Leaving a job during the Great Recession: evidence from the National Longitudinal Survey of Youth 1979

The Great Recession began in December 2007 and continued through June 2009.¹ Well after the recession had ended, the economy remained sluggish,² making it difficult for those who lost a job during the recession to regain their footing in the labor market. For example, after peaking at 10 percent in October 2009, the national unemployment rate dropped to 9.1 percent in June 2011, and fell further, to 8.2 percent, by June 2012.³ This article looks at job leaving (voluntary and involuntary) during the Great Recession and the year after, and examines subsequent employment differences between those who left a job during this period and those who did not. It uses data from the National Longitudinal Survey of Youth 1979 (NLSY79), a nationally representative sample of baby boomers born from 1957 to 1964.

At the start of the Great Recession, sample members were between the ages of 43 and 50, well into their working lives. The NLSY79 contains a wide range of data that allows analysis of job leaving, including a complete work history for each respondent, various background characteristics, and a cognitive test score.

The following three figures use weekly work-history data to provide a picture of employment, unemployment, and out-of-the-labor-force status for men and women before, during, and after the Great Recession. Figure 1 shows employment percentages by gender from January 2005 through mid-June 2014.⁴ About 87 percent of men and 77 percent of women were working in 2005. Employment for both groups begins to trend downward in 2008, as the Great Recession gets underway. The figure shows that even after the Great Recession ended (June 2009), the percent employed did not return to prerecession levels for either group. For women, the decline continues, and in early 2014, about 70 percent of women are working. For men, employment increases a bit through 2010, levels off through late 2013, then once again starts trending downward; in early 2014, about 80 percent of men are working.

The next two figures show the percentages of men and women unemployed and out of the labor force. Figure 2 shows that unemployment for both men and women hovered around 3 percent during the mid-2000s. Around March 2008, unemployment for both groups begins to rise, peaking shortly after the recession ended. Unemployment for men peaked at over 7.5 percent in early 2010. For women, a peak of over 5.5 percent hit in late 2009. After their respective peaks, unemployment for both groups trended downward over the next several years, with both falling below 4 percent by early 2014.

Figure 3, showing the percentages of men and women out of the labor force from January 2005 through mid-June 2014, tells a different story. The percentages for both genders remain mostly level, around 10 percent for men and 20 percent for women, through late 2008, then start trending upward in early 2009. The percentages mostly level off from 2010 through 2011, about 12.5 percent for men and over 22 percent for women, but then start another upward climb. By early 2014, about 16 percent of men and over 26 percent of women are out of the labor force.

Alan Krueger, as well as other researchers, has documented the declining labor force participation of prime-age workers.⁵ He notes that discouragement after job loss during the Great Recession, health problems, and addiction to prescription painkillers are some of the possible reasons for this decline.

As mentioned earlier, figure 2 shows that unemployment for both men and women experienced an upward trend that began around March 2008 and peaked in late 2009 (women) and early 2010 (men). On the basis of this information, I define the study period for this article as running from mid-March 2008 (a few months after the start of the Great Recession) to June 2010 (a year after the end of the recession). To ensure the study sample includes only individuals who were previously attached to the labor market and could be followed through 2014, I limit the sample to those who participated in the 2014–15 NLSY79 interview and worked at least 30 hours per week in the 6 weeks prior to mid-March 2008.⁶

Table 1 displays descriptive statistics of those in the study sample, broken down by gender and whether the respondent left a job during mid-March 2008 of the Great Recession through a year after the recession ended (study period). (Note that correlations found in these data and presented in this article are not intended to imply a causal relationship.) About 21 percent of men and 24 percent of women left a job, either voluntarily or involuntarily, during this period. Voluntarily leaving a job is not necessarily a negative, as it can allow a person to move to a job that is a better match and offers higher wages. However, involuntary job loss because of plant closings or other reasons has been found to have large negative wage effects.⁷ Throughout this article, sample members who left a job during the study period are referred to as “job leavers.”⁸

Both men and women who left a job had lower Armed Forces Qualification Test (AFQT) scores⁹ and were more likely to be divorced. Moreover, they were much less likely to have attained a bachelor’s degree or more—close to 25 percent (job leavers) versus 33 percent (those who did not leave a job). Among men, both job leavers and those who did not leave a job worked at least 90 percent of the weeks in the 25 years prior to March 2008. Women who left a job worked 78 percent of the weeks in the prior 25 years, compared with 85 percent for women who did not leave a job.

The bottom rows of table 1 show health-related measures from the 2006 interview, prior to the start of the Great Recession. As seen in the table, there was little difference between the reported health limitations and body mass indexes for men who would go on to become job leavers and those who would not leave a job. However, for women, there were some differences. About 11 percent of job-leaving women reported that a health condition limited their work, compared with 5 percent of women who did not leave a job.

Figure 4 depicts the probability of leaving a job during the study period, by gender and education level. What stands out is that men and women with less than a high school diploma were much more likely to leave a job than those with higher educational attainment. This is particularly true of women. For example, women with less than a high school diploma were more likely, by almost 30 percentage points, to leave a job than women with a bachelor’s degree or more. The difference for men is only 14 percentage points.

Table 2 focuses on job leavers by gender and shows characteristics of the first job that the job leaver vacated during the Great Recession and year after. Among these job leavers, 94 percent of men and 84 percent of women left a job in which they were working 35 or more hours per week. Tenure averaged 7 years for men and 6 years for women. About 43 percent of men and 36 percent of women reported being displaced by their employer, involuntarily losing the job through no fault of their own, rather than quitting or being fired.¹⁰ About 13 percent of men and 19 percent of women reported having health limitations that might affect work at the interview in which they reported their earlier job loss. The industry distribution shows that turnover was highest in the construction and manufacturing industries for men and in both the trade and the health and social assistance industries for women.¹¹

Figures 5, 6, and 7 show weekly employment, unemployment, and out-of-the-labor-force percentages for men and women, by job-leaving status, from January 2005 through mid-June 2014. These figures use the same study sample as the tables, a sample that includes restrictions on prior work behavior (work at least 30 hours per week in the 6 weeks prior to mid-March 2008), rather than the full sample used for figures 1, 2, and 3.

Figure 5 shows that, leading up to the study period, employment for job-leaving men hovers slightly below that of men who did not leave a job, but it remains fairly high for both. However, during the study period, employment for men who did not leave a job remains above 95 percent, but for job leavers it falls dramatically, to a low of about 62 percent in early 2010. Employment for job-leaving men trends upwards in subsequent years and hovers around 75 percent in 2013 and 2014, far from the prerecessionary level of over 90 percent. Employment for women who did not leave a job mostly tracks with that of their male counterparts. For job-leaving women, employment drops precipitously around the start of the study period, falls to a low of about 64 percent in early 2010, and remains fairly flat through June 2014. 

Figure 6 shows that men and women who did not leave a job had very low unemployment, creating a flat trend line before, during, and after the recession. Both groups of job leavers had low unemployment before the study period, but unemployment rose quickly during 2008, peaking at about 23 percent for men and 17 percent for women by the end of 2009. Unemployment for both groups slowly declined through 2012 and eventually leveled off. It fell below 7 percent for men by 2013, did the same for women by 2013, and remained around that level for both groups through the first half of 2014.

Figure 7 shows the percentages of men and women out of the labor force from January 2005 through mid-June 2014, separating job leavers and those who do not leave a job. For male job leavers, the percentage out of the labor force rises after the start of the study period (March 2008), to over 15 percent by mid-2010, and remains relatively flat through June 2014. By June 2014, 17 percent of male job leavers are out of the labor force, compared with 6 percent for men who do not leave a job during the study period. For female job leavers, the line rises in a similar fashion from mid-2008 to mid-2010, reaching about 22 percent before it levels off. It then begins another upward climb in 2012, peaking at about 28 percent in June 2014. This compares with a peak out-of-labor-force percentage of about 9 percent in June 2014 for women who do not leave a job.

Table 3 shows the percentage of respondents who report health limiting their ability to work, comparing responses from the 2008 and 2010 interviews with responses from the 2012 and 2014 interviews. These data are important because reported health limitations may help explain the continuous decline in employment for job leavers (as compared with those who do not leave a job during the study period) found in the figures.

At each interview, respondents were asked whether health limited their ability to work. As seen in table 3, job-leaving men were more than twice as likely to have reported a health limitation at the 2008 or 2010 interview than their counterparts who did not leave a job (16 percent versus 6 percent); the same holds true for women (21 percent versus 9 percent). Nearly 34 percent of female job leavers reported a health limitation at the 2012 or 2014 interview, compared with 14 percent of women who did not leave a job. Among job-leaving men, 25 percent reported a health limitation at the 2012 or 2014 interview, as compared with 12 percent of men who did not leave a job.

Table 4 breaks down respondents by the percentage of weeks that they were employed during the 2011, 2012, and 2013 calendar years. These data are important because they show whether the downturn in time spent working (shown in figure 5) is attributable to the same group of respondents over the years or to different respondents who churned in and out of employment. To determine this, I calculate the percentage of weeks employed, by calendar year, and then create various indicators measuring the extent of work both during a particular calendar year and over all 3 calendar years combined.

The table shows that in each of the 3 years, the vast majority of men who did not leave a job during the study period (more than 89 percent) worked at least 90 percent of the weeks, with only 2 to 4 percent working fewer than 20 percent of the weeks. In contrast, fewer than 70 percent of job-leaving men worked over 90 percent of the weeks, and about 20 percent worked fewer than 20 percent of the weeks.

To see if the same men worked most weeks, I look at the data for all 3 years. As shown in the table, 84 percent of men who did not leave a job worked at least 90 percent of the weeks and only 1 percent worked fewer than 20 percent of the weeks. Of job-leaving men, 56 percent worked at least 90 percent of the weeks and 14 percent worked fewer than 20 percent of the weeks. This leaves a substantial group (30 percent) that rotated in and out of employment. Of job-leaving women, almost 42 percent worked over 90 percent of the weeks and 19 percent worked fewer than 20 percent of the weeks. This leaves a large group (39 percent) that rotated in and out of employment.

Table 5 shows characteristics of respondents’ current or most recent job at two interview dates, one before the Great Recession (2006) and one after the recession ended (2014).¹² In the 2006 interview, almost all men had a current or most recent job, regardless of whether they would go on to leave a job during the study period. The wage was similar for both groups, about $30 an hour for those who do not leave a job and $26 an hour for job leavers. The vast majority of men in both groups worked over 35 hours per week. Tenure at the job averaged over 10 years for men who do not leave a job and 6 years for job leavers.

By 2014, the difference between the two groups of men was much greater. Ninety-four percent of men who did not leave a job during the study period either had a job or held one during the prior year, while only 80 percent of their job-leaving counterparts reported the same. Further, those who did not leave a job had a wage of $33 an hour, compared with $26 an hour for job leavers. However, both groups are similarly distributed across industries.

Most women had a job at the 2006 interview or prior year, regardless of whether they were future job leavers. Women who would go on to not leave a job had a wage of $22 per hour, compared with $17 per hour for future job leavers. Over 80 percent of women in both groups were working at least 35 hours per week. As with men, there are wider differences between the two groups at the 2014 interview.

In the 2014 interview, about 92 percent of women who did not leave a job either had a job or held one during the prior year, compared with only 74 percent of job leavers. Job leavers were also less likely than women who did not leave a job to be working 35 or more hours per week (74 percent versus 88 percent). Both groups were largely concentrated in health and social assistance industries.

In summary, among individuals in the NLSY79 who were attached to the labor market (worked at least 30 hours per week in the 6 weeks prior to the study period), over 21 percent of men and nearly 24 percent of women left a job at some point during mid-March 2008 of the Great Recession through June 2010, a year after the recession ended. Men and women who left a job during this period (job leavers) generally had lower educational levels and lower cognitive test scores than those who did not. However, the labor market characteristics for both job leavers and those who did not leave a job in the years leading up to the Great Recession looked very similar. The vast majority of job leavers had worked 35 or more hours per week and had accrued about 6 to 7 years of tenure, on average. At the interview in which they reported the prior job loss, 13 percent of men and 19 percent of women reported having a health condition that could limit their work. Job leavers experienced a large shift out of the labor force in the years after the recession ended, whereas those who did not leave a job during the study period experienced a more gradual shift. A substantial portion of job leavers were churning in and out of employment during the postrecessionary years. Job leavers were also more likely to report a health condition that limited their work in the postrecessionary period than those who did not leave a job.  

ACKNOWLEDGMENTS: The author thanks Jeff Groen, Michael Horrigan, Mark Loewenstein, and Chuck Pierret for comments on an earlier draft.