Jan Eeckhout (Professor at Pompeu Fabra) has published a paper titled “Instantaneous Inflation”. Under normal circumstances, we would not review the paper given the low quality. However, because we have received several questions, we have decided to do so. The main point is that Eeckhout (2023) argues that “US and Eurozone inflation is back to the 2% target” (even if his own charts suggest otherwise) because he relies on a metric that is almost identical to the 3m/3m ar of headline inflation. Put it differently, Eeckhout has written a full paper to invent the wheel; and still, the conclusion is wrong given the usual issues when using headline figures. Please, stop here if not interested in the details (we expect the Fed and ECB staff to dismiss the paper entirely).
What the paper does and results
The paper has invented the wheel. As basic as it might sound, Eeckhout (2023) suggests to calculate inflation by putting “more weight on recent observations and less on observations in the distant past” using specific weights (Kernel). Eeckhout (2023) argues that according to this innovative (?) metric, “US and Eurozone inflation is back to the 2% target”.
Comments
2%, 4%.. what’s the difference? Figure 1 shows the 3m/3m ar of headline CPI (US) and headline HICP (EA, nsa). In December, the 3m/3m (calculated using Eurostat approach) were 1.8% and 4.3%, respectively. In other words, Eeckhout (2023) has invented the wheel given that his methodology is approximated well by a basic 3m/3m (ar) metric. Not only, but while in the abstract the author writes “inflation in the US and the Eurozone is back to the target of 2%”, the reader finds out at page 6 that “instantaneous inflation” for the Eurozone “has not reached 2% yet (it is closer to 4%)” (from which the comparison with the 3m/3m at around 4%). It remains a mystery how a 4% inflation rate can be considered “being back to 2% target” (for the record, Eeckhout (2023) uses non-seasonally adjusted data for the Euro area, a choice that raises additional question marks).
Figure 1. 3m/3m ar of headline CPI (US) and headline HICP (EA)
Some of the results in Eeckhout (2023) are, in fact, worrying. Unsurprisingly, when Eeckhout (2023) employs his methodology on core data, he obtains completely different results. For instance, his measure of “instantaneous inflation” is running around 4% (ar) for core CPI in the US with core services around 6-7% (ar). Needless to say, the literature (see for instance Luciani and Trezzi (2019)) has shown that current headline inflation (as opposed to current core inflation) has no predictive power to forecast future headline inflation. Therefore, the very same results of Eeckhout (2023) suggest a completely different picture for the future.
Implications for the Fed and ECB staff
This paper has no implications and we expect the Fed and ECB staff to dismiss it entirely.