torium on inventions", which have found echo in business and scientific circles. Such climate of opinion is hardly healthy for technological advance. On the other hand, in the Soviet Union technological progress is being fostered as a means of achieving the governmental objective of a socialist, planned, large-scale economy for the satisfaction of expanding consumers' needs.76

In summary, resistance to technological change has been so much a part of the texture of the historical process, that it cannot be ignored when the future of

76 Webb, Sidney and Beatrice, Soviet Communism: A New Civilization (New York 1936), pp. 767-71; Moore, Harriet, The Stakhanov Movement in American Russian Institute, Research Bulletin, vol. 1, no. 2 (New York 1936).

technology is charted. There are psychological factors in individual and group behavior which predispose toward inertia in receptivity to innovation, but these may be counterbalanced by potent incentives that promise material and nonmaterial rewards. The basic determinants of the presence or absence of impediments to technological change lie therefore in the nature of the social, and primarily the economic, structure of a society, in the degree to which it offers incentives to the masses of the population and in the manner in which these can be realized through a planned economy. Capitalism has inherent in its structure and functioning, factors which militate against such realization, and thus prevent industrial practice from keeping apace with scientific knowledge.

Introduction

The economic literature of the past two centuries is interspersed with debates concerning the effects of the increasing use of machinery on the volume of employment. The introduction in the eighteenth and nineteenth centuries of the early forerunners of our modern machinery was in many instances considered by workers and government authorities alike as an evil to be averted. This was particularly true in England between 1725 and 1775 when new machines were being widely introduced in the textile industries. The workers involved had no assurance that these laborsaving devices meant anything more to them than lost employment opportunities, and the land owners and sheep raisers saw in the coming rise of the cottonmanufacturing industry the destruction of their own markets for wool. The results were machine-breaking riots, persecution of inventors, and legal restrictions. John Kay, who invented the flying shuttle in 1733, was forced to flee the country; Hargreaves, the inventor of the spinning jenny (1764), was compelled to change his residence; and Crompton, who invented the spinner's mule in 1779, was forced to go into hiding. The prohibition on the manufacture of all-cotton fabrics in England was not lifted until 1774.

Since those early days of our industrial society, every period of widespread unemployment has brought with it a revival of the old protests and the old discussions. The most recent protests against the introduction of labor-saving devices have been no less earnest, albeit less violent, than those of the eighteenth century. Yet we need only to look about us to observe the tremendous multiplication of labor opportunities which past technological improvements have provided in creating vast new industries and new services. The question is whether any economic changes have occured during recent years and especially during the last two decades which justify the dark prophecies of ever-increasing unemployment that have become current of late.

The vast number of persons unemployed since 1929, and the apparent disparity between recent production increases in certain industries and the extent of reemployment in these industries have aroused new interest in the problem of "technological displacement" and unemployment. With recovery in industry and

1 Director of the Works Progress Administration National Research Project on Reemployment Opportunities and Recent Changes in Industrial Techniques. He was assisted by Harold L. Posner.

business manifesting itself in increased production and sales, increased employment, and increased profits and dividends, we face the disturbing realization that estimates of the number of unemployed show no proportionate decrease and that the relief burden remains at a high level. Numerous startling developments in production techniques and some recent dramatic instances of displacement of workers through the introduction of these new techniques have helped to focus attention on the effects of decreasing labor requirements per unit of production on the volume of employment.

A full investigation of the effects of changing technology on the volume of employment and unemployment would involve an analysis of the effects of changing prices of goods and services, of changing costs of capital and labor and the changing proportions of each employed in the production process, of changing demands for goods and services, and of a multitude of other factors which play an important part in determining the profitableness of employing workers. Only such an economic analysis, dealing with the fundamental elements of our economic society, could be expected to arrive at conclusions concerning the underlying causes of unemployment in general and the particular type of unemployment which might be attributed to changing industrial techniques. Such an analysis would have to extend far beyond the size and scope of a report of this nature. Since, however, the net effects of the underlying economic factors find their quantitative expression in the net changes of the volume of production and employment, a brief statistical analysis of the relationship between the total volume of goods and services produced in the country and the number of hired workers engaged in this production offers an approach toward a better understanding of the nature of a problem which has come to be referred to popularly as that of "technological unemployment." While such a statistical analysis may not permit the drawing of any conclusions as to the underlying causes of what occurred during the period under consideration, it at least makes possible an examination of some of the measurable effects in a new light. It is therefore proposed, in this report, to subject the period since 1920 to a bird's-eye view-to examine the available statistical information on the volume of production and employment in the light of the changes in output per man-year which took place during this period, and to bring together data which indicate to what extent employment dislocations have

The statistician faces a great many complex questions, frequently bordering on the metaphysical, when he attempts to construct a measure of the changing volume of total physical production. But inasmuch as such measures, however, crude they may be, are extremely useful as indicators of how well or how badly the economy of the country is faring, a number of attempts have been made to construct them. The results are always hedged around with all manner of qualifications, but the very need for these indexes has resulted in their widespread acceptance. Aside from the usual deficiencies from which such indexes suffer, none of the available indexes of the "physical volume of production" takes into account the volume of commercial services which constitute an important section of our economy. The omission of this growing source of gainful employment has been due primarily to the difficulty of applying appropriate physical measures to this type of economic activity. Since an over-all picture of the changing volume of employment opportunities cannot possibly leave out of consideration this expanding group of industries, it is necessary to use a different, an all-inclusive, measure of the growth of production-the national income.

"Year in, year out the people of this country, assisted by the stock of goods in their possession, render a vast volume of work toward the satisfaction of their wants. Some of this work eventuates in commodities, such as coal, steel, clothing, furniture, automobiles; other takes the form of direct, personal services, such as are rendered by physicians, lawyers, Government officials, domestic servants, and the like. Both types

2 A number of surveys designed to collect and analyze both original and published data on these and other aspects of the relationship between employment opportunities and changes in industrial techniques are now in progress under the direction of the National Research Project.

The term "productivity" as used in this chapter means the ratio: quantity output per employee man-year. For further discussion of the concept, see the section on "Over-all productivity."

Arthur F. Burns, Production Trends in the United States Since 1870 (New York: National Bureau of Economic Research, 1934), pp. 254-261.

of activity involve an effort on the part of an individual and an expenditure of some part of the country's stock of goods. If all commodities produced and all personal services rendered during the year are added at their market value, and from the resulting total we subtract the value of that part of the nation's stock of goods which was expended (both as raw materials and as capital equipment) in producing this total, then the remainder constitutes the net product of the national economy during the year. It is referred to as national income produced, and may be defined briefly as that part of the economy's endproduct which is attributable to the efforts of the individuals who comprise a nation." 5 The "national income produced" thus represents the net value of the goods and services produced through the "efforts whose results appear on the market place of our economy." 6

In order to convert these monetary values of income produced into measures representing the varying quantities of goods and services produced annually, it is necessary to adjust them for changes in the value of money. This may be done by applying an appropriate index of prices to the monetary value, thus obtaining what may be referred to as the "volume of goods and services produced." It is virtually impossible to construct a completely satisfactory index for deflating monetary national income. Since, however, some index must be used in order to make even a rought adjustment rather than "leave the dollar totals completely uncorrected for the striking changes in the purchasing power of the dollar that occurred” during the period since 1920, it was decided to use the index of "prices of finished goods" constructed by Dr. Simon Kuznets and used by him to adjust "income produced" for changes in purchasing power. The index used is of course subject to all the limitations which are inherent in a fixed-weight composite of its nature. Insufficient importance, for example, is given to consumers' services and "nonessentials" which constituted an increasing proportion of the national income produced over the period. Another criticism might concern the fact that the prices included in the index reflect the prices of goods sold during the year but not necessarily produced that year, while the index of output covers goods produced but not necessarily sold. In general, however, the broad outlines of the deflater used here are much the same as those

U. S. Congress, Senate, National Income, 1929-32, S. Doc. No. 124, 73d Cong., 2d sess., 1934, p. 1. For detailed description of the limitations of the concept "national income produced" see ch. I. • Ibid., p. 6.

7 Income Originating in Nine Basic Industries, 1919–34 (National Bureau of Economic Research Bulletin 59, May 4, 1936), p. 5. The index was constructed by combining an index of the price of capital goods with the B. L. S. index of the cost of living, using the weights 1 and 9, respectively.

of other indexes prepared for similar purposes and it is felt that despite the weaknesses inherent in any index of this nature the resulting "volume of goods and services" fairly depicts the growth and decline of the national output it purports to represent. More complete data and more refined techniques of measuring both national income and price changes would undoubtedly improve the accuracy of the year-to-year fluctuations, but it is highly questionable whether such improvements would materially change the general picture.

TABLE 1.-Indexes of monetary income produced, prices of finished goods, and volume of goods and services produced, 1920-35

1 For 1920-29, the index is based on "realized income", less "imputed income" plus "business savings" as shown in Levin, Moulton, and Warburton, America's Capacity to Consume (Washington, D. C., 1934), table 5, pp.152-153; this series was spliced to the "income produced" data, 1929-35, as shown in Survey of Current Business, July 1936, p. 18. Income from work relief was excluded.

S. Kuznets, Income Originating in Nine Basic Industries, 1919-34 (National Bureau of Economic Research Bulletin 59, 1936), table 3, p. 24. An index of the cost of capital goods was combined with the B. L. S. cost of living index with weights of 1 and 9, respectively. The figure for 1935 was constructed by the authors, using the same method.

* Column (2) divided by column (3).

• Preliminary.

NOTE. All figures were rounded after computations were made.

The three statistical series, monetary income, prices, and the quantitative volume of goods and services produced, are shown in table 1. It will be observed that while the total national monetary income produced rose 24 percent from 1920 to 1929, the total quantity of goods and services produced in the country, according to these estimates, increased 46 percent. After the sharp declines during the years 1929 to 1932, the monetary income rose 19 points from 1932 to 1935 to a point equal to 79 percent of the 1920 level, while the physical income climbed 26 points during the same period and reached a level 14 percent higher than the

quantity of goods and services produced in 1920. It should be borne in mind, however, that since the population of the country increased 19 percent from 1920 to 1935, the goods and services produced per capita in 1935 were still equal to only 96 percent of the 1920 production.

The National Labor Force

With this over-all picture of the fluctuations in the national output before us we now turn to the second question: How much hired labor was engaged in the creation of this annual product and how much of the labor available for employment remained unused? Again we run into statistical difficulties. While the obstacles encountered with respect to employment and unemployment statistics are different from those encountered in the field of production statistics, it is perhaps equally difficult to determine the number of workers employed during the years from 1920 to 1935. Although the amount of information concerning the volume of employment has grown immensely during the last two decades, the data available for the period prior to 1930 are rather fragmentary and the statistics for the years since 1930, while representing a vast improvement over the preceding period, still leave many fields of economic activity unexplored. It thus again became necessary to gather whatever information was available and to construct the best estimates of employment and unemployment that could be obtained under the circumstances.

United States Census occupation statistics for 1920 and 1930 present data on the number of "persons 10 years old and over gainfully occupied." Since this chapter concerns itself only with the effects of changing productivity on the employment opportunities of those who depend upon paid jobs for their livelihood and are, therefore, subject to unemployment, all enterprisers, self-employed, and unpaid family workers on farms have been subtracted from the census figures in order to arrive at the number of workers "available for employment." The figures for the years between 1920 and 1930 have been interpolated, taking into account the flow of immigration, emigration, and farm-city movements. For the years after 1930 the estimates are based on changes in the age distribution of the population, as well as immigration and farmcity movements, and therefore include those who under "normal" conditions would have obtained their first employment experience but who, during the period of

8 Gainful workers returning to farms are, like farmers and their unpaid family workers, not considered part of the labor supply so long as they remain on the farm. Their inclusion, however, would make very little difference in the size of the labor supply, since there was a net movement to the farm during the depression of only some 266,000 persons (in 1932), or approximately 100,000 gainful workers. See Farm Population Estimates (U. S. Dept. of Agr., mimeographed, Oct. 27, 1936.)

widespread unemployment, may never have worked." It was not found possible to make any adjustments which would take into account variations in the concept of "employability", although it is known that the drastic changes wrought by the depression have, on the one hand, compelled many persons to seek jobs who would otherwise not have been available for employment and, on the other hand, forced many people out of the labor market because of the realization of their inability to obtain employment.

Including therefore, as far as possible, all persons in the United States who are ordinarily employed by others, we arrive at an estimate of the total of the country's supply of labor for hire each year. It is known, however, that a portion of the labor supply at any given time is actually not available for work because of illness, vacations, voluntary transfers between jobs, labor disputes, and similar reasons for idleness. On the basis of the 1930 census data on unemployment, it is estimated that approximately 22 percent of the total supply of workers is "unusable" for these reasons.10 The remainder of the labor supply represents the country's total manpower available for hire. (See table 2 and fig. 1.)

The next step involved the construction of estimates of the proportion of this available manpower which was actually employed in the creation of the goods and services produced each year. Data available in the United States Bureau of the Census, the United States Bureau of Mines, the United States Bureau of Labor Statistics, and the Interstate Commerce Commission were used for this purpose, together with the data published by the Ohio State Department of Industrial Relations which give rather complete information on employment fluctuations in the various fields of economic activity. The annual estimates based on these sources represent averages of the total number of workers who appeared on pay rolls during each year since 1920. Since the sources make no distinction between part-time and full-time employees, the unad

The Census of Occupations includes only those who either have or at one time had a "usual occupation."

10 The 1930 Census of Unemployment reported that of the total "gainful persons" in 1930 approximately 1,000,000 workers, or 2 percent, were not available for work for these reasons. The groups included in this calculation consist of the following census classes: class C-persons out of a job and unable to work; class D-persons having jobs but idle on account of sickness or disability; class E-persons out of a job and not looking for work; class F-persons having jobs but voluntarily idle without pay. To these classes were also added that portion of those listed under class A (persons out of a job, able to work, and looking for work) and class B (persons having jobs but on lay-off without pay, excluding those sick and voluntarily idle) who were not working because of: voluntary absence, personal disability, family reasons, labor disputes, and dissatisfaction with the job.

Inasmuch as these million workers probably belonged almost entirely to the employee group considered in this chapter as constituting labor supply (instead of the entire census class "gainful persons"), it is estimated that at any time approximately 2 percent of the total labor supply is not available for work.

1 Excludes enterprisers, self-employed, and unpaid labor on farms. 'Allowance of 2.5 percent of the total labor supply for sickness and similar lost time. Column (2) less column (3).

Estimated full-time man-year equivalents of the average annual number of wage and salaried workers employed. The adjustments for part-time unemployment were based on data from 11 studies made in various cities.

Obtained by multiplying column (5) by the percentage found for 1920 (91.9). Column (4) less column (6).

1 Column (7) divided by column (4).

NOTE. All figures except columns (2) and (3) have been rounded off after computation.