Somewhere in the World There Must Be Someone Who Understands the Unemployment Insurance Seasonal Adjustment Factor
Looking at this time series, the seasonal adjustment does look a little crazy, but I promise– it’s not that bad. Let’s go to the ten-year:
So, yes, that spike looks huge, but historically, it’s only really big. Obviously, it’s possible that the seasonal adjustment should be different in recessions. We could test for this– we could set up a Markov-switching model with trigonometric terms, and allow switching in those terms to see whether the adjustment should be smaller or larger in recessions. It’s possible that this wouldn’t be terribly enlightening, as different recessions can be very different, but it wouldn’t take too terribly long, and I’m sure Brad has some graduate students he could put to work.