Everyone knows the story about the two old men in a retirement home. The first one complains that, "the food here is poison." His friend agrees with him then adds, "and the portions are so small."

We have been getting this story the last several years in assessments of the labor market. The economy remains far below full employment by every measure. The employment to population ratio is still more than 4 percentage points below its pre-recession level. We are almost 9 million jobs below trend levels. In addition, many have pointed out that a disproportionate number of the jobs that have been created have been low-paying jobs in sectors like hotels and restaurants.

Some analysts have picked up on the latter point to argue there has been a fundamental change in the economy. They claim we are moving to an economy that produces high-paying jobs for highly skilled people (e.g. doctors, lawyers, financial engineers) and bad jobs for everyone else. I have beaten up on this one elsewhere. (The trick for doctors and lawyers is protectionism not skills.)

However there is no dispute that bad jobs seem to be growing rapidly as a share of employment at present. The question is why. An alternative explanation to the "it just happens" view is that the weak economy itself is responsible for the proliferation of bad jobs. In other words, because the economy is not generating decent jobs in any reasonable number, workers are forced to take bad jobs. In that story the proliferation of bad jobs is the direct result of a weak economy.

I did a very crude test of this story. I regressed the rise in the share of hotel and restuarant employment from 2007 to the first half of 2013, across states, against their unemployment rate in July of 2013. Here's the picture.


While far from conclusive, this looks like pretty good support for the bad labor markets lead to bad jobs story. (For regression fans, the coefficient of the unemployment rate variable was 0.0013 with a t-statistic of 4.65, which is significant at the 1 percent level.) There are of course other reasons than the lack of good jobs that could cause the share of restaurant employment to grow more in states with high unemployment.

For example, if we believe that the rate of growth of restaurant employment is more or less fixed, then states with weak overall job growth would see a rising share of restaurant work. But the substatantial variation in the growth rate of restuarant employment across states would make this story less credible.

Anyhow, I wouldn't claim this simple test seals the case, but it strongly suggest that the story of bad jobs is the story of weak labor markets. In this story the food is poison because the portions are small.