The Layers of Inflation Persistence – A Comment

The NY Fed has published an analysis “The Layers of Inflation Persistence” by Almuzara, Jarocinski, and Sbordone. The analysis is an update of a NY Fed blog post circulated last April (that we reviewed here). We show that the authors’ results are in line with our CI-C model, which can therefore be used as guidance going forward.

Our analysis

The common component across core PCE items is losing steam. The paper estimates the “common component” (the “trend”) across core PCE items using a dynamic factor model as in Stock and Watson (2016). Figure 1 shows the main results of the analysis. The authors find evidence of a decline in the size of the persistent component of core PCE inflation starting in September 2022. The common component (“multivariate core trend” in the Figure) was around 2 percent before Covid. It then started to increase at the end of 2021, it peaked at the beginning of 2022 and it has been declining in the last few months. The intuition is simple: while housing remains strong, other sectors are decelerating; therefore, the “common” component across all sectors is losing steam.

Figure 1. Main result of NY Fed analysis by Almuzara et al. (2022)

The authors’ results are very much in line with our CI-C model. Therefore, our model can be used as guidance going forward. Our CI-C (Common-Idiosyncratic, and Covid) model is based on Luciani (2020, 2021) -a model maintained by the Fed Board staff- and is close in spirit to Almuzara et al. (2022). Figure 2 shows the latest estimate of the common component from our CI-C model (3m/3m ar, 6m/6m ar, and YoY). Our results are very much in line with Almuzara et al. (2022): the common component seems to have peaked at the beginning of 2022 and it has lost steam in recent months. Having said so, the level of the common component, no matter how estimated, remains above the Fed target.

Figure 2. Metrics of the common component from our CI-C model.

Implications for the Fed. The reason why we maintain our CI-C model is because it mimics a model maintained by the Fed (Board) staff. Unsurprisingly, the NY Fed model by Almuzara et al. (2022) is sending similar signals. The question is: are the results in Almuzara et al. (2022) dovish? The answer in our view is “not really” for three reasons. First, the Fed staff (and most FOMC members) were already aware of the results at the time of the December FOMC. Second, as mentioned, there is still no conclusive evidence that the “common component” has returned back to target. Finally, as previously discussed, while there are signs that moderation is on its way, the medium-term forecast (which is crucial for monetary policy) remains unfavorable.

The Fed can start thinking about pivoting when the common component will be back down, not earlier than that. We are going in the right direction but we remain on a minefield. For the time being, be patient and monitor our CI-C model. It will give us early signs.

Contact

Want something more tailored?

We provide tailored consulting on ad-hoc projects.