THOUSANDS

In the latter year the index of quantity output per establishment stood at 185 as compared with 154 in 1929 and 100 in the base year, 1914.

CHART I.-Wage earners and establishments, 1899-1937.

As was pointed out in part I of T. N. E. C. Monograph No. 27, overall averages of this type are somewhat misleading, since such an average is calculated merely by dividing the total number of wage earners by the total number of establishments, etc. In 1937, for example, there were 51.4 wage earners per establishment. Only about one-sixth of all establishments, however, were actually this large, although these relatively few large establishments accounted for nearly five-sixths of all the wage earners employed in manufacturing. In other words, the majority of establishments were small, but the relatively large establishments accounted for the greater portion of the wage earners employed.

In order that the shifts in size of establishments might be measured, establishments were distributed on the basis of the actual number of wage earners employed in each of 4 years (1914, 1919, 1929, and 1937) throughout the period under study. Increases in both the number of very small establishments and the number of very large establishments were recorded in the 1914-19 period; there was very little change in establishment size in the 1919-29 period, while a marked reduction in the number of smaller establishments and a moderate increase in the number of larger establishments occurred after 1929. Detailed statistics for each of 204 industries indicate that these overall fluctuations were not necessarily characteristic of all industries. Approximately one-third of these industries were actually operating on a smaller scale in 1937 than in 1914. On the other hand, there were 31 industries in which employment per establishment more than doubled. It should be noted, however, that not all of these 31 industries

300282-41--4

were really large or that employment per establishment was sufficiently great even in 1937 to influence to any very great extent the overall average cited in preceding paragraphs.

The greater part of the increase in the average size of establishments in all manufacturing was accounted for in very large measure by the growing importance of certain large-scale industries. Total employment in the 50 industries operating the largest establishments increased 37 percent between 1914 and 1937, as compared with a 32.3 percent increase in employment in all manufacturing. Five of these industries-steel works and rolling mills, motor vehicles, motor-vehicle bodies and parts, electrical machinery, and chemicals-accounted for one-third of the entire increase in plant-size during the 23-year period. Not only did the average size of establishment increase in these five instances, but total employment nearly trebled.

Growth in the average size of establishments, however, should not be confused with growth of an industry. Of the 63 industries in which the average size of establishment declined, 22 reported an increase in total employment, while 4 of the 31 industries in which the average size of establishment more than doubled reported an actual decline in total employment.

In summary, there has been a slow but definite increase in the size of establishments due in large part to the increasing importance of certain large-scale industries in our economy. Until the 1929-33 depression, there was no evidence of a change in the relative number of small-scale operators. At that time the decrease in small plants was greater than the decrease in large plants, and since 1933 the come-back in small plants has not been so great as the increase in the number of large establishments.

An analysis was also made of the concentration of operations in large establishments-a concept which refers to the distribution of employment among the various establishments in an industry. The number of establishments employing one-half of the wage earners in an industry was taken as the measure of concentration of operations in big establishments.

In order that changes in the concentration of operations in the various industries might be appraised, two measures of concentration were developed. The first measure, which was called the absolute index, was based on the actual number of establishments required to account for half the wage earners in each industry. The second measure, which was called the proportionate index, was based on the proportion of the total number of establishments which was required to account for half the wage earners in each industry.

The establishments in each industry were arrayed according to the number of wage earners employed by each, and the number of the largest establishments required to account for half the total wage earners was thus determined. This figure was calculated for seven different years throughout the period from 1914 to 1937, inclusive, and for 195 industries for which comparable data were available from the Census of Manufactures. For ease in comparison the figures were converted to an index with 1914 as the base. In this form an increase in the absolute index from one period to another meant that more establishments were necessary to account for half the workers and, thus, that concentration had decreased. A final step was thus necessary in the calcu

lation of the indexes in order that they would reflect directly the changes in concentration. This involved the calculation of the reciprocal of the indexes for each year. The absolute index of concentration used in this study, then, was the reciprocal of the index representing the actual number of establishments employing half the workers. The proportionate index of concentration was calculated in the same

manner.

Concentration of operations in terms of establishments is a measure of the extent to which the business of an industry was done in a small number of its largest plants, or conversely, how evenly it was spread over the various establishments in the industry. Growth of concentration refers to the expansion of some units at the expense of others. Concentration increases when the growth in size of establishments, as measured by the number of wage earners employed, is among the larger establishments and decreases when the growth in size is among the smaller establishments. There are, of course, various combinations of circumstances which account for increases or decreases in concentration.

In measuring the growth or decline of concentration over a period of time, it is important to remember that a good deal depends upon the degree of concentration in the base year. If the degree of concentration was exceedingly high in the base year, the same relative growth or decline in concentration may not be so significant as if the degree of concentration were low at the outset.

When this sort of analysis is applied to the general manufacturing data, the most surprising fact is the high degree of establishment concentration in 1914. A total of 5,950 establishments, or only 3.4 percent of all establishments, employed half the wage earners in manufacturing in that year.

CHART II.-Absolute index and proportionate index of concentration for all industries,

1914-37.

While changes in concentration since 1914 have been comparatively small, concentration has nevertheless exceeded this amount in every year for which there are data. The fluctuations in the absolute and proportionate indexes of concentration are shown graphically in chart II. Both indexes rose sharply in 1919, dropped in 1921, rose somewhat in 1923, and exhibited comparatively little change from 1923 to 1929. Although the fluctuations in the two indexes were not of equal amplitude, the direction of change was approximately the same until 1929. From 1929 to 1933, however, there was a precipitous decline in the proportionate index and a sizable increase in the absolute index of concentration. This movement in opposite directions by the two indexes during the depression was caused by the fact that the decline in the total number of establishments was even greater than the decline in the number of large establishments employing half the wage earners.

From the 1933 level the absolute index declined slightly in 1935 and again in 1937 while the proportionate index advanced abruptly in 1935 and then dropped somewhat with the absolute index in 1937. The point seems to be that the depression reduced the number of enterprises necessary to employ one-half the workers, but did not reduce it as rapidly as the reduction in all establishments. The recovery, on the other hand, increased the number necessary to include one-half the workers, but less rapidly than the increase in smaller establishments.

Nearly all of the 25 industries with an exceptionally low degree of concentration throughout the entire period were relatively small, while most of the industries with an exceptionally high degree of concentration during the period were relatively large. Eight of these industries employed more than 100,000 wage earners-motor-vehicle bodies and parts, ship and boat building, newspaper and periodical printing and publishing, electrical machinery, cigars and cigarettes, bread and other bakery products, rubber tires and tubes and other rubber goods except boots and shoes, and wholesale meat packing. The patterns of fluctuation in the extent of concentration among the individual industries, however, differed widely, and generalizations are difficult.

The growth or decline of concentration of operations in the largest establishments of an industry should not be confused with either the growth or decline in total employment or the growth or decline in size of establishment. Employment in the petroleum-refining industry, for example, increased from 25,400 in 1914 to 83,200 in 1937. The total number of establishments doubled and the average number of wage earners per establishment increased from 144 to 228. The growth, however, was chiefly in the size of the smaller establishment. The number of establishments employing half the workers increased from 9 to 21, and the percentage of establishments employing half the workers increased from 5.1 to 5.8. The absolute index of concentration thus declined from 100 in 1914 to 43 in 1937 and the proportionate index declined from 100 to 88.

There is evidence, however, that concentration of operations in very large establishments declines when a relatively small, highly concentrated industry expands. In general, the 37 industries for which at least a 50-percent decline was recorded in the absolute index from 1914 to 1937 were characterized by the above attributes. The screwmachine products and wood-screws industry, for example, employed 9,071 wage earners in 70 establishments in 1914, and 21,287 wage earn

ers in 311 establishments in 1937. In 1914 three of the 70 establishments employed half the workers and, in 1937, 19 of the 311 establishments employed half the workers. Thus, the absolute index declined from 100 to 16 and the proportionate index declined from 100 to 70.

Most of the 28 industries for which at least a 100-percent increase was recorded in the absolute index from 1914 to 1937 were characterized by a relatively low degree of concentration in 1914 and by a reduction in total employment during the 23-year period. The piano industry, for example, employed 23,861 persons in 242 establishments in 1914 and 5,698 persons in 38 establishments in 1937. In 1914 the 30 largest establishments in the industry employed half the workers and in 1937 four establishments employed half the workers. Thus the absolute index increased from 100 in 1914 to 750 in 1937 and the proportionate index increased from 100 to 109.

Of the 20 industries for which at least a 50-percent increase in the proportionate index was recorded, however, most were characterized by a relatively low degree of concentration at the outset but 16 experienced expansion in total employment. The aircraft and parts industry, for example, employed only 162 persons in 12 establishments in 1914 and 24,003 persons in 92 establishments in 1937. In 1914 half the wage earners were employed in two plants, or 17 percent of all the establishments, and in 1937 half the workers were employed in 5 plants, or 5.4 percent of all establishments. Thus, while the absolute index therefore declined, the proportionate index increased from 100 in 1914 to 315 in 1937.

It should be emphasized that no one characteristic nor even any combination of characteristics was invariably associated with increasing or decreasing concentration of operations as measured by either the absolute or proportionate indexes. Differences arose from industry to industry according to special circumstances affecting production in each industry. For manufacturing as a whole, however, these measures support the conclusions suggested in the analysis of size of establishments-that the tendency toward dominance by a small number of larger establishments is increasing.

CENTRAL-OFFICE CONTROL IN MANUFACTURING

Expansion and growth in size have been common to most phases of our economic life. They are particularly significant in the manufacturing branch of the economy. Here, improvements in machine technology and in the techniques of management and administration have made possible the organization of large masses of men, machinery, and raw materials under central control. Attending these advances of a more technical nature, there have been developments in the field of business organization which have facilitated the concentration of large aggregates of capital under unified control. In many instances, the areas of influence and control have grown so large that they play an important role in economic behavior. An appraisal of the extent of multiplant operations in manufacturing is the subject matter of this section.

The problem of measuring the extent and the nature of integration among manufacturing enterprises and of appraising the structure of manufacturing operations might be approached from any one of its