Reports & Studies

The Committee on Economic Security (CES)

"Social Security In America"

APPENDIXES

APPENDIX I

PROCEDURES FOLLOWED IN ESTIMATING UNEMPLOYMENT COMPENSATION COVERAGE IN THE UNITED STATES, 1922-33

The coverage, or the compensable labor force, of a plan for unemployment compensation might be defined as that portion of the persons able and willing to work (employables) who would be affected by the operation of the unemployment compensation program. This group would include the employed for whom contributions are made to the compensation fund as well as the unemployed who draw benefits from it. The size of the group would necessarily be influenced by limitations included in the provisions of the program. In arriving at the estimates presented in the tables which supplement this discussion, only two general limitations have been imposed, since they constitute the only ones that are measurable. The first of these, which will be referred to hereafter as an "occupational" exclusion, provides that all self-employed workers and workers in certain employments are not eligible to participate in the plan; and the second, termed hereafter a "size-of-firm" exclusion, removes from participation in the program employees attached to firms with less than eight employees in at least 20 weeks in the year. The assumed occupational exclusion eliminates all proprietors, owners, independent operators, and other self-employed, and the following employments : agriculture; domestic service in private homes; officers and crews of vessels on the navigable waters of the United States; persons working for a son, daughter, or spouse, and children under 21 in the employ of a father or mother; employees of the United States Government or of its instrumentalities, of a State government or of its political subdivisions or instrumentalities; employees of nonprofit organizations operated exclusively for religious, charitable, scientific, educational, or literary purposes, or for the prevention of cruelty to children or animals.

These limitations could, in any plan, be modified and other limitations added. For example, workers 65 years and over might be excluded; workers on common carriers engaged in interstate commerce might be excluded; or all employees might be included regardless of occupational or size-of-firm factors. Such modifications in exclusions would, however, change the estimates of coverage.

The most abundant source of data for estimating the coverage of an unemployment compensation plan limited by the exclusions stated above is the fifteenth census of the United States, taken in April 1930. The general report on occupations, which includes a tabulation of gainful workers distributed by industry, occupation, and age, was especially valuable since as a rule it made possible the segregation of most occupational exclusions in absolute numbers. This general report was also the source used as a base for Mr. Nathan's estimates of employment and unemployment so that fortunately these estimates could be utilized in the coverage calculations without additional adjustment which would otherwise have been necessary. The year 1930 was the base year for all the coverage calculations as well as for the employment and unemployment estimates.

386

TOTAL COMPENSABLE LABOR FORCE, APRIL 1930

The first step in the procedure for estimating coverage was to eliminate from the gainful workers those made ineligible by the occupational exclusion. For this purpose it was possible to use the census enumeration in all cases except those in which only a portion of the number in an occupational group, or only a part of an industry, was excluded. The nature of these exclusions and sources of data are indicated in table I-1, and the amount of the reductions by industries is shown in table I-2. It may be seen, for example, from table I-1, that 33 percent of manicurists and barbers, 90 percent of oystermen and fishermen, and all those in the independent hand trades are excluded as self-employed workers. It was estimated that 10 percent of the employees in electric-power plants and 5 percent of the employees in gas works were attached to municipally owned concerns. Consequently, these employees were excluded as public servants. The aggregate of all the occupational exclusions amounts to about 20,134,000, or 41 percent of the gainful-worker group. Variation from industry to industry in the proportion of gainful workers excluded may be seen in table I-2.

(IMAGE OF TABLE I-1)

387

The next step in estimating coverage was to eliminate from the groups remaining after the occupational exclusions those employees made ineligible by attachment to small establishments. For the purpose of this exclusion, the only sources of available data are the 1929 United States industrial censuses of mines and quarries, of manufacturing, and of the distributive industries. Failure of even these censuses to show the enact analysis desired made it necessary to use different methods for different industries in order to find the percentage of workers to be excluded. A description of each procedure follows.

(IMAGE OF TABLE I-2)

Mines and Quarries.-The data available in the census of mines and quarries are limited in two ways: The scope of enumeration excludes establishments having less than $5,000 annual sales volume, thus omitting the number of wage earners working in many of the small establishments; and the only tabulation on the size of firm includes wage earners only, thus excluding salaried workers, the majority of whom are employed by the larger establishments. Since these limitations have opposite influences, and because of the minuteness of the percentages of exclusion, it was assumed that they were offsetting, and, therefore, the distribution found in the census was not adjusted in any way. Since no distribution was given by the census for firms with eight or more employees, it

388

was necessary to interpolate on a smoothed curve of size-of-firm distributions. Tables I-3 and I-4 summarize the distributions found in the census of mines and quarries, and indicate that 33 percent of the establishments and about 1 percent of the workers in the industry are excluded from coverage.

Manufacturing.-The data reported on the distribution of employees by size of firm in the census of manufactures is subject to the same limitations as in the census of mines and quarries and was consequently treated in the same manner. Table I-6 summarizes the estimated distribution, which shows 112,863 or 53 percent of the establishments and 362,388 or 4 percent of the employees excluded from coverage.

(IMAGES OF TABLE I-3 AND I-4)

Distribution Industries.-The censuses of wholesale and of retail distribution show no tabulations of the distribution of employees by size of establishment. However, both report the size of establishment by volume of sales. It was therefore necessary to assume a relationship between the volume of sales and the number of employees, and to make estimates on this basis.

Wholesale distribution.-The following distributions were used from the census of wholesale distribution:

(1) Wholesale trade.- A summary of the number of employees and the volume of sales by type of establishment includes all establishments and employees in the industry.

(2) Wholesalers only.- A distribution of 79,840 establishments, employees attached, and volume of sales, by type of establishment.

(3) Wholesale merchants.- A distribution of 76,590 establishments by type of establishment and volume of sales. This group is included in the group "Wholesalers only."

389

"Wholesale merchants" is the only group classified by volume of sales and this, therefore, was used as the basic distribution. This classification included about 96 percent of "wholesalers only" establishments. The distribution, however, did not include an enumeration of the number of employees attached to the establishments. It was necessary, therefore, to estimate the number of employees in the "wholesale merchants" group on the basis of the number of employees enumerated under the classification of "wholesalers only" before distributing them according to the volume of sales per establishment in the "wholesale merchants" group. This was done by reducing the number of employees in proportion to the difference between total sales. The resulting number of employees was distributed in proportion to sales in each sales classification (by type of establishment) and the number thus placed in each classification was then divided by the number of establishments within that classification to find the average number of employees per store. The establishments were then arrayed according to these averages and the results summarized. How-

(IMAGE OF TABLE I-5)

ever, the "wholesale merchants" group included only 45 percent of wholesale-trade employees. This distribution was adjusted to include the employees and establishments in the entire "wholesalers only" group, as shown in table I-6. To make the distribution more inclusive, those wholesale-trade groups with characteristics somewhat similar to those contained in the "wholesale merchants" group were considered to have the same percentage distribution of employees and establishments as the type to which they were compared. This process increased the size of the distribution to include about 75 percent of all establishments. To the remaining 25 percent, i. e., the establishments for which no comparison could be made, arbitrary estimates were assigned. They are chiefly such types as chain-store warehouses, district-sales houses, general-sales houses, etc., and it was assumed that a large proportion of the employees would be covered by an insurance plan. Consequently, in most cases the entire enumeration was included in the compensable part of the distribution. However, for a few types such as bulk-tank stations, brokers, commission merchants, resident buyers, etc., which have no great significance in terms of the relative number of employees represented, smaller percentages of inclusion were applied. Wholesale manufacturers were entirely excluded from the estimates since they are covered in the 1929 census of manufactures. The summary of the results is shown in table I-7, which shows 117,703, or approximately 70 percent of the establishments, and 263,862, or only about 17 percent of the employees excluded.

390

Retail Distribution.-The 1930 census of retail distribution contained the following data that were. usable in obtaining desired estimates: The total number of employees, the number of owners and firm members, total volume of sales, number of stores classified by annual net volume of sales, and number of stores classified by type of operation. Again, assuming that the average sale per worker remained constant, workers were distributed according to volume of sales, own-

(IMAGE OF TABLE I-6 AND I-7)

ers and firm members were then deducted. It was assumed that stores having a sales volume of less than $10,000 were operated solely by owners or firm members. The stores with sales volume from $10,000 to $75,000 were assumed to have one owner or firm member per store actively engaged in sales work. Results of this procedure are shown in table I-8.

Since many chain stores presumably fall below the $75,000 annual sales-volume classification, it was deemed necessary to make an adjustment to include stores belonging to chain organizations. It was estimated that 70 percent of all chain

391

stores appear in the group with less than $75,000 annual sales, and that this group employed 50 percent of the employees in the chain-store type of retail outlet. Therefore, these proportions were shifted for the purposes of coverage into the higher group with more than $75,000 annual sales. The adjusted summary distribution is given in table I-9, which shows 1,466,000, or about 95 percent, of the establishments as excluded as against 1,681,029, or 42.5 percent, of the employees. Graphic interpolation was used to determine more accurately the volume-of-sales size of the stores having eight or more employees.

(IMAGES OF TABLE I-8 AND I-9)

Other Industries.-The foregoing discussion outlines the procedures for the industries for which the size of establishment was available. These industries, however, included only 60 percent of the total compensable labor force. To the remaining 40 percent a percentage exclusion for each industry on the size-of-firm basis was arbitrarily determined and applied. Assistance was obtained from specialists in the various industries in determining each of these percentages whenever possible. It was assumed that the size-of-firm exclusion would not affect employment in gas works, electric power plants, steam railroads, street railroads, telephone and telegraph companies, and water trans-

392

portation. The estimates of exclusions in the other industries are shown in the tabulation below:

THE EMPLOYMENT STATUS OF THE COMPENSABLE LABOR FORCE, APRIL 1930

Having derived an estimate of the eligible compensable labor force in April 1930, it was next necessary to determine the employment status of the compensable labor force in order to discover how many persons would be covered by the system and how many would be beneficiaries of the fund.

In determining the exclusions from total employment in each industry it was necessary to estimate the employment in the industries by occupations since the incidence of unemployment is decidedly less as the skill of the occupation becomes more significant. Since employment statistics by socio-economic groupings{1} were not available for April 1930, the number unemployed had to be estimated in each grouping and deducted from the gainful workers in each corresponding grouping in order to obtain the number employed therein. Obtaining unemployment by industry and occupation was the chief difficulty in such a procedure. However, in volume II of the "General Report on Unemployment," Fifteenth Census of the United States: 1930, an occupational enumeration is made of gainful workers and of the unemployed in classes A and B. This is summarized by socio-economic groups in table I-11.

{1} These are summary occupational groupings and include the managerial, professional, clerical, skilled, and unskilled groups.

393

Since classes A and B combined constitute over 90 percent of the enumerated unemployed, it was assumed that the occupational distribution of the unemployed in classes A and B would hold for the total estimated unemployed. Accordingly, these ratios were applied to "gainful workers" in the various occupational groups in each industry,and the resulting estimates were adjusted so that their summation would equal the estimated number of unemployed in each industry in April 1930. By deducting these adjusted estimates from the number of gainful workers, the number of employed in each industry by occupational groupings was obtained.

(IMAGE OF TABLE I-10)

The next step was to reduce these estimates of employed gainful workers by the occupational and size-of-firm exclusions. It was assumed that the percentage of employed workers in each occupational grouping within each industry who are barred from eligibility to the plan by the occupational exclusion would be the same as was the percentage of gainful workers eliminated by that exclusion from the corresponding occupational grouping within the industry; these percentages were accordingly applied to employed workers.

After the occupational reduction had been made, the same percentage exclusions were applied to the employed in industrial groups as were applied to gainful workers for the size-of-firm exclusion.

The employed coverage, resulting from the two reductions, is shown by type of exclusion in table I-12, and in somewhat different form in table I-13. It

394

may be seen that the percentage of the employed workers within each industry included in the compensation plan differs slightly from the percentage of total gainful workers covered within each corresponding industry (table I-10). These differences are caused by the higher incidence of unemployment in the lower socio-economic groups, especially among the skilled and unskilled groups, where unemployment is much greater than in the managerial group.

The procedure thus far may be summarized as follows: First, the total compensable labor force was obtained by excluding agricultural workers, domestics, family workers, sailors, public servants, employees of nonprofit organizations, and employees working in small establishments; next, the number of employed who would contribute to the operation of the plan was estimated.

The number of unemployed covered in the plan remains to be obtained. This was estimated as the difference between the total compensable labor force and the covered employed, and amounted to approximately 3,177,000, or about 72 percent of the total number unemployed; it is shown by industries in table I-14. All three are compared in table I-15.

It should be remembered that these estimates of employed, unemployed, and total compensable labor force are all influenced by restrictions specifically defining coverage. Were these restrictions changed, coverage would vary accordingly. The procedure outlined herein, however, is illustrative of the methods that may be utilized in estimating the actual coverage finally determined upon.

EMPLOYMENT STATUS OF THE COMPENSABLE LABOR FORCE, 1922-33

The foregoing discussion applies to April 1930 only and merely serves as a starting point for the determination of the compensable labor force and its employment status during the whole period 1922-33. On the basis of indexes of gainful workers and employment it was possible to extend the estimates for April 1930 forward and backward.

Data available for this procedure were much more meager for the years 1923-28 than for the years 1929-34; consequently these two segments of the period were treated separately. However, the procedure in general was the same and included the following steps: First, total coverage was determined on the basis of changes in the number of gainful workers; second, employed coverage was extended by means of employment indexes; and, finally, by the difference between total and employed coverage, unemployed coverage was found.

(IMAGE OF TABLE I-12)
(IMAGE OF TABLE I-13 and I-14)
(IMAGE OF TABLE I-15)

Compensable Labor Force and Its Employment Status, 1922-28.-Since employment data classified by industrial groups were not available for 1922-28, they could not be used in computing the compensable labor force. For want of a better grouping, therefore, calculations were made for the period on the basis of the total nonagricultural classification. An index of industrial coverage was derived from a yearly estimate of gainful workers in industry obtained as follows: First, the total number of gainful workers was estimated by increasing the 1920 census figure by 773,000 each year after 1920 ; then from these yearly figures the estimated gainful workers in agriculture each year were deducted. The use of a constant yearly increase in total gainful workers probably minimizes the amount of labor brought into the market in the early 1920's, but this underestimate tends to make the figures conservative.

The employment status of the compensable labor force in industry was computed as follows: An employment index was constructed from a series derived by deducting nonagricultural unemployment estimates{2} from the total number of gainful workers in nonagricultural pursuits, with April 1930 regarded as 100. These index numbers were applied to nonagricultural employed coverage as of that date and yielded the desired estimates of employed coverage in Industry for the years 1922-28.

The difference between the employed compensable labor force and the total compensable labor force was considered the unemployed coverage. The results of the foregoing calculations are displayed in table I-16.

Compensable Labor Force and Its Employment Status, 1929-33.-For the period 1929-33, data segregated by industrial groups were available so that it was possible to use more refined procedures than were used for the previous period.

The employed compensable labor force was estimated by applying employment indexes to the employed coverage by industries with April 1930 as a base. Since it was not possible to allocate unemployment by industries in the same manner as employment, because of the mobility of labor, unemployed coverage could be estimated only as the summation of the differences between employed coverage by industry and total coverage by industry for any year as compared

{2} Estimates of nonagricultural unemployment, 1922-27, Committee of the President's Conference on Unemployment, Recent Economic Changes (McGraw-Hill Book Co., New York, 1929), vol. II, p. 478; 1928 estimates by the staff of the committee.

398

to April 1930, plus a portion of the total accumulated increment to gainful workers.

The increase in the size of the covered group was determined on the assumption that the normal number of entrants into the gainful-worker group, by definition, necessarily replaces employed workers or is an addition to the employed worker group, and also on the basis that the increase is distributed evenly throughout all industries. With these two factors in mind the increment of 1,049,000 gainful workers between April 1930 and 1933 is no longer composed of the new entrants but of those workers who have been replaced by new entrants and so are unemployed. Assuming no change in the size of the group it was estimated that the proportion of the increment covered bore the same ratio to the total increment as the covered unemployed bore to the

(IMAGE OF TABLE I-16)

total unemployed. Thus, the difference between total coverage in April 1930 and employed coverage in 1933 is 8,495,297, or 66 percent of the 12,976,000 total number of gainful workers unemployed. This percentage was applied to the total accumulated increment in gainful workers since April 1930 and the results of the calculation were added to both total and unemployed coverage.

These methods may have the weakness of not accounting for the changes in the relationship of employment in the small establishments to total employment and the changes in the incidence of unemployment in the socio-economic groups, but since no data are available to correct for these factors, no adjustment could be made.

Table I-16 summarizes coverage for the entire period displaying the total, employed, and unemployed compensable labor force, and the percent of gainful workers covered, as well as the unemployment rate in the compensable group.

399

The Compensable Labor Force by States.-It is impossible to utilize existing data on nonagricultural employment and unemployment by States in the construction of accurate tables for the compensable labor force. Estimates have been made, however, which will serve to indicate the relative numbers who may be covered by State unemployment compensation systems and the degree of unemployment among these covered workers fn 1930, 1931, 1932, and 1933. Table I-17 presents these estimates. It has already been noted in table 23 (p. 108) that the percentage of gainful workers who will be covered by State unemployment compensation systems varies considerably in proportion to the extent of industrialization of the State. It is also apparent from the estimates in table I-17 that the percentage of unemployment within the compensable labor force of the several States will vary considerably. Table 6 (p. 60) shows the States arrayed by order of the percentage of unemployment within the covered group in April 1930 and in 1933, as compared with the average of 1930-33 and the United States average. In 1930, 14 States had unemployment averages higher than that for the United States as a whole (Nevada, Michigan, Rhode Island, Illinois, Ohio, Oregon, New Jersey, Massachusetts, Montana, Vermont, California, Utah, New York, and Indiana). In the estimates of unemployment in the compensable labor force in 1933, 17 States had higher percentages than the average for the country as a whole, and the States are arrayed in an order different from the 1930 rank (Michigan, Pennsylvania, New Mexico, New Jersey, Arkansas, Nevada, New York, Arizona, Florida, Montana, Rhode Island, Illinois, Colorado, Massachusetts, Wyoming, Utah, and Indiana).

(INSERT TABLE I-17)