In this note, we update FRB-US through Q2. In part II, we investigate what drives core inflation in the model, and under what conditions core can re-accelerate. In part II, we also compare FRB-US with other models.
According to the model, the peak of the FF rate is identical to the previous run (here) at 5.8%, but the strength of the economy implies one cut less in 2024 and one/two less in 2025. The model projects the 10y yield at 4.0% and the 30y yield at 3.9% at the end of this year (the probability of the 30y yield to reach 5.5% by 2023:Q4 is less than 5% according to the stochastic simulations). A re-acceleration of inflation is always possible, but it is not the baseline at this point. Finally, we also assess what would happen if core PCE price inflation falls to 3% and remains there, with and without raising unemployment.
(For questions about FRB-US and the scenarios presented below please, contact Tilda Horvath tilda.horvath@underlyinginflation.com. We remind the reader that we run scenarios on request, free of charge. Please, get in touch for more details about each scenario. The model contains 200+ variables, for obvious reasons we cannot show all of them.)
The new FRB-US forecast
Higher FF rate path. Figure 1 shows the updated FRB-US forecast. As usual, the figure shows a comparison between the latest SEP (June in this case) and the “inconsistent FRB-US” forecast (the blue line and the orange line in Figure 1, respectively). The relevant variables are updated through Q2, except for core PCE price inflation for which we have put our nowcast of Q3 in-sample.
Figure 1. SEP forecast (blue line) and “inconsistent” FRB-US simulation (orange line)
Note: Real GDP growth and core inflation are expressed as YoY. Core inflation is core PCE price inflation. The blue line shows the latest SEP. The orange line shows the “inconsistent FRB-US” forecast (the current baseline), that is the model-based forecast removing the “add-factors” put by the Fed staff to match the latest SEP.
Here are the details:
GDP growth is stronger. In the previous run (see here) we assumed GDP growth at 2.0% (QoQ, ar) in Q2. Funny enough, the (true) model forecast for Q2 was 2.4% but we judgmentally lowered it because it seemed too high. In any case, the upward surprise (the details of the GDP report, and the updated dataset posted by the Fed staff) result in a Q4/Q4 GDP growth close to 2%. For the record: the model never forecasted a recession for 2023, although it did forecast a slowdown in H2 which is now shifted to 2024. (Our previous FRB-US notes are here: June 25, June 1, April 28, and March 19). As for 2024 and 2025, the model continues to deliver a forecast significantly below the SEP.
Lower unemployment rate. Because growth is stronger, by Okun law the path of the unemployment rate is a bit lower than in the previous run. The unemployment rate is now projected at 4.0% in 2023:Q4 (vs 4.2% in the previous run), and at 4.45% in 2025:Q4 (vs 5.0% in the previous run).
Core inflation marginally higher. The path of core PCE price inflation is overall similar to the previous run (and to the latest SEP) projected to be around 2.5% in 2025.
Path of the FF rate is higher. Putting everything together, the model expects the peak of the FF rate at 5.8% (same as previous run), with the usual caveat that the effects of the credit crunch are not reflected (that is, the reader might want to lower it by 25-50bps). Nevertheless, because the economy remains strong, the FF rate is now projected at 5.2% in 2024:Q4 and 4.4% in 2025:Q4. (The 10y yield consistent with this scenario in 2023:Q4 is 4.0%, and the 30y yield is 3.9% according to FBR-US).
Figure 2 shows the stochastic simulations around the baseline (the orange line) shown in Figure 1. Following standard procedures of the Fed staff we draw randomly from historical errors (1970q1 to 2017q4) for 54 variables and 5,000 replications. The simulations continue to show that: (i) calling a recession in this environment is like a flip of a coin, and (ii) conditional on the estimated persistency, the model sees a relatively low probability of getting back to the inflation target by the end of the medium-term. (Note: the stochastic simulations for the 30y yield –here– suggest that the probability of the 30y yield reaching (say) 5.5% by 2023:Q4 is less than 5 percent).
Figure 2. Stochastic simulations around the baseline.
Note: Real GDP growth and core inflation are expressed as YoY. Core inflation is core PCE price inflation.
Alternative scenarios
We have run a set of alternative scenarios. These scenarios should be taken with cautious because rather than let the model delivering its forecast, we have forced the path of one or more variables of interest. The goal of this exercise is to provide a (internally consistent) scenario of what the Fed would do.
Scenario 1. Inflation falls to 3% and stays there. Results are shown here. Bottom line: the path of the FF rate is very similar to the baseline in 2023 and 2024, and a bit higher in 2025. This happens because in the baseline core PCE price inflation is expected to reach 3% only at the end of 2024. Therefore, in FF space, the difference between the baseline and this scenario becomes visible only at the end of the forecasting period.
Scenario 2. Inflation falls to 3% and stays there, and the unemployment rate does not increase. Results are shown here. Bottom line: FF rate stays around 5%. In this scenario we have forced core inflation to stay at 3% indefinitely and the unemployment rate at 3.8%. (please, note that we had to extend the x-axis). The bottom line is that the FF rate does not really fall and remains around 5% even in 2028. Also, the reader should notice that not even in this scenario (which assumes a much stronger economy than the latest SEP), real GDP growth is as high as in the June SEP at the end of 2025. Put it differently, in order to match the June SEP in 2025, the Fed staff had to fold in large “add factors”, because otherwise growth would be lower in every scenario.
Scenario 3. Inflation falls more quickly. Results are shown here. In this scenario, we assume that core PCE price inflation falls to 2% by 2024:Q2, admittedly a pretty optimistic scenario. We take this scenario as the lower-bound of the FF rate path net of recession. In this scenario, the Fed starts cutting soon and the FF rate is projected at 4.6% at the end of next year. One interesting evidence is that despite core inflation being at 2%, the FF rate remains well above “neutral” because the unemployment gap remains negative throughout the entire forecasting horizon. (Note: the reader can see the steady-state values in FRB-US by looking at scenario #2 where we have extended the x-axis. In the model, U* is currently estimated at 4.2% and the “neutral” FF rate is at 2.5%, with r* at 0.5%).
Scenario 4. A less inertial Fed. We have run a set of scenarios in which we have used the non-inertial Taylor rule as a reaction function for the FOMC. The reason is that there is evidence the Fed cut rates more quickly than it raises. Results are available upon request. The main takeaway is that it is possible to have a more dovish path of the FF rate (compared to the ones shown above) but only assuming that (i) inflation falls pretty quickly and/or the US has a slowdown that raises the unemployment rate.
Conclusion
The main message of Part I of this note is that FRB-US continues to estimate the terminal rate at 5.8% but projects a path of the FF rate a bit more hawkish than the previous run. Nevertheless, even considering the strength of the economy, it is hard (in fact, very hard) to project the 30y yield at 5.5% (any reference to open trades by famous traders is purely accidental). In Part II, we explain under what conditions inflation can re-accelerate comparing FRB-US to other models (spoiler: it is pretty hard, although not impossible, for core PCE to remain above 3% in the long-run).