From a new working paper by Jan Eckhout (UPF Barcelona), h/t Torsten Slok:
The current practice of measuring inflation for monetary policy uses the average annual rate of inflation. When inflation changes rapidly, either in increase or decrease, the average annual rate is biased towards data from very far in the past and conveys the real price level with a delay of six months. I suggest using spot inflation as a more appropriate measure of price change. The scale counteracts noise in the data with the accuracy of an instantaneous rate change. Using the latest inflation figures, it shows that spot inflation in the US and the Eurozone has returned to the 2% target and that the period of high inflation is over. This is not the case for the United Kingdom, Japan and Australia.
Note that this is related to the discussion we had on this blog about the relative merits of 12-month inflation (or some people’s desire to use a 17- or 18-month annual inflation rate), 3-month or 1-month inflation. 12 months overweight older data, 1 month too noisy. The figure below shows the traditional 12-month month (in blue) and a single month (or month by month) as black circles.
Using a kernel polynomial with a weight of a = 4 (the traditional scale is 0, that is, an equal weight for each monthly rate of inflation), the author obtains the red line, which he calls “instantaneous inflation”. (See Fig. B1.a in the paper for a graph showing how recent observations are more likely than older observations, compared to the simple 12-month change, a = 0).
The author notes:
I find smooth inflation in December 2022 to be 2%, the inflation target (see Figure 1, where a = 0 corresponds to the conventional measure and a = 4 corresponds to the intraday inflation measure by a bandwidth factor of 4… [year-on-year] Instead, the measure of inflation in December was 6.5%.