Employment, Production, and the National Security Program in the United States

THE NATIONAL SECURITY program of the United States, as outlined in the July 1951 report of the President’s Council of Economic Advisers, envisaged that total defense expenditures would reach, at mid-1952, an annual rate of about 19 per cent of gross national product, which was expected to be at the level of $345 billion (in terms of prices in the first half of 1951). The probable level of civilian employment in the United States, when the current program reaches its planned “peak,” is obviously of direct importance to the stabilization policies not only of the United States but also of other nations. An estimate of this level is provided by an analysis of the relationships between output and employment in the United States during the past two decades—for the economy as a whole and, as far as possible, for important industrial divisions.1

Abstract

THE NATIONAL SECURITY program of the United States, as outlined in the July 1951 report of the President’s Council of Economic Advisers, envisaged that total defense expenditures would reach, at mid-1952, an annual rate of about 19 per cent of gross national product, which was expected to be at the level of $345 billion (in terms of prices in the first half of 1951). The probable level of civilian employment in the United States, when the current program reaches its planned “peak,” is obviously of direct importance to the stabilization policies not only of the United States but also of other nations. An estimate of this level is provided by an analysis of the relationships between output and employment in the United States during the past two decades—for the economy as a whole and, as far as possible, for important industrial divisions.1

THE NATIONAL SECURITY program of the United States, as outlined in the July 1951 report of the President’s Council of Economic Advisers, envisaged that total defense expenditures would reach, at mid-1952, an annual rate of about 19 per cent of gross national product, which was expected to be at the level of $345 billion (in terms of prices in the first half of 1951). The probable level of civilian employment in the United States, when the current program reaches its planned “peak,” is obviously of direct importance to the stabilization policies not only of the United States but also of other nations. An estimate of this level is provided by an analysis of the relationships between output and employment in the United States during the past two decades—for the economy as a whole and, as far as possible, for important industrial divisions.1

With revisions in the current defense program, the “peak” estimates for employment and production required to fulfill a revised program will differ from those calculated for the program as outlined in 1951. The statistical relationships obtained in this paper would not be invalidated, however, and estimates corresponding to a revised program could easily be derived from them.

The period covered by the analysis presented here starts with the year 1929 since many of the statistical series used are not available earlier. The official gross national product statistics compiled by the Department of Commerce provide the most notable example of this restriction. While GNP estimates for earlier years have been made by the Department of Commerce, they have never been officially released and, presumably, they are not so comprehensive and accurate as the later data.2

During the short span of the past twenty years or so, there have been both a severe depression and a world war, followed by a rapid demobilization and an unprecedented prosperity in the United States. Since the drastic changes during this short period have been numerous, satisfactory statistical relationships that are suitable either for explaining the past or for forecasting the future cannot easily be calculated from the recorded data. Therefore, the conclusions reached in this paper are necessarily experimental. Many of the relationships obtained are clear-cut, however, despite the small number of observations covered; they also correspond closely to what would be expected on logical grounds.

“Peak” Employment Estimates

The “peak” estimates of employment in mid-1952 are summarized in Table 1. The assumptions made in obtaining these estimates, and the classification of employment sectors, are explained in detail in subsequent sections of this paper. As indicated by Table 1, the probable level of total employment in mid-1952 ranges between 62.23 million and 62.73 million persons (seasonally adjusted).

From estimates of production and employment, the ratios of the probable percentage changes in employment to the corresponding percentage changes in production from the first half of 1951 to mid-1952 can be derived for both nondurable and durable manufacturing industries and for the economy as a whole. These ratios may be said to measure the “finite elasticities” of employment with respect to production (Table 2). If the finite elasticity is x, an expansion of y per cent in gross national product (GNP) will require an increase in total civilian employment of xy per cent.

The estimates of finite elasticities for the economy as a whole range from 0.33 to 0.46. The reasons which make possible increases in output much greater than the corresponding increases in employment are manifold; they include (1) shifts of manpower between the different fields of activity so as to result in a higher level of gross product for a given level of total employment, (2) a probable lengthening of the working week, and (3) prospective improvements in productivity in many fields.

The employment-GNP elasticity estimates given in the paper are considerably lower than an estimate that might be obtained from an analysis which does not give sufficient attention to shifts of manpower. (See section below, “Methods of Estimation.”) The difference would indicate that, according to our estimates, the increase in employment that is required for the expansion of GNP from the 1951 (first half) level of $323.8 billion to a higher given level (see Table 3) is smaller than the increase that might be indicated by other elasticity estimates. While the absolute magnitude of the difference might not be large, the difference becomes greater as GNP increases. It is also important that any required increase in employment from the present level is a marginal increase from practically full employment. Hence any small increase in this marginal increment means a proportionately greater increase in the strain placed on the economy.

In working out the estimates used in this paper, full consideration has been given to the fact that there are certain discernible “normal” relationships between GNP and its distribution among various industrial divisions, and that these relationships change when there is a shift from peacetime to wartime output. Evidence of these changing relationships is afforded by the experience during World War II, and similar changes can be expected during the present shift toward rearmament production. But since the manpower control measures that may be introduced in the next year or so are likely to be less severe than those during World War II, the shifts in manpower resulting from the rearmament program may be less pronounced. Therefore, the derived employment-GNP elasticity may be a little too low. However, it is in any case an indication of the level of GNP that is attainable at any level of employment.

The results described in this paper have been obtained from a small number of observations. The risks involved in drawing conclusions from such limited information are well known. In this particular case, however, the information that is available as a basis for the analysis is unavoidably limited, and a choice has to be made between a pure guess and a combination of bold “intuitive” judgment and quantitative extrapolation from recent experience. Most of the conclusions presented here are the result of an attempt to explore the second of these alternatives. They are not therefore the result of statistical inference in the ordinary sense of that term.

Table 1.

Summary of U.S. Civilian Employment Data—Reported and Estimatedby Industrial Divisions, 1950-52 1

(In millions)

article image

The reported data are those of the Department of Commerce and the Department of Labor. The estimated data are those obtained by the author in accordance with the methods described briefly in the notes following this table; see text for a more detailed description of these methods and also for details concerning the “alternative” estimates. Data are monthly averages, by years. Totals do not always equal the sum of the items because of rounding.

For a description of “classified” and “unclassified” employment, see text.

This estimate is not based on the estimates for the various industrial divisions. See text, section “Alternative Estimate of Classified Civilian Nonagricultural Employment as a Whole.”

Total of estimates for nondurable manufacturing, mining, trade and service, government, transportation and public utility, contract construction, and finance given in column 7, plus alternative estimate for durable manufacturing in column 8.

Total of estimates for trade and service, government, transportation and public utility, contract construction, and finance given in column 7 plus alternative estimate for industrial production as a whole given in column 9.

Includes estimate for agricultural employment given in column 7.

Relationships on the Basis of which the Estimates in Table 1 Are Made

Nondurable manufacturing

Estimates of nondurable manufacturing production were obtained on the basis of the statistical relationship between nondurable production and GNP in 1941-43 and 1946-49. Employment estimates were then obtained from the relationship between employment and nondurable production during the period 1949-51, first half. In spite of the short period covered, this relationship may be regarded as significant, because it is similar to the relationship between employment and production during World War II.

Durable manufacturing

Estimates of durable manufacturing production were obtained from the 1939-49 relationship between durable production and three independent variables (GNP, ratio of defense expenditure to GNP, and machine tools shipments). Employment estimates were then obtained on the basis of the relationship between employment and production in 1939-45 and 1949.

An alternative estimate was obtained from the direct relationship in 1939-49 between employment and two independent variables (GNP and the ratio of defense expenditure to GNP).

Mining

The relationship between employment and production in mining is rather haphazard, and it is assumed that “peak” employment will be the same as employment in the first half of 1951.

Alternative estimates for industrial employment as a whole

Estimates for employment in industrial production as a whole were also obtained from the direct relationship in 1939-49 between industrial employment and two independent variables (GNP and the ratio of defense expenditure to GNP).

Trade and service

A good relationship between employment in this field and GNP was obtained for the years 1929-40 and 1947-49. The 1949-51 (first half) points, however, reveal a tendency similar to that during World War II (1941-45), i.e., very little change in employment in this field in spite of expansion in GNP. The 1949-51 (first half) relationship was therefore used in estimating the “peak” employment in this field.

Government

Government employment is estimated from the 1932-49 relationship between employment in this field and GNP.

Transportation and public utilities

These data were estimated from the relationship between employment in this field and GNP during the period 1949-51 (first half).

Contract construction

The “peak” estimate is assumed to be the same as the 1951 (first half) figure, since expansion in construction of defense plants is likely to be balanced by restriction of civilian construction.

Finance

“Peak” employment is assumed to be the same as at the end of the first half of 1951.

Alternative estimates for total “classified” nonagricultural employment as a whole

These estimates were obtained from a simple regression relation between total “classified” nonagricultural employment and GNP in the period 1929-49. They are not based on estimates for the various industrial divisions.

Total “unclassified” nonagricultural employment

The postwar relationship between employment in this group and total “classified” nonagricultural employment has tended to return to the prewar (1933-40) relationship. In view of the severe decline in this relationship during the earlier years (1940-43) of World War II, however, the “peak” estimate was obtained by extending the 1950-51 (first half) declining tendency.

Agricultural employment

On the assumption that the effects of any increase in production on employment would be offset by further increases in efficiency, the “peak” employment in this field is estimated as being the same as in 1951 (first half).

Table 2

Measurement of “Finite” Elasticities of Civilian Employment with Respect to Production

article image

Based on data in Table 1.

Data for nondurable and durable manufacturing industries are based on indices of Board of Governors of Federal Reserve System; those for total agricultural and nonagricultural, i.e., the total economy, are based on GNP figures given in The Midyear Economic Report of the President (Together with a Report to the President by the Council of Economic Advisers) (Washington, D. C., July 1951).

The low and high figures relate to the 1952 figures given in the two columns “estimate” and “second alternative estimate” in Table 1.

Table 3

Estimates of Employment (Seasonally Adjusted) Required for Several Levels of Gross National Product1

article image

All the comparisons are in terms of seasonally adjusted employment figures. Though this is desirable, as the GNP figures are all seasonally adjusted annual rates, the differences will be proportionally the same whether adjusted or unadjusted employment figures are used.

Monthly averages, by years.

Actual figure as reported by the Department of Commerce.

Actual figure of the Department of Commerce, 60.19 million, seasonally adjusted.

Estimate of President’s Council of Economic Advisers.

Industrial Distribution of Civilian Labor Force

As a step preliminary to the construction of “peak” estimates, the changes during the last two decades in the industrial distribution of the total U.S. civilian labor force were first summarized, and an over-all picture was thus obtained of the utilization of manpower in the United States.

With the increase in population in the United States, the total labor force expanded during the past two decades by about 30 per cent, from 49.4 million persons in 1929 to about 64.7 million at the end of 1950. After military personnel are excluded from these figures, the increase is about 28 per cent—from 49.2 million in 1929 to 62.8 million in the first half of 1951 (seasonally adjusted).

The past two decades or so can be meaningfully divided into four periods: (1) the prewar period 1929-40, (2) the wartime period 1941-45, (3) the postwar period 1946-50, and (4) the first half of 1951, when the effects of the current defense program on the economy began to reveal themselves. A comparison of the changes in the industrial distribution of the total civilian labor force during these four periods provides a logical starting point for a study of the output-employment relationships during these twenty years.

The average civilian labor force during the four periods may be divided into four groups (Table 4): (1) “classified” civilian nonagricultural employment, (2) “unclassified” civilian nonagricultural employment, (3) agricultural employment, and (4) unemployment. The group, “classified” civilian nonagricultural employment, corresponds to the coverage of the Department of Labor’s monthly reports on full and part-time wage and salary workers in nonagricultural establishments who worked or received pay during the pay period ending nearest the fifteenth of the month. It is described as “classified” in this study, because the data are available for further subdivision into important industrial subgroups (Table 5).

Table 4

Civilian Labor Force and Percentage Distribution by Kind of Employment1

article image

Based on data in The President’s Economic Report (see Table 2, fn. 2), pp. 235-36.

Seasonally adjusted data obtained by taking moving averages. The data, after rounding, checked with data supplied by the Bureau of the Census.

Annual averages of monthly data.

Details do not necessarily add to 100 per cent because of rounding.

Table 5

“Classified” Civilian Nonagricultural Employment and Percentage Distribution by Industry Groups1

article image

Based on data in The President’s Economic Report (see Table 2, fn. 2), pp. 235-36, and the Federal Reserve Bulletin, August 1951, p. 1006.

Seasonally adjusted data.

Annual averages of monthly data.

Details do not necessarily add to 100 per cent because of rounding.

Averages of 1939 and 1940 figures only. Earlier data are not available for durable and nondurable industries separately. Average for manufacturing industries as a whole is 32.4 per cent for 1929-40.

The second group, “unclassified” civilian nonagricultural employment, is covered by reports of both the Department of Commerce and the Department of Labor. In addition to the Department of Labor’s reports based on the survey for the pay period ending nearest the fifteenth of the month, the Department of Commerce compiles reports on the labor force, based on a survey made during the week including the eighth of the month, which also give estimates for nonagricultural employment.3 This includes not only the “classified” salary and wage workers reported by the Department of Labor, but also the following four groups: (1) proprietors, (2) self-employed persons, (3) those who had jobs but were not at work for such reasons as vacation, illness, bad weather, temporary lay-off, or industrial disputes, and (4) domestic servants. Differences between the Commerce and Labor nonagricultural employment figures presumably represent the sum of these four groups, which is described in Table 4 as “unclassified” civilian nonagricultural employment.4 The agricultural employment and the unemployment figures given in the table are also based on Department of Commerce estimates.5

As shown in Table 4, the percentage of “classified” nonagricultural employment in the total labor force increased greatly, at the expense of the other groups, from the prewar period to the wartime period. Undoubtedly, defense activities made the heaviest demands on this “classified” group. The reduction in employment for this group during the postwar period, however, was very slight, and the further increase in the first half of 1951 reflects the effects of the current defense program.

The percentage of the total civilian labor force in the “unclassified” nonagricultural group fell greatly from the prewar period to the wartime period, but after the war it more than regained its losses. There was a further gain in the first half of 1951, compared with the period 1946-50.6

The most striking feature of Table 4 is the consistent decline of agricultural employment, as a percentage of the civilian labor force. This was due mainly to the tremendous strides made in the improvement in agricultural technique and equipment and to the diminishing number of less-than-fully-occupied farm workers.

Unemployment, of course, declined greatly from the prewar period to the wartime period. The unemployment percentage was kept low by the unprecedented prosperity after the war; and the expansion of defense activities brought a further decline in the first half of 1951.

“Classified” civilian nonagricultural salary and wage workers represent the most important sector of the labor force. Among the industrial divisions of this group (Table 5), employment in each division, as a percentage of the group total, declined from the prewar period to the wartime period, except the percentages for the durable manufacturing industries and for government which gained greatly. This is what would be expected, since the demands from the mobilization process and from the expansion in the production of heavy military equipment during the war were greatest in these two fields. This tendency was reversed during the change from war to demobilization and reconstruction, except in the mining industries where a small percentage decline continued. Increases in the percentages for the durable manufacturing industries and government are again observed when the first half of 1951 is compared with the five preceding years. The changes in the other divisions during the same period will be brought out more meaningfully in the year-to-year analysis in later sections.

Methods of Estimation

While the changes in the industrial distribution of the labor force throw some light on the effects of the last war and the current defense program on employment, they do not provide a sufficient basis for deriving relationships between output and employment, or for estimating production and employment at mid-1952.

These estimates can, of course, be made in several ways. For example, a “peak” estimate of output could be obtained from some “expected” employment figure: by deducting “estimated” unemployment from the civilian labor force that is “expected” to be available at mid-1952, an estimate could be made of total civilian employment at that time, and then on the basis of some average productivity estimates, and after adjustments for expected weekly manhour changes, an estimate of gross national product could be obtained.7

This approach, however, overlooks the probable shifts of manpower under the impact of mobilization requirements and policies and the associated changes in productivity which were discussed in an earlier section of this paper. Therefore, in our analysis the relationships between the levels of GNP8 and production in the different fields9 are first investigated wherever production data are available. For some industries, other variables have to be introduced in order to obtain a stable relationship over a sufficiently long period of time. The production-employment and the production-manhours relationships are then studied for these fields. In the cases for which production data are not available, or where the distinction between employment and production is rather unimportant,10 the direct relationship between GNP and employment is investigated.

These relationships should be of interest in their own right. In addition, estimates for output and employment (both total and in the different fields) at mid-1952 can easily be derived from them. In a few instances where it appeared that the relationships in the last two years or so were similar to those during the earlier stage of World War II, “peak” estimates have been based on projections from these short-term trends, instead of those obtainable from the “normal relationships.”

Nondurable Manufacturing Industries

The numbers of salary and wage workers employed since 1939 in nondurable manufacturing industries (Chart 1) indicate a long-term trend toward increased employment in this group. This increase, however, no more than reflects the growth of the total labor supply. The changes in this group as a percentage of “classified” employment have not been substantial, though a downward trend during the early war years, a brief reversal of this trend in 1944-46, and a subsequent decline are clearly observable.

Chart 1.
Chart 1.

Nondurable Manufacturing Industries: Employment, 1939-51, and Relationship between Production and Gross National Product

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

Production

From the relationship between nondurable manufacturing production and GNP since 1929, shown in the lower half of Chart 1, it is evident that, with the exception of 1944 and 1945, a good linear relationship held throughout the period, though the path followed during the prewar downswing years 1929-32 was somewhat different from that of the subsequent upswing. It might have been expected that the severe restriction of civilian consumption during the war and its subsequent large-scale expansion would have upset the normal relationship which had become apparent before the war. A large part of the defense requirements during the war were, however, for nondurable goods, so that the reduction in civilian consumption was largely compensated by an increase in military consumption. The postwar points indicate a tremendous expansion in the production of nondurable goods, but their relation to total output has been substantially the same as in the prewar upswing years. The period from 1941 to 1949, with the same exceptions, also presents a good linear relationship which is only slightly different from the relationship for the longer period.

The years 1944 and 1945 were exceptional, no doubt because the share of munitions and military equipment in GNP was at that time at its highest, and the price deflation used in obtaining the real GNP figures has not been adequate to adjust the result to the well-known overvaluation of these items.11

From the data recorded in Chart 1, linear regression equations may be deduced, from which hypothetical production figures may be calculated for subsequent years. Two equations, covering the periods 1933-49 and 1941-49, have been fitted for this purpose, 1944 and 1945 being omitted in both cases. The year 1950 and the first half of 1951 have also been omitted, so that the actual values for these years may be compared with the extrapolated values based upon the regression equations.12 These computed values turn out to be very close to the recorded values, as shown in Table 6.

Table 6

Index of Nondurable Manufacturing Production

(1935-39 = 100)

article image

From Federal Reserve Bulletin.

Equation in Appendix.

The “average” elasticity of nondurable industrial production with respect to GNP, i.e., the slope of equation 1 or 2 multiplied by the ratio of the mean values of GNP and the index of nondurable manufacturing production, is 1.01 or 0.93. In other words, an increase in GNP means an approximately equal percentage expansion in nondurable manufacturing production.

If GNP reaches $345 billion by the middle of 1952, the probable level of the index of nondurable manufacturing production at that time, based upon these regression equations, will be 215 or 212 (1935-39=100). In making the employment estimate below, the latter figure has been used, because equation 2, upon which it is based, gives a slightly better fit for the first half of 1951.

Employment

From the data, shown in Chart 1a, for employment (data being available only since 1939) and production in nondurable manufacturing industries, it is clear that the postwar points for 1946-49 form a good linear relationship with the prewar points for 1939-41. For the war years, however, the points form another line, with a much less pronounced slope, increases in weekly working hours and improvements in efficiency per manhour making possible a greatly increased production without any significant change in the number of workers. A similar tendency was apparent in 1950 and the first half of 1951. If it is assumed that this trend will continue, employment in mid-1952 may be estimated by extending the linear relationship indicated by the figures for 1949, 1950, and the first half of 1951.13 This gives an estimated “peak” employment in nondurable manufacturing industries of about 7.18 million. The elasticity of employment with respect to production is 0.31, which means that a one per cent expansion in production requires an increase of only 0.31 per cent14 in the number of workers.

Chart 1a.
Chart 1a.

Nondurable Manufacturing Industries: Relationship between Employment and Production, and between Manhours Per Week and Production

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

Manhours

The number of man hours worked is an element directly relevant to this study, especially in view of the fact that the wartime expansion of nondurable manufacturing production was obtained with an almost constant number of workers. The relationship between weekly manhours worked and production in this field is shown in the lower half of Chart 1a. A linear relationship could be fitted to the data for the period 1939-49, but closer study reveals more meaningful relationships.

Data for the earlier years of World War II, 1940-43, and the period 1949-first half of 1951, which are the initial years of the two defense programs, form a nearly perfect parabolic curve. As the respective defense programs unfolded, the efficiency of nondurable manufacturing production improved in much the same way. There is also a fairly good linear relationship for the three immediate postwar years, 1946-48, and another linear relationship for the last three points, 1949-first half of 1951.

If 212 (1935-39 = 100) is taken as the “peak” estimate for the index of nondurable manufacturing production, a reasonable estimate of the number of weekly manhours worked in nondurable manufacturing industries can be obtained by using either the parabolic curve or the linear relationship for 1949-51 (first half).15 The estimate obtained from the parabolic curve is 283 million manhours per week, and that secured from the linear relationship is 287 million.

Durable Manufacturing Industries

The phenomenal increase in employment in durable manufacturing industries from 1939 to 1944, reflecting the heavy demand during the last war, and the subsequent decline both in actual numbers and in percentages of the total civilian labor force and of “classified” civilian nonagricultural employment are shown in Chart 2. The effects of the current defense program on these industries are indicated by the increases in all three measurements in 1950 and the first half of 1951.

Chart 2.
Chart 2.

Durable Manufacturing Industries: Employment, 1939-51, and Relationship between Production and Gross National Product

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

Production

Durable manufacturing production is the field in which the impact of the defense program is expected to be greatest. With real GNP16 at any given level, any increase in production for defense may be expected to require a higher level of production of durable manufactured goods. Therefore it is not surprising to find many shifts in the relation between durable manufacturing production and GNP in the period 1929-first half of 1951 (lower half of Chart 2). When, however, the peacetime and wartime periods are distinguished, a meaningful interpretation of the changes in the relationship becomes possible.

As observed above for the nondurable manufacturing industries, the prewar downswing years (1929-32) followed a somewhat different path from the upswing years (1933-37). During the period 1938-43, the steady expansion of the defense program shifted the relationship, with a steeper slope than in the peacetime upswing years. With the decline in GNP in 1944-46 there was a steady reduction of durable production. It is rather unexpected that, in spite of the heavy backlog of demands for durable civilian goods, the postwar points 1947-49 are actually below the prewar upswing relationship (1933-37).

The 1949-51 (first half) points form a perfect straight line. It would be tempting simply to project this line to forecast the level of durable production when the defense program reaches its planned “peak” at the middle of 1952; but there are some other variables which should be taken into account in explaining the behavior of durable manufacturing production over a longer period of time. One of these relevant variables is, obviously, the ratio of defense expenditure to total product. Another is machine tool shipments. Both in the earlier years of the last war and at the present time, machine tools have been a bottleneck checking expansion in the production of durable military equipment. In spite of heavy orders, production is not possible if adequate machine tools are not available for retooling the plants.

A multiple regression equation covering the period 1939-49, in which GNP, the ratio of defense expenditure to total product, and the index number of machine tool shipments17 are regarded as three independent variables, has accordingly been constructed,18 and the extrapolated figures for production in 1950 and the first half of 1951 based upon this function correspond reasonably well to the recorded figures for these periods (Table 7). The probable index of durable manufacturing production, based upon this regression equation, in mid-1952 when the defense program reaches its planned “peak,” is 324.19

Table 7

Production, Employment, and Weekly Manhours for Durable Manufacturing Industries

article image

From Federal Reserve Bulletin.

Based on equation 6 in Appendix.

Based on data of the U.S. Department of Labor. The figure for the first half of 1951 is a seasonally adjusted figure obtained by the method described in footnote 13, p. 293.

Based on equation 7 in Appendix.

Based on data of the U.S. Department of Labor.

Based on equation 8 in Appendix.

Employment

The relationship between employment and production in the durable manufacturing industries since 1939 (the first year for which separate employment figures for this industrial group are available) is shown in the top section of Chart 2a. While a fairly good linear relationship is observable for the period as a whole, it is particularly good for the 1939-45 and 1949-51 (first half) points, the points corresponding to the years during which the two defense programs were expanding. A linear regression equation,20 covering 1939 to 1945 and 1949, again gives satisfactory results when the values for 1950 and 1951 (first half) obtained by projection are compared with the actual record of employment in those periods (Table 7). On the basis of this regression equation, the “peak” employment in this field, corresponding to the “peak” production index of 324, is 10.14 million.

Chart 2a.
Chart 2a.

Durable Manufacturing Industries: Relationship between Employment and Production, and between Manhours Per Week and Production

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

Manhours

During World War II there was a substantial increase in the weekly hours worked per man, so that the relationship between weekly manhours and production, shown in the lower section of Chart 2a, might be expected to be even better than that between employment and production.

As expected, the correlation coefficients of this relationship, both for 1939-49 as a whole, and for the years 1939-45 and 1949, are slightly higher than the corresponding coefficients for the employment-production relationships. The figures for 1950 and 1951 (first half) projected from the equation for 1939-45 and 194921 are, however, not so close to the actual figures as in the employment case (Table 7). The estimate of “peak” weekly manhours, corresponding to a “peak” production index of 324, is 464 million.

Alternative estimates of employment and manhours

These results may be compared with those obtained by a simpler method of estimation, which directly derives employment and manhours worked in durable manufacturing production from their relationships with GNP and the ratio of defense expenditure to GNP during 1939-49. By extrapolating for 1950 and 1951 (first half) from the employment equation,22 an estimate is obtained of 8.55 million workers in 1950, compared with the reported figure of 8.01 million; and for the first half of 1951, the estimate is 9.37 million workers, against the actual figure of 8.97 million.23 These estimates are not so close to the actual magnitudes as those obtained above. The corresponding “peak” estimate is 10.47 million workers, somewhat larger than the previous estimate of 10.14 million. (The estimate of 10.47 million workers is given in Table 1 as the “first alternative” estimate for mid-1952.)

The extrapolated values derived from the weekly manhours equation,24 however, are closer to the actual data than those obtained above: the estimate for 1950 is 340 million weekly manhours, against the reported figure of 330 million; for the first half of 1951, the estimate is 378 million, compared with the reported figure of 372 million. The “peak” estimate obtained from this equation is 437 million weekly manhours, somewhat below the earlier figure of 464 million.

Mining Industries

In terms of percentages of the total civilian labor force and of total “classified” civilian nonagricultural employment, a slight long-term downward trend in mining employment is observable (Chart 3). Even the heavy demand for mineral production during World War II did not significantly affect this tendency.

Chart 3.
Chart 3.

Mining Industries: Employment, 1929-51, and Relationship between Production and Gross National Product, and between Employment and Production

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

Production

The elasticity of mining production with respect to GNP based upon a regression equation25 for the period 1929-40 was 1.08. The wartime points followed a path with a less steep slope (Chart 3). The 1949-51 (first half) points form a straight line with a slope similar to that of the prewar years. By a simple extension of this relationship, the “peak” index of mineral production is estimated at 176 (1935-39 = 100).

Employment

While the relationship between mining employment and production is haphazard, it can be seen from the bottom section of Chart 3 that the years 1929-42 may be divided into three periods, 1929-35, 1936-38, and 1939-42, with output per man tending to increase in each successive period. These increases are not to be interpreted as evidence of improvements in efficiency, since there must also have been substantial changes in weekly manhours worked in this field.26 The important point is that the 1949-51 (first half) points show an increase in the production index from 135 to 163 (1935-39=100) without any increase in the number of workers. The “peak” employment figure is assumed to be the same as in the first half of 1951, 0.92 million workers;27 but, since a considerable expansion is expected in this field, this estimate may be too low.

Alternative Estimate of Industrial Employment

An alternative estimate of industrial employment as a whole can be obtained by relating this variable directly to GNP and to the ratio of defense expenditure to GNP for the period 1939-49. The 1950 employment estimate obtained from this equation 28 is 16.66 million workers, compared with the actual figure of 15.79 million; for the first half of 1951, the estimate is 17.78 million, against the reported figure of 16.93 million.29 The “peak” employment in industrial production as a whole is estimated from this equation at 18.99 million workers. (This estimate is given in Table 1 as the “second alternative” estimate.) This may turn out to be an over-estimate, if the trend of negative “residuals” in 1950 and the first half of 1951 is maintained.

Trade and Service Industries

The trade and service industries rank next to the manufacturing industries in the number of salary and wage workers employed, averaging roughly 14 million per month in 1950.30 As shown in Chart 4, the proportion of trade and service employment in total classified nonagricultural employment has been fairly constant: it averaged about 32 per cent both in the thirties (1920-40) and after World War II (1946-49). However, it dropped to about 28 per cent during the war. As a percentage of the total civilian labor force, its importance has steadily increased since 1933.31

Chart 4.
Chart 4.

Trade and Service Industries: Employment, 1928-51, and Relationship between Employment and Gross National Product

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

Trade and service employment had an excellent linear relationship with GNP during the period 1929-40 and 1947-49, the prewar relation being restored after the war (lower section of Chart 4). The “average” elasticity of employment in this field with respect to GNP is about 0.7; an expansion in GNP in peacetime would therefore result in a less-than-equal percentage increase in trade and service employment.

The number of workers employed in this field during World War II was held down at an almost constant level of about 11 million in spite of the tremendous increase in total real output. The 1949-51 (first half) points reveal the same tendency. The three points form a perfect straight line with a very slight slope.32 By projecting this line, the “peak” employment in this field is estimated at 14.64 million. This may be an underestimate, however, since with less drastic military draft and manpower control measures, employment in this field may have a better chance to expand with GNP during the defense program than during the war.

Government

There has been a long-term tendency during the past two decades for government employment to grow, both in terms of actual numbers of workers and as percentages of the total civilian labor supply and of total “classified” civilian nonagricultural employment (Chart 5). A consistently good linear relationship between government employment and GNP is shown throughout the period 1932-49.33 The “average” elasticity of government employment with respect to GNP is estimated at 0.6. The levels of employment extrapolated from the regression equation for 1950 and the first half of 1951 differ very little from the actual figures: an estimated 6.05 million persons in 1950, compared with the actual figure of 5.91 million; and an estimated 6.43 million for the first half of 1951, against the actual figure of 6.26 million (seasonally adjusted). By further projection of this linear relationship, the “peak” government employment is estimated at 6.78 million.

Chart 5.
Chart 5.

Government: Employment, 1929-51, and Relationship between Employment and Gross National Product

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

Transportation and Public Utilities

Employment in transportation and public utilities has failed to show any long-term growth commensurate with the increase in total civilian labor supply (see Chart 6). In actual numbers, employment in this field was only slightly larger in 1951 (first half) than in 1929. In terms of percentages of total “classified” civilian nonagricultural employment, there has been a rather clear downward trend.

Chart 6.
Chart 6.

Transportation and Public Utilities: Employment, 1929-51, and Relationship between Employment and Gross National Product

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

From the relation between transportation and public utilities employment and GNP, shown in the lower half of Chart 6, it is clear that the prewar upswing years 1933-39 followed a different path from the downswing years 1929-32. The early war years, 1940-44, form a good linear relationship among themselves. Throughout the postwar years the curve seems definitely to have shifted upward.

It is, however, very difficult to construct from these frequently changing relationships a regression equation that can be used for forecasting. In view of the good linear relationship during the early years of the last war, the straight line connecting the three points, 1949-51 (first half), may be taken to represent the relationship likely to be maintained during the next year or two.34 The “peak” employment in transportation and public utilities thus estimated is 4.20 million.

Contract Construction

The generally accepted view that fluctuations in construction employment are more severe than in employment as a whole is confirmed by the data recorded in Chart 7. The severe drop during the prewar depression years and the sharp increase after World War II are clearly revealed. The great expansion in the earlier war years for the construction of defense plants was followed by an equally sharp decline during the later years of the war as a result of the restriction of civilian construction. From the relationship between employment in contract construction and GNP (lower half of Chart 7), it is again seen that the downswing trend (1929-32) was different from the upswing (1933-40). The rapid building-up of defense plants shifted the relationship sharply upward in 1941 and 1942. During the years 1943-45, defense efforts were concentrated upon producing the final products from these plants, and the corresponding downward tendency in construction employment shown in the chart clearly reveals the effects of the severe restrictions in those years on construction for ordinary purposes. The high positions of the postwar points (1947 onward) correspond to the fulfillment of civilian demands pent-up during the war.

Chart 7.
Chart 7.

Contract Construction: Employment, 1929-51, and Relationship between Employment and Gross National Product

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

Many, though not all, of the defense plants required for the current preparedness program are already available. The experience of the last war suggests the probability of a downturn in the near future similar to, though not so sharp as, that in 1943. The restriction of ordinary construction will, however, probably be much less severe than in 1943-45. For a forecast of construction employment at the planned “peak,” there is therefore a choice between assuming approximately the same level of employment as in the first half of 1951 and assuming a continuation of the linear trend of 1949-51 (first half). Since it is probably more prudent to make the former assumption, the “peak” employment in the field of contract construction is estimated at the 1951 (first half) level of 2.55 million (seasonally adjusted); this, however, may turn out to be an underestimate.

Finance

A long-term growth of employment in finance, commensurate with the increase in the civilian labor force, is seen in Chart 8. The reduction during the war years of employment in this field in terms of percentages of total “classified” civilian nonagricultural employment corresponds to the transfer of manpower to more essential fields.

Chart 8.
Chart 8.

Finance: Employment, 1929-61, and Relationship between Employment and Gross National Product

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

In finance, as in the two preceding fields studied, the relationships between employment and GNP during the prewar downswing years (1929-32) and the upswing years (1933-40) followed different paths, as shown in the lower section of Chart 8. Reflecting the effects of policies designed to ensure the rational utilization of manpower, employment in finance, in spite of the increase in GNP, showed a clear downward trend during World War II.

The postwar points indicate that the prewar upswing relationship has been more or less restored. A regression equation35 covering the period 1933-40 and 1947-49 gives an “average” elasticity of employment in finance with respect to GNP of about 0.43.

The levels of employment for 1950 and the first half of 1951 projected from this equation correspond quite closely to the actual figures: an estimate of 1.78 million persons for 1950, against the actual figure of 1.81 million; and an estimate of 1.86 million for the first half of 1951, which is the same as the actual figure.

In view of the downward trend during the last war, the “peak” employment in finance should not, however, be estimated from this equation. It is more reasonable to assume that it will be maintained at 1.87 million, the level prevailing at the end of the first half of 1951.36 Because of the less drastic military draft and manpower control measures so far adopted, this may be an underestimate.

Alternative Estimate of “Classified” Civilian Nonagricultural Employment as a Whole

Data for the total numbers of “classified” salary and wage workers37 employed in civilian nonagricultural establishments from 1929 to the first half of 1951 are plotted in the upper section of Chart 9, together with the percentage relations of this group to the total civilian labor force. The trends are of course a summation of the component tendencies described in preceding sections. A supplementary estimate of the total “classified” workers can be derived directly from the past relationship of these figures with GNP, shown in the lower section of the chart. A correlation coefficient as high as 0.977 is obtained between these variables for the period 1929-49 as a whole.

Chart 9.
Chart 9.

Total “Classified” Civilian Nonagricultural Employment, 1929-51, and Relationship between Employment and Gross National Product

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

The employment figures estimated from the regression equation38 are 44.07 million workers for 1950 and 46.62 million for 1951 (first half), compared with the actual figures of 44.12 million for 1950 and 46.25 million (seasonally adjusted) for the first half of 1951. The probable “peak” level of total “classified” civilian nonagricultural employment at midyear 1952, as estimated from this equation, is 48.93 million—the figure given in Table 1 as the “third alternative” estimate.

While the estimated figures for 1950 and 1951 (first half) are very close to the actual data, there was a considerable divergence in 1944-45 and 1947-48. The 1944-45 deviations could be explained by the “flattening out” tendency in the employment-output relationship, especially pronounced in those years, in nondurable manufacturing, trade and services, contract construction, and finance, which has already been explained in earlier sections. No really satisfactory explanation can be given here for the deviations in 1947-48.

“Unclassified” Civilian Nonagricultural Employment

Consideration should now be given to the group “unclassified” civilian nonagricultural employment,39 which covers the four categories (1) proprietors, (2) self-employed persons, (3) those who had jobs but were not at work for such reasons as vacation, illness, bad weather, temporary lay-off, or industrial disputes, and (4) domestic servants. While detailed figures are lacking for workers in these categories, classified by industrial divisions, it is possible to estimate the number in these four groups as a whole for the period studied.

As shown in Chart 10, this group of workers, both in actual numbers and as a percentage of the total civilian labor force, decreased in the depression years; the numbers then rose gradually and by 1940 were almost the same as in 1929. During World War II there was a severe decline. In the postwar years, the actual number exceeded that before the prewar depression; in percentage terms, however, it just about returned to the previous level.

Chart 10.
Chart 10.

“Unclassified” Civilian Nonagricultural Employment, 1929-51, and Relationship between “Unclassified” and “Classified” Civilian Nonagricultural Employment

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

For this group the number of persons employed is related to total classified nonagricultural employment (lower half of Chart 10) instead of to GNP. The prewar upswing trend (1933-40) is again different from the prewar downswing (1929-32). The number of persons decreased sharply during the war as a result of wartime manpower policies, whereas in the postwar years there was a tendency toward re-establishing the prewar upswing relationship.

The 1951 (first half) point corresponds closely to the point projected according to a regression equation for 1933-40. On account of the severe decline in 1941-43, this equation should not be used to estimate the “peak” number of people in this group. In the absence of any better alternative, it may be assumed that the “peak” figure can be located by projecting the line connecting the 1950 and 1951 (first half) points. The figures thus obtained corresponding to the four estimates given in Table 1 are 6.90 million, 6.79 million, 6.65 million, and 6.68 million. These may all be underestimates, since the military draft and manpower control measures adopted so far under the defense program are by no means so drastic as those of World War II.

Agriculture

There has been a consistent long-term decline in agricultural employment during the past two decades, both in the actual number of workers engaged and as a percentage of the total civilian labor force (Chart 11).

Chart 11.
Chart 11.

Agriculture: Employment, 1929-51, and Relationship between Agricultural Production and Gross National Product, and between Agricultural Employment and Production

Citation: IMF Staff Papers 1952, 001; 10.5089/9781451949391.024.A004

Production

A study of both the relationship between agricultural production and GNP and that between agricultural employment and production is of interest. As shown in Chart 11, the former relationship is so scattered that GNP clearly cannot be regarded either as a factor explaining past variations in agricultural production or as a “predictor” for the future, even though a positive functional relationship is discernible between the two variables. Agricultural production is subject to a host of random influences (such as weather conditions, floods, insects, diseases, etc.), so that the absence of any clear relationship is more or less to be expected. The chart permits little more to be said than that, after a rapid expansion in the earlier war years, agricultural production was more or less constant during 1942-45 owing to manpower shortages, in spite of the tremendous increase in GNP, and that there was a relative increase in agricultural production in the postwar years 1948-49.

Employment

The increase in labor productivity in agriculture between 1937 and 1950 was so great and continuous (bottom section of Chart 11) that the relationship between employment and production is actually negatively inclined. This development was due mainly to two factors. First, there were great improvements in agricultural technique and equipment during that period. Second, the number of farm workers less than fully occupied was declining. For agricultural employment, it is therefore desirable to construct a multiple regression equation with two independent variables, agricultural production and time.40 Such an equation is of little value, however, in forecasting the “peak” employment in agriculture, for there is no estimate available of the probable level of agricultural production when the defense program reaches the planned “peak.” There is only the assurance that “domestic supplies of foods in the coming year will probably be at least as great as those of the last year.”41 Since the effects of a probably larger production on employment may be compensated by further improvement in efficiency, the bold assumption is made that the “peak” agricultural employment will be at about the same level as that during the first half of 1951, i.e., about 7.05 million (seasonally adjusted).

APPENDIX

The regression equations from which the extrapolated values recorded in the text have been derived are listed below. GNP is measured throughout in billions of 1939 dollars; P represents the index number of production (1935-39=100) in the relevant industrial sector;42 E the volume of employment, in millions of workers; and M the weekly manhours worked, in millions.

Nondurable Manufacturing Industries

  • 1. Period, 1933-43 and 1946-49

    • P=1.22GNP-1.79

    • R=0.996

  • 2. Period, 1941-43 and 1946-49

    • P=1.13GNP+11.06

    • R=0.968

  • 3. Period, 1949-51 (first half)

    • E=0.012P+4.740

  • 4. Period, 1940-43 and 1949-51 (first half)

    • M=2.96P-0.65P2-54.03

    • R=0.999

  • 5. Period, 1949-51 (first half)

    • M=0.62P+154.74

Durable Manufacturing Industries

  • 6. Period, 1939-49

    • P=1.85(0.39)GNP+2.05(0.60)D/GNP+0.19(0.07)T-77.84

      (where D = defense expenditure and T= index of machine tools shipments)

    • R=0.973

  • 7. Period, 1939-45 and 1949

    • E=0.025P+1.960

    • R=0.996

  • 8. Period, 1939-45 and 1949

    • M=1.36P+22.62

    • R = 0.999

  • 9. Period, 1939-49

    • E=0.053GNP+0.056D/GNP+0.076

    • R=0.954

  • 10. Period, 1939-49

    • M=2.01GNP+4.02D/GNP+4.29

    • R=0.956

Mining Industries

  • 11. Period, 1929-40

    • P=1.27GNP-7.30

    • R = 0.967

Industrial Employment as a Whole

  • 12. Period, 1939-49

    • E=0.082GNP+0.034D/GNP+3.720

    • R = 0.943

Trade and Service Industries

  • 13. Period, 1949-51 (first half)

    • E=0.013GNP+12.30

Government

  • 14. Period, 1932-49

    • E=0.032GNP+1.106

    • R=0.993

Transportation and Public Utilities

  • 15. Period, 1949-51 (first half)

    • E=0.007GNP+2.951

Finance

  • 16. Period, 1933-40 and 1947-49

    • E=0.0063GNP+0.8123

    • R=0.991

“Classified” Civilian Nonagricultural Employment

  • 17. Period, 1929-49

    • E=0.21GNP+11.45

    • R=0.977

Agriculture

  • 18. Period, 1937-50

    • E=0.00674-0.1840t+9.4068

    • R=0.984

      where A=index of agricultural employment (1935-39 = 100) and £=time (1 for 1937, 2 for 1938, etc.)

*

Mr. Liu, economist in the Statistics Division, is a graduate of National Chiao-Tung University, and received graduate training at Cornell University. He was formerly Assistant Commercial Counselor, Chinese Embassy, Washington, D. C., and Professor of Economics, National Tsing-Hua University, Peiping.

1

The estimated levels of production and employment at midyear 1952 will be referred to briefly, in this paper, as the “peak” estimates. The use of this phrase does not, however, imply that peak levels will actually be reached at that time.

2

Most of the statistical data used are found in The Midyear Economic Report of the President (Together with a Report to the President by the Council of Economic Advisers) (Washington, D. C., July 1951), especially Tables B-2, B-3, B-ll, B-12, B-13, B-16, and B-17 in Appendix B. This publication will be cited below as The President’s Economic Report. The total national security expenditure corresponding to the 19 per cent of GNP cited in the text is given as $64 billion in the President’s Economic Report, Table 2, p. 69. The GNP figure obtained by dividing $64 billion by 19 per cent is somewhat smaller than the annual rate of $345 billion cited elsewhere in the Report (p. 60) as the expected level of GNP at the end of the second quarter of 1952. In deriving the “peak” estimates presented in this paper, the figure of $345 billion has been used. The figures obtained represent estimates for the end of the second quarter of 1952 in terms of annual rates.

3

The President’s Economic Report, Table B-11, p. 235.

4

These differences give only an approximate measure of “unclassified” civilian nonagricultural employment, since the concepts and methods used in the two surveys are not exactly comparable. In any case, the “unclassified” group roughly corresponds to the residuals left after “classified” salary and wage workers have been subtracted from total civilian nonagricultural employment.

5

The President’s Economic Report, Table B-11, p, 235.

6

In comparison with 1950 alone, however, the figure for the first half of 1951 shows a sharp decline. See below, p. 312.

7

This seems to be the method used by the Council of Economic Advisers in deriving GNP figures for the various periods in 1952.

8

The GNP figures used in deriving the statistical relationships in the later sections are “real figures” in 1939 dollars (as given in The President’s Economic Report, Table B-2, p. 226), instead of in 1951 (first half) dollars, since the latter were not available until after this study was well under way. The “peak” GNP figure in 1939 dollars that corresponds to the figure of $345 billion in 1951 (first half) dollars is $177.3 billion.

9

The fields covered are industrial production—which includes nondurable goods manufacturing industries, durable goods manufacturing industries, and mining—trade and services, government, transportation and public utilities, contract construction, finance, “unclassified” civilian nonagricultural employment, and agriculture.

10

For instance, the trade and service industries.

11

It can be seen from the chart that, if the GNP figures had been smaller in these years, the two points would have been closer to the linear relationship between the two variables.

12

See equations 1 and 2 in the Appendix.

13

See equation 3 in the Appendix. A seasonally adjusted employment figure is available for the first half of 1951 only for manufacturing industries as a whole (16.01 million). An estimate, 7.04 million, has been obtained for nondurable manufacturing industries by multiplying 16.01 million by the ratio of seasonally adjusted factory employment in nondurable manufacturing industries to that in manufacturing industries as a whole.

14

This figure, which represents the “average” elasticity of the equation, is slightly higher than the “finite” elasticity given above (p. 284).

15

See equations 4 and 5 in the Appendix.

16

See footnote 8, p. 288.

17

This index is in terms of physical volume, obtained by deflating the Department of Commerce index of value of shipments by the wholesale price index of metal products.

18

See equation 6 in the Appendix.

19

The figure used for machine tool shipments in this computation was obtained by a simple linear trend extrapolation of the index to mid-1952. This is a reasonable procedure, for during the earlier stages of the last war there was also a linear trend in machine tool shipments.

20

See equation 7 in the Appendix.

21

See equation 8 in the Appendix.

22

See equation 9 in the Appendix.

23

See footnote 3, Table 7.

24

See equation 10 in the Appendix.

25

See equation 11 in the Appendix.

26

No comprehensive data are available for manhours covering the entire field of mining.

27

Seasonally adjusted.

28

Equation 12 in the Appendix.

29

Seasonally adjusted.

30

A change in the reporting routine (i.e., a shift since 1939 of the automobile repair service industry from the trade to the service division) makes it impossible to investigate the trade and service industries separately. See The President’s Economic Report, Table B-12, note 2, p. 236.

31

The unclassified nonagricultural employment also includes a large number of proprietors and self-employed persons in the trade and service division. (See section below on Unclassified Civilian Nonagricultural Employment.)

32

See equation 13 in the Appendix.

33

See equation 14 in the Appendix.

34

See equation 15 in the Appendix.

35

See equation 16 in the Appendix.

36

Seasonally adjusted figure for June 1951. The June figure, instead of the average figure for the first half of 1951, is used because employment in this field during these six months tended to increase.

37

For a description of this term, see section above, “Industrial Distribution of Civilian Labor Force.”

38

See equation 17 in the Appendix.

39

For a description of this term, see section above, “Industrial Distribution of Civilian Labor Force.”

40

See equation 18 in the Appendix.

41

The President’s Economic Report, p. 169.

42

Various components of the index of industrial production prepared by the Board of Governors of the Federal Reserve System.

IMF Staff Papers: Volume 2, No. 2
Author: International Monetary Fund. Research Dept.
  • View in gallery

    Nondurable Manufacturing Industries: Employment, 1939-51, and Relationship between Production and Gross National Product

  • View in gallery

    Nondurable Manufacturing Industries: Relationship between Employment and Production, and between Manhours Per Week and Production

  • View in gallery

    Durable Manufacturing Industries: Employment, 1939-51, and Relationship between Production and Gross National Product

  • View in gallery

    Durable Manufacturing Industries: Relationship between Employment and Production, and between Manhours Per Week and Production

  • View in gallery

    Mining Industries: Employment, 1929-51, and Relationship between Production and Gross National Product, and between Employment and Production

  • View in gallery

    Trade and Service Industries: Employment, 1928-51, and Relationship between Employment and Gross National Product

  • View in gallery

    Government: Employment, 1929-51, and Relationship between Employment and Gross National Product

  • View in gallery

    Transportation and Public Utilities: Employment, 1929-51, and Relationship between Employment and Gross National Product

  • View in gallery

    Contract Construction: Employment, 1929-51, and Relationship between Employment and Gross National Product

  • View in gallery

    Finance: Employment, 1929-61, and Relationship between Employment and Gross National Product

  • View in gallery

    Total “Classified” Civilian Nonagricultural Employment, 1929-51, and Relationship between Employment and Gross National Product

  • View in gallery

    “Unclassified” Civilian Nonagricultural Employment, 1929-51, and Relationship between “Unclassified” and “Classified” Civilian Nonagricultural Employment

  • View in gallery

    Agriculture: Employment, 1929-51, and Relationship between Agricultural Production and Gross National Product, and between Agricultural Employment and Production