Bernake: The Challenge of Long-term Unemployment
From Federal Reserve Chairman Ben Bernake's remarks at the National Association for Business Economics Annual Conference, Washington, D.C.
A full copy of the Chairman's remarks is available on the Federal Reserve website
Discussions of the labor market at this juncture necessarily have a "glass half-empty or half-full" tone. Recent improvements are encouraging, but, as I have noted, in an absolute sense, the job market is still far from normal by many measures, and millions of families continue to suffer the day-to-day hardships associated with not being able to find suitable employment. Although most spells of unemployment are disruptive or costly, the persistently high rate of long-term unemployment we have seen over the past three years or so is especially concerning. In this episode, both the median and average durations of unemployment have reached levels far outside the range of experience since World War II (figure 11). And the share of unemployment that represents spells lasting more than six months has been higher than 40 percent since December 2009 (figure 12). By way of comparison, the share of unemployment that was long term in nature never exceeded 25 percent or so in the severe 1981-82 recession.
Those who have experienced unemployment know the burdens that it creates, and a growing academic literature documents some dimensions of those burdens. For example, research has shown that workers who lose previously stable jobs experience sharp declines in earnings that may last for many years, even after they find new work. Surveys indicate that more than one-half of the households experiencing long unemployment spells since the onset of the recent recession withdrew money from savings and retirement accounts to cover expenses, one-half borrowed money from family and friends, and one-third struggled to meet housing expenses. Unemployment also takes a toll on people's health and may have long-term consequences for the families of the unemployed as well. For example, studies suggest that unemployed people suffer from a higher incidence of stress-related health problems such as depression, stroke, and heart disease, and they may have a lower life expectancy. The children of the unemployed achieve less in school and appear to have reduced long-term earnings prospects.
In addition, unemployment - especially long-term unemployment - imposes important economic costs on everyone, not just the unemployed themselves. Elevated unemployment strains public finances because of both lost tax revenue and the payment of increased unemployment benefits and other income support to affected families. People unemployed for a long time have historically found jobs less easily than those experiencing shorter spells of unemployment, perhaps because their skills erode, they lose relationships within the workforce, or they acquire a stigma that deters firms from hiring them. Loss of skills and lower rates of employment reduce the economy's overall productive capacity over the longer term. In the shorter term, because the process of matching the long-term unemployed to jobs typically takes more time, the currently high level of long-term unemployment might in itself be a reason that further progress in reducing the unemployment rate, and thus in achieving a more complete recovery, could be slow.
A pessimistic view is that a large share of the unemployment we are seeing, particularly the longer-term unemployment, is structural in nature, reflecting factors such as inadequate skills or mismatches between the types of skills that workers have and the skills that employers demand. If this view is correct, then high levels of long-term unemployment could persist for quite a while, even after the economy has more fully recovered. And it appears true that over the past two decades or so, structural factors have been responsible for some increase in long-term unemployment. For example, because an older worker who loses a job typically takes longer to find a new job than does a younger worker in the same situation, the aging of the baby boom generation has probably contributed to a gradual rise in long-term unemployment. Factors such as globalization, technological change, and the loss of lower-skill manufacturing jobs have likely reduced the employability and earnings potential of some groups of workers. To the extent that higher rates of unemployment, especially long-term unemployment, result from structural factors, the scope for countercyclical policies to reduce unemployment would be impaired, and the benefits of a more complete economic recovery for many workers who are unemployed or discouraged would be more limited.
However, although structural shifts are no doubt important in the longer term, my reading of the research is that, at most, a modest portion of the recent sharp increase in long-term unemployment is due to persistent structural factors. Consider, for example, rates of job finding by those unemployed for varying amounts of time (figure 13). Unsurprisingly, the rate at which the long-term unemployed find work is lower than that of those who have been unemployed for only a short time; on average over the period from 1994 to 2007, a bit more than one-third of those already unemployed for one to four weeks found employment within the next month. In contrast, over that same period, only about one-sixth of those already unemployed for more than 27 weeks managed to find a job within a month.
If the recent increase in long-term unemployment were being driven by structural factors rather than, say, the severity of the recession, then the job-finding rates of the long-term unemployed should have fallen sharply relative to those out of work for only a few weeks. But that's not what we're seeing. Rather, as figure 13 shows, the job finding rates of the more recently unemployed and the long-term unemployed all fell over the recession in roughly the same proportion, and they remain low. This pattern is consistent with cyclical factors accounting for the bulk of the recent increase in long-term unemployment. Similarly, the fact that labor demand appears weak in most industries and locations is suggestive of a general shortfall of aggregate demand rather a worsening mismatch of skills and jobs. Counterexamples like the energy boom in the upper Midwest, where there may be some mismatch in the geographic location of suitably skilled workers or an overall shortage of potential workers with relevant skills, might best be interpreted as the exceptions that prove the rule; a mismatch story would suggest that strong labor demand would be appearing in more sectors or geographical areas by now.
An empirical relationship that economists have long used to interpret developments in the labor market is known as the Beveridge curve (figure 14). That curve - named after the British economist William Beveridge - compares unemployment (the number of workers looking for employers) to job vacancies (the number of workers that employers are seeking). In good times, when the unemployment rate is low, businesses are growing and workers are harder to find, so job vacancies tend to be high. Similarly, in bad times, unemployment is high and few jobs are available (vacancies are low). Thus, the Beveridge curve, the relationship between unemployment and vacancies, is downward sloping.
On the usual interpretation, a recession is a period in which the economy is moving down along the Beveridge curve; as output and the demand for labor fall, job vacancies decline and unemployment rises. In contrast, changes in the structural determinants of unemployment are thought to be reflected in shifts of the Beveridge curve to the left or right. For example, suppose that, because of changes in technology or in the mix of industries and jobs, the mismatch between the skills of the unemployed and the needs of employers worsens. Then, for a given number of job openings, the number of the unemployed who are qualified for those jobs is smaller and the unemployment rate is higher than it would have been before the mismatch problem worsened. Graphically, an increase in a skills mismatch would be reflected in a shift of the Beveridge curve up and to the right.
From figure 14, we can see some outward shift in the relationship between job vacancies and unemployment, consistent with some increase in structural unemployment since the onset of the recession. However, a more in-depth analysis of the evidence suggests that the apparent shift in the relationship between vacancies and unemployment is neither unusual for a recession nor likely to be persistent. Research has found that during and immediately after the serious recessions of 1973 to 1975 and 1981 to 1982, the Beveridge curve also shifted outward, but in both cases it shifted back inward during the recovery. This temporary outward shift during a deep recession may be the result of a particularly sharp increase in layoffs, which raises unemployment quickly, even as vacancies adjust more slowly. Another possible explanation for a temporary shift in the Beveridge curve is extended and emergency unemployment insurance, which induces unemployed workers who might otherwise consider leaving the labor force to continue searching for work. Or employers may be more selective in hiring when their need for workers is not pressing and take more time to fill vacancies in an effort to find especially qualified hires. In any case, the data appear consistent with the shift in the vacancy-unemployment relationship in recent years having been relatively modest and likely to reverse, at least in part, as the economy recovers further. When historical experience is taken into account, these patterns do not support the view that structural factors are a major cause of the increase in unemployment during the most recent recession.