Dante’s seventh ring of hell, as I recall, is being forced to serve on a large committee. But imagine that you come across a large committee that somehow functioned effectively. Implausible certainly, but if you force yourself to imagine a well-functioning committee, I think you’ll agree that you’ll need some special feature to account for this minor miracle.
The FOMC is a large committee that functions pretty well. You cannot understand FOMC choices without understanding how consensus develops, is maintained, and evolves. This is a feature, not a bug. The Fed’s framers intended that the FOMC function as a diverse committee working toward compromise. (Note, for example, that an even number, 12, votes determine policy.) Despite today’s political environment, folks working together to form consensus can be a good thing, in part because the consensus-based approach has desirable effects on the policy outcomes.
The FOMC is made up of 19 policymakers—12 reserve bank presidents and 7 governors. But only 12 vote any given time—5 presidents and 7 governors. The governors and presidents are appointed according to elaborate rules that differ for the two groups. The rules were designed with the express intention that the presidents have dispersed views and views that are systematically more hawkish than the governors. This seldom-emphasized historical point might warrant more discussion, but I’ll just provide one anecdote and some references.
Getting the mix of hawkish and dovish views right was viewed as a primary issue throughout the Fed’s formative years. As Paul Warburg, one of the most astute participants in the debate put it, a “formula had to be found” in which in which hawkish and dovish tendencies would be balanced. Congressmen Glass and Steagall were at the center of the debate as the FOMC as formed, and when Glass proposed that only the governors should vote, Steagall responded that Glass was “without peer” in his advocacy of inflation. In the end, the ratio of 7 governors to 5 presidents was chosen.
So we have a solid voting majority of 7 governors and nearly twice as many presidents chosen to have eclectic views that are tilted to the hawkish. The design, I think you’ll agree, had the intended effect: throughout the Fed’s history, we’ve gotten an eclectic mix of Presidents, and on average they’ve been more hawkish than the governors. 
You might suspect that the unique features of the FOMC design would not matter very much, and in simple times this suspicion is probably correct—policymakers will largely agree and policy will look like the typical view of the 19. But in contentious times, two factors give the design a major role in shaping policy.
First, the main differences of opinion over policy in contentious times in recent decades have stemmed from very persistent differences of opinion about the how the economy is working. In what has been the standard case, the hawks, based on their implicit or explicit models, see structural reasons why inflation will surely rise unless there is a pronounced policy response. In the doves’ views/models, inflation forces are more subdued. Second, the FOMC meets every six weeks to consider adjusting policy. And one thing everybody on the FOMC agrees on is that the particular timing of any given policy action (almost) never matters.  Under these two conditions, the Chair has a powerful consensus building tool: “Stick with the consensus for one more meeting; then if the data more clearly support your view, we’ll switch.” So long as the FOMC members trust each other, this argument is, and probably should be, quite persuasive.
The Greenspan Fed in the 1990s provides an example of how this can work. Toward the second half of the 1990s, the economy was booming and essentially the entire FOMC—hawks and doves—saw a need to raise rates to stem incipient overheating. Greenspan perceived an unprecedented productivity surge that would defuse any inflation pressures. Accordingly, at each FOMC he said, “Let’s put off tightening one more meeting; then if that inflation materializes we’ll tighten.” This went on for many FOMC meetings as the inflation failed to appear.
If the FOMC had been publishing its Survey of Economic Projections at that time, Greenspan’s dots might well have been the lowest on the FOMC over this entire period. And yet this is not an example of the Chair running roughshod over the Committee. The Committee preferences are fully and accurately reflected in the credible promise that the moment inflation appears, the FOMC will tighten. With disagreements about the working of the economy lasting multiple years, the Chair’s consensus can for a very long time look quite different from the typical dot of the 19.
Of course, we are once again in a period of persistent difference of opinion over how the economy is functioning. Almost since the beginning of the recovery, a hawkish contingent on the FOMC has been concerned that excessive inflation and other excesses are just around the corner. Jon Hilsenrath of the Wall Street Journal beautifully documented this as of 2013 under the title ‘Doves’ Beat ‘Hawks’ in Economic Prognosticating. Since 2012 when the dot plot appeared, the hawkish perspective on inflation dynamics has been observable in hawkish dots, and the typical dot has been much more hawkish on average than subsequent policy. Of course, those dots do reflect policy views, and those views are affecting policy: if the inflation appears, the group will indeed tighten. But for seven years or so, those clear signs of overheating that concern the hawks have remained just around the corner, and those dots have had little to do with actual policy outcomes.
If you listen to policy communication on behalf of the consensus, you’ll see this conditional treatment of the hawkish views quite clearly. For example, the bulk of the FOMC signed onto the famous thresholds for raising interest rates. The hawks could agree because the thresholds basically said that extraordinary accommodation will stay in place unless inflation takes off. Hawks and doves could disagree mightily on which threshold was more likely to be hit first, but they could agree on the conditional policy. In the period since the thresholds, we’ve had regular reassurances that if signs of overheating become clear, the Fed will tighten.
Let me re-state how the key parts fit together. In the face of unresolved differences about how the economy is working, the chair can build consensus—indeed, near-unanimous support—for policy today based on the mutual trust that if one or the other view becomes more clearly true in the future, the group will ‘do the right thing.’ This appealing and simple fact about policymaking has a paradoxical effect under conditions, such as those we observed in the late 1990s and recently, when pronounced differences of opinion about the functioning of the economy last for years. The key in both these episodes is that the more hawkish on the committee persistently believe that excessive inflation is right around the corner, but the inflation does not appear. And ‘stick with me for one more meeting’ leads to years of near-unanimously-supported policy that is far from what the typical dot of the 19 depicts as most likely.
I am not attempting to argue that the FOMC design is ideal in any sense or that policy has been ideal. In the messy real world of policymaking, ideal seldom belongs in the discussion. From the standpoint of policy actions, I am arguing that the FOMC committee structure matters, and, in particular, that it has resulted in a reasonably coherent policy that evolves systematically and predictably: whether nor not the policy has been right, it has been systematic and predictable. The next piece of the argument is that the communications implications of the committee design are neither simple nor salutary.
Let me end this post with a parable that I hope cements the ideas. The FOMC is going to re-paint the Board room. In an effort to be open with the public, the FOMC polls the 19 for their preferred color choice and announces the results. The issue, it turns out, is contentious: the poll results show a rainbow of colors. Thoughtful observers find no value in the poll results because they shed no light on how the 19 views, as digested by the 12 voters, will come together in one outcome. But some eager observers form a color prediction by stirring the preferred colors together to get what they view as a ‘central tendency’ of the rainbow. Sensible folks know that this is silly: the consensus choice will depend on the strengths of the views of each voter about each of the colors.  And the choosers may be pretty accommodating, given that they can always re-paint in 6 weeks.
It would be nearly impossible to discern how they’ll get from 19 preferred choices to one outcome. But fortunately, nobody should care much. Remember the Greek tragedy point from the previous post: even when there is a circuitous path that is filled with drama, the outcome may, nonetheless, be pretty much inevitable. The color choice? They’ll pick beige. That’s what large committees do.
Bottom line: The evolution of the consensus of a large, well-functioning, committee may be systematic and predictable, even when the component views are less so. Choices by committee are one thing; communications by committee quite another. Stay tuned.
5. Susan Beldon presented one of the first analyses of dissenting FOMC votes documenting this fact. As for the eclectic part, having presidents chosen by the 12 Reserve Banks scattered around the country was intended to help get a wide range of views and avoid Washington groupthink. This too has worked, in that many ideas that have become mainstream—at least for a time—have entered the Fed’s discussions through one or another reserve bank. The St. Louis Fed’s introduction of monetarism to the system is often mentioned in this context. [back]
11. A key aspect of the parable that is paralleled in the monetary policy case is that the decision is multi-dimensional. This makes it particularly hard to see how the various considerations will be combined in reaching a decision. [back]