Job data came in better than expectations, with the US economy only shedding about 250K jobs in July. The unemployment rate did much better than expected, falling for the first time since the recession began.

A few bloggers have taken a stab at trying to explain this. How can unemployment be falling when we are still loosing jobs. This is especially confusing since every econ blogger has be going on and on about how unemployment is a lagging indicator and will keep rising long after the recession is over.

My answer is pretty simple. One, a drop of .1% is not that much and could be noise but two, to the extent this is real it is probably just mean reversion. That is the unemployment rate was unusually high and now it is dropping back.

Here is the unemployment rate measured against job growth.



Unemployment has risen more dramatically that the log term relationship between unemployment and growth.


We can look at the change in unemployment more directly – sort of an Okun’s Law for job growth.


In this case we are not looking at slopes so much as the actual deviation from the the trend line. Prior to this month unemployment had risen 1 percentage point higher than one would expect given the job losses over the prior year.

This month gives us slightly more jobs lost over the 12 months, 4.2% vs. 4.1%, and slightly lower unemployment, 9.4% vs. 9.5%. Thus we are trending back towards the historical relationship.

My theory last month was that fewer people than expected were dropping out of the labor market. In fact, I suspected that some households increased labor supply due to the uncertainty surrounding the financial crisis.  This certainly happened in my household and I recommended it to friends and family.

As the financial crisis has abated this process has unwound in my small sample. There may be something similar going on in the economy at large.