In the past, there was a fairly close contemporaneous relationship between capacity utilization and unemployment. However, much like the relationship between output and unemployment, a lag in the relationship has developed in the last two recessions (see graph). That is, in past recessions an upturn in capacity utilization was matched by an upturn in employment, there was no delay in the relationship, but in recent recessions there has been about a half year delay before unemployment reacts to changes in capacity utilization (or perhaps even a bit longer).
[Note: To highlight the relationship, the graph shows 100-(UN rate) on the right-hand scale. Thus, an upward movement in the red line represents a decline in unemployment.]
I’ve been relatively pessimistic about the recovery of employment, but I don’t want to just present evidence that supports my views, and there are two things about this graph that are encouraging. First, the “V” in capacity utilization seems steeper than it was in the last two recessions. If the steep recovery of capacity utilization continues and employment follows, the recovery could be a bit faster than I’ve been anticipating (though the recovery of capacity utilization could certainly flatten out, and that possibility has to be factored into any policy response — in the past two recessions the initial change in capacity utilization was also steep for the first few months, but it didn’t last).
Second, the lag between changes in capacity utilization and the change in employment appears to be shorter than the last two recessions. If so, then employment will recover faster. But the word “appears” here is important. Looking at the response of unemployment in the last (2001) recession, there were initial encouraging signs for unemployment just like this time, but then the recovery of unemployment stalled and actually increased a bit more before finally beginning to decline consistently. It’s certainly possible that will happen again.
Thus, while there are some encouraging signs here — the steepness of the recovery for capacity utilization and the apparently shorter lag between improvements in capacity usage and improvements in employment — but neither of these are unqualified, the steepness could change and the shorter lag isn’t yet certain, so I’m hesitant to change my priors about the recovery too much based upon this one piece of evidence. But good news is better than bad news, even if the good news is tentative at best.