Why friends don’t let friends calculate 18-month non-seasonally adjusted growth rates

Bruce Corridor appears to suppose that calculating 18-month inflation charges (both annualized or not) is okay. It is It is all okay. Except you do that with non-seasonally adjusted information. In case you do, you have to be actually clear. Illustrative instance of CPI under.

Figures 1: Annualized 18-month inflation for the seasonally adjusted city CPI, all customers (blue), and for the non-seasonally adjusted city CPI, all customers (mild brown), calculated utilizing logarithmic variations. NBER outlined recession dates shaded grey from peak to finish. Supply: BLS by way of FRED, NBER and personal calculations.

Final time I recall the numbers, my imaginative and prescient is 20-400 (as a substitute of 20-20) uncorrected, however even I can see the 2 sequence differ considerably at completely different instances.

In homes Reader CoRev will accuse me of getting manipulated the info, let me notice that I take advantage of the FRED sequence CPIAUCSL for seasonally adjusted CPI and CPIAUCNS for seasonally adjusted values. All different calculations are fairly simple in case you take a look at the formulation within the legend field – except you are unfamiliar with utilizing pure logs, proper distrusts them. (Right here is Jim Hamilton’s contribution to using logs and log variationsin case you do not belief Menzie Chinn by title or worldview.

I’ve to say 18 months is a little bit of an odd alternative. 1 month, 3 months, 12 months, 2 years, 5 years, okay. 1.5 years, properly, you need to have a cause.