Thursday, September 10, 2009

Judging Downturns II

In response to my previous post, an economist at a financial firm emails me the following analysis:

I am following the discussion on measuring the severity of recessions.

I prefer the unemployment rate for historical analysis. In olden days, it was likely better measured than real GDP. And before 1947, there is no quarterly real GDP. We have plenty of data on industrial production, but as the US becomes less of a manufacturing economy, its relevance is fading.

The point a reader raised about changing composition of the labor force is good. We can adjust for that. The CBO creates a quarterly estimate of NAIRU going back to 1947. The changing composition of the labor force is reflected in the estimate of NAIRU.

The first chart below compares the unemployment rate (quarterly average) with the CBO's NAIRU estimate.

The second chart shows the difference between the two series:

As of Q2:2009, the gap is just a touch lower than in Q4:1982. The gap is likely to expand for two or three more quarters, making it the worse gap of the post World War Two Period.

Another way to gauge the slack is to focus on a single demographc. Let's take married men, spouse present. They are the most stable segment of the labor forcce, and here is their unemployment rate:

It has been terrible this time, but not as bad as the early-1980s. Why this time seems worse is the unemployment rate for teenagers is a record high. I think we should give some blame to 3 consecutive annual hikes in the minimum wage.

A final way to compare the rise in the unemployment rate from bottom to top; the dates of the cycle in the unemployment rate need not coincide with NBER peaks and troughs. Here are my calculations for real GDP and the unemployment rate.

The rise in the unemployment rate in 1945 was much less than in the recent recession. But the rise in the unemployment rate in 1937-38 was much worse than the current recession.

If we combine the 1980 and 1981-82 recessions into one super-recession, the rise in the unemployment rate was 5.0 percentage points, slightly bigger than what we have seen so far.

Finally, critics like to make remarks that the US has artificially reduced its unemployment rate by putting people in jail, claiming that those people would otherwise be unemployed. First, there is no evidence they would be otherwise unemployed. But more important, secondly, the total institutionalization has not changed. Mental hospitals have been emptied out. Sad to say, many of these mentally ill people are in jail. The point is that the sum of all institutionalization is probably stable. As for comparing the US to Europe, many European countries reduce their unemployment rates by putting the long-term unemployed on disability, so they drop out of the labor force.

Bad as things are now, they don't feel as bad as in 1982. With good reason. The ratio of employment to working age population for those 20 years old or older has fallen a lot, from a peak of 65.4% in Q1:2007 to 62.0% in Q2:2009. But by contrast, in Q4:1982, that ratio was only 58.9%. In recovery, the rate did not hit 62.0% until mid-1986. Our current bad rate is the same as a good then-record high in 1986. More people working is a type of income insurance.