آیا تنظیمات یا ارزیابی های خصوصی در کمیته سیاست پولی است؟

Using Bank of England voting data, we show empirically that members’ votes are driven by heterogeneous individual assessments of the economy as well as their individual policy preferences. Estimates indicate that internal committee members form more precise assessments than externals and are also more hawkish. The estimates allow the first quantification of the gain due to information aggregation on monetary policy committees. The marginal gain from additional committee members tapers quickly after five members. There is no evidence of gains through externals’ moderating internals’ preferences. A relatively small committee of highly informed internal members emerges as a desirable committee structure.

What role do private assessments of economic conditions play in explaining individual voting behavior on the Bank of England's Monetary Policy Committee (MPC)? While empirical work focuses largely on preference differences (hawkishness or dovishness) as the primary driver of individual voting differences, both committee designers and theoretical models emphasize that policymakers may differ in their assessments of economic conditions. Even if they share the same preferences, different beliefs on such conditions can lead to different votes. Pooling of private assessments also provides a natural channel for understanding why committees outperform individuals, as found in experimental work (Blinder and Morgan, 2005 and Lombardelli et al., 2005). Our empirical approach allows both differences in preferences and in private assessments to play a role in committee decisions. The first major contribution of this paper is to estimate a model of equilibrium voting behavior using the individual voting record of the MPC. In the model, the individual structural parameters of interest are preferences and the precision of private assessments, which we call expertise. The estimation follows the two-step approach of Iaryczower and Shum (2012), who estimate these quantities for US Supreme Court justices. The model is estimated both under the assumption that voting is sincere, in which case members behave as if their votes determine policy, and strategic, in which case members condition their votes on being pivotal. These assumptions do not change estimates of expertise, but do change those of preferences. A model evaluation exercise based on out-of-sample prediction shows that sincere voting explains the data modestly better than strategic voting. 1 The expertise estimates show that private assessments play an important role in individual voting behavior. In meetings with the greatest uncertainty about the state of the economy, an individual member is up to 40 percentage points more likely to make the correct decision relative to a model in which private views play no role. Modelling private views is therefore important for explaining voting behavior, and differences in these views can be expected to lead to different votes even when members share the same preferences. The estimates also allow one to explore differences between externally appointed experts—members appointed solely to make the decision each month—and internal members—those who also have executive responsibilities in the central bank. Some countries (for example, the UK, Poland, and Hungary) make use of externals, presumably to take advantage of diversity, while others (for example, the USA and Sweden) do not. In line with the existing literature (e.g. Gerlach-Kristen, 2003, and references below), internals' estimated preferences are significantly more hawkish, but our novel finding is that they have higher estimated expertise. This finding implies that the justification for externals’ inclusion from a purely informational perspective requires further thinking. To try to uncover the source of these differences, we examine the behavior of members who have worked in central banking prior to their MPC appointment—insiders—in order to see whether prior experience or current position matters most for the behavior of internals. Estimates show that insiders are more hawkish and have more expertise than outsiders, which suggests that it is prior experience as central bankers rather than holding a senior position in the central bank that drives the internal-external differences.2 Optimal committee design is an important and open area in monetary policy (Reis, 2013). The second major contribution in this paper is to use the estimated parameters to assess how different committee structures affect the quality of decision making via counterfactual simulations. The first aspect of committee design considered is the effect of size. A well-known hypothesis is that committees outperform individuals because they aggregate dispersed private information (see Gerlach-Kristen, 2006 or Blinder, 2007 for recent discussions of this in the context of monetary policy). A major advantage of the estimation approach is that one can measure the extent of these gains in a real-world committee. In periods in which the economic environment is most uncertain, moving from an individual expert to a committee structure improves decision making substantially, although the gains decline rather quickly. For example, a committee of five internal members is between 7 and 11 percentage points more likely to make the correct decision than an individual internal member. This indicates that the gain from committees’ pooling imperfect knowledge is potentially significant. However, the marginal benefit of an additional member tapers quite quickly after about five members; the marginal gain from adding four additional internal members, making a committee of nine internal members, is only about 1–3 percentage points. The overall message is that the additional members on large committees, such as the 24-member Governing Council of the ECB, may improve decision making only marginally if expertise is sufficiently high. Given the potentially large costs from adding more and more members, such as information exchange problems and free-riding by members (see Sibert, 2006, for a discussion), a smaller committee is likely to be better. The second design issue explored is whether, as argued by Blinder (2007), external members can add value through moderating internals’ hawkish preferences. This would support their inclusion in committees despite their lower expertise. When the economic data broadly points to the high interest rate likely being correct, the preference moderation effect indeed arises because internals are less willing to follow private signals that low rates should be chosen. But, the moderation effect is not found to be large enough to fully overturn the effect of lower expertise, and internal and external members perform equally well. Moreover, when economic data suggests that the low interest rate is likely correct, internals’ hawkish preferences work to offset externals’ rigid decision making, amplifying internals outperformance of externals. Overall, the gains and the losses from preference moderation should roughly cancel out on average. In short, a relatively small, homogenous committee of members with high expertise performs very well even if they have a hawkish bias. Of course, further work remains to be done on the exact nature of deliberation in the committee meeting, and alternative sources of value that external members might bring. Still though, the paper is the first to separately identify the role of preferences and individual assessments in monetary policymaking, and so provides potentially valuable facts for extending our knowledge of both the how and why of policymaking by committee. The paper is related to two main strands of the existing literature. First, there is a growing literature on all aspects of the use of committees to make monetary policy decisions: this includes important summaries of the state of the knowledge (Gerling et al., 2005 and Blinder, 2007), issues of agenda-setting (Riboni and Ruge-Murcia, 2010), reputation-building on monetary policy committees (Sibert, 2003 and Hansen and McMahon, 2013), credibility of committees (Mihov and Sibert, 2006), and the desirability of a committee over an individual when there is uncertainty about the economic situation (Gerlach-Kristen, 2006). Committee decision making has also been extensively studied by social psychologists and Sibert (2006) provides a discussion of the main findings and how it applies to monetary policy committees. Secondly, there are a number of other papers that specifically study the MPC and internal-external differences within it including Gerlach-Kristen (2003), Bhattacharjee and Holly (2005), Spencer (2006), Besley et al. (2008), Harris et al. (2011) and Hix et al. (2010). As mentioned above, these all focus exclusively on preference differences, with the general conclusion that external members are more dovish the internals. The rest of this paper is organized as follows. Section 2 provides background information on the MPC. Then, Section 3 presents the voting model that serves as the basis of the empirical analysis. Section 4 details the estimation strategy and Section 5 describes the data used to implement the strategy. The estimation results are presented in Section 6. Section 7 presents counterfactual results on the committee structure and Section 8 concludes.

8. Conclusion Taken together, our results give an empirically novel view of monetary policymaking by committee and address important issues in committee design. Reis (2013) argues that committees are preferred to individual decision makers for any of a four main virtues: (i) pooling of private information, (ii) providing a diversity of views which generates discussion of the evidence for and against different views, (iii) guarding against autocratic power, and (iv) making less volatile decisions. Our paper directly addresses the first of these, and shows that groups significantly outperform individuals due to information aggregation. Another insight is that adding dovish members (externals) to a group of hawkish ones (internals) does not necessarily improve average decision making quality through preference moderation. The results and analysis can also shed light on the structure of other important committees like the Federal Open Market Committee (FOMC) and European Central Bank Governing Council (ECB). At the Bank of England, a small committee appears desirable because the marginal benefit of additional members due to information pooling declines rather quickly. From this perspective, the 19-member FOMC and 24-member ECB are potentially too big when one considers the costs of more members. Of course, in economies with more regional heterogeneity like the US or euro area, drawing on a greater number of views may be more important for determining the right interest rate. We find one type of appointee (internals) forms, on average, more precise assessments than another (externals). This raises the question of whether such differences exist on other committees with different member types, and whether the composition of these committees is optimal. For example, Reis (2013) argues that one justification for the inclusion of regional Fed Presidents on the FOMC is that they contribute new ideas. An approach similar to ours using policy preference data from the FOMC could help determine whether Presidents indeed have more (or less) expertise than the Fed Governors based in Washington DC. One argument in favor of Presidents having more expertise is that each has a reasonably large staff and budget which might encourage competition in the market for economic analysis (Goodfriend, 1999). The Fed Governors, by comparison, have to rely mostly on the analysis of the Board of Governors' staff, which is shared amongst FOMC members. At the Bank of England, it is the externals who only have a small staff and these are the group found to have less expertise. One note of caution is that the correct decision might depend on multiple dimensions, some of which externals have more expertise on than internals. A model that explicitly built in such diversity would be a natural extension of our work. Moreover, the committee structure may affect deliberation and therefore the evolution each member's views. We leave these issues to future research. Nevertheless, the paper is, to our knowledge, the first to decompose voting heterogeneity on a monetary policy committee into distinct preference and signal components, and as such provides potentially valuable facts for moving the debate on monetary policymaking by committee forward.