characteristics of Income-Tax Data and Method of
Measuring Income Concentration

This note is intended as a brief discussion of some of the reasons
why the method of measuring income concentration used in this study,
as compared with other methods, seemed well adapted to overcome
certain limitations of the income-tax data. Largely due to the effects
of imperfections of the data on the accuracy of the measures and to
the difficulties involved in interpreting them, certain statistical indexes
of income ''inequality" have not been utilized. For the present study
the absence of ambiguity and assumptions as to the meaning of

CONCENTRATION OF ECONOMIC POWER 105

"inequality" were governing considerations in selecting a method of
measuring income concentration.

Statistical devices such as the indexes or coefficients of Pareto and
Gini and other measures of dispersion are commonly utilized to study
changes in the degree of inequality in the distribution of income among
the higher income recipients. These methods do not generally relate
the indicated changes in the inequality in the upper ranges of the
income distribution to the total income of all individuals. It is
possible, for example, that a decUne in the degree of income inequality
among the individuals with relatively high incomes may be accom-
panied by no change in the proportion of the total income received by
them. It seems that such a shift in income inequality occurred in
1931. (See ch. II, p. 33.) The simple method of taking propor-
tions of total income received by fixed percentages of income recipients
seems to have an advantage for the present purpose in that it relates
the incomes of the higher income recipients to the total income of all
individuals. However, as the data only cover the highest 2 percent
of income recipients, this measure, in common with the others, is
limited by the fact that changes might take place in the distribution
of income among the lower income recipients without being reflected
in the income concentration measures.

In view of certain characteristics of the income-tax data, the method
adopted for approaching the question of income concentration sur-
moimts in large part the obstacles arising from the manner in which
individuals are classified in the income-tax statistics. Individuals
were classified according to the size of their "statutory net incomes."
This classification is unsatisfactory partly because of peculiarities in
the definition of net income arising from the exclusion of certain
deductions from net income, and the exclusion of tax-exempt interest.
Of greater importance is the alternative provision in effect from 1924
thi'ough 1933 for treating losses from the sale of assets held more than
2 years as a tax credit rather than deducting the amounts from other
income. This peculiar definition of net income, therefore, places many
individuals in the wrong income class. This defect in the data is
largely overcome in the measures of income concentration presented
in this study. ^ Devices for measuring income inequality which
depend on the distribution of the individuals by detailed income classes
are, therefore, largely unsuited for these statistics.

In addition, the relationship between the net income concept of
the income-tax law and the concept of economic income for which
measures of income concentration are derived is such that year-to-
3'"ear changes in the concentration of economic income are exaggerated
somewhat when measures are based on the net income concept. Ap-
pendix note B-1 is concerned with the relation of statutory net
income to economic income. While the transformation of the dis-
tribution of statutory net income to economic income is not entirely
satisfactory, it is believed that the measures of the concentration of
economic income presented are considerably superior to measures
based on statutory net income when these are used to indicate changes
in the concentration of economic income.®

' See appendix note A-1, pp. 71-3.

« For limitations of the transformation from statutory net income to economic income see appendix note
A-1, pp. 75-7.

106 CONCENTRATION Of ECONOMIC POWER

The method used also minimizes the importance of certain practices
designed for tax avoidance that have arisen, particularly in the
higher-income brackets, in connection with the operation of the
income-tax law. These practices take the form of excessive or arti-
ficially created deductions and division of income between members
of a family, especially between husband and wife. As all the deduc-
tions, except business and realized capital losses, are added to statu-
tory net income to secure economic income, any deductions of this
type are inxjluded in economic income. It should be noted that
the deductions which appear on the schedule for business income are
not added to statutory net income. Only those deductions appearing
on the face of the income-tax return are taken into account.

With regard to division of income between members of a family
for the purpose of tax avoidance, the effect of the practice is also
minimized by the method adopted for measuring income concentra-
tion. Such devices are most often used in the high-income brackets
where the surtaxes rise sharply. A simple example will serve to illus-
trate how tl\e effect of this practice is minimized. Let it be assumed
that through the division of property, two incomes of $250,000 appear
in the income-tax data instead of one $500,000 income which appeared
in the data during earlier years. The additional "income recipient"
so created '^ould probably not be reflected in the estimates of the total
number of income recipients. The effect of this change is that there
is included with the highest 1 percent of income recipients one less
income at the lower limit of the group when the $500,000 income is
divided than when there is only one $500,000 income. As the mini-
mum incomes of the highest 1 percent are relatively small, varying
from $5,375 to $10,140, it is readily appreciated that the effect of the
practice on the measures of income concentration is minimized by
taking proportions of total income received by fixed percentages of
income recipients.'^ The effect of division of income between members
of a family would have a more important effect on measures of income
concentration which depend upon a detailed distribution of income
by income classes. The measures for the highest 1 and 2 percent of
income recipients are least affected by this practice as the increase in
such divisions of property is believed to have taken place largely in
the higher-income brackets.

Tc some extent the division of income within a family may take the
form of the creation of trusts as well as the outright division of prop-
erty. The aspect of this procedure that is of present interest is the
effect of any increase in this practice, brought about by a desire to
reduce taxable mcome, on the size of individual incomes and, hence,
on the measures of income concentration. The same general consid-
erations outlined immediately above apply to this'practice. In addi-
tion, the effect on the data of any increase in the practice of creating
trusts for the purpose or reducing income taxes tends to be offset by
the double-counting involved when the income distributed to bene-
ficiaries is included both in the income of the trust and in the individual

' This ran be demonstrated with use of data on the nurnber of wives filing separate returns. It may be
indicated here that the percentage of incomes included in the highest 1 percent of income recipients repre-
sented by wives filing separatt^ly from husbands has varied from 3.2 percent in 1919 to 7.4 percent in 192S.
The average for the years 1934 through 193fi was .i.3 percent only slightly higher than the average of 4.7
percent for the years 1918 through 1924. The increase has been somewhat larger for the smaller proportions
oj income recipients. Part of this small increase is due to rising proportion of income with gainful occu-
pations. According to the 1920 census 23.6 percent of the women between the ages 20 to 64 reported an
occupation and the 19.30 figure is 26.2. Another factor is the changing status of women as this is reflected
in the separate maintenance of property, either held before marriage or inherited after marriage.

CONCENTRATION OF ECONOMIC POWER 107

income of the beneficiary. As noted at an earlier point (footnote 1,
p. 71) the income of trusts are included in the data on individual
income. This double-counting is true only of the data based on
economic income as the statutory net income of trusts represents
undistributed income and the amount distributed is included with
"other deductions" which was added to statutory net income in order
to obtain economic income. The distributed amounts are reported
by the beneficiaries as individual income under the heading of fiduciary
income. It is not possible to secure data on the statutory net income
of trusts over a period of years or even for recent years. In 1935 the
statutory net income reported in Statistics of Income by all types of
trusts and estates was $208,000,000 m 1935 and $347,000,000 in 1936.
Presumably a substantial portion of these amomits represents the
incomes of estates and of trusts created for purposes other than
reducing tax payments.

In connection with the division of property for the purpose of reduc-
ing taxable incomes, it should be mentioned that the gift tax in effect
in 1924 and 1925 and since 1932 considerably reduces the feasibility
of dividing income between members of the family. In addition, the
change in the law effective in 1937 modifying the personal exemption
for trusts operates in the same direction. In view of the foregoing
discussion, it is not believed that the findings on income concentration
presented in this report are appreciably affected by any increase in
the practice of dividing incomes between members of a family.

Other devices of reducing income in order to avoid taxes apparently
cannot be taken account of by any method of measuring income con-
centration. In a list of methods of avoiding income taxes prepared
by the Treasury Department ^ the only method of reducing individual
income for the purpose of tax avoidance not mentioned above is the
personal holding company which takes various forms. This device
enables the individual to accumulate income from property or per-
sonal services without having it included along with his other income
in his income-tax return to the Bureau of Internal Revenue. How-
ever, legislation adopted in 1934 and amended in 1936 and in 1937,
which was designed to force personal holding companies to distribute
their incomes and to disallow certain deductions, appears to have
accomplished its purpose. The new provisions probably result in
individual incomes being larger in recent years than in earlier years
when resort could be had to these practices of reducing incomes.®

' See p. 7, Tax Evasion and Avoidance, Hearings Before the Committee on Ways and Means, House of
Representatives, 75th Conp., 1st sess. The committee report also mentions artificial deductions for losses
from sales or exchanges of property resulting from transactions between several corporations under common
control and incomes of nonresident aliens.

' For data on the incomes of personal holding companies see the issues of the Statistics of Income, pt. 2
for 1934 through 1937.

INDEX

Page
BROOKINGS INSTITUTION. America's capacity to consume; cited

(n.) 15

BURR, S. S., joint author. See EBERSOLE, J. F.

COMPETITION. Importance of in bringing about more equal distribu-
tion of income 2

CONFERENCE ON RESEARCH IN NATIONAL INCOME AND

WEALTH. Studies; cited (n.) 10

EARNING POWER. Definition of 9

Transition from earning power to purchasing power 55

EARNING POWER CONCENTRATION. Differences between trend

in concentratioii of purchasing and of earning power 64-65

EBERSOLE, J. F., S. S. BURR, and G. M. PETERSON. Income

forecasting bv the use of statistics of inconie date; cited (n.) 71

ECONOMIC POWER, CONCENTRATION OF. Relation to concen-
tration of income 3

EMPLOYEES OF STATE AND LOCAL GOVERNMENTS:

Exclusion from income concentration data 81-82

Statistics:

1918-37. Compensation and number of employees of State and

local governments; table A-5 83

FEDERAL INCOME TAX DATA:

Limitations of 6

Questionable reliability prior to 1918 14

GAINFUL WORKERS. '.S^e INCOME RECIPIENTS.
GOLDENTHAL, ADOLPH J. Concentration and composition of indi-
vidual incomes, 1918-37; \\ritten by iii

GREAT FORTUNES. Relation to monopolistic practices 1

INCOME:

Definition of, changes in 14

Economic, definition of -- 10

Passage from statutory net income to economic income 73-77

Relation of statutory net income to economic income 99

Stages at which income may be measured 9

Transfers of income 58

INCOME COMPOSITION:

Description of available data for high incomes 36

Description of Statistics of Income data on income sources 89-96

Relation to income concentration 49-53

Statistics :

1918-36. Composition of incomes of the highest one percent of

income recipients : chart 3 42

1918-36. Composition of incomes of the highest one percent of

income recipients, dollar amount; table 12 39

1918-37. Composition of total individual income; dollar amount;

table 10 37

1926-35. Percentage composition of incomes by income classes,

1926, 1929, 1932, and 1935; table 15 48

INCOME CONCENTRATION:
Changes in:

Business activitv and 17-18

During 1918-37". 16-17

Size of total income per recipient and 18-19

Year-by-year summary 19-20

Characteristics of income-tax data and method of measuring income

concentration 104-106

Concentration by type of income- 44-46

Differences in variability of income shares. _- 31-32

Divergent shifts in , 1 32-33

109 •

UQ INDEX

INCOME CONCENTRATION— Continued Page

Effect of relief and veterans' adjusted-service payments on 70

Effect on measures of income concentration of income concept in use

during the years 1934 through 1937 101-104

Effect of work relief wag on 68-69

Essential to take account oi" income structure in interpreting data on. 27

Estimates of actual degree ■ 15-16

Fluctuation analysis 49-53^

Notes on data presented in tables 1-3 71

Relation to concentration of economic power 3^

Relative mean deviations of the two measures of income concentration,

1 926-37 ; table B-2 100

Selected proportions of income recipients 31

Statistics:

1918-36. Percentages of each type of income received by the

highest one percent of income recipients; table 14 45

191&-37. Concentration of income and total income per recip-
ient; chart 1 18-

1918-37. Shares of total individual income received by selected

proportions of income recipients; chart 2 30'

1918-37. Shares of total individual income received by the

highest one percent of income recipients; table 1 16

1918-37. Shares of total individual income received by selected

proportions of income recipients; table 2 22

1920, 1928. Percentage increases in shares of total individual
income from year of lowest (1920) to year of highest (1928);

table 9 . . _ 32

1926-37. Shares of total individual income received by selected

proportions of incom recipients; table 3 23^

1934-37. Shares of otal individual income, excluding relief
and veterans' adju' ed-service payments, received by selected
proportions of income recipients; percentages and indexes,

table 21 68

1934-37. Shares of total individual income, including relief
and veterans' adjusted-service payments, received by selected
proportions of income recipients; percentages and indexes,

table22 70

Stock market situations in twenties and 3

INCOME DISTRIBUTION:

Problems before Temporary National Economic Committee and 1-6

INCOME INEQUALITY. Changes in, by eradication of monopolistic

elements in our economy 2

Methods of measuring 104-105

INCOME RECIPIENT. Definition 12

INCOME RECIPIENTS:

Notes on data presented in table 6 79

Selected proportions of income recipients 20

Statistics:

1918-37. Number of individuals in the selected proportions of

income recipients; table 6 27

1918-37. Number of persons with gainful occupations; table A-4- 80
INCOME TAX:

Effect of income tax in reducing incomes 61-63

Effect of increase in surtax rates in 1936 64

Flexible rate structure for personal incomes 5

Variable rates for different types of income 35

States imposing personal income taxes 57

Statistics:

1918-37. Federal income taxes paid by all income recipients and

by the highest 1 percent; table A-6 98

1918-37. Proportions of economic income of highest 1 percent of

income recipients paid in Federal income taxes; table 19 63

1926 and 1936. Federal income taxes paid by selected propor-
tions of income recipients; table A- 7 ^- 98

1930-36, 1938. Estimated collections from State income taxes;

table 56

1930, 1934 . 136, and 1938. Collections from selected taxes;
Federal individual income .tax as percent of total each year;
table . 57

INDEX 111

Page
KING, WILLP'ORD I. The national income and its purchasing power;

cited (n.) 60

LABOR MOBILITY AND EFFICIENCY. Effect of governmental ex-
penditures on 5

LEVEN, M., H. G. MOULTON and C. WARBURTON. America's

capacity to consume; cited (n.) 15

LOUGH, W. H. High level consumption; cited (n.) 59

MONOPOLY. Relation between concentration of income and monopoly. 2

MOULTON, HAROLD G., joint author. See LEVEN, M.

NATHAN, ROBERT R. Monograph No. 4, T. N. E. C, prepared

under general supervision of in

NATIONAL BUREAU OF ECONOMIC RESEARCH. Income in

the United States, cited (n.) 14

NATIONAL INCOME. Comparisons influenced by income distribu-
tion 4, 5

NATIONAL INDUSTRIAL CONFERENCE BOARD. State income

NATI ON AL RESOURCES COM MITTEK " 'Consiamer expenditureV in

the United States ; cited (n.) 59

NEWCOMER, MABEL. Estimate of the tax burden on diff"erent income

rtjocGfic" oif"f*(i ('n ) Oo

PETERSON, G. M",' joinVauthor." ' See'EBERS'OLE," L F.

PUBLIC EXPENDITURES. Influence on future income distribution. __ 5

PURCHASING POWER. Definition 11,55-59

PURCHASING POWER CONCENTRATION:

Differences between trend in concentration of purchasing and of

earning power 64—65

Measures of 1918-37 55

Note on methods used in deriving the data in tables 16 and 18 97

Statistics:

1918-37. Effect of adjustments for "purchasing power" on the
income shares of the highest one percent of income recipients;

table 17 61

1918-37. Shares of total purchasing power available to the

highest one percent of income recipients ; table 16 60

1926, 1936. Shares of "purchasing power" available to selected
proportions of income recipients and effect of adjustments for

purchasing power; table 18 62

REAL INCOME. Definition 11

REALIZED CAPITAL GAINS AND LOSSES:

Change in definition 102

Inclusion in various income definitions (n.) 10

RELIEF PAYMENTS:

Direct and work relief has introduced new elements into the Nation's

income structure 65-67

Effect of relief payments on income concentration 70

Statistics. 1 929-37. Direct and work-relief payments, dollar amount

and percent of total income; table 20 67

SAVINGS:

Influence of income distribution on 4

Importance of high-income recipients in determining volume of 14

SIMONS, HENRY C. Personal income taxation; cited (n.) . 10

SMALL-SCALE ENTERPRISES. Determination of importance of 3

TAX AVOIDANCE, EFFECT ON DATA OF 106-107

TAX-EXEMPT INTEREST, ESTIMATES OF - 74-76

TAX RESEARCH FOUNDATION. Tax Systems of the World, edition

7, 1938; cited (n.) . 56

TAXATION:

Incidence of 58

See also INCOME TAX.
TAXES, PERSONAL. Deducted from income for measures of purchas-
ing power 56-58

TWENTIETH CENTURY FUND. Facing the tax problem; cited (n.). 10,56

Studies in current tax problems; cited (n.) 56

UNEMPLOYMENT. Attributable to causes ultimately traceable to

enervation of competitive forces in the economy » 2

]^12 INDEX

Page
UNITED STATES TREASURY DEPARTMENT. Bulletin, August

1939; cited (n.) 58

Selected State-imposed taxes, 1 930- 35 ; cited (n.) 57

VETERANS' ADJUSTED SERVICE COMPENSATION:

Effect on income concentration 70

New elements introduced into the Nation's income structure by 65-69

Statistics:

1929-37. Dollar amount of payments and percent of total in-
come; table 20 67

WARBURTON, CLARK. Studies in income and wealth, cited (n.) 10

Joint author. See LEVEN, M.
WELFARE. Relation to income distribution 4

o

^OSTON PUBLIC LIBRARY

3 9999 06351 9

3 4