The Quality of Jobs
The Challenges Created by A Dynamic Labor Market
The news is encouraging: employment has grown disproportionately in the industry/occupation job categories paying above-median wages. Even in the traditionally lower-paying service industry, a majority of the net employment growth has been managerial and professional specialty positions, which typically pay above-median wages. Contrary to conventional wisdom, the new jobs are not disproportionately part-time, low-skill positions.
The second purpose of the study is to examine job displacement. Although the economy is generating millions of net new jobs, it is clear that the speed of transformation in the U.S. labor market has left many American workers anxious about their economic futures. A dynamic and growing labor market can impose costs as well as offer opportunities, and policies to help workers deal with job transitions are critical to reducing these adjustment costs.
The unemployment rate has fallen from over 7 percent in January 1993 to 5.6 percent in March 1996, and has been below 6 percent for 19 consecutive months. Given current demographic trends, the Bureau of Labor Statistics projects the labor force to continue growing by approximately 1.1 percent annually between 1994 and 2005. Therefore, to keep unemployment low, the economy needs to average a net increase of about 120,000 new jobs per month. Employment is now expanding at a pace consistent with steady, sustainable growth and low unemployment.
Higher-paid Jobs in the Service Sector. The "service sector" is quite diverse. It includes many low-wage positions, but also many high-wage positions in financial services, hospitals, and computer and accounting services. For this reason, it is important to determine whether employment growth within services has occurred primarily in the high-skill managerial and professional specialty occupations or in low-paying occupations. The Current Population Survey provides evidence on employment growth by occupation.
Growth of Higher-paid Jobs by Industry and Occupation. An even more detailed picture of the nature of the new jobs created emerges from an examination of industry/occupation categories. Using data from the February 1994 and February 1996 Current Population Surveys, we sorted full-time workers into 45 detailed occupations in 22 major industries. A quarter of the sample reported earnings in addition to the industries and occupations in which they worked. Although many of the possible 990 industry/occupation cells were small, only 6 percent of the population-weighted sample was found in cells with 10 or fewer sample members reporting earnings data for both surveys. In order to avoid the high sampling variability associated with insufficient numbers of observations, we eliminated these small cells from our analysis. There were 287 job categories in each year after eliminating the cells with 10 or fewer sample members.2
The first step in our analysis was to rank the 287 occupation/industry cells by the median weekly earnings of full-time workers. Approximately half of all full-time employment in February 1994 was found in cells with median weekly earnings above $480 (in February 1996 dollars). The employment growth in these "high-wage" job categories can then be compared to overall employment growth. Our key measure of job quality is the percentage of total employment growth that occurred within the occupation and industry categories that paid above-median wages in February 1994. The results were striking.
Another way to summarize the results from our industry/occupation analysis is shown in Figure 3. Here we ranked the 287 industry/occupation categories by their median weekly earnings for full-time workers, and sorted them into 10 ordered groupings -- each with 10 percent of employment in February 1994 -- by their earnings ranking. If all 10 groups had grown proportionately to their share of employment in February 1994, each would have accounted for 10 percent of the net new employment. But rather than accounting for their proportional share of total employment growth (30 percent), the top three deciles accounted for much more (over 50 percent).
In sum, the data indicate the following about the nature of recent job growth:
The New Jobs are Mostly Full-Time. Data from both the Current Population Survey and the BLS establishment survey indicate that most of the net new jobs are full-time. The Current Population Survey includes data on part-time employment. Figure 4 portrays the proportion of employed persons reporting that they worked part-time for "economic" as well as "non-economic" reasons. Despite a shift in both series corresponding to a redesign of the survey in January 1994, the proportion of employed persons reporting to be employed part-time has actually declined slightly. The declines have been even larger for those working part-time for "economic" reasons, often referred to as the "involuntarily underemployed."
The establishment data indirectly support the conclusions from the household survey. If the net new jobs were disproportionately part-time, we would expect average hours worked per job to fall.5 But the employment data show that average hours worked for all jobs (including the new jobs) remained roughly constant: the number of nonfarm payroll positions and the total number of hours worked both grew at about the same rate over the past three years (see table). This suggests that the new jobs have not been disproportionately part-time.
|Hours worked (annual basis, billions)||202.2||217.8||7.7|
Little Change in Multiple Job Holding. Some Americans decide to hold more than one job, in order to save for a house or to meet unexpected expenses. Nonetheless, multiple job holding would raise concerns about the quality of jobs if an increasing number of Americans have to work two or three jobs to make ends meet. A frustrated worker is said to have reacted to the news that 8.5 million new jobs have been created by replying, "Yeah, and I have three of them." But the data simply do not indicate any significant movement in multiple job holding. The percentage of employed persons working multiple jobs has remained in the neighborhood of 6 percent since the late 1980s.
Impact on Wages and Income Inequality. Between the 1970s and the early 1990s, average real wage growth slowed and income inequality widened. In recent years, however, there are some encouraging signs that the tide may be turning on these labor market challenges. In 1994 -- the most recent year for which data are available -- real median family income rose and the poverty rate fell for the first time in 5 years. Improving job quality can enhance these recent gains, although the effects may only become manifest after an extended period of time. As discussed above, most of the recent net increase in employment has occurred in occupations and industries that typically pay above-median wages. But the additions to the workforce have had only a marginal effect on aggregate wage data, since net employment growth represents only a relatively small percentage of total employment for the U.S. workforce. Nevertheless, the news about the quality of net job growth is encouraging, and bodes well for the future. Although there is still much left to be done, recent trends show that the labor market is on the right track.
While the anxiety felt by many workers is real and important, it is also important to take an objective look at the evidence. Not all sources demonstrate increased economic anxiety. For example, the Michigan and Conference Board surveys of consumer sentiment recently have been above their historical averages. Respondents to those surveys apparently do not view employment prospects as poor. Nevertheless, considerable evidence suggest that many Americans are concerned, some very concerned, about job displacement. In order to know how best to respond to these concerns, we need a more precise assessment of the nature of the displacement problem. Has job displacement in fact increased? Is it affecting different categories of individuals today than it did ten years ago? This section of the report examines these questions.
Evidence from the Displaced Worker Survey. The BLS conducts a survey of displaced workers every two years, with the most recent published data from February 1994. The table below summarizes the displacement rates (defined as the number of workers displaced per 100 employed) for the 1981-82 and 1991-92 periods.
The overall number of workers displaced was roughly the same proportion of the workforce in 1991-2 as in 1981-2, although the recession in the early 1980s was more severe than the one in the early 1990s. However, it is difficult to determine precisely how to account for the business cycle in assessing displacement rates. A comparison of aggregate displacement rates also conceals a fundamental change in the incidence of job displacement. The table shows that older, white-collar workers were considerably more at risk of displacement in 1991-92 than during the previous recession. And further analysis shows that job displacement rates rose for more educated workers. These changes in the incidence of job displacement may be a reason for the reports of heightened anxiety regarding job loss. Although blue-collar and less educated workers remain more likely to be displaced than others, displacement rates have clearly risen among those workers who had previously been largely immune from the threat of job dislocation.
25-34 years of age
|35-44 years of age||3.8||3.9|
|45-54 years of age||3.0||3.8|
* Expressed as a percent of workers with three or more years of tenure on their current job.
Based on data from the Bureau of Labor Statistics.
Indicators of Recent Job Displacement. As noted above, the Displaced Worker survey is conducted only once every two years, and the most recent published data are from the 1994 survey, which covers the 1991-93 period. Unfortunately, the official displacement data for the period after 1993 are not yet available.6 (The results of the 1996 Displaced Worker Survey, conducted in February, should be available later this summer.) Until the official displacement data are available, other measures can be used to get an indication of how the labor market has been changing since 1993.
One indicator comes from unemployment data on job losers. Figure 5 shows the job loss rate, defined as the ratio of recently unemployed job losers -- those who are unemployed due to job loss (as opposed to job leavers or labor market entrants), unemployed less than 5 weeks, and not on temporary layoff -- to total employment in the Current Population Survey. This job loss rate roughly approximates the net "flow" into unemployment due to job loss, since it considers only those who have los their jobs recently. As shown in Figure 5, the job loss rate has continued to fall since 1992.
The Costs of Job Displacement. The Displaced Worker Survey provides information on the impact of job loss