The Corner

Structural or Cyclical? A Dispersion Test

Are our labor-market troubles cyclical or structural? This is a burning policy question, and I find it disagreeable for at least two reasons: First, the line between these two is often blurry. Second, there is no question but that both structural and cyclical weaknesses plague the labor market.

But put aside that nuance and allow me to answer the question. Our problems are primarily cyclical.

The most recent evidence supporting this conjecture is a research note by my friend Kris Dawsey at Goldman Sachs.

Mr. Dawsey investigates the degree to which there is “mismatch” in the labor market: Are unemployed workers who used to work in a shrinking industry simply unemployable in different industries? Are firms demanding skills that workers don’t have? Do the workers of a certain skill type disproportionately live a long way away from firms demanding those skills?

If you have a lot of mismatch, and if it has endured for a long time, then you should have a lot of “dispersion” across unemployment rates — some areas or industries or occupations should have relatively high unemployment, while other areas, industries, or occupations should have relatively low unemployment.

Of course, you would expect dispersion at the beginning of a recession. The question is whether the dispersion has remained.

Mr. Dawsey:

Our dispersion measure in any given month is the standard deviation of unemployment rates calculated across the cross-section of observations. Exhibit 3 shows that unemployment dispersion across three different dimensions — industry, occupation, and region — spiked sharply during the recession. This was consistent with the extreme underperformance of the construction sector (and to a lesser extent the auto sector) which was to some extent geographically concentrated. However, most of the increase in dispersion has already reversed, and the general trend appears to be continuing downward.

This lack of dispersion combined with labor-market weakness suggests that we still have a (primarily) cyclical problem. And that we should be more worried about getting people back to work than inflation.

— Michael R. Strain is a resident scholar and economist at the American Enterprise Institute. You can write to him on Twitter at