Joseph Lawler writes,

The beauty of the structural change theory of jobless recoveries is that it is testable. If hiring is determined by shifts in what the economy produces, then the data will show that permanent layoffs outnumber temporary layoffs — that is, workers aren’t returning to their old companies. Also, there would be noticeable differences in hiring and firing patterns between different industries before and after the recession. The market would determine in which activities firms were overinvested, and which were ripe for further growth.

The data does seem to conform to the theory. In a 2003 study, Federal Reserve economists Erica L. Groshen and Simon Potter found that in past recessions, temporary layoffs spiked. They also found that the two recent jobless recoveries involved clear winning and losing industries, a conclusion they arrived at by looking at payrolls in different industries before and after the recession. The industry that best exemplifies this trend is the dot-com boom of the the late ’90s. During the recession of 2001 it became clear that there was a bubble in web companies, and the industry began rapidly to shed trained workers. The extreme over-investment in the dot-com bubble is an identifiable reason why the job reports were so bleak so long after the end of the downturn. It simply took a long time for the economy to shake out where the former dot-com workers belonged. One industry, notably, that did take off in the wake of the ’01 recession was…housing.