Noam's been recently floating the wacky idea that the coming employment recovery could be here much sooner than the last two jobless recoveries have led most of us to expect. I was skeptical about his argument, but here's some more evidence that maybe we shouldn't expect a worst-case scenario for job growth.

Riccardo DiCecio and Charles Gascon of the St. Louis Fed have put together this chart showing how unemployment and job vacancies have deviated from their long-run averages over the past four cycles.

The solid lines show a cycle's recessionary period and the dotted lines show the ensuing expansionary stage. Here's how to interpret the chart:

In a recovery, typically vacancies increase faster than the number of unemployed workers decreases, generating a counterclockwise pattern. This pattern is evident during the 1983-90 expansion, where the curve moves quickly away from the bottom- right corner of the chart. However, the “jobless recoveries” following the 1990-91 and 2001 recessions are different: During both recoveries, vacancies remained low and the number of unemployed increased for a considerable time.

What was unique about the jobless recoveries, say DiCecio and Gascon, is that the preceding recessions were structural ones. 75% of jobs lost in the 1990-91 recession and 50% of the losses in the '01 recession were suffered by the manufacturing sector. That number is down to 25% during this recession. The assumption here is that it's easier for service workers to find jobs in the growing service economy than for former manufacturing workers to make the shift into the service sector. And that makes sense to me.

P.S. DiCecio and Gascon also make a very interesting observation about how technology has improved job search:

The inward and outward “shifting” of the curve is of particular interest. Among other things, increased labor market efficiency in matching unemployed workers with open positions shifts the curve toward the origin. The most recent points lie on a curve closer to the origin.