Long-Term Unemployment in Recent Recessions
Ilg, Randy E., Monthly Labor Review
During the 1990-91 recession, the long-term jobless rate was much lower than that associated with the 1981-82 contraction; however, unemployment of more than 6 months has been very slow to recover from the recent downturn
The number of unemployed persons and the unemployment rate are among the most visible and politically sensitive economic statistics. But while these aggregate measures certainly are important, policies undertaken to lessen the extent of unemployment, or its impact after the fact, must be based on a detailed knowledge of the makeup of the jobseeking population. For example, joblessness among teenagers can be quite different, in both cause and effect, from joblessness among adult men and women with families. Similarly, policy implications of short-term unemployment, such as that which occurs for seasonal reasons or because of temporary fluctuations in product demand, are much different from those associated with longterm unemployment stemming from chronic deficiencies in demand or from structural problems.
In recent years, the amount of time that persons go jobless has been a critical aspect of the discussion regarding whether, and for how long, extended unemployment insurance benefits might be provided for those whose normal coverage has expired. This brief analysis focuses on the extent of long-term unemployment associated with the 1990-91 recession and its aftermath, and compares it with conditions related to other major recessions of the past two decades. (The minor recession that occurred in 1980 is not addressed separately here.)
Data on the long-term unemployed - those jobless 27 weeks and longer (or more than 6 months) - suggest similarities among the last three major recessions, but also indicate some differences.(1) In each case, the incidence of longterm unemployment continued to increase following the official end of the recession.(2) Levels peaked and began to improve slightly more than a half year after the official end of the 1973-75 and 1981-82 recessions, but were much slower to peak following the 1990-91 recession, as were other major labor market indicators.(3)
As business picks up around the end of a recession, changes in personnel needs initially lead to increases in work hours and then to reductions in unemployment, particularly among persons who had most recently become unemployed. On average, unemployment does not begin to decline until 2 months after the official end of a recession - essentially, a concurrent change. However, there is little immediate effect on the long-term unemployed. Data comparing workers' labor force status in one month to that of the prior month have shown that the longer a worker has been jobless, the less likely he or she is to get a job in a particular month.(4) This is referred to as a "sorting effect," whereby the unemployed persons with the most recent employment and most marketable skills (or perhaps, those with the most flexible job demands or "reservation wages") are more likely to find jobs. Those who have been jobless for long periods, may, as a group, have relatively high reservation wages or be a poor match for available job openings. A sustained period of job growth typically is required for the number of long-term jobless to decline - that is, for the number of persons leaving long-term unemployment to exceed the number of persons entering the category.
How did long-term unemployment in the most recent recession compare to that in the prior two major recessions (and their aftermath)? To answer this question, the key additional questions to ask are - how severe was each recession; for how long did labor market conditions deteriorate; and how quickly did they improve?
Because the labor force normally is growing, the best way to compare long-term joblessness across recessions is to view it as a "proportion" of the labor force - that is, to calculate a long-term unemployment rate representing those jobless 27 weeks and over as a percentage of the labor force. …