Since the mid-1990s, there has been a trend towards greater transparency in economic policymaking – particularly with respect to monetary policy – and a number of central banks, including Sweden’s Riksbank and Britain’s Bank of England, have adopted a very transparent monetary policy regime known as inflation targeting. The United States does not subscribe to inflation targeting, but the Fed has also become much more transparent about its policymaking and operations over the past 15 years.
Economists have argued that greater transparency is beneficial, improving democratic accountability by making it easier to judge whether a central bank is committed to its announced policy and improving policy effectiveness by facilitating the interpretation of policy changes (see the very accessible review of the discussion by Posen 2002).
But greater transparency of central bank policymaking – in which committee deliberations are made more open to the public – may prevent the full and frank discussion needed to make the best decisions. In a recent paper (Meade and Stasavage 2008), we compare discussions of the Fed’s Federal Open Market Committee (FOMC) before and after committee members knew that all statements would eventually be made public. Our empirical results indicate that after 1993, when FOMC participants knew that their deliberations would be made public, they were less likely to challenge then Fed chairman Alan Greenspan. This suggests that greater transparency hindered free deliberation and may have permitted Greenspan's views on interest rates to dominate US policymaking. In the discussion of the current crisis in credit markets, some have suggested that US interest rates were too low for too long.
Closed doors and open minds
Concern about the effects of open deliberations is not new. Speaking about the secrecy rule that prevailed during the US Constitutional Convention of 1787, James Madison emphasised that full publicity would have made members more reluctant to express their true opinions freely. Madison saw secrecy as having been critical to the Convention’s ultimate success.
Fed policymakers expressed similar concerns when, in 1993, the US Congress pressured them to become more open about their decision-making process. At issue was whether the Fed would agree to publish verbatim transcripts from meetings of its Federal Open Market Committee (FOMC). In Congressional testimony, Alan Greenspan argued against publication, saying that the FOMC “could not function effectively if participants had to be concerned that their half-thought-through, but nonetheless potentially valuable, notions would soon be made public” even with a publication lag of five years. Greenspan noted further that the character of the meetings would change with transcript publication, from lively, useful sessions to bland, sterile ones. In the end, the Fed made the change and subsequently decided to make all of its meeting transcripts available. Transcripts of all FOMC meetings and conference calls from 1978 through 2002 are currently available on the Fed’s web site.
In our paper, we look at whether more information provided by the central bank to the public about monetary policy deliberations can affect the deliberation process itself and ultimately stifle useful debate. We employ a theoretical model in which policymakers care both about making the right policy decision and about how they are viewed by the public. We show that it is possible that policymakers could hold back during deliberations for fear of looking uninformed or incompetent if they know that the content of their discussions will eventually be released to the public. If this happens, then monetary policy might be adversely affected. Thus, the benefits of increased transparency would need to be assessed against the resultant damage to the policy process.
Unique aspect of FOMC transcripts: the tapes that weren’t destroyed
Our research employs a unique aspect of the situation and the transcripts themselves to analyse whether the publication of meeting records has affected the Fed’s deliberations. FOMC meetings have been recorded for more than 30 years so that the Fed staff can write meeting minutes after every meeting. (An account of each FOMC meeting has been published in some form following each meeting since 1936, but these published accounts have been short, non-attributed records of meeting discussion.) Policymakers knew about the recording but thought the tapes were destroyed after the minutes were written. Thus, transcripts exist from a time when policymakers did not know that their deliberations would be made public. We compare deliberations before 1993, when Fed officials believed their remarks were private, with deliberations after 1993, when officials knew that all statements would eventually be made public.
In our empirical analysis, we use a dataset collected from the transcripts themselves that codes verbal messages of each meeting participant and characteristics of the participants, including their name, Fed district, years of experience, and whether they are an official voter at the meeting and, if so, whether the official vote cast agreed with the verbal message sent during discussion. (In the Fed system, votes rotate in a fixed fashion for some policymakers, so that only 12 of the 19 officials vote at any given meeting).
Over the time period that we examine (1989-1997), Chairman Greenspan presented his proposal for the setting of the policy interest rate first and then solicited other meeting participants for their views. After all the participants had expressed their opinions, an official vote was taken on the policy proposal. We focus our analysis on the willingness of the meeting participants to express verbal disagreement with Greenspan’s proposed policy before and after 1993.
The empirical results provide clear evidence of a change in the character of FOMC deliberations – policymakers were less likely to express verbal disagreement with Greenspan’s proposal after 1993. This remains the case even after other potential influences on officials’ views, such as a variety of measures of the current economic environment as well as Fed forecasts for inflation, are taken into account.
We tested the robustness of these empirical results using supplementary hypotheses. First, while the publication of transcripts may have affected FOMC deliberations, it should have had no impact on the votes of policymakers because the votes were published both before and after 1993. We tested the votes in our empirical model and found that votes were unaffected by the release of the transcripts. Moreover, a policymaker should have been less likely to switch his view on the policy proposal between the discussion and the vote – for example, verbally disagreeing with the proposal but voting in favour of it – once the transcripts became public. Our empirical results also accord with this.
Our empirical findings are supported by a number of parallel observations about the changing character of FOMC debate since 1993. While before 1993, FOMC discussions were characterised by frequent “off the cuff” remarks and interruptions, since 1993 there has been an increase in prepared statements that may result in less real deliberation. Our results have significant implications for the design of monetary policy institutions, as well as for the operation of committee-based government decision-making more generally.
Meade, Ellen E. and David Stasavage (2008). “The Dangers of Increased Transparency in Monetary Policymaking,” Economic Journal April, 695-717.
Posen (2002). “Six Practical Views of Central Bank Transparency”, http://www.iie.com/publications/papers/posen0502.pdf