When it comes to central bank communication, we have a simple guiding principle: the financial system and the real economy should respond to data, not to policymakers.
Monetary policymakers cannot reduce uncertainty associated with things like evolving technology, natural disasters, wars, changing preferences, fiscal surprises and the like. However, they can avoid adding to these fundamental sources of uncertainty by communicating clearly in advance how they will react to unexpected events. Put another way, central bankers can best achieve their inflation and economic objectives if policy is largely systematic and everyone understands their reaction function (see, for example, Cieślak, McMahon and Pang).
Ideally, this means that policymakers’ public statements will have an impact on financial markets and conditions only when they contain information about either the reaction function itself or about policymakers’ updated assessment of economic conditions. It also means that the reaction function should be stable most of the time, so that observers do not need to wait for a central bank to act to understand how monetary policy will respond to shocks. Clearly, the FOMC’s delay in hiking rates in 2020-21 even as inflation surged—the “patience” that we highlighted in the first part of this series on strategy—added unnecessarily to uncertainty about policymakers’ commitment to price stability (see, for example, Bauer, Pflueger and Sunderam). Put simply, policy seemed far from systematic.
With this in mind, we urge that those organizing the strategy review reflect on how the FOMC can best improve the predictability of their policy actions.
Turning to the more general issues associated with central bank communication, we contributed a paper on the topic to the FOMC’s 2019-2020 strategy review. At the time, we recommended that the Committee simplify its public statements, while conveying any divergence of views; clarify how its policy will react to changing conditions; and highlight policy uncertainty and risks.
On the first, public statements and the divergence of views, we discussed how to promote group accountability. In our view, the Committee’s statements should focus more on fundamental policy drivers—like prices and employment—rather than on the prospective policy rate. FOMC participants should employ the first person plural when speaking about the consensus, while providing a full rationale for any dissents. If anything, given the magnitude of shocks in recent years—in a period when there were a multitude of reasons for both differing assessments regarding the state of the economy and proper policy responses—what stands out is the remarkable lack of dissent. Given the risks of “groupthink” in such circumstances, better communication would facilitate and normalize the kind of thoughtful dissent that reinforces the credibility of the FOMC as a policymaking body.
With respect to the second, we suggested explaining the reaction function in much greater detail. When growth, unemployment, inflation, and financial conditions deviate from what they expect, how will policymakers react systematically and predictably? As we wrote then, when people (not just financial market participants) understand the FOMC’s reaction function, policy actions will simply validate financial conditions that react to changing economic conditions. There would be no policy surprises.
Finally, we emphasized the need to clearly communicate uncertainty about how the economy is likely to evolve and the consequences for current and future policy. Among other things, we advocated use of scenario analysis where the Committee would communicate its views of the likely reaction to specific sets of unlikely, yet severe, events.
In this post, we discuss two related aspects of the FOMC’s communication policy: forward guidance and the quarterly Summary of Economic Projections (SEP).
Forward guidance: Evans, Fisher, Justiniano and Campbell usefully distinguish two types of forward guidance: “Delphic” and “Odyssean.” These terms are taken from Greek history and literature. The oracle at Delphi delivered predictions about the future (prophecies). These involved no commitment to act. In contrast, Odysseus, the hero of Homer’s epic poem, tied himself to the mast of his ship – an irreversible commitment – to resist the lure of the Sirens and avoid shipwreck and death.
In the context of monetary policy, Delphic guidance is merely a forecast of the likely evolution of the economy that takes account of the likely interest rate path. It involves no future policy rate commitment. In contrast, Odyssean guidance includes some commitment to a future interest rate or balance sheet path (perhaps based on time or conditioned on economic developments). Obviously, these serve quite different functions. In our view, the latter should be reserved for periods when there are no alternatives, while the former is a generally welcome means of providing information both about policymakers’ assessment of the economy and about their reaction function.
The following chart – created for the 2019-20 strategy review – highlights phrases from FOMC statements that were commonly interpreted as providing forward guidance. In the chart, the target range for the federal funds rate is in blue and the 10-year Treasury yield is in red.
FOMC forward guidance about the federal funds rate, 2007-19
Looking at these phrases, we can see that “commitment” is a matter of degree. When interest rates were at their effective lower bound – from December 2008 to December 2015 – phrases like “extended period” and “at least through” were designed to imply a time commitment for keeping interest rates low. A (very) casual look at the 10-year Treasury yield suggests that forward guidance may have contributed to a flattening of the yield curve starting in mid-2011, consistent with a credible commitment to “low for longer” policy rates.
There is now substantial theoretical and empirical work on forward guidance (see here). Simulations of theoretical models provide implausibly large effects – at least 10 times what more conventional forecasts suggest (see here). That said, empirical evidence provides support for the use of forward guidance involving some degree of commitment at the effective lower bound. For example, Swanson concludes that it is more powerful than large-scale asset purchases.
Nevertheless, in our view, forward guidance should almost always be Delphic, taking the form of a conditional forecast. Central bankers should limit their use of interest rate commitments to episodes when there are virtually no alternatives. In practice, when aggregate demand falls dramatically, alternatives such as fiscal stimulus and balance-sheet policies are usually available.
The key reason to limit policy rate commitments is to preserve credibility and to limit the costs of policy errors in periods when heightened uncertainty renders the future particularly obscure. For example, we saw in 2021-22 how policymakers—wishing to sustain the credibility of their commitments to keep interest rates low for longer—delayed hiking rates even when conditions rendered their earlier commitments wholly obsolete.
Based on the FOMC’s past behavior, many policymakers may disagree with our cautious attitude toward policy rate commitments. Even so, the FOMC’s strategy review is a key opportunity for the Committee to clarify its approach to forward guidance. When and how is it appropriate to use guidance, and in what form?
The SEP and the reaction function: This brings us to the Summary of Economic Projections (SEP). The FOMC is very careful to never refer to the SEP as forecasts. Instead, they describe these as “projections for GDP growth, the unemployment rate, inflation, and the appropriate policy interest rate.” Each FOMC participant provides the policy rate path that they believe would be consistent with their economic and inflation projection. In most periods—at least when policy is not stuck at the effective lower bound—the SEP is a form of Delphic guidance. It provides information about the Committee’s view of the likely evolution of the economy and appropriate policy over the coming two to three years.
However, because of the way the SEP reports participants’ projections, it is not possible to extract a reaction function from the information. With that in mind, we hope that the Committee considers two fundamental changes to the SEP. The first concerns the relationship among the four variables included in the SEP: inflation, real growth, unemployment, and the policy interest rate. Current practice is to publish in real time the median and the ranges for these, but not the matrix of responses that reveals each participant’s set of responses .Consequently, we cannot answer a simple question like: does a relatively high interest rate projection from an FOMC participant reflect some mix of a higher-than-average equilibrium real interest rate (r*) estimate, a high inflation projection, and a low unemployment forecast; or is that the FOMC participant favors a more aggressive reaction to a shared forecast of the fundamentals?
To address this clear shortcoming, in our 2019 paper we recommended that the FOMC publish immediately with each SEP the “matrix” that links the projections for growth, unemployment, inflation and interest rates for each FOMC participant. (Current practice is to publish these projections in matrix form only with a 5-year lag.) By clarifying where there are agreements and disagreements, the matrix would help observers understand the Committee’s collective reaction function, in part by facilitating inference about the nature and stability of the consensus. More information about the reaction function might also encourage an earlier and more systematic policy response to economic shocks, in contrast with the FOMC’s extraordinary delays of 2021-22 (see also Cieślak presentation at the June 2024 Hutchins conference).
Importantly, even a complete matrix would leave some key aspects of the FOMC reaction function opaque. To further improve transparency, we encourage the Committee to consider supplementing the current SEP with the publication of the distribution of participant responses to specific scenarios that deviate substantially from the current outlook for the economy and financial conditions. These scenarios would focus on, but not be limited to, prominent tail risks. Collectively, information on the likely reaction to such specific circumstances ought to enhance the SEP and FOMC deliberations and foster more systematic policy. Consistent with Bordo, Levin, and Levy, such scenario analysis would be particularly helpful in communicating the FOMC’s reaction function when uncertainty is unusually elevated, as it was at the start of the COVID pandemic.
Summary. This completes our four-part series on the FOMC’s upcoming strategy review (see also Part 1, Part 2, and Part 3). To recap, we suggest that the review focus on four major issues.
How to eliminate the inflationary bias arising from the current strategy that seeks to counter undershoots – but not overshoots – of inflation, combined with a shift away from preemption toward “patience.”
Whether to raise the inflation target from 2 percent.
Whether the federal funds rate remains the appropriate target interest rate and how to view the size of the balance sheet and large-scale asset purchases.
When and how to use forward guidance, and how to further improve public understanding of the Committee’s reaction function.
Our answers to these questions are:
In comparison with the 2020 framework, make clear that the new commitment is to a strategy of preemption in which overshoots and undershoots are treated far more equally.
Retain the 2 percent target, as the costs of a higher target are likely to far exceed the benefits.
Consider a shift to a framework that targets the U.S. Treasury repo rate, and that lets the size of the balance sheet be determined by banks’ demand for reserves.
Make clear that large-scale purchases are reserved for very special and rare circumstances.
Reserve the use of forward guidance as a policy rate commitment only in extreme conditions when alternatives for stimulating aggregate demand—such as balance-sheet tools and fiscal stimulus—are not available.
Expand the information provided with each Summary of Economic Projections to improve communication of the Committee’s reaction function.