The outcome of the September SEP was more dovish than expected by FRB-US. How should we interpret that? And what should we expect from the FOMC going forward? In this note, we offer three scenarios for the US economy to rationalize the outcome of the September SEP and suggest how to use the model going forward.
The bottom line is the following: forcing the model to match the new SEP path of the unemployment rate results in a nearly identical path of the FF rate. Translated: Powell is concerned about the unemployment rate, possibly more than he can admit publicly. Conditional on that, it seems that the Fed reaction function has not changed (yet), contrary to our post-FOMC note.
The simulations
Figures 1 to 3 show three simulated scenarios. In each figure, the red dashed line shows the current model baseline, the solid gray line shows constructed scenarios, and the blue dots show the September SEP. The idea behind this exercise is simple: can we match the path of all variables in the new SEP by modifying some of the model parameters and/or assuming some shock? If so, what can we learn? Here are the scenarios and the results:
In scenario #1 (Figure 1) we forced the model to follow the path of the FF rate in the SEP. For this reason, in top-left panel in Figure 1 the gray line overlaps with the blue dots. The takeaway is that if one assumes such a judgmental path of the FF rate, growth remains well above the SEP and the unemployment rate does not reach 4.4% as in the SEP but declines gently. Put it simply: in the model, it is not possible to match the new SEP by forcing the path of the FF rate.
Figure 1. Matching the SEP FF rate path
In scenario #2 (Figure 2), we repeat what we did for scenario #1 but we force the model to match the path of the unemployment rate. Specifically, we assume that the output gap in 2024-2027 will result in the path of the unemployment rate consistent with the new SEP. Again, for this reason, in bottom-left panel in Figure 2 the gray line overlaps with the blue dots. The surprising result is that the implied path of the FF rate is very similar to the new SEP. In this scenario, the model has the FF rate at 4.66% in 2024:Q4, 3.3% in 2025:Q4, 2.7% in 2026:Q4, and 2.6% in 2027:Q4. Please, remember that in Figure 2 we forced the model to match the path of the unemployment rate only; despite this, the path of the FF rate and core inflation are very similar to the new SEP. Put it simply: the model suggests that the reaction function of the Fed has not changed. Rather, this simulation is telling us that the FOMC (meaning Powell) has become more pessimistic than the model about the labor market.
Figure 2. Matching the SEP unemployment rate path
In scenario #3 (Figure 3), we tried to match the new path of the FF rate by lowering the inertial coefficient of the Taylor rule. The idea is the following: is the new SEP consistent with a less inertial FOMC? According to the model, the answer is “no” for two reasons. First, to match the path of the FF rate in the new SEP, we had to lower the inertial coefficient in the Taylor rule to an unreasonably low level (at least, out of recessions): from 0.85 to 0.50. Second, even if we match the FF rate, growth and the unemployment rate remain off. In other words: according to the model, the reaction function of the FOMC has not switched to “non-inertial” (as it generally does during recessions), at least not yet.
Figure 3. Assuming a lower inertial in the Taylor rule
Conclusion and how to use the model going forward
Yes, it is possible to make sense of the new SEP: Powell and the FOMC have become worried about the labor market, possibly more than they can admit publicly. As shown, a judgmentally lower FF rate (scenario #1) makes little sense without some real concern about the labor market (scenario #2). Not only, but an aggressive Fed (scenario #3) also makes little sense without recessionary concerns. Bottom line: to us, the exercise suggests that Powell is concerned about the labor market, possibly more he can admit publicly.
The good news is that conditional on the path of the unemployment rate in the new SEP (scenario #2), the model does match the SEP FF rate. For this reason, going forward we will try to provide model-based scenarios under different paths of the unemployment rate (as usual, we invite scenarios on-demand). This should give us hints about the future SEPs.