For more than two years, the rates of long-term unemployment have been historically high. Even as
the overall unemployment rate inched down at the end of 2011, more than 40 percent of the thirteen million unemployed had been out of work for six months or longer. To put this in perspective, the previous all-time peak in the long-term unemployment rate—in June 1983—was just 26 percent.

Even so, this standard long-term unemployment rate conceals an important part of the story of long-term hardship in the labor market. Using the terminology of the Bureau of Labor Statistics, in 2011, almost one million people were “discouraged” workers, another 1.5 million were “marginally attached” to the labor force, and many millions more had left the labor force altogether. Over eight million workers were stuck in part-time jobs because they could not find full-time work or because their previously full-time hours had been cut. Two million more were among the nation’s prison and jail population, and are not counted, by design, in official labor-market statistics. Many of the people in these circumstances had been there for six months or longer, but none were included in the official tally of the long-term unemployed.1

Whatever its form, long-term, involuntary joblessness takes an enormous toll on those who experience it. In recent congressional testimony, Columbia University economist Till von Wachter summarized research connecting unemployment with: substantial and long-lasting earnings losses,2 enduring employment and earnings instability,3 a higher incidence of poverty,4 higher rates of divorce,5 reduced physical and mental health,6 increased rates of disability,7 declines in life expectancy,8 and adverse impacts on the children of the unemployed (including poor educational outcomes9 and lower adult earnings10). “All of these costs,” von Wachter noted, “are likely to be larger for workers unemployed for a longer period of time.”11

Long-term joblessness also inflicts a substantial cost on the rest of society. In addition to the loss of goods and services caused by leaving productive economic resources idle, unemployment—and especially long-term unemployment—also imposes other direct and indirect costs on the economy, including unemployment insurance payments and a deterioration in unemployed workers’ job skills.

In this essay, we attempt to paint a demographic portrait of long-term hardship in the labor market.12 We display various measures of long-term hardship by race and gender, education, and age using data from the official monthly Current Population Survey for 2011. In addition to the conventional long-term unemployment rate, we also show a broader measure that captures further dimensions of long-term hardship. This additional measure is the “U-6” alternative unemployment rate of the Bureau of Labor Statistics (BLS), which adds “discouraged” workers, the “marginally attached,” and workers who are “part-time for economic reasons” to the official unemployment rate.

“Discouraged workers” are people who are currently not in the labor force, but want a job, are available to work, and have looked for work in the last year but have since stopped looking because they don’t think any work is available. Discouraged workers are a subset of the “marginally attached,” a group that also would like to work, is available to work, and has searched for a job in the last year, but has stopped searching without specifying a lack of jobs as the reason for giving up. Discouraged
workers made up about one-fourth (26.2 percent) of the marginally attached workers in 2007, before the recession, and more than one-third (37.8 percent) of the marginally attached workers in 2011. Those who are “part-time for economic reasons” worked fewer than thirty-five hours per week because their employers cut their hours (due to a lack of demand) or because they were not able to find a full-time job.

We include the U-6 data here because we believe that, under reasonable assumptions, a large share of those fitting this expanded definition of unemployment are experiencing long-term hardship in the labor market. Unfortunately, the BLS does not ask respondents to the official labor-market survey how long they have been “discouraged,” or “marginally attached,” or working “part-time for economic reasons”—or whether, or for how long, they were unemployed before ending up in their current situation. Nevertheless, most discouraged and marginally attached workers were at one time unemployed by the official definition, and many—especially in the current downturn—were likely long-term unemployed before becoming discouraged or marginally attached. In addition, a portion of workers who are part-time for economic reasons will have been in that state for long periods and some may have taken a part-time job, after a spell of unemployment during which they were looking for a full-time job. A reasonable guess in the current labor market would be that half of the 8.3 million people who were part-time for economic reasons in 2011 had been in that state (or unemployed before that) for at least six months.

For each of the demographic groups (organized by race and gender, education, and age), we first  resent the standard and the U-6 unemployment rates. We next display the conventional long-term  unemployment rate, which reports the share of the unemployed in each group that has been  unemployed for six months or longer. Side by side, we show an alternative measure of long-term unemployment, which reveals the share of the total labor force (in each group) that has been  unemployed for six months or longer. As we’ll see later, the two measures can give a very different picture of the incidence of long-term unemployment.

Finally, for each group, we show the demographic composition of the short- and long-term unemployed, the U-6 unemployed, and those who are not in the labor force. By comparing the demographic composition of each of these labor-market categories with the composition of the corresponding, overall population, we can easily see which demographic groups are over- or underrepresented in the various categories of labor market hardship.13

Race and Gender

We start with race and gender. Figure 1 shows the standard unemployment rate and the broader U-6 measure14 for eight racial, ethnic, and gender groups in 2011. Black men (18.1 percent) and black women (14.3 percent) had the highest unemployment rates. Latino women (11.9 percent) and Latino men (11.3 percent) followed. White men (7.8 percent) were next, with a rate that was lower than the national average (9.1 percent). Asian women (7.5 percent) and Asian men (7.3 percent) had unemployment rates that fell between those for white men and white women (6.8 percent), the group with the lowest unemployment rate of these eight race-and-gender categories.

For all of these same groups, U-6 rates were much higher than the standard unemployment rate. More than one-fourth of black men (25.5 percent) were unemployed or underemployed by this broader definition. U-6 rates were almost as high for black women (21.9 percent), Latino women (21.1 percent), and Latino men (19.8 percent). Rates were lower for Asian women (13.6 percent), white  men (12.8 percent), white women (12.4 percent), and Asian men (12.3 percent), but still substantially higher than the corresponding conventional unemployment rates.

For these eight race-and-gender groups, Figure 2 looks at two different measures of long-term unemployment (LTU): the standard measure, which reports the share of the unemployed that has been out of work for more than six months (defined as LTU/U, where, for each group, LTU is the number of long-term unemployed and U is the number of unemployed); and the alternative measure,
mentioned earlier, which reports the share of each group’s total labor force that has been out of work for six months or longer (defined as LTU/(U+E), where (U+E) is the total labor force, comprised of the number unemployed, U, and the number employed, E, within each group). In Figure 2, the solid bars show the standard measure (LTU/U) and the solid line shows our alternative measure (LTU/(U + E)). Using the standard measure, black men (50 percent) and Asian men (49.3 percent) had the highest long-term unemployment rates, followed closely by black women (49 percent) and Asian women (48.6 percent). Rates were somewhat lower, but still high by historical standards, for white men (42.6 percent), white women (40.9 percent), Latino women (40.9 percent), and Latino men (38.5 percent).

The alternative measure of long-term unemployment shows greater variation. Using the conventional measure of long-term unemployment, the race-and-gender groups cluster between roughly 40 and 50 percent. Using the alternative measure, black men were substantially more likely to experience long-term unemployment (9 percent) than all other groups; black women (7 percent) had the next-highest rate. Despite having the lowest share of long-term unemployment under the standard  measure, Latinos’ higher overall unemployment rate meant that Latino women (4.9 percent) and Latino men (4.3 percent) were more likely to experience long-term unemployment than Asian women (3.7 percent), Asian men (3.6 percent), white men (3.3 percent), and white women (2.8 percent).

Figure 3 displays the composition of the short-term unemployed (STU), the long-term unemployed, the U-6 unemployed, and the “not in the labor force category (NILF),” by the same eight race-and-gender categories. For comparison purposes, the figure also shows the distribution of these same categories in the civilian, non-institutional population between the ages of sixteen and sixty-four.
The figure illustrates that black and Latino women and men are all over-represented in the pool of  the short-term unemployed, long-term unemployed, and the U-6 unemployed. White women and men are under-represented in the pool of short-term unemployed, long-term unemployed, and the U-6 unemployed. Asians are under-represented in the shortterm unemployed and U-6 unemployed. Asian women are also under-represented in the long-term unemployed, while Asian men are perfectly represented in this category.

Education

The experience of long-term labor-market hardship varies substantially by education level. Figure 4 presents the standard unemployment rate and the broader U-6 measure for four educational categories: less than a high school degree, a high school degree, some college (including an associate’s degree), and a four-year college degree or more. The unemployment rate decreases substantially with education. Workers with less than a high school degree had the highest unemployment rate (18.4 percent), followed by those with only a high school degree (11.6  percent). Education past high school, but short of a four-year college degree, was also associated with  a reduction in unemployment (to 8.6 percent). In 2011, workers with a four year college degree had an unemployment rate (4.5 percent) about equal to the overall unemployment rate in 2007, before the  recession began. The much higher unemployment rate for less-educated workers is a longstanding feature of the labor market, which holds in better economic times as well. Less-educated workers tend to have fewer of the skills that can give workers bargaining power with  respect to employers. As a result, less-educated job seekers face the highest degree of competition (with other less-educated workers) in the labor market, a predicament that is exacerbated in bad times when more-educated workers may begin to compete with the less-educated for available jobs.

Once again, the broader U-6 measure shows much higher rates of unemployment unemployed for more than six months—shows no obvious pattern by education. All the rates, except for those without  a high school diploma, fall within a few percentage points of the 2011 national average of 43.3 percent. In fact, the people who are least likely to experience long-term unemployment by the conventional measure are those with less than a high school diploma. At first glance, it may seem odd that the group with the highest unemployment rate also has the lowest long-term unemployment rate. But the apparent contradiction is fairly easy to untangle. Workers with less than a high school degree are much more likely to be unemployed, but they are also more likely to give up  looking for work when they lose their job. Less-educated workers typically have fewer resources to finance a job search (they are less likely to qualify for unemployment insurance and typically have lower savings); they may also realistically conclude that there are, indeed, few jobs available for them when even the overall unemployment rate is high.

The alternative measure of long-term unemployment, which we generally prefer to the standard measure, sheds further light on the apparent paradox. The likelihood that any given worker in the labor force is long-term unemployed—which is what the alternative long-term unemployment statistic measures— does vary strongly with education. Only 2 percent of workers with a college degree or more were long-term unemployed, compared to 3.8 percent of workers with only some college education, 5.2 percent of workers with only a high school degree, and 6.9 percent of workers with less than a high school degree. So even though a smaller share of unemployed less educated workers has been out of work for more than six months, a much higher share of all workers who are less-educated has been out of work for that long.

Figure 6 presents the educational composition of the various unemployment measures, alongside the educational composition of the working-age population, for purposes of comparison. Workers with less than a high school diploma are over represented among the short- and long-term unemployed, as well as the U-6 expanded unemployment rate. The over-representation of the less-than-high-school-educated workers in the pool of the unemployed reinforces our conclusion that the alternative measure of long-term unemployment is a more reliable way of representing the problem of long-term unemployment. Workers with a high school degree are also over-represented in short-term, long-term, and U-6 unemployment. Meanwhile, workers with a four-year college degree or more are strongly under-represented in all of the unemployment categories.

Age

All the unemployment measures also vary systematically with age. As Figure 7 demonstrates, younger workers—those ages eighteen to twenty-four— are more than twice as likely (16.3 percent) to be unemployed as older workers are (7.3 percent for thirty-five- to forty-four-year-olds, 7 percent for forty-five- to fifty-four-year olds, 6.6 percent for fifty-five- to sixty-four-year-olds, and 6.5 percent for those over sixty-five).15 A similar two-to-one ratio also roughly holds for the U-6 measure, with 26 percent of eighteen to twenty-four-year-olds experiencing U-6 unemployment, compared to about 12 percent of workers thirty-five and older. Workers ages twenty-five to thirty-four are doing marginally better than the youngest cohort and slightly worse than older cohorts.

Workers in the eighteen to twenty-four age group are also more likely (5.3 percent) than older workers to be long-term unemployed (3.7 percent for forty-five- to fifty-four-year-olds and fifty-five- to sixty-four-year-olds, see Figure 8). The age gap, however, is much smaller for long-term
unemployment than it is for the official unemployment rate or for the U-6 measure in Figure 7. As was the case for workers with less than a high school degree, the youngest age group has the highest unemployment rate, the highest U-6 rate, the highest rate of long-term unemployment by the alternative measure, but also has the lowest long-term unemployment rate using the conventional long-term unemployment measure. As with less-educated workers, younger workers are the most likely to give up looking for work after they become unemployed. As a result, the share of eighteen- to twenty-four-year-olds unemployed for six months or longer is lower than for older age groups. But, because so many eighteen- to twenty-four-year-olds are unemployed, a higher share of  all eighteen- to twenty-four-year-olds in the labor force is actually unemployed than is the case for older workers.

As Figure 9 shows, workers in the eighteen to thirty-four age groups are over-represented in all of the unemployment categories. Older workers, fifty-five and over, are underrepresented in the long-term unemployment, standard unemployment, and U-6 measures.16

Costs and Consequences

More than two years of historically high long-term unemployment rates have already exacted substantial individual and societal costs. The recovery—officially underway since the summer of 2009—has provided almost no relief for those experiencing long-term hardship, even as the overall unemployment rate has started to fall. The available data suggest, however, that even this bleak picture is too rosy. The standard measure of long-term unemployment ignores large groups of workers experiencing long-term hardship, including “discouraged workers,” the “marginally attached,” and many of those who are in part-time work, but want to work full-time. Measures such as the U-6 rate, which incorporate these workers and potential workers, show that “long-term  hardship” in the labor market is even more widespread than the official measure of long-term unemployment suggests.

Moreover—whether we use the standard long-term unemployment rate or an expanded “long-term hardship” measure that’s based on the U-6 rate—the data also show that the burden of long-term joblessness is borne unevenly. Blacks and Latinos, less-educated workers, and younger workers are all much more likely to be unemployed, long-term unemployed, “discouraged,” “marginally attached,”  or involuntarily part-time, with terrible consequences for these groups’ current and future economic, social, and personal health outcomes.
Notes:
1. For a more detailed discussion of
the limitations of the standard measures
of long-term unemployment, see John
Schmitt and Janelle Jones, Down and Out:
Measuring Long-Term Hardship in the
Labor Market (Washington, D.C.: Center
for Economic and Policy Research, 2012),
available at www.cepr.net/documents/
publications/unemployment-2012-01.pdf.
2. Louis Jacobson, Robert LaLonde,
and Daniel Sullivan, “Earnings Losses of
Displaced Workers,” American Economic
Review 83, no. 4 (1993): 685-709; Till von
Wachter, Jae Song, and Joyce Manchester,
“Long-Term Earnings Losses Due to Mass
Layoffs During the 1982 Recession: An
Analysis Using U.S. Administrative Data
from 1974 to 2004” (unpublished paper,
Columbia University, 2009), available at
www.columbia.edu/~vw2112/papers/
mass_layoffs_1982.pdf.
3. Ann Huff Stevens, “Persistent Effects
of Job Displacement: The Importance of
Multiple Job Losses,” Journal of Labor Economics
15, no. 1 (1997):165-188; von
Wachter et al., “Long-Term Earnings
Losses.”
4. “Family Income of Unemployment
Insurance Recipients” (Washington, D.C.:
Congressional Budget Office, 2004), available
at www.cbo.gov/ftpdocs/51xx/
doc5144/03-03-UnemploymentInsurance.
pdf.
5. Kerwin Charles and Melvin Stephens,
“Disability, Job Displacement and
Divorce,” Journal of Labor Economics 22,
no. 2 (2004): 489-522.
6. Sarah A. Burgard, Jennie E. Brand,
and James S. House, “Toward a Better
Estimation of the Effect of Job Loss on
Health,” Journal of Health and Social
Behavior 48, no. 4 (2007): 369–384.
7. Kalman Rupp and David Stapleton,
“Determinants of the Growth in the Social
Security Administration’s Disability Programs:
An Overview,” Social Security Bulletin
58, no. 4 (1995): 43-70.
8. Daniel Sullivan and Till von Wachter,
“Job Displacement and Mortality: An Analysis
Using Administrative Data,” Quarterly
Journal of Economics 124, no. 3 (2009):
1265-1306.
9. Ann Huff Stevens and Jessamyn
Schaller, “Short-Run Effects of Parental Job
Loss on Children’s Academic Achievement,”
Economics of Education Review 30,
no. 2 (2011): 289-99.
10. Philip Oreopoulos, Marianne Page,
and Ann Huff Stevens, “The Intergenerational
Effects of Worker Displacement,”
Journal of Labor Economics 26, no. 3
(2008): 455-483.
11. Till von Wachter, “Testimony Before
the Joint Economic Committee of the U.S.
Congress on ‘Long-Term Unemployment:
Causes, Consequences and Solutions,’”
April 29, 2010, available at www.columbia.
edu/~vw2112/testimony_JEC_
vonWachter_29April2010.pdf.
12. For another recent look at some of
these same issues, see Sylvia Allegretto
and Devon Lynch, “The Composition of
the Unemployed and the Long-Term
Unemployed in Tough Labor Markets,”
Monthly Labor Review 133, no. 10 (October
2010): 3-18.
13. The demographic data we show
don’t include the prison and jail population
or, with the exception of data referring
to age groups, the population that’s
sixty-five and older. By design, the official
Current Population Survey (CPS) data that
we analyze here don’t include those in the
nation’s federal and state prisons (which
house criminals serving sentences of one
year or longer) or local jails (which typically
house criminals serving sentences of
less than one year). In the most recent
data, incarcerated adults made up over 1
percent of the working-age population,
and 2 percent of working-age men. See
One in 100: Behind Bars in America 2008
(Washington, D.C.: Pew Center on the
States, Pew Charitable Trusts, 2008), available
at www.pewcenteronthestates.org/
uploadedFiles/8015PCTS_Prison08_
FINAL_2-1-1_FORWEB.pdf; One in 31:
The Long Reach of American Corrections
(Washington, D.C.: Pew Center on the
States, Pew Charitable Trusts, 2009), available
at www.pewcenteronthestates.org/
uploadedFiles/PSPP_1in31_report_
FINAL_WEB_3-26-09.pdf; and John
Schmitt, Kris Warner, and Sarika Gupta,
The High Budgetary Cost of Incarceration
(Washington, D.C.: Center for Economic
and Policy Research, 2010), available at
www.cepr.net/documents/publications/
incarceration-2010-06.pdf. Even after
release, this group faces enormous challenges
in the labor market (see John
Schmitt and Kris Warner, “Ex-Offenders
and the Labor Market,” WorkingUSA 14,
no. 1 (March 2011): 87-109). In order to
isolate any retirement-related effects, we
also exclude those sixty-five and older,
except when we explicitly look at differences
across age groups.
14. The unemployment rate in Figure 1
is expressed in the conventional way as a
share of the total labor force; the U-6 rate
is expressed as a share of the total labor
force expanded to include the marginally
attached. For ease of presentation, both
measures are presented on the same axis
here (and later), even though the denominators
differ slightly.
15. Those over the age of sixty-five are
only included in this portion of the analysis
and have been excluded elsewhere.
16. We also analyzed differences in
unemployment, long-term unemployment,
and the U-6 rate by four major U.S.
regions and by nativity (U.S. versus foreign-
born). Differences in these measures
were generally small across regions and
nativity. Workers in the West fared slightly
worse than those in the rest of the country.
Immigrants had almost identical
unemployment and long-term unemployment
rates as the U.S.-born, but somewhat
higher U-6 rates (primarily because immigrants
were more likely to be working
part-time for economic reasons).