2007-33

November 2, 2007

Labor Force Participation and the Prospects for U.S. Growth


Growth in the labor force is one of two key determinants of the nation’s maximum sustainable, or potential, rate of economic expansion. For more than five decades, a growing labor force provided a sizeable boost to the potential rate of expansion in the U.S. economy. Driven by the emergence of the baby boom generation and the entry of women, growth in the labor force added about 1.7 percentage points per year to the average annual growth in potential real GDP from 1948 to 2001 (CBO 2007).

The current period and foreseeable future look quite a bit different. Since 2001, labor force growth has contributed an average of just 1.1 percentage points to potential real GDP growth, with the contribution tending to diminish over time. In fact, as baby boomers age into retirement and key drivers of rising participation rates over the past 50 years (in particular the entry of women) stabilize, the Congressional Budget Office (CBO) expects labor force growth to add only 0.9 percentage point to potential real GDP growth from 2007 to 2012 and 0.5 percentage point from 2013 to 2017. Absent faster gains in productivity—the other key determinant—these developments translate into projected potential growth in real GDP of 2.7% per year compared to the annual average rate of 3.4% between 1950 and 2006.

This Economic Letter reviews the factors contributing to the projected slower pace of labor force growth over the next decade and focuses in particular on the challenges and uncertainties surrounding one aspect, labor force participation behavior.

Determinants of labor force growth

Labor force growth is determined by growth in the native-born population, net immigration, and the labor force participation rate, which is the percentage of the working-age population (16+) working or looking for work. While no forecast is certain, projections of some components of labor force growth are more uncertain than others. Among the easiest to predict over relatively short time horizons (20 years) is the growth in the native-born working-age population; it is defined as the number of new entrants (those turning 16) less the number of exits (deaths and emigrants), and it depends on decisions made long ago (e.g., past birth rates), and on changes in behaviors that pass through to outcomes over long periods (e.g., the effects of vaccinations or smoking on mortality rates). Over the next 20 years, growth in the native-born working-age population is expected to stabilize at a rate of just under 1% per year. A little harder to project is net immigration, which is affected by demand for entry into the country as well as by the willingness of the U.S. to satisfy this demand. While most analysts believe the demand for entry will remain strong, there is considerable debate about how many new immigrants will be allowed to enter. Therefore, most government agencies responsible for forecasts simply assume that current law, which allows 600,000 entrants per year, will prevail. Some agencies, including the CBO and the Social Security Administration, also include an allowance for illegal immigration. The uncertainty around projections about immigration, legal or illegal, is primarily political; the numbers could change dramatically and immediately if U.S. immigration policy became stricter or more lenient. Holding current law constant, however, immigration is as straightforward to forecast as growth in the native-born population.

Labor force participation, in contrast, is less easy to predict. It is defined as the fraction of the working-age population currently employed or actively looking for employment, and it is a behavioral variable that reflects choices that working-age adults make about how to spend their time. In the short run (one to three years), labor force participation fluctuates in response to unemployment and wages. In the long run, it is driven by cultural expectations and roles, incentives and rules associated with entry into and exit from the labor market, returns to education versus experience, tax rates on capital and labor, and the long-run productivity of the economy. Thus, forecasts of labor force participation incorporate judgments based on studying historical changes in behavior as well as differences across individuals and cohorts in the context of institutions, policies, employment rules, rewards for work, and culture that might have affected observed behavior. Not surprisingly, such judgment-based forecasts entail considerable uncertainty. That said, the current consensus forecast among government agencies is that labor force participation will level off at around 66% and stay flat for the foreseeable future.

Trends in labor force participation

Figure 1: Labor force participation rate, 1948-2007
Figure 2: By major segment

For most of the postwar period, the U.S. labor force participation rate rose steadily, pushed up by baby boomers entering prime working age and women entering the workforce (Figure 1A). Then, broadly concurrent with the decline in the economy at the end of 2000, labor force participation rates began to fall. While the downturn in the economy surely had an effect on participation, a look at patterns in key subgroups in the workforce suggests that more permanent changes were also underway. These patterns are displayed in Figure 1B, which plots trends in labor force participation for prime-age (25-54) men, prime-age women, younger (16-24) workers, and older (55+) workers.
Prime-age males. The long-run trend has been a gradual but consistent reduction in labor force participation; it has been linked to early withdrawals from the labor market associated with access to disability benefits and incentives embedded in defined benefit retirement programs. During the 1990s, the trend towards early withdrawal from the labor force slowed and even flattened for some time.

Looking forward, participation for prime-age males could be affected by at least two conflicting influences. First, the decline in the number of workers covered by rule-based defined benefit programs is slowing, lessening incentives for early withdrawal. Second, Autor and Duggan (2006) show that rising wage inequality among males is pushing up the value of disability benefits, increasing incentives for some prime-age males to withdraw from the labor market. Most projections assume that labor force participation of prime-age males will level off a little above its current rate of 91%, but acknowledge some uncertainty in this forecast.

Prime-age females. The most important driver of the rise in aggregate labor force participation over the past 30 years was the dramatic increase in labor force participation of prime-age women, including married women and women with children. In 1948, 31% of prime-age women were in the labor force, and by 1999 this number had risen to 76.8%. Into the 1990s, many expected that women’s participation rates would continue to climb, albeit more slowly, and perhaps converge with participation rates of prime-age males. However, data since 1998 suggest that participation among prime-age women may have plateaued. Moreover, recent research by Blau and Kahn (2007) suggests that responsiveness of married women to their own and their husbands’ wages has declined, meaning that wages get less weight in the labor market decisions of women than before. For projections, a decline in the responsiveness to price means that decisions are being influenced more by less predictable factors, such as changes in social norms and cultural expectations. On balance, forecasts of labor force participation for prime-age females generally show a modest recovery from current levels with a relatively flat trajectory going forward.

Younger males and females. In contrast to the trends for prime-age men and women, the participation rates of workers age 16-24 have been more volatile. For this segment of the labor force, the participation rate rose rapidly from 1962 through 1980, boosted by the entry of young female workers, then fell sharply during the deep recessions of the early 1980s and recovered to its previous peak very late in the 1980s expansion. More recently, however, the pattern has changed, as participation rates fell in the 1990-1991 and 2001 recessions but then failed to recover their previous levels.

One explanation for this changing pattern comes from evidence on school enrollment, indicating that this population is choosing school over work. This seems to be occurring both among those who used to go to school and work but now invest only in school and among those who choose to continue their education beyond high school and attend college. This explanation is consistent with an environment where, increasingly, returns to schooling are rising rapidly and opportunities for those without advanced education are declining. Given this environment, forecasts for this population’s participation rate generally show a flat trajectory from current levels.

Older workers. Until the early 1990s, projections of labor force participation among older workers (men and women, 55+) were nearly all pointing downward. But through that decade, the participation rate for these workers held steady and even began to rise. Some explained it as a transitory response to the strong economy and rising wages of the 1990s expansion. But the continued rise in participation rates among older workers through the 2001 recession and beyond led researchers to look for factors that might suggest the trend would continue to be upward.

Several such factors have been identified—some that may allow them to work more than previous generations, and some that may compel them to do so (Urban Institute, 2006). One factor that might allow older workers to work longer is that they are getting healthier. In addition, the number of physically demanding jobs is falling in the U.S. economy. Indeed, these two developments may explain why fewer older workers cite health as a reason for not working. Institutional constraints to working at older ages also have declined. Many employers have replaced traditional rule-based defined benefit pension plans (which penalized older workers for staying on the job) with defined contribution plans which increase in value with extra years of work. Older workers may also be compelled to work longer. A declining share of employers offer retiree health insurance, discouraging retirement until Medicare begins. In addition, Social Security reforms affecting those born after 1938 are starting to kick in. These reforms gradually raise the normal retirement age from 65 to 67, increase the rate at which monthly payments rise with delayed benefits, and eliminate the reduction in benefits for those working beyond the full retirement age. Combined, these factors may very well push up participation rates among older workers.

Putting it all together

Forecasting future behavior is never easy, but several changes in institutions, culture, and the economy appear poised to make it even more difficult than usual to project the fraction of working-age adults who will want to participate in the labor market. The task will be especially challenging for those projecting the work efforts of older workers. Flexibility and attention to short-run changes may be the only way to improve the accuracy of long-run projections in these changing times.

Mary Daly
Group Vice President

Tali Regev
Economist

References

[URLs accessed October 2007.]

Autor, David H., and Mark G. Duggan. 2006. “The Growth in the Social Security Disability Rolls: A Fiscal Crisis Unfolding.” Journal of Economic Perspectives 20(3) (Summer) pp. 71-96.

Blau, Francine, and Laurence Kahn. 2007. “Changes in the Labor Supply Behavior of Married Women: 1980-2000.” Journal of Labor Economics 25(3), pp. 393-438.

Congressional Budget Office. 2007. The Budget and Economic Outlook: An Update.

Urban Institute. 2006. “Work and Retirement: Facts and Figures.” August.

Opinions expressed in FRBSF Economic Letter do not necessarily reflect the views of the management of the Federal Reserve Bank of San Francisco or of the Board of Governors of the Federal Reserve System. This publication is edited by Sam Zuckerman and Anita Todd. Permission to reprint must be obtained in writing.

Please send editorial comments and requests for reprint permission to

Research Library
Attn: Research publications, MS 1140
Federal Reserve Bank of San Francisco
P.O. Box 7702
San Francisco, CA 94120