Should we be worrying about hysteresis or not?The chart above depicts that the decline of the unemployment rate from its 10 percent recessionary high hasn't been good news. The "improvement" is caused by declining labor force participation, not by people actually being hired, on net. This could very much be a dangerous phenomenon, as economists have worried that in an extended recession, those who become unemployed due to cyclical phenomena, after a significantly long duration of unemployment, may enter the ranks of the structurally unemployed. They refer to this as hysteresis.
This "hysteresis" story is quite easy to tell from the data above. Gloomy employment prospects--total private hires are now at 3.87 million a year, well below their pre-recession values in the neighborhood of 5 million a year--are causing the unemployed to give up their job search. Given that the Bureau of Labor Statistics defines the most common measure of "unemployment" as having searched (unsuccessfully) for a job in the last four weeks, these drop-outs vanish from unemployment numbers. Yet the employment-population ratio, so often invoked by Brad deLong, has not budged.
Economists care about labor force participation because the labor force is one of the most basic resources, and labor force participation growth is of key importance in long-run output. Low labor force participation, moreover, inflicts all sorts of social costs.
Yet a recent working paper from the San Francisco Fed suggests that we might be worrying about this just a little too much. They find that, because of flows from nonparticipation to unemployment, prolonged high levels of structural unemployment--"Amerisclerosis," as they call it--is unlikely.