The latest run of FRB-US shows a stronger economy, more persistent inflation path and the peak of the FF rate at 5.9%. We discuss how this result can inform about the upcoming June SEP, why we do NOT expect a hike, and why in fact the Fed can even be done hiking.
(For any question about this post or FRB-US, please #AskTilda, our expert, at: tilda.horvath@underlyinginflation.com)
Our previous research
In our previous FRB-US notes (March note, March note part II, and April note) we explained how to use FRB-US to assess the risks around the SEP. Based on the “unrestricted” FRB-US forecast we signaled: (i) very low probability for the Fed to cut rates in 2023, (ii) some upside risks for the FF rate, (iii) more persistent core PCE inflation path than in the SEP, (iv) lower path of the unemployment rate in 2023, (v) low statistical power to predict a recession in 2023:H2, (vi) likely stagnation in 2024/2025.
In the next paragraph we show the new run of FRB-US. Overall, we expect the new SEP to be closer to the model and signal: higher growth, more persistent / higher core inflation, lower unemployment rate, and possibly a higher terminal rate. A very big win for the model, should this happen.
The latest run
Figure 1 shows the latest run of FRB-US using the recently released dataset consistent with the March SEP. The blue line is the path of the variables fitting the SEP. The orange line is the “inconsistent” (or “unconstrained”) FRB-US forecasts, that is the path of the variables after removing the add-factors that the Fed staff built in to match the SEP. In other words, the orange line is the “true” forecast of the model. (For the record, in the last dataset we have found 20 add factors, half of which were built by the Fed staff to match the short-term dynamics of the variables and the other half for the medium-term).
Compared to the previous run, the US economy is expected to be a touch stronger and core inflation is more persistent. Consequently, the model delivers a higher terminal rate of 5.9% in 2023:Q4. The first cut of the FF rate is projected in 2024:Q2. (The path of the 5y yield, 10y yield, 30y yield, and real expected return on equity can be seen here) Figure 2 shows the FRB-US stochastic simulations around the orange baseline in Figure 1 (the stochastic simulations of the financial variables are here). The main takeaway is that calling a recession in 2023:H2 remains a “flip of a coin” and we can easily remain in positive GDP growth territory.
Figure 1. FRB-US model-based forecast (orange) and March SEP (blue)
Note: the figure shows the FRB-US forecasts (orange line) of the main variables. The blue line is the path produced by FRB-US that interpolates the latest (March) SEP. Real GDP growth and core inflation is expressed as YoY. Core inflation is core PCE price inflation. The reader can find an explanation of the publicly available FRB-US and the “unconstrained” version (the orange line) here.
Figure 2. Stochastic simulations using FRB-US
Note: Following standard procedures of the Fed staff we draw randomly from historical errors (1970q1 to 2017q4) for 54 variables and 5,000 replications. Real GDP growth and core inflation is expressed as YoY. Core inflation is core PCE price inflation.
How to use FRB-US for the upcoming June SEP
The evidence in Figure 1 and 2 can be taken as a way to assess the risks around the SEP. The SEP is a judgmental forecast and the median dots are not (necessarily) consistent across variables and across SEP rounds. Therefore, one should always be very careful. Having said that, our exercise aims at assessing the risks around the SEP precisely because we take the FRB-US SEP-consistent dataset and we run the model unconstrained (that is “what the model would deliver without being constrained to match the SEP”). One important caveat is that FRB-US does not have a banking sector (although it does have a financial sector). Therefore, the path of the FF rate in Figure 1 and 2 should be taken as an upper bound. Model simulations show that the credit tightening is equivalent to about 2 hikes. Therefore, the “true” peak of the FF rate is about 50bps lower than in Figure 1. This implies that the Fed is already pretty close to the “true” peak of the FF rate; there is (in theory) room for another hike but the Fed can easily skip a meeting (or being done at all, unless we get other upward surprises going forward). For this reason, on our May 25th note (here) we wrote “We are agnostic about a hike at the June meeting right now (gun pointed to our head we would say no hike)”. We continue to think that the Fed will take time because at this point the marginal gain from an extra hike is probably very small. The FOMC can just take time and see (as we think) if in H2 the data will finally moderate.