John Roberts (former Feb Board officer) has a new blogpost titled “Interpreting the increase in the long-term interest rates”. Please, find below our review and comments.
To keep in mind
An analysis by John Roberts using the Fed staff FRB-US DSGE model suggests that the Federal Funds Rate should peak at about 4 percent (3½ to 4½ according to the scenarios) to control inflation. In all scenarios core PCE price inflation converges back to target only gradually while the unemployment rate is expected to increase in the next 2-3 years.
Translated: Given the size of the inflationary shock, the Fed has to make a complicated choice. We continue to think the Fed will choose to moderate inflation, even at the cost of risking a significant slowdown.
What the analysis does
Using a modified version of the Fed staff Dynamic Stochastic General Equilibrium FRB-US model, Roberts simulates the path of the Federal Funds (FF) Rate as well as other macroeconomic indicators (10-year treasury yield, the unemployment rate, and core PCE price inflation), under two scenarios.
First scenario. In the first scenario Roberts allows the inflation situation to be worse than is implicit in the FOMC’s March projections while keeping fixed the term-premium component of longer-term interest rates. Specifically, Roberts assumes that (i) long-run inflation expectations are no longer well-anchored but evolves endogenously according to actual inflation, and (ii) the slope of the Phillips curve is more sensitive to the unemployment rate than would be consistent with recent history (although not as steep as in the ‘70s).
Second scenario. In the second scenario, Roberts assumes that the inflation process remains similar to that of the pre-Covid period (an assumption that, according to Roberts, is aligned with the median FOMC participant’s view at the time of the March FOMC). In this more sanguine scenario, Roberts introduces an increase in the term premium to account for the recent increase in longer-term interest rates that cannot be explained by the path of core inflation.
Results
Roberts’ results are presented below. Figure 1 shows Roberts’ results under the first scenario, while Figure 2 shows Roberts’ results under the second scenario.
First scenario. In this first scenario, the funds rate peaks at 4½ percent in 2024, while the long-term yield reaches about 3 percent in 2022 and about 3¼ percent in 2024. In this scenario, the increase in long rates to near 3 percent in the current quarter is driven entirely by higher expected future short-term interest rates. The other component of the ten-year Treasury yield, the term premium, is held fixed at a level consistent with the March SEP baseline. As for the labor market, after a momentum-driven near-term decline, the unemployment rate rises steadily; it exceeds 4 percent by the end of next year and rises to 4½ percent in 2025.
In summary, under the first scenario, because inflation is more persistent than embedded in the March SEP, monetary policy ends up being more aggressive than communicated. However, by doing so, the labor market comes under stress and the unemployment rate increases. Finally, inflation gradually converges toward target but only at the very end of the projection horizon (which in Roberts’ exercise is 2030).
Second scenario. In the second scenario, inflation is somewhat higher over the next few years than in the March SEP baseline but with longer-term expectations assumed to remain well-anchored, inflation comes down more quickly toward 2 percent. Under this setting, the funds rate is around 2½ percent in the fourth quarter of this year and reaches 3½ percent at the end of next year. The more-modest policy change in the second scenario, by itself, explains only a portion of the recent increase in longer-term interest rates. In the FRB/US model, long-term interest rates are a key determinant of financial conditions. With conditions less tight in this scenario, the increase in the unemployment rate is more gradual, and it remains below 4 percent until the end of 2024. Nevertheless, even in this second scenario, the unemployment rate ends up ticking up higher in the next few years, before gradually converging to its long-run level.
Which scenario is more plausible? According to Roberts’, the first scenario is more likely than the second given the growing evidence that longer-run inflation expectations are becoming un-anchored.
Comments and implications for the Fed
Roberts’ exercise is extremely interesting and very well in line with our analyses. In our view, the FRB-US model is used by the Fed Board staff to forecast macroeconomic variables (in the so-called Tealbook “extension”), as well as for stochastic simulations. In both scenarios, the inflation shock is so large that the path of the FF rate ends up being steeper than in the March SEP projection. Taking an agnostic view on long-term inflation expectations (that is, roughly an average of the two scenarios presented above), the FF rate needs to reache about 4 percent at the peak. Not only, but in both scenarios the unemployment rate ends up trending higher in the next few years before converging to its long-run average.
(Please note that in the first scenario, core PCE price inflation remains way above target in 2024 despite the more aggressive monetary policy. This evidence largely confirms the one from our “main” Phillips curve model)
The increase in the unemployment rate does not imply a recession in the model (using conventional metric of falling real GDP). However, increases in the unemployment rate of this magnitude have not historically occurred outside of a recession in the real world.
Bottom line: the Fed has almost no space for a policy mistake. Landing the 747 safely is looking increasingly difficult because of the growing evidence of de-anchoring expectations. The Fed has a dual mandate but its reputation is at risk. We strongly doubt the tone and actions can change, until there will be robust evidence that inflation and expectations have moderated. At this point, we expect the Fed to be determined, no matter what the cost in terms of growth unless inflation surprises significantly on the downside.