There are a couple of things I want to look at in regards to unemployment growth over the last recession. The first is taking the relationship between unemployment and job growth examined in a previous post and breaking it out by educational attainment.

Examining educational attainment means that we are going to need at least two graphs to keep things from getting completely messy. First lets look at the relationship before the current recession. The data only go back to 1992 but the pattern is clear.


The color coded lines represent the linear relationship between job growth and unemployment since 1992.

  • Blue is less than high school diploma
  • Red is high school but no college
  • Green is some college
  • Purple is college graduates

The lines are stacked in our expected order but the slopes are nearly identical.  That is, as job loss increases the same number of unemployed are added to each category. Its just that there seems to be more structural unemployment among those with lower educational attainment.

The pattern from this recession seems different.


In this case we have divergent relationships. Workers with lower educational attainment start out with higher unemployment and add more unemployed workers as the economy worsens. This would imply not just greater structural unemployment but greater cyclical sensitivity.

Now there are a couple of serious caveats. First and most obviously the 2001 recessions was concentrated in Information Technology. This may mean that the relationship in the first graph is artificially weak.

Second, and more technically there may be an errors in variables problem in the first graph that is downwardly biasing the coefficient. Examining that is probably more than I want to get into but keep in mind that all of this very rough.