“Higher for longer” continues to be the genuine baseline. A new Fed paper (“Post-COVID Inflation Dynamics: Higher for Longer” by Randal J. Verbrugge and Saeed Zaman) suggests that conditional on the December SEP path of the unemployment rate, core inflation is set to remain above target (and above the SEP forecast) despite recent improvements. A severe recession accomplishes the mission in the simulations to reach target, but it is far from optimal. The implication of the paper for monetary policy -our take here- is that pushing the FF rate above the level signaled in December should not be on the table. However, the Fed has little to celebrate because the medium-term forecast remains unfavorable. “Higher for longer” is the baseline, until the models will say otherwise.
What the paper does
The paper employes SVAR techniques to model core PCE price inflation using three components that align with those noted by Chair Powell in his December 14, 2022, press conference: housing services, core goods, and core services ex housing. The model assumes non-linearities in slack and it is conditional on the SEP unemployment rate path and a rapid deceleration of core goods prices. The authors construct forecasts and error bands via counterfactual simulations as in Kilian and Lütkepohl (2017) with shocks bootstrapped from estimated residuals.
Results
According to Verbrugge and Zaman (2023) core inflation is more persistent than in the SEP and can remain well above target in the medium-term. Core PCE price inflation is forecasted to remain higher for longer: rather reaching 2.1 percent by the end of 2025 as in the most recent SEP, the model projects that it will be at 2.8 percent, with the 70 percent confidence interval spanning 2.4 to 3.2 percent (see Figure 1). The key to this result is the fact that inflation is estimated to be more persistent than commonly believed and assumed in the SEP. (The reader should pause here and think that the model results are conditional on the recent deceleration of core goods. In other words, the model already “knows” that inflation has moderated but still fails to deliver 2% in the medium-term)
Figure 1. Forecast of core inflation: SEP vs Verbrugge and Zaman (2023) model.
A severe recession disinflates the US economy, but it is not optimal. The authors simulate the path of core inflation under different scenarios for the unemployment rate (in the “severe recession” case the unemployment rate is assumed to peak at 7¾%, in the “moderate recession” at 5¾%, and in the “soft landing” a touch above 4%). The results of these simulations are shown in Figure 2. The main takeaway is that only in the “severe recession” case core inflation gets close to target by 2025. In all other cases, core inflation remains well above target (and above the December SEP). Having said so, using a reduced-form welfare analysis, the authors also show (not reported in this note for brevity) that a severe recession is far from optimal because the welfare losses (that is, the weighted average of the deviations of inflation and the unemployment rate from their targets) is much higher in this case than in any alternative. Put it differently, the paper suggests that it is preferable to accept a higher inflation rate because otherwise the cost on the labor market would be too high.
Figure 2. Forecast of core PCE price inflation under different scenarios
Implications for the Fed staff and the FOMC
The horses are ready but the Fed cannot open the gates. Can core inflation converge back to target faster than the models suggest? The answer is “yes” for the usual reasons (omitted variables, etc..). However, until the models suggest that the probability of reaching target is very low, it is unlikely (in our view) the FOMC will signal a lower path of the FF rate. In our experience, the Verbrugge and Zaman (2023) paper has been heavily scrutinized and discussed by the Fed staff. The results of the paper are in line with the signals from other models including our “main” Phillips curve model. In other words, the Fed is not ready to open the gates. The horses shall wait.