We expect the Fed staff to revise up its 2021 core PCE price inflation forecast, as the incoming data have been a bit stronger than expected since the closing day of the July Tealbook. Going beyond this year, we do not expect any material change in the staff inflation forecast for 2022 and 2023. As such, the Fed staff should remain more dovish than the median FOMC participant.
Our set of Phillips curve models suggest that the risks around the staff forecast are skewed to the upside, with increased uncertainty. Our set of trend inflation models took signal from the recent strength of the data and suggest that the level of underlying inflation is now closer to target. We interpret the evidence from the trend models as a “significant progress” towards the Fed average inflation target goal.
As for the SEP, we expect an upward revision to the median forecast in 2021, 2022, and 2023 as the incoming data have been strong and the median FOMC participant appears to be more hawkish than at the time of the June FOMC. Having said so, we continue to believe that the actual future path of the federal fund rates will be more dovish than implied by the SEP, as the median FOMC participant is currently more hawkish than the median FOMC voter.
The current forecast
Table 1 show the evolution of the Fed staff forecast (bottom panel) for core PCE price inflation, as well as the evolution of the median SEP forecast (mid panel). In the last few rounds, the Fed staff has revised up the 2021 forecast in reaction to higher-than-expected incoming data. At the same time, the forecast for 2022 (a bit below 2 percent) and 2023 has remained largely unchained (at 2 percent or a touch higher).[1] The evolution of the SEP forecast has followed a similar path, although the median SEP forecast for 2022 and 2023 was set above the Fed target at the June FOMC.
The top panel of Table 1 shows our calibrated decomposition of the contributions to core PCE price inflation (history and current Fed staff forecast) according to the Fed staff framework.[2] In our view, the staff continues to assume that the level of underlying inflation is flat at 1.8 percent.[3] Also, in our view the staff is assuming that the elevated unemployment rate will put downward pressure on core PCE prices in 2021 but expects resource utilization to tighten further and put upward pressure on consumers’ prices in 2022 and 2023. Switching to core import prices, the passthrough from the Dollar is expected to add about 25bps this year, given the depreciation of the Broad dollar.
Going forward, in our view the staff forecast assumes a small Dollar appreciation (about 2 percent per year) which results in a small negative contribution of core import prices to core PCE price inflation. As for energy prices, the passthrough from crude oil prices to core PCE prices is expected to be small, if any. Finally, the portion of observed inflation that cannot be attributed to the fundamentals (the “other factors” in Table 1) is expected to be large this year but should fade next year, as the currently elevated inflation should prove to be transitory. Figure 1 shows core PCE price inflation and its estimated contributions.
[1] The Fed staff forecast is inferred from the FOMC minutes. For instance, this is the relevant part from the July 2021 FOMC minutes: “The staff expected the 12-month change in PCE prices to move down gradually over the second part of 2021, reflecting an anticipated moderation in monthly inflation rates and the waning of base effects; even so, PCE price inflation was projected to be running well above 2 percent at the end of the year. Over the following year, the boost to consumer prices caused by supply issues was expected to partly reverse, and import prices were expected to decelerate sharply; as a result, PCE price inflation was expected to step down to a little below 2 percent in 2022 before additional increases in resource utilization raised it to 2 percent in 2023.”
[2] The staff framework is explained in Detmeister et al. (2014).
[3] The staff assumption about underlying inflation is explained in Laubach et al. (2014) – “[…] according to the staff’s framework, if long-term inflation expectations remain unchanged, inflation will gravitate toward a fixed level in the absence of resource slack or other shocks. In many macroeconomic models this resting point is numerically identical to the long-term inflation expectations of price and wage setters. […] In the staff baseline, notional long-term inflation expectations are currently estimated to be anchored at 1.8 percent”.
Table 1
Note: Details may not sum to totals because of rounding. The “other factors” line includes the contribution of core non-market inflation, as well as the contributions of factors that are unrelated to other fundamentals. The “Board staff” forecast is inferred from the FOMC minutes. The “FOMC” forecast refers to the median SEP.
Figure 1
Note: The Fed staff framework of core PCE price inflation is explained in Detmeister et al. (2014). “Underlying inflation” is the inflation rate that would prevail in the long-run, net of transitory shocks. For a discussion of “underlying inflation” see Rudd (2020). Slack is captured by the unemployment gap. The estimate of the natural rate of unemployment is provided by CBO. “Import prices” and “Energy prices” are expressed in deviation from core PCE prices, market-based.
Risks around the forecast
We assess the risks around the Fed staff / SEP forecasts in two ways. First, we assess whether the staff assumption about the level of underlying inflation (1.8 percent, flat in history and forecast) is supported by econometric evidence. Table 2 shows the estimated trend inflation from our set of econometric models. Our models suggest that trend inflation was about 1.8 percent before the pandemic and at the beginning of this year. However -on average- the models took some signal from the recent incoming data and in the last two quarters they estimate a higher level of trend inflation (1.9 percent in 2021:Q1, and 2.0 percent -to rounding- in 2021:Q2). We interpret the evidence from these models as a “significant progress” towards the Fed average inflation targeting goal, although additional confirmation will be required in the next few quarters before concluding that trend inflation has shifted to a permanently higher level. Nevertheless, this evidence is an upward risk around the Fed staff inflation forecast.
As a second exercise, we use our collection of Phillips curve models to generate a distribution of point forecasts and assess the risks around the staff inflation framework. The results of this exercise are reported in Figure 2. Our set of Phillips curve models include about 100,000 different specifications, each differing from the other according to the variables included and the restrictions applied to the model.
The results of our exercise are presented splitting between “anchored” and “unanchored” models: in the former, inflation evolves over time but remains “anchored” to a trend (which itself can be flat as in the staff framework or evolve overtime), while in the latter any innovation to actual inflation results in a permanent shift. The top panels of Figure 2 show the distributions of point forecast in 2021-2023 for the “anchored” models. The top left panel shows the distributions of point forecasts stopping the estimation in 2021:Q1, while the top right panel includes 2021:Q2 in the sample. The bottom two panels show the same results for the “unanchored” models.
Three main conclusions arise. First, if long-term inflation expectations (or alternatively trend inflation) remain stable, the behavior of actual inflation is expected to average close to the Fed target. Therefore, monitoring the evolution of the anchor remains crucial. Second, if the estimation is stopped in 2021:Q1, the models are predicting an inflation rate just a touch below 2 percent over the forecast horizon with moderate uncertainty around it. However, if the second quarter of the year is included in the sample, the means of the distributions shift upward. In this sense, the models are suggesting that core PCE price inflation might remain above target over the entire forecast horizon. Finally, if 2021:Q2 is included in the sample, the dispersion (uncertainty) around the mean forecast increases significantly. The same evidence applies to the “unanchored” world, although in this case the means of the distributions ends up being much higher. Overall, our set of Phillips curve models suggest that the risks around the Fed staff forecast are skewed to the upside, and that the uncertainty around the staff forecast has increased significantly in the last quarter.
Table 2
Note: “Average” refers to the simple mean across all models.
Figure 2. Point forecast distributions from our set of 100,000 Phillips curve models
End of sample 2021:Q1
End of sample 2021:Q2
Anchored models
Unanchored models
Note: whiskers show the distribution of point forecasts in each year over the relevant forecast horizon from the set of Phillips curve models we maintain (total number of models is around 100,000). The models are split between “anchored” and “unanchored, according to the restrictions imposed. Each model differs according to several dimensions: (i) lag structure of each variable, (ii) variable capturing economic slack (i.e. unemployment gap, output gap, etc..), (iii) type of non-linearity (i.e. linear in slack, quadratic in slack, etc..), (iv) variables capturing passthrough from the Dollar, (v) variables capturing the passthrough from Brent, (vi) type of trend (i.e. flat, long moving average, etc..). Each whisker shows the relevant descriptive statistics of the distribution: minimum, maximum, 25th percentile, 75th percentile, the mean (green horizontal bar), and the median (green triangle). The y-axis is expressed in percent.
Conclusion
Our shadow forecast suggests that the Fed staff will revise upward its 2021 core PCE price inflation forecast but will not materially change the 2022-2023 outlook as the fundamentals in those years remain similar to last Tealbook. Nevertheless, our set of econometric models suggest that the risks around the Fed staff forecast are now skewed to the upside with significant uncertainty. As for the SEP, we expect an upward revision to the median forecast in 2021, 2022, and 2023 as the incoming data have been strong and the median FOMC participant appears to be more hawkish than at the time of the June FOMC. Having said so, we continue to believe that the actual future path of the federal fund rates will be more dovish than implied by the SEP, as the median FOMC participant is currently more hawkish than the median FOMC voter.