Thank you, Governor Ferguson. I’ll be referring to the material called “The Committee’s Communication Strategy” that was handed out to you before the start of the meeting.
Next week marks the tenth anniversary of the Committee’s first public explication of a change in its intended federal funds rate not associated with a change in the discount rate. But increased openness has not been a U.S. phenomenon alone. Across the industrial world and in many emerging-market economies, monetary policy makers now provide real-time information on their forecasts, the risks to those forecasts, and their goals and objectives, often accompanied by narratives explaining their strategies. These come in the form of statements, policy (often called inflation) reports, and regular press conferences.
The Committee has been riding that wave of increasing communications. In some ways, it has been ahead of other central banks. For example, no other policymaking body immediately releases the vote tally at its meetings. However, in other ways, the Committee has not been at the leading edge. On page 2 of your handout, I provide a reason why you may have felt it important to be measured in changing your information policy: Your words matter. The bars plot the market response—as measured by the change in the two-year Treasury yield in narrow windows surrounding the events—around the release of the Committee’s statement (the top left panel), its minutes (the bottom left panel), and the Chairman’s semiannual testimony (the top right panel) since 1997. (The reason I describe the top left panel as exclusively the response to the statement is that I have used federal funds futures rates to strip out the portion of any market surprise about the level of the funds rate.) The responses both to statements and to testimonies have been sizable. Except for the most recent release, though, minutes have tended to be much less of a market event—presumably because of their publication lag. As shown in the table at the bottom right, your statements and the Chairman’s testimonies have had effects on market prices that, on average, have been as large as those associated with releases on the employment situation and larger than those associated with the publication of data on the ISM index or consumer prices.
Another point to make about your ten-year experience with statements: At least as can be judged from the transcript of your meeting on February 4, 1994, Committee members may not have fully appreciated that they were setting a precedent with a public announcement. In the event, the response to that short paragraph was so positive, you could not go back to signaling policy changes through open market operations. That might be a lesson to remember as I outline various potential changes to your communications strategy in the rest of my briefing.
Page 3 gives a roadmap of the rest of my remarks. The options confronting the Committee are interrelated. To facilitate your discussion, my briefing will have five parts. I intend to lay out the range of plausible options before the Committee at the outset and then speak about them in more detail. Those include alternative formulas for the risk assessment of your statement—that is, I will report on the progress of the working group chaired by Governor Ferguson. Then I will discuss the pros and cons of expedited release of the minutes, which were given more completely in the memo by Brian Madigan that I circulated, and of an enhanced role for the FOMC projections, which were examined in a memo by Benson Durham. I will end about where I start by repeating the “Chinese menu” of possibilities so that you can begin the hard work of determining how the pieces might fit together.
Your first look at that Chinese menu is provided on page 4, which gives a broad overview of options across three main communication issues on today’s agenda. The left column lists the possibilities for the paragraph in the Committee’s statement that gives an assessment of risks to the outlook expressing a probabilistic assessment of the uncertainty surrounding the dual goals of inflation and output. To some, this serves to hint at the direction of future policy without offering a commitment. This being a central bank, it is always an option to maintain the status quo in which the drafters start from a formulaic assessment of the risks to inflation and output but have some discretion in the exact wording—with the expectation that the result will preserve the three-part assessment of risks used since May. This approach allows some consultation with Committee members through the Bluebook.
The option dubbed “gradual evolution” involves expanding both aspects of the process by allowing more discretion to the drafters to alter the words according to evolving economic conditions and to provide more discussion of the words of the statement in the Bluebook. However, there would be an explicit expectation of including a forward-looking aspect to the statement and an assessment of risks, except on those rare occasions, as last March, when no such assessment could reasonably be made. If you would like, we could structure the forward-looking portion of the Bluebook around two or three draft announcements keyed to conveying to the public possible changes in either the outlook or policy. Those alternatives would provide the basis for your policy discussion and vote, and one of them—with whatever last-minute changes you wanted—would be released to the public. If events between the publication date of the Bluebook and the Committee meeting changed the situation sufficiently, we would circulate a Bluebook supplement that suggested more appropriate wording.
Another set of possibilities is to adopt new formulaic language, as you had with the “balance of risks” from 2000 to early last year. In just a bit I will report on the three formulas that the working group forwarded to the full Committee. Lastly, you could choose to discontinue the assessment of risks portion of the statement, accepting that substantial criticism would ensue for backsliding away from openness. In principle, that need not be backsliding if the Committee decided to alter some of the other margins of its communication policy. Of course, such other changes also may be viewed as desirable in conjunction with retention of the risk assessment. One possibility would be to expedite the release of the minutes, the subject of the middle column. About two weeks is the barest minimum required for the Secretariat to draft, circulate for comments, and release finished minutes, but you may see some benefit in providing a bit wider cushion by releasing minutes, say, three or four weeks after each meeting. Another possibility is to enhance the role of the economic projections of the Governors and Bank Presidents as outlined in the right column. You might want to increase the frequency with which the survey is conducted, lengthen the projection period, or increase the number of variables projected. You might want to continue to release those projections in the Monetary Policy Report, but they could also be part of expedited minutes or the subject of a separate release.
As you could probably tell from the thirty-page memo that circulated last week, the working group and its staff devoted considerable attention to alternative formulaic language. Page 5 reviews some of the key lessons learned. In particular, members settled on six important design principles that any formula should satisfy: (1) Good governance suggests that the Committee should vote on the exact wording of the risk assessment. (2) The statement should not hamper the policy discussion, nor should the Committee feel constrained in its choice of action by a formula. (3) The statement should be flexible enough to encompass the Committee members’ views about the operation of the economy and the concepts that can be usefully measured. (4) The statement should be clear to the public, or at least as much as it is possible for central bankers to endeavor to be. (5) The statement should cover the range of feasible contingencies so that the formula would last, if not for the ages, at least a few years. (6) The statement should avoid the use of potentially charged terms, such as “risk,” when referring to outcomes that may be seen as positive. For example, the statement should not convey the mis-impression that the Committee is against strong economic growth for its own sake.
While the working group settled on these six principles fairly quickly, three questions remained unresolved, which are provided in the column at the right. In particular:
• Should the wording of the risk assessment be in terms of the levels of output and inflation or their changes? Conveying a sense of levels may make it easier to describe economic performance relative to the benchmarks of potential output and the Committee’s inflation goal. However, such specificity may be seen as a drawback because there is a range of opinion among members about appropriate benchmarks.
• What conditioning assumption for monetary policy should be employed? The forecast could be based on the assumption of unchanged interest rates, as was the case with the old balance of risks language, but that presents problems for projections far into the future should that assumption diverge significantly from what is built into financial market prices. The alternative would be to assume a historically normal policy path; but if the Committee does not reveal that assumption, the public may have difficulty understanding the message of the risk assessment.
• Over what period should the outlook and risks be considered? A shorter period—that is, the next several quarters—has the advantage of limiting the factors that influence the outlook. However, such a horizon may not be sufficient if you think the lags in the transmission of policy are substantial, suggesting that you should look forward to the foreseeable future.
Confronted with different potential reasonable answers to these questions, the working group narrowed the field of potential formulas to three: two stated in terms of levels but differing as to the policy assumption and period, and a third stated in terms of changes. I should note that the three alternatives do not explicitly mention the conditioning policy assumption. In principle, that design feature could be conveyed to the public when the new policy is announced and mentioned periodically in the minutes as a reminder. The “levels A” alternative is provided on page 6, with the left column repeating what was in the memo you received on January 20. As noted in red in the text and highlighted at the right, this alternative is stated in terms of levels of inflation and economic activity and makes explicit reference to benchmarks for inflation and output. While the paragraph is silent about the conditioning assumption for policy, it is based on an assumption of “normal” policy. In addition, as shown in blue, the horizon is “several quarters,” but outcomes are described relative to paths that extend further into the future. The “levels B” alternative provided on page 7 shares a focus on levels and explicit benchmarks (as noted in red at the right) but differs in two important dimensions. First, while it is also silent about the policy assumption, the drafters assumed that the implicit forecast is based on an unchanged policy stance. Second, as shown in blue, the horizon is “the foreseeable future,” that elastic concept borrowed from the old balance of risks language. The “changes” alternative given in the left column of page 8 focuses, as indicated in red at the right, on changes in output relative to potential and changes in inflation. Presumably, if the Committee saw fit, the prior paragraph of the statement that describes economic conditions could refer to levels and benchmarks. As in the “levels B” alternative, this formula is based on the undisclosed assumption of unchanged policy and applies to the foreseeable future.
Staff subjected these three alternatives to a variety of stress tests to gauge their performance both in scenarios that seem likely in the near future and that have challenged the Committee in the past. I am not revealing any secrets of the working group by indicating on page 9 that problems emerged with all three formulas. In particular, measurement issues arise when words with vague or multiple meanings are used, such as the “long-run trend of potential” or “long-run sustainable pace” or even an unquantified notion of “price stability.” In addition, you may feel uncomfortable about the clarity of any statement that relies on words such as “path consistent with,” which may admit many possible trajectories, or “the foreseeable future,” which, for some, appears to have meant a period as short as the intermeeting period and to others a period measured in years. Lastly, the conditioning assumption used in the implicit forecast introduces a fundamental question as to the purpose of the risk assessment. Do you want to send hints about the future direction of rates (which seems easier to do with the assumption of unchanged policy), or are you conveying qualitative information about the fan chart surrounding your forecast (as seems the case with the assumption of “normal policy” or the related assumption of “appropriate policy”)? The working group decided that its mandate was to present information to allow the Committee to determine an acceptable tradeoff among the various pitfalls, recognizing that abandoning an attempt to arrive at a formula is also an option.
No doubt, this decision also interacts with any you might make about expediting the release of the minutes, the subject of page 10. Central banks that do release minutes before their next scheduled meeting, such as the Monetary Policy Committee of the Bank of England, stress advantages such as a desire to provide more timely information to the public consistent with their responsibility to be transparent and accountable. In addition, minutes provide a platform to present the outlook and policy strategy in a more nuanced fashion than is possible in an immediately released statement. And relevant to my earlier discussion, expedited minutes may make you more comfortable in trimming the statement.
The disadvantages mostly seem related to possible effects on Committee dynamics and other aspects of your communication policies over time. You might be concerned that, as the minutes increase in market importance because they are released in real time, mention in the minutes would be viewed as a chip in the bargaining process in arriving at a policy decision. The quality of the minutes might suffer over time if they were “scrubbed” to mitigate potential market reaction. In particular, you might be hesitant about including conditional statements in the minutes, such as a desire to change policy at the next meeting if intervening data run in a certain direction or to hold an intermeeting conference call. The bottom line is that experience has shown that market participants tend to overreact to contingent statements, and fear of possible inappropriate market response might be viewed as a decided negative. In addition, the Committee, and future Committees, would have to exercise some self-discipline so as not to succumb to the temptation of using the minutes to send signals in response to developments after the meeting.
Some thought would also have to be given as to the timing of the release of the minutes, particularly on those occasions twice a year when the Chairman delivers the Monetary Policy Report to the Congress. Indeed, as a general principle, the prospect of creating an additional news event in which market participants and reporters await some fixed release from the FOMC may seem especially daunting. And in that regard, the Committee may want to review its policy regarding blackout periods so that the release of the minutes marks the first news about the thinking of policymakers at the previous meeting.
Twice a year, the minutes include the central tendencies and the ranges of the Governors’ and Bank Presidents’ forecasts of key macroeconomic variables. Early release of the minutes naturally raises the question of the role of those projections, as is discussed on page 11. There are several margins over which the Committee could enhance the role of those forecasts. In particular, you might want to (1) increase the frequency of the survey to every meeting or every quarter; (2) increase the length of the projection period, thereby giving the public some sense of the Committee’s views of the longer-term prospects for the economy; (3) increase the number of variables in the projections, perhaps to include core PCE inflation, which seems central to the deliberations of the Committee. Indeed, you might view the projections as a means of conveying to the public your view of the longer-term elements essential for policy, such as the growth rate of potential output and your individual working definitions of price stability, and perhaps the degree of uncertainty about that outlook, by including the width of uncertainty bands about each variable forecasted. The latter might be particularly attractive, as wide bands would reinforce that monetary policy is as much about controlling potential risks when a forecast, almost inevitably, does not come true as it is about making projections. At a practical level, the Committee will also have to decide whether it is appropriate to separate its projections from the Monetary Policy Report and testimony.
The pros of an enhanced role of the projections given at the top right should seem familiar. By getting them out earlier, you would be providing more-timely information to the public, both about your central tendencies and range of opinion, thereby increasing transparency and accountability. Some might find a particular attraction in substituting a survey of numbers that is mechanical for the risk- assessment portion of the statement, which requires consensus. The cons relate mostly to the Committee’s process and market participants’ interpretations. You will have several decisions as to whether the forecasts should be included in the minutes of the meetings, can be revised after meetings, or their release augmented with explanatory text. The public might not understand the conditional nature of the projections and ascribe to them more weight than they actually receive in the policy process, thereby creating another news event in intermeeting periods and impairing your reputation as forecast errors cumulate.
With the last page in your package, I deliver on my promise to end where I started. Page 12 repeats the Chinese menu and underscores that I am seeking direction from the Committee on all aspects of its communications policy. Many of the options that are listed will require much more staff work to arrive at a solution acceptable to the Committee. That is, the discussion to follow will not be the end of work on communications policy, but I am hopeful that it is the end of the beginning of work on communications policy.