What Is Unemployment Telling Us?

It's important to keep in mind that economic data, even though their definitions and methods of collection may not have greatly changed, do not necessarily give the same signals or tell the same story over time.

Take unemployment. Comparable employment losses in different recessions have resulted in disparate increases in unemployment.

Let's look at the experience of the six most recent labor market recessions (which in some cases lasted longer than the officially dated recessions). The recession periods run from the monthly cyclical peak to the monthly trough of employment and unemployment separately, taking into account that sometimes unemployment has lagged employment.

The employment measure used is total civilian employment as a percent of the working-age population, an indicator of labor utilization. The jobless measure is the conventional (official) unemployment rate. (Comparisons based on changes in the level of jobs and unemployment produce similar results.)

In the 1974-75 labor market downturn, the decline in the employment-population rate pushed up the jobless rate by 3.9 points to 9.0 percent in May 1975. The employment loss carried a lot of punch. A one percentage point drop in the employment rate resulted in a sizeable 1.6 point rise in the unemployment rate.

In the subsequent labor market recessions of 1980, 1981-82, and 1990-91, the sensitivity of the unemployment rate to the employment rate was only slightly less, with each one point drop in the employment rate leading to a 1.3 to 1.5 point increase in the jobless rate.

In the extended job market decline of 2001-03, however, the picture changed. The fall in the employment rate, similar to the drops in 1974-75 and 1981-82, resulted in boosting the jobless rate by two points, to six percent. Here a one point decline in the employment rate led to slightly less than a one percentage point rise in the jobless rate. The weakness in employment failed to increase unemployment as much as in earlier recessions.

The same dampened effect occurred again in the latest employment recession. A sharp fall in the employment rate pushed up the jobless rate by five points into double-digit territory. But for every one point decline in the employment rate, the unemployment rate again rose by only one point. Employment losses in the new century have not been lifting the jobless rate by anywhere near as much as previously.

If the sensitivity of unemployment to employment in the latest recession had been equal to the average sensitivity in the four recessions of the 1970s - 1990s, the official unemployment rate today would be about two percentage points higher.

What accounts for the unemployment rate becoming less sensitive to falling employment?

Part of the answer has to do with the shifting behavior of labor supply. When the labor force participation rate is compared to the employment rate during recessions, the result is a downward shift in the sensitivity of participation to a given drop in employment. There has been a greater tendency for jobless workers to withdraw from or not enter the measured labor force in the two latest recessions, compared to prior ones.

Thus the discouragement effect appears to have strengthened. This could be due to a growing structural mismatch between job requirements and workers' available skills or to the disappearance of more of the jobless into the underground economy.

The large-scale exodus from the labor force in the latest recession has overwhelmed the effect that extended unemployment benefits have had in keeping the jobless in the work force, and counted as unemployed. This makes the recent sharp decline in labor force participation all the more startling.

It's hard to escape the conclusion that this labor market recession is not just bigger than others, but a different animal altogether.