When Will the Fed Pivot?

Keep in mind

We expect the FOMC to show lower SEP dots starting from next June. Our expectation is based on the fact that the risks around our “main” Phillis curve model seem to be skewed to the downside in the coming months. Depending on the data, the model might signal around 20% probability of being at or below 2% in 2025 by the first quarter of 2023, and up to 40% by the second quarter of 2023. Based on this, the chance the FOMC will lower the dots starting in June 2023 is not negligible anymore.

Results

In our view, the Fed staff and the FOMC will revise down their forecast when our “main” Phillips curve model will. Figure 1 shows the evolution of the 2023 Q4/Q4 forecast comparing the SEP to three alternatives. The December SEP was unsurprising, as the “main” model (which mimics the Fed staff framework described in Detmeister et al. (2014)) correctly anticipated the upward revision.

Going forward, will the “main” model revise down its forecast? If so, why, to what extent and when?

Figure 1. Evolution of 2023 Q4/Q4 core PCE price inflation forecast (%)

Note: for the “main model” the “Dec ‘22” forecast refers to the forecast at the time of the December Tealbook (about one week before the FOMC). The latest run of the model is about 2-3 tenths lower, as it includes the November CPI as well as revised assumptions for the exogenous variables. Similar evidence applies to the 2022Q4/Q4 and 2024Q4/Q4 forecast. The “Fed staff” forecast for “Dec ‘22” is our best guess at the moment, given that the December 2022 FOMC minutes are still not available.

Our “main” Phillips curve model is likely to be surprised on the downside in the coming months. Table 1 shows the latest forecast of the “main” model for the next two quarters. The model expects core PCE price inflation at 3.8% QoQ in 2023:Q1 and 3.5% in 2023:Q2. These figures are consistent with an average MoM of 32bps in Q1 and 26bps in Q2. However, the incoming data (narrative, bottom-up approach, CI-C model and monthly distributions) seems to suggest that the strength of the data is a bit lower. For this reason, the second set of columns in Table 1 show our current baseline (judgmental) forecast in which we expect the average MoM of core PCE prices to be a bit lower (4bps) than implied by the “main” model in both quarters. The last 3 columns of Table 1 show a more favorable scenario in which core PCE prices grow 5bps less in each month than in our judgmental forecast.

Table 1. Main model forecast of core PCE price inflation and alternative scenarios

Note: the table shows the “main” model forecast (QoQ and implied average MoM) in the first two quarters of the year. The remaining columns show our baseline (judgmental) forecast, and a more favorable scenario. “Av MoM” indicates the average MoM growth rate in each quarter consistent with the respective QoQ.

The “main” model is likely to signal around 20% probability of be at (or below) 2% in 2025 in the first quarter of 2023, and up to 40% in the second quarter of 2023. The top panels in Figure 2 show the model forecasts under our baseline including 2023:Q1 or 2023:Q2 in the sample, while the bottom panels show the forecasts under the more favorable scenario. The main takeaway is that under our baseline, the model is expected to signal about 15% probability of inflation returning to target (or below) by 2025 in the first quarter of 2023, and 20% in the second quarter (as opposed to 5%-ish in the latest run). In other words, because we expect the incoming data to be a bit lower than the model, the inclusion of Q1 (or Q2) in the sample lowers the model forecast. Should the data be a touch lower than the baseline, then the model is expected to signal about 25% probability of inflation returning to target (or below) by 2025 in the first quarter of 2023, and about 40% in the second quarter. In this favorable scenario, putting Q2 in sample delivers a probability of core inflation at or below 3% by 2025 of about 85%.

Figure 2. Main model forecasts under different scenarios and samples

Note: the figure shows the “main” model forecast under different scenarios. The left panels show the model forecasts including 2023:Q1 in sample, while the right columns include 2023:Q2 in sample. The top panels use our baseline forecast for the first two quarters of next year, while the bottom panels use the more favorable forecast.

Based on the above results, we expect the FOMC to show lower  SEP dots starting from next June. In the next few months, we expect FOMC members to monitor the incoming data and reiterate the message of the latest FOMC which should remain broadly unchanged in the March SEP. The road to normalization is still quite long and history teaches that disinflating an economy is a minefield (given the current distribution, we would not be surprised to get a 0.4% core CPI in the next couple of months).

In our view, given what happened in the last 12 months, the Fed staff and the FOMC will be very careful before “pivoting” (that is, signaling a lower FF rate path). But as shown, there is now a real chance the June SEP might surprise to the downside.

(We will update the Taylor rules following the usual schedule. However, the reader can refer to the recent FRB-US model-based simulations by J. Roberts. According to Roberts, if inflation will follow a trajectory similar to what have assumed in the “favorable scenario”, at the end of 2023 the FF rate could be below 3% under a moderate policy response and below 2% under an aggressive policy response. These figures should be taken as a lower bound for the FF rate, given that inflation in the “favorable scenario” is below our baseline).

Contact

Want something more tailored?

We provide tailored consulting on ad-hoc projects.