Seasonal Adjustment Time Bomb

In the 3 most recent months, the cumulative seasonally adjusted increase in the U.S. CPI has been just 0.5%. Yet the actual CPI increase, which is to say the non-seasonally adjusted increase, has been a full 1.8%.The Bureau of Labor Statistics has in other words assumed a full 1.3% seasonal increase in prices for the last 3 months. This assumed seasonal increase is 0.3%:points higher than it was last year during February to April.

One can argue whether this seasonal adjustment is right or not. I personally suspect it is too high, meaning that the seasonally adjusted increase for the last few months has been underestimated. However, it should be remembered that seasonal adjustments are a zero sum game on a yearly basis. This means that during the coming 9 months, the reported seasonally adjusted increase should be 1.3% higher than the actual increase. Looking at the data for seasonally adjusted and non-seasonally adjusted increase, you can see that this month (May), the actual and the seasonally adjusted increase will be equal. After that, the seasonally adjusted increase will mostly be either equal or higher than the actual increase, with the biggest difference in favor of the seasonally adjusted measure (When in other words the biggest seasonal price cuts are assumed) being in July and December.

The point here is that while the official inflation numbers have probably been understated by a too great seasonal adjustment, this will not be repeated in the coming months. And as actual inflation remains high that in turn means that there will almost certainly be a lot higher headline inflation numbers in coming months than in the previous months.