“We need models in which the credibility of a central bank is endogenous to its actions. The assumption that expectations are determined by the target is misleading at best and dangerous at worst.” Mervyn King, Inflation Targets: Practice Ahead of Theory, June 2024.
Four years ago, at the conclusion of its first policy strategy review, the Federal Reserve’s Open Market Committee (FOMC) expressed the intention “to undertake roughly every five years a thorough public review of its monetary policy strategy, tools, and communication practices.” As Chair Powell highlighted recently, that review is set to begin in the next few months. With this prospect in mind, now is an excellent opportunity to highlight those elements of the current strategy that are most in need of attention. (We benefit from the June 2024 Hutchins Center conference which gathered suggestions for the strategic review.)
Our focus is on the following four issues that we will address in a series of posts:
Asymmetry and inflation bias. The asymmetry of the strategy introduced in 2020 imparts an upward bias to inflation relative to the target. That is, the commitment to make up for undershoots – but not overshoots – of inflation, combined with a shift away from preemption toward “patience,” means that average inflation is more likely to be above than below target. How can the Committee reduce or eliminate this bias?
Inflation target. Can the FOMC enhance the credibility of its inflation commitment by raising its inflation target? The combination of the effective lower bound on policy rates and several years of struggling to raise inflation to the 2-percent target was a key motivation for the troubling asymmetry of the 2020 strategy. Would a higher inflation target be less troubling?
Policy toolkit. The breadth and complexity of the Committee’s policy toolkit raises a host of questions. For example, does the federal funds rate remain the appropriate target following the virtual disappearance of unsecured interbank lending? What is the role of central bank lending in the future operational framework? How should we view the balance sheet and tools including large-scale asset purchases? Under what conditions should the FOMC commit to a future policy rate path? And can we clearly distinguish between the monetary policy usage of a policy tool and other uses of the same tool?
Communications. There is always room for improving the FOMC’s communication strategy. Currently, there are two major unresolved issues: how best to use forward guidance and how to enhance the primary tool for signaling the FOMC’s reaction function, namely, the Summary of Economic Projections (SEP).
Consistent with the opening citation, simply announcing an inflation target does not make it credible. The entire policy framework—including the strategy, the tools for implementation, and the communication arrangements—also must be credible.
In the remainder of this post, we briefly characterize the functioning of the 2020 strategy and then turn to the first issue on our list: the asymmetry and bias in the current policy strategy.
Performance of the 2020 strategy. Since this post is the initial one in the series, we begin with some general considerations regarding the performance of the 2020 strategy. The chart below suggests that the strategy is inconsistent with maintaining an average inflation target of 2 percent over the long run. The figure depicts the effective federal funds rate (black dashed line) and the deviation of the effective rate from a simple Taylor rule with a 2 percent inflation target (π*) and a 2 percent equilibrium real interest rate (r*). Importantly, that deviation – the amount by which the policy is too accommodative relative to the rule – was largest, nearly 10 percentage points, in the first quarter of 2022. This “excessive accommodation” surpassed even that of the mid-1970s, when actual inflation was double-digit and the FOMC was not yet aiming at a 2-percent target.
Federal funds rate and the gap between a Taylor Rule-implied rate and the federal funds rate (quarterly), 1960-2Q 2024
Notes: The dashed line is a quarterly average of the daily effective federal funds rate. The shaded area is the excess of the effective federal funds rate (rff ) over the Taylor-rule-implied funds rate: rff – [2 + π + 0.5(π-2) + 0.5(u-u*)], where π is the percent change from a year ago of the personal consumption expenditure (PCE) price index excluding food and energy (JCXFE), u is the unemployment rate (UNRATE), and u* is the noncyclical rate of unemployment (NROU). All data are quarterly from FRED.
To understand the consequences of the 2020-21 deviation, note that periods when the policy rate is far below the one implied by the Taylor Rule – accommodative episodes when the shaded area is well below the zero line – are generally associated with increases in inflation. Conversely, when policy is notably contractionary, so the current rate is well above that implied by the Rule and the gap is positive, inflation typically declines.
These observations are consistent with the notion that there is a useful relationship between inflation outcomes and the stance of monetary policy. So, it should come as no surprise that the extremely accommodative policy in the two years following the March 2020 COVID shock supported rapid inflation: cumulative inflation since 2021 has exceeded the long-term trend by a whopping 9 percentage points. The following chart highlights this acceleration of prices. Starting in 1995, it plots the cumulative increase of prices (the red line) as well as the price level implied by a 2-percent target (the dashed black line).
Cumulative inflation since 1995: Actual personal consumption expenditures (PCE) price Index vs.
2-percent trend (quarterly), 1995-2Q 2024
Notes: The chart plots the natural log of the personal consumption expenditures price Index (PCEPI), with January 1995=0 (solid red) along with a 2-percent inflation trend (dashed black) . Data are from FRED.
The chart reveals two very different inflation episodes. First, starting in 2013, inflation fell persistently below target, helping to motivate the 2020 strategy shift. In contrast, inflation since 2021 has been consistently above target. Even as of this writing, trend PCE inflation—measured by its six-month annualized trimmed mean—is still running at a 3 percent pace (see here). As a result, the red line continues to rise further above the dashed black line. Overall, we do not see how this pattern is consistent with a credible strategy of long-run price stability.
Key components of the 2020 strategy. Turning to details of the 2020 strategy, two highly accommodative elements are contributing to the ongoing inflation overshoot. The first is the shift from preemption to patience. That is, the 2020 strategy calls for the FOMC to wait until inflation rises before tightening policy rather than to act in anticipation of rising inflation. At the same time, the FOMC need not wait until inflation falls to its target to ease its policy stance. The second is the introduction of an asymmetric posture in which the FOMC appears far more committed to make up for undershoots of its inflation target, than it is to make up for overshoots. Policymakers refer to this 2020 framework as a “flexible average inflation-targeting” (FAIT) strategy. However, an asymmetric posture is unlikely to be consistent with maintaining average inflation at the target over an extended period. (In addition to fostering inflation, Kiley (2024) argues that an asymmetric strategy also can add to economic volatility.)
From preemption to patience. The shift from preemption to patience reflected the FOMC’s loss of confidence in their ability to predict inflation. Given the obvious failures of inflation models over many years, it is surely appropriate to be humble. Policy cannot be more precise than the very imprecise models on which it is based. At the same time, as we wrote in 2020, “the Committee still needs a model, if for no other reason to guard against the danger of significantly overshooting their long-run average objective. The inherently backward-looking nature of the patient shortfall rule raises this risk.”
Experience since 2021 validates the concern that the “patient” approach is vulnerable to price surges that raise the price level permanently above its multi-decade trend of 2 percent. To be sure, the supply shocks emanating from the COVID pandemic played a key role in driving inflation higher, but the Fed’s willingness to spur aggregate demand well beyond the economy’s capacity also played a role. Amid an unprecedented peacetime fiscal expansion, the FOMC kept policy rates near zero for two years while implementing a record expansion of its balance sheet. By the spring of 2022, when the FOMC finally began to raise interest rates, trend inflation exceeded 7 percent. Conversely, the FOMC is about to ease policy even though trend inflation remains well above its target (see Chairman Powell’s 2024 Jackson Hole speech). That is, the Committee seems more confident about predicting inflation declines than increases.
FAIT’s asymmetry. Turning to the asymmetry of the FOMC’s flexible average inflation targeting strategy, in our view the Committee’s posture since 2020 is inconsistent with maintaining average inflation over an extended period. To see why, we need first to explain how a true averaging regime works, which we call TRUE, and then distinguish TRUE from FAIT.
Unlike a very simple inflation targeting framework that ignores past policy misses, a TRUE regime aims to correct for past misses, making it more like price-level (or price-path) targeting. Under TRUE, the actual short-run target for inflation varies by the amount required to bring average inflation back to the long-run target over a specified period. Fashioning a TRUE regime requires dividing the averaging period into two parts, a historical look-back period and a target restoration window. If inflation averaged above the target during the look-back period, then the short-run inflation target during the restoration window would be below the target, and vice versa.
Such a TRUE regime could require policy actions that no central bank today seems willing to accept. Consider, for example, applying a TRUE regime to recent U.S. experience with a 10-year averaging period, divided into a 5-year look back and a 5-year restoration window. From June 2019 to June 2024, PCE inflation averaged 3.5 percent. To restore average inflation to 2 percent over the 10 years ending July 2029, actual inflation during the restoration window would have to average roughly 0.5 percent per year! If we were to shorten the full averaging period to 5 years, again divided evenly between the look-back and the restoration windows, then getting back to the 2 percent target would require that inflation average barely ¼ percent per year through the end of 2026!
Clearly, the FOMC is not going to target one-half percent inflation on average for the next five years, much less near-zero inflation for the next 2½ years. In practice, when inflation overshoots as it has, policymakers let bygones be bygones, targeting 2 percent inflation going forward. But this means that price surges become permanent additions to the price level, imparting an upward bias to average inflation over time. That is, the Committee does not appear willing to make up for overshoots, while their 2020 strategy appears purposely aimed at countering undershoots.
Put differently, part of the current strategy is to encourage expectations that policymakers will make up for price undershoots when policy rates are near the effective lower bound. These expectations lower the real interest rate at every level of the nominal policy rate, thereby adding further policy stimulus. The converse – that policymakers will make up for price overshoots by targeting sub-2 percent inflation temporarily – does not hold.
To summarize, regardless of how flexible the “F” in FAIT is, it is not average inflation targeting. While TRUE appears too restrictive for any central bank today, the FOMC’s FAIT is unlikely to achieve long-run price stability in a world that faces repeated adverse supply shocks. In that world, observers will rightly expect that large, sustained, upward deviations of inflation from the target will result in permanent increases in the price level. In our view, that renders FAIT unfit for purpose.
How to fix the strategy. Going forward, the Committee should aim for a strategy that is less backward looking and more symmetrical. Admittedly, the FOMC needs to be humble about its ability to predict inflation. Yet, preemption requires prediction, at least when there are large inflation risks. For this limited purpose, even a flawed model beats no model at all. To illustrate the point, consider the behavior of money and the labor market during the first half of 2021, when policy rates were at zero and the Taylor rule deviations in our first chart were peaking. M2 growth exceeded 20 percent – a record – through early 2021, while the job vacancy rate (v) surged from 4.5 percent at the end of 2020 to a phenomenal 6.6 percent in mid-2021, surpassing the unemployment rate (u) of 5.9 percent (see Benigno and Eggertson on the key role of this v/u ratio in U.S. inflation surges). Even in an unsophisticated model, these indicators of rapidly expanding aggregate demand anticipated inflation risks inconsistent with the FOMC maintaining a near-zero policy rate.
Reforming FAIT requires that the FOMC be more explicit about what averaging means and to sharply reduce the upward bias built into the current framework. For example, in a TRUE regime, the Committee would specify the time frame, both in terms of the historical look-back period and the recovery window. Yet, as recent experience shows, with a long-run inflation target of 2 percent, such a TRUE regime could lead to periods when the FOMC would need to target an unacceptably low inflation rate during the restoration window.
We see two approaches to limiting this “extremely-low-inflation target” risk, while maintaining an average inflation target of 2 percent. One is to aim at keeping inflation (somewhat) below the long-run target most of the time, leaving room for upward surges when large, adverse supply shocks inevitably occur. However, since such a strategy means missing the target most of the time, it could pose a serious risk to the Committee’s credibility.
Alternatively, the FOMC could specify a framework that limits, but does not eliminate, the deviation from the TRUE regime. For example, the Committee could set a floor for the recovery window inflation target – say, 1 percent. While a recovery-window floor may still allow for some upward drift in the price level over the long run, this upward bias would be considerably smaller than under the current strategy, which appears to abandon any effort to correct inflation overshoots by setting a 2 percent target whenever overshoots occur.
A very different approach to reducing the asymmetry under FAIT would be to raise the long-run inflation target itself. A higher inflation target reduces the risk of hitting the effective lower bound in the face of an adverse demand shock. We will address this approach in the next post in this series. To foreshadow our conclusion, we doubt that a moderate increase in the inflation target – say to 3 percent – would materially reduce the asymmetry problem, while an even larger increase in the target almost surely would be highly unpopular and face considerable political resistance.
To conclude this first part of our series on the FOMC’s upcoming review, we urge the Committee to reexamine two critical parts of their strategy: the shift from preemption to patience and the operation of their FAIT framework. In our view, the first means that policymakers will almost always be late in addressing inflation, while the second imparts a long-run upward bias to inflation. Together, this means that inflation will average above target, putting the credibility of the policy at risk.