Federal Reserve Bank of San Francisco

FRBSF Economic Letter

1997-05 | February 21, 1997

More Economic Letters

Job Loss during the 1990s

Rob Valletta

Although the U.S. economy is well into a sixth year of solid expansion and national unemployment remains at low levels, concerns about corporate downsizing, job displacement, and job security continue to affect the national mood. These concerns were highlighted early in the recent presidential election campaign. Indeed, the latest figures suggest that displacement rates have stayed high through 1995, particularly when compared to displacement rates during earlier periods of low unemployment. Moreover, wage growth continues to be moderate, despite a national unemployment rate that is hovering around its lowest level since the late 1980s.

In this Economic Letter, I discuss recent evidence on job displacement and its implications. The data are consistent with the view that high displacement rates in recent years were largely due to corporate downsizing, with much of the consequent burden falling on skilled white-collar workers with substantial job seniority. However, current strong labor market conditions suggest that any wage-moderating effects of these trends are likely to be temporary.

Recent Evidence from the Displaced Workers Survey

The primary source of information on current labor market conditions is the Current Population Survey (CPS). This monthly survey is conducted by the U.S. Bureau of the Census for the U.S. Bureau of Labor Statistics (BLS); the survey collects information on a wide variety of demographic and economic information for a rotating representative sample of 50,000 to 60,000 U.S. households. The CPS is the official source for estimates of the national unemployment rate, and it also provides information on the characteristics of the unemployed. However, its usefulness in accounting for job loss is limited, largely because it does not provide information on unemployed workers’ previous jobs or on the experiences of displaced workers who find new jobs without an intervening spell of unemployment.

To remedy this situation, the BLS began the Displaced Worker Survey (DWS) in 1984. This survey is a biennial supplement to the CPS, conducted early in even-numbered years. It provides information on a nationally representative sample of displaced workers. The DWS supplement begins by identifying respondents who suffered a permanent job loss during the preceding three years due to a plant closure, slack work, abolition of their position or shift, or “other” (which includes completion of a seasonal job, failure of a self-operated business, and “some other reason”). Individuals who fall into the first three categories are asked a series of questions about their pre-displacement job, subsequent unemployment experience, and post-displacement job (if one was found by the survey date).

The most recent survey, conducted in February 1996, updated the displacement data through 1995. For the results reported here, I focus on the rate of displacement during the overlapping three-year intervals that correspond to the DWS response periods, dating back to 1981. The 1993-95 figures reflect the fall revision to the initial August DWS release; the revision accounted for problems in the original survey that led to a significant undercount of the number displaced. Unless otherwise indicated, I exclude the “other” category because it has only limited implications for loss of permanent jobs that might entail some form of job security and because the BLS did not provide tabulations for this category.

The latest DWS documents strikingly high displacement rates for the period 1993-95. Figure 1 shows the sum of displacement rates (displaced workers as a share of total employment) for the three key categories defined above. Displacement rates have remained high since their peak during the 1990-91 recession. In fact, they barely declined despite a decline in the unemployment rate from 7.5 percent in 1992 to 5.6 percent in 1995.

A useful comparison is between the displacement rates during 1985-87 and 1993-95. Although the average unemployment rate was lower during the later period (6.2% versus 6.8%), the displacement rate was higher (7.9% versus 6.3%). Furthermore, as described in Kletzer (1996), inclusion of the “other” category in these calculations raises the relative displacement rate during 1993-95 and reveals a considerable increase in displacement since the 1990-91 recession.

Figure 2 shows displacement rates in the three main categories: plant closed, slack work, and position or shift abolished. After reaching a peak during the recessionary period of 1989-91, the rates of displacement due to plant closure and slack work have declined somewhat. In contrast, displacements due to abolition of a position or shift have increased since the recession of the early 1990s.

This latter finding is particularly noteworthy. Job displacement tends to be countercyclical, as firms shed workers during bad times and maintain or expand their workforces during good times. Displacements due to plant closure or slack work, in particular, are likely to arise from such demand-driven considerations; the figure is consistent with this, although the relationship was somewhat weak during the 1990s recovery. In contrast, displacements due to position or shift abolition are likely to reflect corporate restructuring activity, in which firms reorganize their internal operations by redefining job responsibilities and eliminating specific groups of workers. Firms may be as likely to restructure during good times, when business demands are clear, as during bad times, when changes appear necessary but long-run business needs are obscured. To the extent that this form of displacement reflects the effects of corporate restructuring, it appears that the early 1990s recession coincided with a sharp increase in restructuring which has been maintained (or even strengthened) through 1995. This pattern is different from the pattern during and after the 1980s recession, as displacements arising from this source stayed relatively flat throughout that decade.

The pattern of displacement by broad occupational group, shown in Figure 3, also is consistent with the corporate restructuring interpretation for the sustained high displacement rates in recent years. This figure shows displacement rates in four broad occupational categories which contain either blue-collar workers or skilled white-collar workers (the “services” occupation category is omitted due to ambiguity about its blue-collar versus white- collar content). Workers in the two blue-collar categories at the top face higher displacement rates than do workers in the white-collar categories at the bottom. However, whereas the rates for blue-collar workers have come down since their peak during 1989-91, the rates for the white-collar categories have remained approximately at their recessionary levels.

The media in recent years have made much of a presumed increase in displacement among highly paid white-collar workers. The data presented in Figure 3 largely bear out this view. In addition to its implications for total displacement rates, this trend is notable for its apparent divergence from past practice, when skilled white-collar workers enjoyed substantial job rights and the possibility of service at the same firm until retirement.

Other breakdowns of the data (not shown in figures) also are consistent with the view that high skill and high wage workers have experienced a disproportionate increase in displacement rates. One indicator of job-based or industry-based skill accumulation is worker tenure (seniority). During 1991-95, displacement rates for workers with 15 or more years of tenure at their current firm have remained above their 1989-91 level, while displacement rates for workers with less than 15 years of seniority have come down substantially over the same period. Also, workers with better than a high school education have experienced sustained high rates of displacement since the recession, compared to declining rates for workers with a high school education or less (Kletzer 1996). Finally, the recession and post-recession pattern by industry reveals a substantial decline in displacements for manufacturing workers but relative constancy at high levels for workers in most other industries.


The data presented here suggest that despite the prolonged recovery from the early 1990s recession, job displacement rates have remained high for certain groups of workers. Survey evidence shows that workers are aware of these trends. For example, a recent New York Times article (Lohr 1996) revealed that in surveys conducted in December 1995 and December 1996, approximately 50 percent of respondents indicated that they were “very concerned” or “somewhat concerned” (versus “not at all concerned”) that they or someone in their household might be laid off in the next two or three years.

In an earlier Weekly Letter (Valletta 1996), I argued that the recent furor over declining job security has arisen because narrow groups of workers no longer enjoy the job security they once did. The underlying reasons for job security formed an important part of this argument; the primary reasons are accumulation of firm-specific skills and the use of rising wage profiles as a worker motivation device. These reasons point to highly skilled workers, particularly those with substantial firm-specific skills and seniority, as those most likely to enjoy some form of job security and those for whom declining job security is likely to be most salient. The evidence from the most recent DWS is consistent with this view of declining job security: skilled, white-collar workers with substantial seniority have endured relatively high displacement rates since the recession, with the key source of these displacements (position or shift abolition) being aimed at specific groups of employees rather than a firm’s entire work force (as is the case with the plant closing or slack work categories).

The exact magnitude and nature of any underlying changes in employment relationships remain to be fleshed out. One key question for now is whether these changes have affected workers’ ability to push for higher wages. Some analysts have argued that job insecurity arising from high displacement rates may be moderating workers’ wage demands and helping to keep inflation in check. However, given the restriction of these trends to white-collar workers, and the small percentage of the work force who experience displacement, any wage moderating effects of job insecurity are likely to be small. Moreover, the sustained low unemployment rate indicates that recently displaced workers on average find new jobs quickly, and recent impressionistic evidence reveals slowing job displacement. These generally strong labor market conditions suggest that any moderating effects of job insecurity on wage growth and inflation are likely to be temporary, as workers and firms adjust to the new employment regime and wage-setting behavior catches up.

Rob Valletta


Kletzer, Lori G. 1996. “Job Displacement: What Do We Know, What Should We Know?” Mimeo. University of California, Santa Cruz. Forthcoming, Journal of Economic Perspectives, 1997.

Lohr, Steve. 1996. “Though Upbeat on the Economy, People Still Fear For Their Jobs.” New York Times (December 29).

Valletta, Robert G. 1996. “Has Job Security in the U.S. Declined?” FRBSF Economic Letter 96-07 (February 16).

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