Arrow, Schelling, and the Fed

Does the Fed want to loosen monetary policy? If so, why doesn't it do it?

There are two answers (or at least, two simple answers) to this question:

1. The Fed doesn't want to loosen monetary policy.

2. The Fed wants to loosen monetary policy, but thinks it can't.

I'm trying to come up with a third answer.

(Scott Sumner goaded me into writing this post; I would caricature his position as being closer to answer 1.)

Arrow said that you can't (or you can't always) aggregate preferences. The Fed's decisions are not made by a single individual. It is different from the Bank of Canada, where the Governor has ultimate responsibility for monetary policy. So it may not make sense to talk about what the Fed "wants".

Schelling said that zero is a magic number. And he said that in coordination games, magic numbers matter. The equilibrium that players choose will be at a magic number, or "focal point".

If the Fed were voting only on what Federal Funds Rate to set, we wouldn't have to worry about Arrow. Voting members of the Fed would have single-peaked preferences over that one dimension. Hawks would want a high FFR; doves would want a low FFR; and the outcome would be the FFR chosen by the median voter. (This is how the CD Howe Monetary Policy Council chooses its overnight rate recommendation.)

Once we move outside that one-dimensional choice set, I don't think we can safely assume single-peaked preferences. Given a choice between three options: {do nothing; do unconventional policy one; do unconventional policy two}, individual voters' preference-orderings could be all over the map. Majority voting on the three options, taken pairwise, might not give a transitive ordering. The actual outcome of majority voting would depends on the order of motions in the agenda.

How is the Fed agenda set? (I mean "agenda" in the Roberts Rules sense, not in some weird conspiracy theory sense.)

Dunno. From my experience, in far too many university committees, the agenda is really set in some sort of coordination game, not by any "official" procedure.

Zero is a magic number in any coordination game. So is the status quo (which is just another way of saying "zero", because it means zero change). Zero unconventional policy one. Zero unconventional policy two. Zero FFR (OK, 0.00%, 0.25%, whatever).

Most committees want to reach a consensus, which makes it a coordination game. It is hard to get a committee to move away from zero. Some other particular number has to stand out as a focal point, so that players can coordinate on that focal point.

Even if only a minority of voters want zero, you can't necessarily get a coalition of voters to move away from zero. Even if the doves outnumber the hawks, there is no magic dimension on which the doves can coordinate to create an agenda for a majority coalition to move away from zero. There are too many different "unconventional" policies. The FFR dimension, being the historical dimension, is the only magic dimension.

(Which is one more reason to destroy the "monetary policy IS the FFR" idea, and its hold on our thinking.)

I think I understand Schelling. I know I'm a bit shaky on Arrow. I know I'm ignorant of the Fed's internal workings. And putting all three together is hard, anyway. I'm just throwing this out there, to see if anyone who knows better picks it up and runs with it. But it makes some sort of sense to me.

Comments

[Phil Koop]

If you take Meyer's account at face value, the Fed did not reach decisions by consensus in his day (which was also Greenspan's.) Perhaps things have changed under Bernanke.

[jsalvati]

It's important to note that Arrow's point is specifically about a traditional voting system. There are many other ways you might try to aggregate preferences, including fractional voting.

[Roadrunner]

http://www.princeton.edu/~ceps/workingpapers/198blinder.pdf

[David Andolfatto]

Nick,

It would have been helpful if you had defined what you meant by a "looser" monetary policy. In my mind, there are only two practical options: [1] lower the interest rate paid on reserves, [2] increase purchases of longer dated government bonds.

It seems clear enough that option [1] would have almost trivial consequences, given that this rate is already close to zero. Given the very low interest rate on even longer dated maturities, it is not even clear that option [2] would have any quantitatively important effects. (If people think it would, I would be interested in hearing what sort of mechanism they have in mind).

There is a third option, discussed in an earlier post (which I haven't had the time to read) of purchasing other asset classes. The Fed got away with the MBS purchase program under emergency conditions, but at some political cost (Congressional pressure threatening Fed independence, or even its existence). The Fed does not want to go there again, if at all possible. (And if you think it should, you should be talking about decision-making in Congress, not the Fed).

It could be that the majority of Fed policymakers view our fundamental economic problems as stemming from conditions that the Fed cannot control (fiscal policy, structural reallocation, uncertain lending environment that prevents banks from lending, etc.). If this is true, then the best the Fed can do is to try to manage inflation expectations to the best of its ability, promising to act in some manner in the event of deflationary pressure, while at the same time promising not to let inflation get out of control in the event of improved economic conditions (in short, trying to maintain its implicit inflation target of around 2%).

Now that I write this, the thought did occur to me that the Fed recently did "loosen" its policy when it announced that it would replace the unexpectedly rapid runoff in its MBS holdings with longer dated government securities, so as to maintain the size of its balance sheet. So again, I guess it depends on how one defines a "loosening" of MP.

[david]

Hi Nick.

It seems to me that there may be other factors, operating at the individual level, i.e., apart from the game theoretic or aggregation issues.

I've spent a long time either as an advisor to committees of public economic decision-makers or as an advocate trying to influence them in highly technical and often high profile areas. It won't come as a surprise to you that their policy preferences are determined not only by the public interest but by their private interests and in particular the fact that they may be held to account, reasonably or not, for any failures. They know that it is better to be right than wrong but they also know that it is far better to be wrong by having been "careful" and "prudent" by adhering to some sort of respected consensus (which is often the safe haven of past practice) than to be wrong while having departed from the consensus. They know that they will either not be faulted or be faulted to a lesser degree for following the "best independent and academic advice" or whatever the source of the consensus.

Also, given uncertainty, it is possible, in the minds of decision-makers, perhaps reasonably, for a non-consensus policy to simultaneously have not only a high probability of success (if the consensus is completely wrong) and a high probability of catastrophic failure (if the consensus is completely right), with the difference made up by a low probability of a mediocre outcome. So there may be a strong personal incentive for choosing the "wrong" policy or for convincing themselves that the consensus policy really is the prudent course after all.

Even if decision-makers are completely unconcerned with their personal reputation, their objective may be, for the noblest of public policy reasons, to minimize the probability of a really bad outcome rather than to maximize the probability of really great outcome.

In addition, the more important and higher profile the decision and the higher the personal and public stakes, the greater the chance that decision-making paralysis will set in the hopes that things will become clearer or that events take the matter out of their hands.

[Nick Rowe]

Phil: My very cursory reading of Fed-watchers suggests they do try to get consensus now. Perhaps that ties in with david's point about individuals' being very wary of departing from the consensus when there's a risk of the non-consensus policy failing badly. Do committees always buy IBM, while individuals buying for their own use don't? It sounds plausible to me.

jsalvati: like probabilistic voting? Not sure how well that model would fit the actual Fed, or any committee. But maybe people who think most strongly about some option get extra weight in the outcomes. Dunno.

Roadrunner: I skimmed a quick read of Blinder's paper. Didn't see any immediate connection. Maybe I missed it.

David A (must call you that now, with another david commenting): Ignore this post (where I really don't know much about what I'm talking about). I would really like you to read my other post, where I've deluded myself into thinking I do. Plus, it's much more up your street.

Yes, I should have defined "loosening". I meant "shifting the AD curve to the right in {P,Y} space". And I think they can do it. By purchasing pro-cyclical assets, for example, announcing they are doing this, announcing they intend to raise those assets' prices by doing this, and announcing a price level path target (or whatever) they aim to hit by doing this as well.

[David Pearson]

Nick,

Bernanke told us there is no support for a "higher inflation target".

The Fed seems to believe that, to influence AD, it would have to raise inflation expectations above their recent historical level. In other words, that there is an explicit trade-off between using an inflation target to influence AD, and higher inflation in the long run. Most targeting proponents believe this trade off is, at most, very temporary.

The question for targeting proponents is whether a higher long-run target is necessary to dislodge deflationary inertia. In other words, does the Fed have to promise 4% (long run) inflation to spark the kind of behavior that will enable it to reach a 2% target? And if it promises only 2%, will it continue to fall short given that its unlikely actors will, for instance, run out and buy houses, or reduce their savings rate, as a result. Of course, a short term target of 4% might help on the margin, but the expectation that the Fed will soon tighten to reduce that rate to 2% is not exactly a tonic for expectations.

So it seems the two groups are talking past each other. Targeting proponents say, "just meet your existing target, plus a little higher catch-up for two years. There is little cost to this." To which the Fed replies, "we are not going to raise our long term target, as the cost is too high relative to the benefit."

[Andy Harless]

The single policy dimension should be the maximum maturity at which the Fed targets a zero interest rate.

[Nick Rowe]

David P. : "So it seems the two groups are talking past each other. Targeting proponents say, "just meet your existing target, plus a little higher catch-up for two years. There is little cost to this." To which the Fed replies, "we are not going to raise our long term target, as the cost is too high relative to the benefit.""

That way of putting it sounds plausible to me. It's similar to the argument for indexation of minimum wages. Or whether fiscal policy should be done through tax cuts or spending increases. The countercyclical argument always gets opportunistically obfuscated by the long run argument, even though they are quite distinct.

Andy: I disagree. For one, that means binding the hands of future committee decisions, and pre-commitment is hard enough for an individual and harder for a committee. Second, lowering long-term bond rates (or committing to keeping the FFR at 0% for a long period) is not the only "unconventional" monetary policy, and not the best one in my view. But most importantly, even permanently zero nominal interest rates may not be enough to escape a black hole.

[Phil Koop]

Nick: There's no question that Greenspan would never have tolerated the sort of running off at the mouth going on under Bernanke. I don't have a view on how to interpret this. Does Bernanke genuinely seek consensus? Does he just want cover from inflation hawks so that he seems reasonable by comparison? Or is it just as you suggest: it is easy to rule an independent Fed with an iron fist when you are doing what everyone likes anyway, harder when you have to demand the unconventional.

I don't know. But I think you might profit by reading Meyer's book. It is simultaneously brief, easy to read, and deceptively subtle. Don't be fooled by the LCD language.

[Frances Woolley]

David: "It won't come as a surprise to you that their policy preferences are determined not only by the public interest but by their private interests..."

This is a crucial point. I agree with what you say but would add to it. Private interests might also include how one is likely to be affected personally by, say, unanticipated inflation, a higher rate of unemployment, etc. 'The public interest' is not a single well-defined concept either - a reduction in the rate of unemployment, for example, has less of an impact on people in secure jobs, an increase in government debt passes costs onto future generations, etc.

I would love to see a public choice model of monetary policy. Actually doing a search on google scholar it seems there's a fairly big literature on the political economy of monetary policy , mostly looking at issues like credibility of policy etc. I don't see anything analyzing rules and institutions in the way that Nick is suggesting they be understood.

[JP Koning]

Another answer: Central banks are odd bastions of Soviet-style central planning within a broadly free market. They're destined to do a poor job for the same reason that a centrally-planned grocery monopoly will tend to have shortages, lineups, spoilage, and bad service. The managers of such a grocery monopoly can hold whatever odd beliefs they wants about stocking shelves (ie stock tuna on Tuesdays since they each start with the letter t) as there is no competitive system to discipline their irrationality. Ditto for central bankers.

[Simon van Norden]

JP Koning: If I replace the term "central bank" in your post with that of any other government institution (parliament, the army, the dept of highways, the court system, food inspectors, tax collectors, etc.), your argument still applies, right?

[Simon van Norden]

Nick:

You (and other commenters on this list) have studied Bernanke's remarks much more carefully than I have. Is it reasonable to interpret Bernanke's remarks with a different standard than those of the other board members? I'm tempted to think that when someone like Bullard or Kocherlakota speaks publically, they are trying to influence the concensus. However, I wondering whether Bernanke denies himself that freedom. Instead, is it realistic to think that he's restricting himself in public comments to reflecting the view of the (very divided) board?

[Determinant]

Greenspan recounted in his memoirs that his legendary opacity and obfuscation weren't natural; he learned it as part of his job as Fed Chairman. His aides and assistants were adamant that he had to speak like that to avoid influencing markets.

It seems Bernanke disagrees with that, plus I think the fact that he has had to do extraordinary things in a short period of time, much of which required the agreement and co-operation of Congress (bailouts, etc) meant he had to take a more active and open stance than his predecessor did.

[Nick Rowe]

Frances: Yes, most of the political economy/public choice/positive models of monetary policy I have seen just black box the central bank. The central bank has a utility or loss function that obeys all the normal stuff like transitivity, and maximises/minimises it subject to a constraint. Or explores the interaction between monetary and fiscal authority using game theory. I don't remember much on the internal workings. Though I do vaguely remember seeing some.

JP: an argument there for Scott's NGDP futures market?

Simon: actually, my guess is that you have probably read as much of Ben Bernanke's speeches as I have. I try to be a Bank of Canada watcher, but even there I fail. I can never remember who's who.

"Is it reasonable to interpret Bernanke's remarks with a different standard than those of the other board members? I'm tempted to think that when someone like Bullard or Kocherlakota speaks publically, they are trying to influence the concensus. However, I wondering whether Bernanke denies himself that freedom. Instead, is it realistic to think that he's restricting himself in public comments to reflecting the view of the (very divided) board?"

Yes, I think that is reasonable, based on my experience of committees (I'm absolutely convince that the Fed is really no different from the Carleton Timetable Committee ;) ). And because of Arrow, it's really logically impossible (in general) to give a coherent account of what a committee wants and believes. That account cannot make sense.

Determinant: if Greenspan wanted to avoid moving markets, I think that Bernanke needs to move markets.

[Nick Rowe]

Kevin Moran at Laval sends me a link to a very recent paper "Monetary Policy by Committee: Consensus, Chairman Dominance, or simple majority" by Alessandro Riboni and Francisco Ruge-Murcia, both at Universite de Montreal.
http://www.mitpressjournals.org/doi/abs/10.1162/qjec.2010.125.1.363

From the abstract: "For all central banks, results indicate that the consensus model fits actual policy decisions better than the alternative models."

[JP Koning]

"JP: an argument there for Scott's NGDP futures market?"

No, I'd say an argument for George Selgin and Larry White's free banking system.

Simon: Yep.

[Scott Sumner]

Nick, First of all, I accept your answer to my previous comment in the last post---if the fed funds rate was 2% they would cut it now. But I think that actually supports my point. They are not trying to boost NGDP with unconventional tools, so there is no point in us trying to dream up unconventional tools to help them boost NGDP. (BTW, I am about to reject my own advice, and write a post telling them what to to, as if your argument is right.)

(Interestingly, the fed funds rate was 2% when Lehman failed, and the world economy fell off the cliff. And at the next Fed meeting they refused to cut their fed funds target.)

I think this post does illustrate one of the problems. But another is that Bernanke doesn't like badly split decisions. That gives three or four hawkish voters an effective veto over affirmative policy moves in any unconventional direction. So it's not really a majority rule institution, de facto. De jure it is, and I now wish Bernanke would run it that way.

It's like our Senate filibuster. America is run on the implicit assumption that the minority is smarter than the majority. How's that working out for us?

[JP Koning]

I second Phil Koop on the Fed not reaching decisions by consensus in Greenspan's day. If you listen to the following Econtalk interview, ex-Fed economist Michael Belongia describes how Greenspan often skirted the FOMC to decrease the fed funds rate. A majority vote of the 12 member FOMC is required for ff rate changes. Greenspan pressured the 7 member Board of Governors to lower the discount rate, and piggybacked a "technical pass through" from the discount rate to the fed funds rate without asking for an FOMC vote. He definitely got his way.

http://www.econtalk.org/archives/2010/01/belongia_on_the.html

[Nick Rowe]

William: Sorry. You are faking it on the topic. I deleted your comment.

[Nick Rowe]

William: what is fake is your pretending to be on topic. If you want to talk about your own topics, start your own blog.

[Will-i-am]

Nick: Not a fake. Quite sure. I look like a jackass. Sorry. I’ll go to the back corner of the room with a self-imposed D hat and spend time counting on my thumb. 1.2.3::4? That darn whole for pinky is always so hard too find!

[William]

While Nick’s ultimately mistaken, in fact, to say I’m a fake pretending to be on topic, I have been a jackass. Heck, a naive highjack of many topics perhaps. To have someone abruptly pull a topic out of a body of topics (thematic) go another way to suit a creative purpose is disruptive and bad mannered, (especially when royalty drops by – and reads notebooks. Would shock any teacher). Anyway, won’t happen again. I’ll sit at the center rear silently, listen, no rude noises and won’t touch a topic for another purpose until everyone else has finished probing.

For 0 cents cost to all here’s the Typology of Focal Points rooted by the generating substance theory for sensorial quality to create a Philosopher’s Stone:
A. Schelling Point :: Earth: well grounded, good pivot. Strong. Still flawed.
B. Shelling Point :: Fire!: perfect. Strongest.
C. Shilling Point :: Air: culturing propaganda, ok, but weakest perhaps
D. Selling Point :: Water: makeup beauty, weak too but fits pinky well.
I’ll give the typology a few spins to demonstrate its potential, especially to touch an embedded topic (so still not off-topic): the “hawk-dove” dialectic with endpoint in view. Yep. It’s Hegel. Thought I’d need a handful of Philosophers’ Stones to knock him (why I’ve kept Marx alive in a container) off the sovereign horse. Perhaps this One in a sling will be enough. I’ll aim at his pivot foot. Don’t want to really hurt him or anyone, after all.

[Nick Rowe]

William: please stop posting. I deleted your totally off-topic and bizarre comments because I wanted to save you embarrassing yourself further.