Hoover Institution Conference: Getting Global Monetary Policy on Track

Central Bank Communications Beyond “How Many?”

Thank you for the invitation out to the conference. It’s a special delight for me to see John Cochrane, my old neighbor from across the street. John used to keep our emergency house key for us, and the alarm company would call him if things went wrong and we weren’t there. I’m hoping that if my talk goes sideways here, he’s still willing to be my emergency contact.

And I should add now, to the great relief of my colleagues, that these are my own views and not those of the Federal Reserve System or the Federal Open Market Committee (FOMC).

Today, I want to talk a bit about central bank communications, but I came to it in an odd way.

When I started at the Chicago Fed in January ’23, it felt like we had just lived through a really crazy time in our economic history and there were so many fundamental economic issues we needed to discuss and resolve.

But then at my first appearances with the press, the questions seemed to always center on the same thing: So, how many rate hikes do you see this year?

I was actually a bit dismayed. By itself, with no economic context or rationale or even what a person thinks will happen with economic conditions, the answer to “how many?” is mostly just speculation. There is not much economic content in a question like that.

Why not ask about what is actually informative about future decision-making—like reaction functions or economic worldviews or risk assessments?

In my puzzlement about why they seem so interested in context-free numbers, I have concluded that this may be a bit of our own doing in that some aspects of our current communications approach actually encourage this kind of speculation.

So today I would like to make three points: where I think Fed communications, especially in the Summary of Economic Projections (SEP), fails to deliver economic information that it could; how it might be improved; and the importance of participants’ public speaking and writings for communicating their reaction functions and risk-management considerations. And I will conclude with a bit of a defense of cacophony, the much-maligned bugaboo of FOMC communications.

Everyone knows that communications issues have always been important for central banks and are closely tied to inflation expectations, bank credibility, and even public trust. I’m not sure I’m fully on board with Ben Bernanke’s comment that monetary policy is 98 percent talk and 2 percent action, but we all know the importance of a clear understanding of a central bank’s goals and strategies in monetary policy transmission.

The Fed and many other central banks’ approaches to communications have changed dramatically in the past 30 years with the shift to much greater transparency.

One of the leading components of that has been the Summary of Economic Projections and its “dot plot” of FOMC participants’ interest rate projections. In it, each person gives an answer (anonymously) to the “how many?” question and does so for the next several years.

It is widely followed and provides a dose of transparency about a diverse collection of views. And its influence may be spreading. Recently there have been been high-level considerations of whether the European Central Bank (ECB) should adopt a dot plot.

As a starting thought, though, it is worth noting that in the 2020 Hutchins Center survey of academic and private sector Fed watchers, only about 50 percent of respondents reported that they find the dot plot useful.¹

One thing that might be pulling that number down, in my opinion, is that the SEP is not actually that useful for identifying FOMC participants’ monetary policy reaction functions. I’m defining the reaction function broadly here—how someone would react to changes in economic conditions—rather than in a technical way, like what their policy function coefficients are in a Taylor rule.

I mostly agree with the old Mervyn King (2000) argument that it’s a sign of success when markets react to new data releases more than to statements of central banker opinions.² One key to achieving that outcome is that people understand FOMC participants’ reaction functions.

The SEP isn’t that helpful in communicating them. Each participant’s rate path projection is made with a forecast of economic conditions, but the SEP does not report which goes with which. Even the median doesn’t give you a coherent observation because the person making the median rate projection is not likely to be the same person making the median inflation or gross domestic product (GDP) projection.

Without a connection of economic conditions to rate projections, the dot plot is just a collection of opinions without economic content. No wonder the press doesn’t ask about anything else. Our own major communications document implicitly tells them that the economic rationale isn’t needed. Just tell us “how many” and we’ll put them in a plot.

Because it can’t be connected to the economic conditions the participant thinks will justify that interest rate, there is nothing to tell us why they think this a reasonable choice.

We can do better. In a way, this is really a cousin to the old debate about time-dependent versus state-contingent forward guidance. State-contingent guidance is more useful because it conveys economic content and a reaction function.

Saying in 2011 that the Committee doesn’t expect to raise rates for two years is not as helpful as an Evans rule that it won’t consider moving until unemployment is below 6.5 percent as long as inflation is below 2.5 percent. The time-based guidance is basically just opinion. The Evans rule clarifies what the Committee is watching and conveys information about its reaction function.

Now think about the dot plot. It’s a lot like time-dependent guidance. It’s not tied to anything. We would be better off if it could be more like state-dependent guidance—more like a reaction function.

So let’s consider how the SEP could be made more useful in this way: First, I am going to echo some comments that Mickey and Charlie just made. We could go a long way by publishing the matrix. It doesn’t need to personally attribute each projection, but at least a matrix that anonymously matches the economic forecasts to the rate path for each participant would answer some important questions.

For example, suppose you observe a high dot for next year’s funds rate. You could tell in the matrix whether it was that they think the economy is overheating and they want to rein in inflation or they think faster, noninflationary growth implies a higher equilibrium rate. Today you can’t.

Over time, more would be revealed about reaction functions as observed changes in the economic environment show the relationships between participants’ views on the outlook and appropriate policy.

Another, more direct form of state-contingent information might be to add something like the stress test scenarios that are done for banks or the alt sims in the Tealbook. Of course this wouldn’t be as easy to implement as publishing the matrix, but it is food for thought on how to convey reaction functions.³

But the goal is to convey information about how participants might respond to developments. It’s also helpful when their speeches and writings explain their thinking. And this idea also highlights the important piece of communications—which is conveying the implications of unusual circumstances and risk management for setting policy. In times of heightened uncertainty or unusual risks, the policy playbook can be complicated and requires more than just dots. And such times seem to happen with uncomfortably high frequency.

The Committee has communicated unusual circumstances and practiced risk-management policy for a long time. This shows up in formal voted-upon FOMC statements, policy tilts, and other communications. Examples over the past 30 years include the 1990s policy responses to the Asian Financial Crisis and the Russian default; the recognition in 2003 of the then-novel risk that inflation in the U.S. could be too low; and the heightened uncertainty in 2007–08 as the financial crisis unfolded even while some inflation risk remained. And we are all familiar with examples surrounding the Great Recession and the pandemic.

In all of these cases, communications was an important tool to get us beyond context-free “how many?” discussions and to provide the public with better state-contingent information that underlies our policy decisions.

Conclusion

So my plea is that we try to push our communications and especially the SEP to better convey how we respond to economic conditions. This does mean a large number of voices. Alan Blinder (2007) and others have called this the cacophony problem and would probably view it as a negative of having this much information available to the public before every meeting.

I don’t agree, though. Yes, it would be easier to understand if there is just one voice talking about policy. But the FOMC is a committee with lots of views, and the policy gets forged in that environment. Without offending the Senate, in my view, in the twenty-first century the FOMC is the world’s greatest deliberative body. Having folks with different worldviews makes for better decision-making. There’s nothing wrong with the public knowing these views. It’s reality.

Today we are conducting policy during a fascinating and uncertain time. How we communicate will be critical. We should endeavor to convey our deliberations and our thinking, not just our opinion. It’s why I say that “how many?” is not enough.  

Notes

¹ Powell and Wessel (2021).

² As King (2000) stated, “A transparent monetary policy reaction function means that the news should be in developments of the economy not in the announcements of decisions by the central bank.”

³ People have been thinking about these kinds of ideas from the earliest days of the SEP releases. The FOMC discussed them as alternatives when it debated and ultimately rejected producing a consensus forecast back in 2012. Among external commentators, a prominent paper by Steve Cecchetti and Kim Schoenholtz written for the public Fed Listens portion of the 2019 policy framework review explicitly called for publishing the matrix (Cecchetti and Schoenholtz, 2019).

References

Blinder, Alan S., 2007, "Monetary policy by committee: Why and how?," European Journal of Political Economy, Vol. 23, No. 1, March, pp. 106–123. Crossref

Cecchetti, Stephen G., and Kermit L. Schoenholtz, 2019, “Improving U.S. monetary policy communications," paper at Fed Listens Event: Conference on Monetary Policy Strategy, Tools & Communication Practices, Federal Reserve Bank of Chicago, June 4, available online.

King, Mervyn, 2000, "Monetary policy: Theory in practice," speech by deputy governor, Bank of England, given to the joint luncheon of the American Economic Association and the American Finance Association, Boston, MA, January 7, available online.

Powell, Tyler, and David Wessel, 2021, "Survey of academic and market observers gives high marks to Fed communications," Brookings Institution, Hutchins Center on Fiscal and Monetary Policy, report, January 14, available online.

The views expressed today are my own and not necessarily those of the Federal Reserve System or the FOMC.