row limits. Production of wheat flour, which generally accounts for 70 to 80 percent of the total value of products, varied between 110,000,000 and 130,000,000 barrels.

TABLE 2.—Indexes of Production, Employment, Man-Hours, Output, Pay Rolls, and Unit Labor Costs in the Flour and Other Grain-Mill-Products Industry, 1919–40 1

1 Source: Indexes of production, employment, man-hours, and output per wage earner and per man-hour for 1919-37 were prepared by the Work Projects Administration, National Research Project on Reemploy ment Opportunities and Recent Changes in Industrial Techniques, David Weintraub, Director. Data for 1919-36 were published originally in the project's report, Production, Employment, and Productivity in 59 Manufacturing Industries, 1919-36, by Harry Magdoff, Irving H. Siegel, and Milton B. Davis. The 5 indexes have been extended through 1940 by the U. S. Bureau of Labor Statistics, using methods and sources similar to those of the National Research Project. Preliminary data from the 1939 Census of Man. ufactures has been used in preparing the production index for the years 1938-40. The employment and payroll indexes, on the other hand, are the regular Bureau of Labor Statistics indexes, and have not been adjusted to the 1939 Census of Manufactures. These indexes show a slight increase from 1937 to 1939, whereas the Census indicates a slight decrease. This decrease is probably attributable to the fact that in 1939 wage earners engaged in distribution, construction, etc., were for the first time reported to the Census separately from wage earners in manufacturing and excluded from the tabulations of wage earners. Since relatively small adjustments have been required previously to make the employment and pay-roll indexes for this industry conform to Census levels, the indexes probably provide a reasonably accurate picture of trends since 1937 in wage earner employment and pay rolls as defined by the Census prior to 1937. The index of unit labor cost is derived by dividing the index of pay rolls by the production index. Not available.

3 Preliminary.

Since production was not increasing and the workweek (averaging approximately 50 hours) was almost as long at the end as at the beginning of the period, large numbers of flour-mill workers were

The index of production is based on census statistics for six classes of products: Wheat flour; bran and middlings; feed, screenings, etc.; rye flour; corn meal and corn flour; and buckwheat flour. These account for over 99 percent of the value of products of the industry. Total production of each of these products in each census year was weighted by the average unit-value of each product in 1929. For intercensal years, production was estimated by means of an annual wheat-consumption series (bushels of wheat ground). ⚫ Estimate of working hours based on prevailing hours reported by the Census of Manufactures and the results of a field study conducted by the National Research Project and the National Bureau of Economic Research. (See Work Projects Administration, National Research Project: Production, Employment, and Productivity in 59 Manufacturing Industries, part 3, p. 54.)

displaced during this period of increasing output per man-hour. From a peak of about 45,500 in 1919, the number of wage earners employed declined to about 22,800 in 1931.

Information on actual hourly earnings is not available for the years. prior to 1932. Comparison of the movements of the index of pay rolls with the man-hour series suggests, however, that wage earners in flour mills derived some advantage from the rapidly rising productivity. There appears to have been a fairly steady increase in hourly earnings from the end of the war through 1929, followed by a decrease in the first years of the depression. However, since this decrease was less marked than the reduction in living costs, as shown by the cost-of-living index of the Bureau of Labor Statistics, the real hourly earnings continued at least as high.

The manufacturer at the same time benefited from a steady decline in labor costs per unit of output. The index of unit labor cost was over 130 in 1919, as compared with 100 in 1929. By 1931 it was less than 84.

During the period from 1931 to 1937, the general trend toward increasing output per man-hour in the flour-milling industry was reversed. There were a number of shifts in the relations between physical production, working hours, employment, and pay rolls.

The decline in output of flour and other grain mill products, begun in 1929, continued during the period. Production was curtailed sharply in 1932 and again in 1933, then leveled off for 2 years, averaging almost 20 percent lower than in 1929. Although production was somewhat greater in 1935 and 1936, it remained substantially below the 1931 level.

Average hours of work were reduced substantially from about 48 per week in 1932 to 43 in 1933, as a result of the President's Reemployment Agreement, which was made effective in August 1933. The establishment in June 1934 of a NRA code for the industry resulted in a further reduction in average hours to 39 per week. In 1936, however, weekly hours reverted to the 1933 level, following the invalidation of the National Industrial Recovery Act.

Employment, which had been declining steadily since 1919, continued downward between 1931 and 1932, but recovered slightly in the following year, largely as a result of the sharp reduction in hours of work. In 1934, the number of wage earners in the industry rose to a point only slightly below the 1929 level. Thereafter, employment remained about the same from year to year.

During this period of declining productivity, average hourly earnings rose steadily, with the most marked increase occurring between. 1933 and 1934, under the impetus of the NRA code. Weekly earnings declined between 1932 and 1933, since the reduction in hours more than offset the slight rise in hourly earnings. From that year on, however, there was a steady improvement in weekly earnings.

The processor's labor costs per unit of output, on the other hand, rose substantially after 1932. By 1937, when labor productivity was at its low point for the decade of the thirties, they were slightly above the 1929 level.

Productivity has again been increasing since 1937, but since production has been at slightly higher levels and the workweek has been further shortened, employment has remained fairly constant. Year by year, the wage earners in the industry have enjoyed slightly higher hourly earnings. The total pay roll for the industry has not, however, changed appreciably. Labor costs per unit of output, while slightly lower than in 1937, have remained higher than at any other time during the thirties.

TABLE 3.-Hours and Earnings in the Flour and Other Grain-Mill-Products Industry, 1932-40 1

1 Source: U. S. Bureau of Labor Statistics. The results obtained by multiplying average weekly hours by average hourly earnings do not necessarily produce the average weekly earnings shown, since the latter are based on reports from all cooperating establishments, whereas the former are computed from a slightly smaller sample of establishments which report man-hours.

Factors Affecting Productivity

Changes in labor productivity are frequently the result of changes in the mechanical equipment or manufacturing methods used by an industry. However, changes in over-all unit labor requirements may also result from shifts in the proportion of the total output of the industry contributed by establishments at different productivity levels. The latter effect may, of course, occur when there have been no changes whatsoever in manufacturing techniques.

Virtually all basic machinery used in flour milling was fully developed prior to the period covered by the present study. The few improvements in milling equipment which first became available after 1919 were for the most part of minor importance. Nevertheless, a wide range of plant efficiencies is found in practice.

The requirement, under the Fair Labor Standards Act, that hours over 44 per week be paid for at one and one-half times the regular rate after October 1938, and the subsequent reductions in this limit to 42 and 40 hours 1 and 2 years later, have undoubtedly been important factors in the decline in average weekly hours. The act has probably not had a material effect on the wage scale, to date, except perhaps in certain sections of the South, since most mills were paying above the statutory minimum (25 cents per hour beginning in October 1938 and 30 cents since October 1939). According to a job analysis of characteristic mill occupations in Kansas in 1939, for common laborers (both sweepers and truckers) average earnings were 45 cents per hour, with the minimum 30 cents, and the maximum 60 cents. (See National Youth Administration in Kansas, Industry Study No. 1: The Milling Industry in Kansas, Topeka, 1940, pp. 35-42.)

Mills were slow to take advantage of labor-saving techniques, largely because labor costs in the flour-milling industry are small and do not constitute a critical competitive factor. For this reason, modernization of existing plants is infrequent; more advanced processing equipment is installed mainly as new plants are constructed. More important than improved techniques as a factor in the rise in man-hour output has been the shift in production from small to large plants. Even assuming full utilization of the most recent technical developments, labor requirements per barrel of flour are in general lower the greater the capacity of the plant, because flourmilling machinery is so nearly self-operative. In other words, a large mill can utilize its workers more efficiently than a small mill, since the labor force cannot be reduced below a certain minimum. A plant with a 400-barrel daily capacity, for example, probably requires about as many workers, except for packaging, as does a 500-barrel mill equipped with similar machinery. Moreover, the larger mills in the industry are in general of more recent construction than the smaller ones, and hence have taken fuller advantage of the labor-saving possibilities of more recent equipment.

TABLE 4.-Man-Hour Statistics for Selected Establishments in the Flour and Other Grain-Mill-Products Industry, by Size of Plant, 1937 1

Source: U. S. Bureau of the Census and U. S. Bureau of Labor Statistics. Census of Manufactures, 1937: Man-Hour Statistics for Selected Industries. Information relates only to establishments reporting satisfactory data on number of man-hours worked during each month of the year. The Census of Manufactures for 1937 covered 2,238 establishments with 26,390 wage earners. The total number of establishments. covered, by size, is shown in table 5.

* Includes the cost of materials actually used in manufacture, the cost of supplies and containers for prod. ucts, fuel, purchased electric energy, and contract work.

The sum of wages. salaries. and cost of materials.

The direct relationship between plant size and the efficiency of labor utilization is graphically illustrated by man-hour statistics compiled from data collected by the 1937 Census of Manufactures (table 4). These data show that in 1937 the ratio of electric energy used (measured in kilowatt-hours) to man-hours of labor was approximately 5.5 times as great in plants with 101 to 2,500 workers as in plants

• All but 1 of the 42 plants in this size class reported less than 500 wage earners.

employing 1 to 5 wage earners. Per wage-earner man-hour, the cost of materials was 3.2 times as great in the largest as in the smallest establishments, the value of products about 3 times as great, and the value added by manufacture 2.3 times as great. The value of products minus prime costs, which represents the nearest approach to operating revenue obtainable from data collected in the Census of Manufactures, was about 2.8 times as great per man-hour in plants with more than 100 wage earners as in those with less than 6 wage earners.

7

Earnings per man-hour are also directly related to size of plant, but wage earners obtain less advantage than the manufacturer from an increase in plant size. In 1937, average wage earnings per manhour were less than twice as high in the largest as in the smallest plants, ranging down from 65.8 cents to 35.9 cents.

There was a tremendous decrease in the number of small milling establishments between the end of the first World War and the early thirties. The total number of mills in the industry reporting products valued at $5,000 or more declined from 6,485 in 1921 to 1,932 in 1933, but this was mostly due to a drop from 5,417 to 1,251 in the number employing less than 6 wage earners (table 5). At the same time, the very large mills contributed a progressively greater share of the total output of the industry (table 6). The proportion of the total value of products derived from establishments with products valued at $1,000,000 or more was 20 percent greater in 1937 than in 1921, while the proportion from mills producing less than $100,000 was more than 50 percent lower.

TABLE 5.-Establishments in Flour and Other Grain-Mill-Products Industry, by Number of Wage Earners Employed, for Selected Years, 1919–37 1

1 Source: U. S. Census of Manufactures. Since 1921, only establishments with a value of products of $5,000 or more have been included; in 1919, establishments with a value of products of $500 or more were included. There were 1,499 establishments in 1919 with a value of products between $500 and $5,000. The change in coverage thus accounts for a considerable part of the decline in the number of small plants between 1919 and 1921. These 1,499 plants had products valued at only about $4 million, however, while the value of products for the industry as a whole, in 1919, amounted to more than $2,052 million.

The lesser efficiency of labor utilization in the smaller mills has not been the only factor contributing to their decline in importance. The following also have been significant: the tendency toward concentration of ownership and control of flour mills; a similar tendency in

? Prime costs include wages, salaries, and material costs.