TECHNICAL NOTES

1. ADJUSTMENT OF 1950 DATA1

The Current Population Survey (CPS) for the week of April 2-8, 1950, provides detailed data by age and sex on the employment status of the civilian noninstitutional population. But this source does not provide a breakdown by educational attainment. The latter breakdown is provided by 1950 census data, which relate to March 1950. The census also provides data on employment status by age and sex. There are important discrepancies between some of the census data and the CPS data on the same matter. The two surveys are relatively close concerning total civilian noninstitutional population; for the category 18 years and older, the difference is only 1.5 percent. But the census figures for labor force and for unemployed are substantially lower than the CPS figures. It seems very likely that the relatively inexperienced census enumerators had difficulties with the questions relating to employment status and that this resulted in an undercount of the labor force, especially the unemployed members of it. The CPS enumerators used the same questionnaire as the census, but they were considerably more experienced, and it seems most likely that their classification of respondents as between "in the labor force" and "not in the labor force," and between "employed" and "unemployed" was more accurate than that of the census enumerators. (Cf. C. D. Long, "The Labor Force Under Changing Income and Employment," Princeton University Press, Princeton, 1958, p. 417.)

The data on employment status used in this statement are essentially those of the CPS, but the breakdown by educational attainment is derived from the census data. Briefly stated, the methodology was as follows: From census data, percentage distributions by educational attainment, age and sex were derived for four major categories—the civilian noninstitutional population, the civilian labor force, the unemployed, and the employed. These percentage distributions were then applied to the appropriate CPS data on employment status by age and sex. The result is essentially CPS data broken down according to census education attainment distributions for each age-sex classification. Another adjustment of the 1950 data was necessary because of the new definition of "unemployed" adopted in 1957. Prior to that year, persons with a job but on layoff, and persons hired but not yet working on a new job were excluded from the unemployed; starting in 1957, they were counted as unemployed. Ewan Clague has estimated that the change in definition added about 250,000 to the unemployment figure. For the purposes of this statement, the unemployed total for 1950 was increased by that number to adjust for the change in definition and thus to make the 1950 figure comparable to those for later years. These additional unemployed were prorated according to the percentage distribution (derived from the census) of the main body of the unemployed classified by age, sex, and educational attainment.

Clarence D. Long has also undertaken to adjust the 1950 male unemployment figures for census undercount and for the 1957 change in definition. His results have been published (hearings before Subcommittee on Economic Statistics, Joint Economic Committee, 87th Cong., 1st sess., Dec. 18, 19, and 29, 1961). His methodology has not been published to the best of my knowledge, and the slight differences between his results and mine suggest some differences in methodology. But, on the whole, I believe that his adjusted figures are not significantly different from mine, as the following comparison shows:

No problems of adjustment arise with regard to the years 1957, 1959, and 1963. All of the data for those years are the official statistics or are computed directly from them. One minor exception may be noted. The 1957 data combine four age classifications into two; the four classifications are separate in all other years. For this study, these 1957 combined classifications were broken apart on the basis of percentage distributions estimated from observed relationships of 1950, 1959, and 1963. This particular adjustment was of importance solely in connection with the analysis of trends in labor force participation rates by age. sex, and educational attainment, an analysis undertaken primarily for the

1 I wish to acknowledge the extensive statistical computations and related work done by George Gott, a graduate assistant in the School of Labor and Industrial Relations, Michigan State University, in connection with the adjustment described here.

purpose of judging the consistency of the adjusted 1950 labor force figures (classified by age, sex, and educational attainment) with the 1950 civilian noninstitutional population figures (similarly classified).

TABLE 23.-1950 unemployment rate, males, 18 and older

2. THE COMBINING OF THE "0 TO 4" AND "5 TO 7" YEARS OF SCHOOL CLASSIFICATIONS

In most of the text tables and in figure 8, the "0 to 4" and "5 to 7" years of school classifications are combined into one. This was done partly to simplify the graphic presentation. The main reason, however, was the dwindling size of the two separate classifications, which can be seen in appendix tables A1 and A2. Finally, the "5 to 7" years of school classification appears to behave somewhat erratically-possibly because of its relatively small size-in virtually all of the series studied. The combined (and larger) "0 to 7" years of school classification does not exhibit such erratic tendencies.

3. LIMITING DISCUSSION TO DATA ON MALES

All of the data discussed in the text of this statement relate to males 18 years and older. The employment status of females has long been influenced by some factors in addition to those affecting males. The most important factor appears to be the continuing "emancipation" of women. During the 1950-62 period, there was a large influx of mature women, largely in the middle and upper educational levels, into the labor market. This influx probably had some effect on male unemployment rates at the same educational levels, although men and women obviously do not compete across the board for exactly the same jobs. After studying the data for males only, females only, and the two combined, I concluded that it was essential for analytical purposes to segregate the sexes. believe that one can discern, despite the complexities of female labor market behavior, the effects of the twisting discussed in the text of the statement; but to demonstrate this would require another essay.

I

86-510-64-vol. 1-15

[From the Monthly Labor Review, December 1963]

WORK HISTORY, ATTITUDES, AND INCOME OF THE UNEMPLOYED

(By Robert L. Stein, Division of Employment and Unemployment Analysis, Bureau of Labor Statistics)

In April 1962 a detailed survey of the characteristics of the unemployed was conducted. For the first time, factual evidence was provided on many of the controversial issues that relate to the nature of unemployment in the United States. Although there is still room for wide differences of interpretation, the new facts gathered and presented in this article make it more difficult to cling to extreme positions. The unemployed were found to be a very diverse group of workers with significant variations in financial resources, need for work, attachment to the labor force, and qualifications and prospects for steady employment in the future. In general, they cannot be regarded as personally responsible for their own difficulties, unwilling to accept suitable jobs, more or less voluntarily unemployed, and only casually interested in an occasional job. The basis for these generalizations is a Bureau of Labor Statistics survey of an estimated 9.6 million workers 18 years old or older who experienced at least 5 full weeks of unemployment in 1961 (counting all spells). This nationwide study, conducted to obtain information not readily available from the regular monthly surveys of the labor force, also included a 5-year labor force history of these unemployed workers for the period 1957-61.

Of the 9.6 million unemployed, nearly 70 percent were men-the vast majority in the prime working years of 20 to 64 (table 1). Family heads accounted for three-fifths of the unemployed, and nonwhite workers for one-fifth. A fourth of the group suffered more than 6 months of unemployment in 1961. Only a third of the unemployed, compared with over half the civilian labor force, were high school graduates.2

Labor force attachment

WORK HISTORY

In April 1962, 67 percent of the 9.6 million workers who were unemployed 5 weeks or more in 1961 were back at work, while another 26 percent were looking for work. Only 7 percent had left the labor force. As the tabulation on the following page indicates, the proportion of women who were no longer working or seeking work was larger than that of men-15 percent compared with 3 percent.

1 The survey included persons unemployed 5 weeks or more in 1961 who were 18 years old or older in April 1962, able to work, not in school, and who had some prior work experience. Collection and processing of the data were conducted by the Bureau of the Census under contract with the BLS. About 3,000 personal interviews with sample respondents were completed. In each case, the person who actually experienced the unemployment was interviewed. For a more detailed account of this survey, see Monthly Report on the Labor Force, March, May, and August 1963, pp. xiv-xxi, 16-24, and 15-27, respectively.

See "Educational Attainment of Workers, March 1962," Monthly Labor Review, May 1963, pp. 504-515.

The fact that a high proportion of the unemployed were back at work should be interpreted with caution because this status relates only to a single week. Preliminary data for a subsample of 1,000 persons interviewed in both 1962 and 1963 about their unemploy ment experiences in 1961 and 1962 revealed that some 52 percent had at least a month or more of unemployment in both years.

NOTE.-Because of rounding, sums of individual items may not equal 100.

TABLE 1.-Percent distribution of persons unemployed 5 weeks or more in 1961,1 by selected characteristics, April 1962

1 Includes persons 18 years old or older in April 1962 who were able to work, not in school, and who had some prior work experience.

NOTE. Because of rounding, sums of individual items may not equal 100.

TABLE 2.-Labor force attachment during 1957-61 of persons unemployed 5 weeks or more in 1961, by sex and duration of employment in 1961

1 Includes persons 18 years old or older in April 1962 who were able to work, not in school, and who had some prior work experience.

"All year" is defined as 50 to 52 weeks.

NOTE: Because of rounding, sums of individual items may not equal 100.

Even among men 45 years old and over who were jobles for more than 6 months during 1961, only 6 percent were not in the labor force at the time of the survey.

Of the 650,000 workers who had left the labor force, half reported that they intended to look for work again-most of them within a month of the survey interview date. Thirty-six percent did not plan to rejoin the labor force, mainly because of ill health or household responsibilities, and 13 percent were undecided as to whether they would seek employment. The proportion of those not planning te look for a job because they thought no work was available amounted to only 3 percent of those not in the labor force in April 1962 and was a negligible fraction (two-tenths of 1 percent) of the total with a month or more of unemployment in 1961.*

To determine whether the unemployed were regular or occasional members of the labor force, the survey attempted to obtain a 5-year work history of these workers. The data indicate that this group's attachment to the labor force was relatively strong. Nearly 80 per cent were either employed or seeking employment all year (defined as 50 to 52 weeks) during 1961 (table 2). Over the 5-year period. nearly half the group were in the labor force all year every year.

Persons who reported "no work available" or "could not find work" were so classified. regardless of other reasons reported.

A study of persons claiming benefits under the Temporary Extended Unemployment Compensation Program (TEUC), conducted by the Bureau of Employment Security of the U.S. Department of Labor in 1961 and 1962, found that 75 precent had been in the labor fore continuously for the past 3 years. The difference results from several factors: (1) since the period covered by the BLS study was longer, a lower proportion with continuos labor force activity would be expected. (2) a much larger proportion of the BLS sample of all unemployed were under 25 years of age 23 percent compared with 12 percent of the TEUC recipients-and had therefore entered the labor force only recently, (3) certais unemployed workers not covered by the unemployment insurance system and not eligible for TEUC (agricultural and domestic workers) are much more likely to move into and ou of the labor force, and (4) unemployed workers with insufficient wage credits in covered employment to qualify for benefits would typically be intermittent members of the labor force.