Measuring Available and Underutilized Labor Resources

• Who is “unemployed”?
• Accounting for nonparticipants who “want work”
• Alternative measures of available labor
• Conclusions
• References

The U.S. unemployment rate averaged 4.2% in 1999, and dropped to 4.0 % in January 2000, the lowest rate recorded since January 1970. The sustained labor market tightness in this expansion has raised concerns that a shrinking pool of available labor may constrain firms’ ability to expand employment and output further. These concerns have drawn attention to alternative measures of labor availability produced by the U.S. Bureau of Labor Statistics (BLS). In this Economic Letter, I provide a comparative discussion of the development and interpretation of these measures of available and underutilized labor.

Who is “unemployed”?

The official unemployment rate was developed by the federal government to provide a consistent measure of unused available labor resources in the economy. These and related labor force statistics are based on the BLS’s Current Population Survey (CPS), which uses a rotating monthly sample of approximately 50,000-60,000 households that is designed to be representative of the entire U. S. population. Among its broad array of questions, the survey elicits information about the work-related activities and preferences of individuals in these households who are age 16 and older.

The CPS survey classifies individuals as “employed” if they worked for pay or profit during the survey week or if they had a job but did not work that week due to temporary reasons such as vacations or labor disputes. It classifies individuals as “unemployed” if they are not employed but are currently available for work and actively searched for work during the four weeks prior to the survey (individuals on temporary layoff also are included). “Active search” refers to activities that involve contact with potential employers, such as job interviewing, sending out resumes and filling out applications, or contacting employment agencies (for more details on how the BLS measures unemployment, see U.S. BLS 1994).

The labor force consists only of individuals who are employed or unemployed by these definitions. The unemployment rate is the unemployed as a share of the total labor force. Table 1 lists a breakdown of the civilian population aged 16 and over by labor force status, in January 1999 and January 2000. During this twelve-month period, civilian employment expanded by 2 million workers, which is about equal to the increase in the population. The increase in employment was supported by an increase in the labor force of 1.7 million and a decline in unemployment of 300,000. Even with historically high labor force participation rates, about one-third of the population age 16 and over are not actively participating; we will refer to this group as “nonparticipants.”

Accounting for nonparticipants who “want work”

The basic definition of unemployment has been essentially unchanged since the CPS survey began in 1940, as has reliance on the unemployment rate as the primary indicator of underutilized labor resources. Unemployment as officially measured is a relatively objective concept because it is based on observable behavior–work or active search (Bregger and Haugen 1995).

The unemployment rate is not beyond criticism as an indicator of underutilized or available labor resources, however. Since the 1960s, a number of observers have criticized its exclusion of “discouraged workers” (sometimes referred to as the “hidden unemployed”). These are nonparticipants who want work but did not search for a job in the four weeks prior to the survey because they believed that their job search would be in vain (Castillo 1998). Such individuals appear to reflect available yet unused labor resources, because many of them would join the labor force if labor market conditions locally or nationally changed. More generally, the pool of available labor perhaps should be expanded beyond the unemployed through inclusion of nonparticipants who want work now but are not participating for various reasons in addition to discouragement. Like discouraged workers, members of this broader group are likely to enter the labor market if local or national economic conditions change.

Statistics on this broad group have been tabulated on a quarterly basis since 1967; in 1994, however, the underlying questions and definitions were redesigned. Although this change means that the pre-1994 data are not fully comparable to the post-1994 data, this change does have the benefit of enabling the BLS to distinguish between two main groups of nonparticipants who want work–the “marginally attached” and the “not marginally attached” (Castillo 1998). The marginally attached group includes individuals who want work and are available at the time of the survey, have actively searched for a job in the past twelve months, but did not engage in active search during the four weeks prior to the survey. Two subgroups of the marginally attached can be identified. The first is discouraged workers, who are not currently searching because they believe no work is available to them. Among the key reasons for discouragement are workers’ beliefs that they lack adequate skills for available jobs or are likely to face employer discrimination. The second subgroup is other marginally attached workers, who face (presumably) temporary obstacles to working and active search. The primary obstacles are child care and other family responsibilities, transportation problems, and ill health. The remainder of nonparticipants who want work–referred to as “not marginally attached”–are individuals who are not available to work at the time of the survey, who have not actively searched in the past twelve months, or both.

Table 2 lists a breakdown of individuals who were nonparticipants in the labor market but wanted jobs in January 1999 and January 2000. Although the total number represents a small share of those not in the labor force (see Table 1), this group as a whole is about 75% of the size of the unemployment pool in either period; thus, adding them in with the unemployed substantially increases the measure of available labor. The marginally attached constitute only about 25-30% of the total wanting jobs, and discouraged workers are only a small share of the marginally attached. Nonparticipants who are not marginally attached constitute the bulk of all nonparticipants who want work.

Alternative measures of available labor

Several alternative measures of available and unused labor are based on combining the unemployed with some portion of the pool of nonparticipants who want work. Since 1994, the BLS has tabulated rate-based measures of unemployed and underutilized labor, referred to as U-1 through U-6 (see Bregger and Haugen for detailed discussion of these measures and their pre-1994 predecessors). Among these measures, U-3 is the official unemployment rate. U-4 and U-5 are the key measures that incorporate information on nonparticipants who want work. For U-4, the BLS adds discouraged workers to the unemployed, and for U-5 the BLS adds all marginally attached workers to the unemployed; the series are then expressed as an “underutilization” rate relative to the labor force plus the added group. For the time being, these series are not available in seasonally adjusted form. Inclusion of either nonparticipant group raises the estimate of the underutilized pool and the corresponding underutilization rate.

The BLS measure U-5 is closely related to the “pool of available labor,” a measure that was cited in last July’s Humphrey Hawkins report from the Board of Governors (1999) and by Chairman Greenspan in several speeches and in testimony since then. This measure adds together the unemployed and the entire group of nonparticipants who want work, not just the minority identified as “marginally attached.” As of January 2000 the BLS has started publishing a seasonally adjusted estimate of the number of nonparticipants aged 16 and over who want work. When combined with the number of unemployed, the resulting series is essentially identical to the Chairman’s series (which sometimes is calculated only for individuals aged 16 to 64). Analyses using these data will be somewhat limited by series breaks in 1994, although the main constraint is the availability of quarterly rather than monthly data prior to 1994. All of the underutilization and availability rates that combine the unemployed with nonparticipants who want work have declined steadily in the last few years and are at very low levels.

In general, the usefulness of these measures depends largely on the labor force attachment of the sub-groups that constitute the pool of nonparticipants who want work. Castillo (1998) provides relevant evidence. She tracked the labor force status of a set of CPS respondents from all twelve monthly surveys in 1994 who also were in the survey sample exactly a year later. She found that among individuals identified as marginally attached in 1994, 47.5% were labor force participants in 1995 and 30.8% were employed. Within the marginally attached group, discouraged workers were slightly less likely than other marginally attached workers to be in the labor force or employed a year later. Among those not marginally attached but wanting work, 38.9% were participating a year later and 30.8% were employed. On the other hand, Castillo also found that the unemployed showed a greater degree of labor force attachment than either of these nonparticipant groups. Among those unemployed in 1994, 72.5% were in the labor force a year later and 53.1% were employed.

To sum up, Castillo’s figures suggest, as expected, a hierarchy in terms labor force attachment: the unemployed show the greatest degree of labor force attachment, marginally attached nonparticipants show a lesser degree (with little difference between discouraged workers and others), and nonparticipants who want work but are not marginally attached show the lowest degree. These behavioral differences between the marginally attached and not marginally attached suggest that the BLS’s U-5 series and the pool of available labor may differ somewhat in their pattern over the business cycle. However, because the marginally attached and not marginally attached behave in a fashion more similar to each other than to the unemployed, it is likely that U-5 and the pool of available labor will exhibit very similar patterns under most labor market conditions.

Conclusions

In times of very tight or very loose labor markets, information on the group of nonparticipants who want work but face obstacles to active participation at the time of the survey may provide a useful supplement to information about labor market conditions contained in the standard unemployment rate. The BLS series U-5, which adds marginally attached workers to the unemployed, and the broader series called the pool of available labor both take a step in that direction. The BLS measure U-5 is slightly more appealing on a behavioral basis. However, substantive differences between the two series are likely to be quite limited, and for analysts interested in business cycle patterns, the broader measure now possesses the key advantages of being available prior to 1994 (in quarterly form) and in seasonally adjusted form through the BLS.

Rob Valletta
Senior Economist

References

Board of Governors of the Federal Reserve System. 1999. “Monetary Policy Report to the Congress Pursuant to the Full Employment and Balanced Growth Act of 1978.” July 22. <http://www.federalreserve.gov/boarddocs/hh/1999/July/FullReport.htm>

Bregger, John E., and Steven E. Haugen. 1995. “BLS introduces new range of alternative unemployment measures.” Monthly Labor Review (October) pp. 19-26.

Castillo, Monica. 1998. “Persons outside the labor force who want a job.” Monthly Labor Review (July) pp. 34-42.

U.S. Department of Labor, Bureau of Labor Statistics. 1994. “How the Government Measures Unemployment.” Report 864 (February). <http://www.bls.gov/cps/cps_add.htm> (accessed February 7, 2000)