The Fed Board has published a FEDS Notes by Danilo Cascaldi-Garcia, Francesca Loria, and David López-Salido titled “Is Trend Inflation at Risk of Becoming Unanchored? The Role of Inflation Expectations”. Please, find below our review and comments including the implications of this note for the Fed staff (and FOMC) and how it relates to (and confirms) what we circulate in our Pre FOMC Meeting Package.
To keep in mind
An exercise of the Fed staff shows that underlying inflation -the most important variable to set the medium-term forecast and consequently monetary policy in this moment- has increased significantly since March 2020. Underlying inflation is now estimated to be well above 2 percent and increasing. Not only but it is now at a level that Fed staff members take as an indication of a “level de-anchoring” in long-term inflation expectations. Long story short: the FOMC has lost some battles but it is now at risk of losing the war.
What the paper does
The analysis estimates the level of underlying inflation using the (extremely technical) empirical framework presented in Mertens (2016). Unfortunately, the FEDS note is short in nature and does not clarify several important points on the identification strategy. Nevertheless, for what it is possible to understand from the note, the intuition is that measures of realized inflation and measures of long-term inflation expectations (as well as nominal yields in the original Mertens (2016) paper) should share a common trend (“underlying inflation” or pi*). Deviations of actual inflation from long-term expectations are assumed to be temporary under anchoring and a committed central bank. However, co-movements of all measures might signal a genuine move in pi*. The model extracts precisely this comovement.
Results
The extracted trend for core PCE price inflation is reported below for convenience. The estimated level of underlying inflation is about 1¾ percent in the years before Covid, close to the 1.8 percent staff assumption (as derived from Laubach et al (2014)) during those years. Since the beginning of the crisis, pi* has moved up significantly and it is now well above (and statistically so) 2 percent.
The authors also simulate what would happen to trend inflation should measure of long-term expectations move up significantly (2 standard deviations) and stay high up to September 2022. Results are reported below: unsurprisingly, trend inflation would move up even more and could be as high as 2½ percent. Worth emphasizing the authors’ interpretation of this exercise: “This exercise highlights the kind of magnitude and persistence of a shock to survey measures of inflation expectations that would reflect trend inflation significantly above the 2 percent inflation target and would prompt substantial risk to de-anchoring long-run inflation.”
How this FEDS note is related to what the Fed staff does routinely (and what we circulate in our Pre-FOMC Meeting Package)
None of the authors work on the inflation desk of the Fed Board. Therefore, we should take this exercise as complementary evidence produced by some staff members in the context of personal research. Nevertheless, in our experience, these notes are discussed by the staff internally and in this specific case do complement well the evidence of other Fed staff models.
As explained in private meetings and previous notes, in our view the Fed staff relies on three pillars to estimate underlying inflation (pi*): (i) measures of trend inflation, (ii) measures of long-term inflation expectations, and (iii) the general equilibrium approach. The FEDS note of Cascaldi-Garcia et al. (2022) can be seen as a model that combines (i) and (ii), while being mute on (iii) (which would require augmenting the model with measures of labor compensation). While Cascaldi-Garcia et al. (2022) ignores (iii), the note is a nice confirmation of what we have been circulating recently.
As a reference, below we report slide 28 of our latest Pre FOMC Meeting Package (March 2022) showing what the three pillars are suggesting. According to trend-based measures, pi* is estimated at 2.6%, while according to the Fed CIE pi* is estimated at 2.0%. Therefore, a linear combination of the first two pillars suggests that pi* is around 2.3%, perfectly in line with what Cascaldi-Garcia et al. (2022) estimate. Finally, the last pillar (which takes into account the general equilibrium of the economy) also suggest a reading of pi* above target.
Implications for the Fed staff
In our view, the Fed staff assumption about pi* is -by far- the most important in the judgmental Tealbook decomposition. The Fed staff has revised up its assumption of pi* from 1.8 to 2.0 percent (at the end of the medium-term horizon) last December. However, the evidence already suggests that pi* is possibly higher than the current Fed staff assumption. In our view, Cascaldi-Garcia et al. (2022) has already triggered an internal discussion about revising pi* again.
The bad news is that since mid-2021 the Fed staff (and the FOMC) has been too slow in adapting to the current environment and ended up being stubborn by the data. In our opinion the evidence is already very clear. In a sense, it is a repetition of what happened last October/November but this time pi* is above target and we are now at risk of de-anchoring expectations. Therefore, the current situation is worse than October/November.
The FOMC has lost some battles. The staff should recognize the risks and prevent losing the war.