Fed Board Common Inflation Paper

The Fed Board has released today a FEDS Note by Eli Nir, Flora Haberkorn, and Danilo Cascaldi-Garcia titled “International Measures of Common Inflation”. Given the subject and the results, it is worth takin a moment to review it.

What the paper does. The paper applies Luciani (2020) “Common-Idiosyncratic” model to Canada, the Euro area, Japan, the U.K., Brazil and Mexico. The idea is to estimate the common component for each country and evaluate its forecasting properties.

Results. Figures 1 to 6 (reported below) show for each country headline inflation, core inflation, and the estimated common component (in the case of Canada, the Figure also shows the BoC CPI-Common which is very similar to the common component itself). The proposed series are less volatile and better represent underlying inflationary forces than traditional core inflation measures. The common component is very persistent for all counties and remains close to but below target for most advanced economies (please note that the estimated common component for the Euro area is nearly identical to our own judgmental estimate of pi*, as discussed in private meetings). When assessing the forecast performance of the common component vs the index excluding food and energy items, the paper shows that the common component is a better predictor of future total inflation for nearly all countries (see Figure 8 below that shows the “score”, that is the ratio of the RMSD of the common over the index ex FE).

Comment and implications for the Fed Board staff. Very interesting exercise that confirms priors for some countries. The fact that the staff has extended Luciani (2020) model to other countries can be a signal that the staff relies heavily on this measure as a “true” measure of “core” inflation. The good news is that we already have in production a common-idiosyncratic-Covid model for the US so we can estimate it in real-time.