I might as well put this here. Brad DeLong writes:
Put me down as somebody who does not believe that the seasonal factor in the unemployment rate is twice as big today as it was four short years ago, or was half as big four short years ago as it was in the early 1990s…
Not that I am complaining about the BLS, you understand. If I could do better, they would already have done better. Nevertheless this is a source of nervousness…
My first thoughts:
At a first glance, the size of the seasonal adjustment factor looks like it is countercyclical to the business cycle, which immediately raises the question: Why would seasonality-based volatility in unemployment increase during a recession?
Could it just be that seasonal employment is less susceptible to business cycle movements than regular employment, so that during a recession the (relatively constant) seasonal movements look larger relative to the smaller total employment number?