Model Update: Main Phillips Curve Model

Results

The model took signal from the higher than expected incoming data and revised up the forecast. The model is now forecasting core PCE price inflation at 4.4% in 2022, 3.3% in 2023, and 2.9% in 2024 (see Figure 1 below).

Based on these results, we expect the FOMC to raise its core PCE price inflation forecast in its June SEP forecast.

Comparison with previous forecasts

Compared to the time of the May 2022 FOMC meeting, the model forecast is stronger in each year of the forecast. Since the day of the May 2022 FOMC, we received the April CPI report and labor market data. The incoming data have been stronger than expected. At the time of the May FOMC (that is, stopping the estimation in Q1), the model expected core PCE prices to growth at 4.1% (a.r.) in Q2 but based on the April CPI our Q2 nowcast is 6 tenths higher. Consequently, the starting point is higher and the model revised up the entire medium-term path.

A word of caution about the model forecast

Our “main” model is a workhorse augmented Phillips curve model, extremely popular in central banks, including at the Fed. Nevertheless, at any given point in time the information set of the econometrician is larger than the model. In this moment the information set of the econometrician is based on the narrative of the incoming data, especially about the level of durable goods. Put it differently, there is a reason to believe that the model might be overestimating future inflation if one thinks that the level of (durable) goods will drop significantly in the upcoming quarters.

An alternative, less favorable scenario

Our “main” model assumes (as in Detmeister et al. (2014)) that pi* (underlying inflation) is flat both in history and in the forecast. A less favorable scenario for the Fed would be if pi* is allowed to evolve over time, for instance if agents will re-set their long-term expectations after a protracted period of high inflation. Using our model, we have simulated a scenario with adaptive expectations in which pi* evolves according to a slow-moving average of realised inflation (for the record: under this scenario, the model is technically still “anchored” but the anchor is allowed to evolve over time). 

Under this scenario, the model forecasts core PCE price inflation at 4.6% in 2022, 3.7% in 2023, and 3.3% in 2023. The intuition is simple: the forecast is higher because the current inflationary shock is large enough to push up pi* in 2022/2024.

The main takeaway from this exercise is that the risks around the forecast are still skewed to the upside, especially at the end of the medium-term. In the baseline (flat pi*), inflation remains above the Fed target. In the less favorable scenario (adaptive expectations), core PCE price inflation is almost half of percentage point higher compared to the baseline at the end of 2024.

Forecasting performance of the “main” model

We are often asked what is the forecasting performance of our “main” model. Figure 2 below shows the root mean squared forecasting error of the model at different horizons against two alternatives: a random walk process and an autoregressive AR(2) model. The model beats the alternative models at all horizons. The performance of the model is particularly remarkable at longer horizons (10-12 quarters ahead) but the model performs significantly better than the RW and AR(2) also 2-4 quarters ahead.

In a nutshell, this is the reason why we rely so much on this model. We do so because the evidence tells us so.

Implications for the Fed Board staff

We employ our “main” model as a way to assess the risks around the Fed staff forecast. The Fed Board staff forecast (as inferred from the FOMC minutes) has been revised up significantly in the last few rounds but remains below the model (in our view, the latest Fed staff for core PCE price inflation was at 3.6%-3.7% in 2022). 

On the other hand, the SEP forecast has evolved closely to the model predictions (in March, the SEP forecast for 2022 core PCE price inflation got revised up to 4.1% from 2.7% in December). Given the evidence of our “main” model, we now expect the June SEP forecast to show again an upward revision (to 4.5% in 2022).

All told, we continue to expect the Fed communication to remain very hawkish in the next weeks.

Figures

Figure 1. Main model forecast (YoY of core PCE, %) - latest run (2022:Q2 in sample)

A. C.I. calculated from parameters distributions

Note: the confidence intervals (C.I.) are estimated using 100,000 draws from the estimated parameters distributions to simulate the path of core PCE price inflation going forward.

First quarter of forecast: 2022:Q3.

B. C.I. calculated using quasi-out-of-sample methods

Note: the confidence intervals (C.I.) are estimated using quasi-out-of-sample methods (estimate the model over a sub-sample, forecast, and calculate the root mean squared forecast errors).

First quarter of forecast: 2022:Q3.

Figure 2. Root mean squared forecasting error of the "main" model

The figure shows the root mean squared forecasting error  (RMSFE) of the “main” model. The RMSFE is computed in a quasi-out-of sample exercise. The RMSFE refers to the difference between the QoQ (a.r.) of the model and the realized (latest vintage) core PCE price inflation figure. “Horizon” is measures in quarters. 

“Model” refers to our “main” model. AR(2) refers to an autoregressive model of order 2.

Contact

Want something more tailored?

We provide tailored consulting on ad-hoc projects.