Friday, August 2, 2013

Hawks, Doves, and Hyenas

People seem fond of categorizing central bankers as "hawks" and "doves." I think what people have in mind when they talk about the hawk/dove divide is the Phillips curve as a policy menu, much as in Samuelson and Solow. Then, hawks and doves are people who agree that what they are doing is choosing a point on the Phillips curve, but disagree about what point to pick. Hawks are more willing to substitute unemployment for inflation than are doves. You can see why someone who perceives himself or herself as a dove in this context would like the nomenclature. A hawk is someone who is trying to screw the unemployed -- worse than a vulture, who just eats what dies of natural causes. The hawk kills, and then eats.

Currently, that's not a helpful way of thinking about, for example, the members of the FOMC, and why they might disagree with each other. Fundamentally, these people don't agree on the theories they are using to guide their decisions. Some people might argue that preferences determine the choice of theories (that seems to be the crux of what Krugman has to say), but I don't think so. Here's a better guide to how people think:

New Keynesians: These are people who have absorbed modern macroeconomics, and have bought into the framework developed by Mike Woodford and others. That framework was an outgrowth of the neoclassical growth model (Cass-Koopmans), Brock and Mirman, and Kydland-Prescott, incorporating monetary factors, and with a role for monetary policy. This is the theory that is dominant on the committee. It's what's principally driving the forward guidance aspect of current monetary policy, i.e. promises about future policy that are supposed to have beneficial effects today. The hard-core New Keynesians are Williams (San Francisco), Evans (Chicago), and Kocherlakota (Minneapolis). Bullard (St. Louis) has some New Keynesian sympathies, as does Plosser (Philadelphia), I think.

Old Keynesians: This is some version of IS-LM, AD-AS, Phillips curve, much like what you would find in some undergraduate textbooks (not in my favorite one, as Mankiw would say). This is pretty loosey-goosey, and can incorporate almost anything. Fortunately for the Old Keynesians, the New Keynesians (for just this reason) have taken pains to describe what they do in the language of Old Keynesians - you can often find words like "IS curve" and "Phillips curve" in there. The rationale for quantitative easing (QE) - a cornerstone of current policy - is pretty much Old Keynesian. To the extent anyone justifies it, it's done using pre-1970 theory. The "transmission" mechanism that Fed officials describe for QE - purchase assets, interest rate falls, investment goes up - sounds like something straight out of a principles of economics textbook. The most prominent Old Keynesian on the FOMC is Janet Yellen.

Old Monetarists: This is literally a dying breed. Hard-core monetarism is represented best by the views of Milton Friedman: (i) monetary factors are very important, but attempts by the central bank to intervene to influence real activity in a good way are likely to go haywire; (ii) there exists a stable money demand function; (iii) central banks should be targeting money growth so as to control inflation. Views (ii) and (iii) died in most, if not all, central banks in the 1980s. But Plosser has some old monetarist views.

There are some outliers in there. George (Kansas City), Fisher (Dallas), and Lockhart (Atlanta) are hard to place. Rosengren is a Keynesian with some background in banking, and Lacker doesn't fit into a neat macro pigeonhole. The others I don't know much about.

Increasingly important, and I think surprisingly overlooked in light of the financial crisis, are the backgrounds of monetary policymakers in general economics. What do they know about banking theory, incentives, and information problems? What do they know about banking and monetary history and how the financial systems in different countries work? In the next financial crisis, we want a person running our central bank who understands what too-big-to-fail is about, the nature of long-run moral hazard, and how moral hazard can work against the central bank even during a crisis.

As well, in the context of run-of-the-mill macro monetary policy, different views about the mechanism by which monetary policy works can be irrelevant. Everyone appears to accept that there is some short-run nonneutrality of money at work. Everyone accepts that it is important that the central bank control inflation. The key differences are in views about the persistence of monetary nonneutrality, and how to spot inefficiencies that the central bank is capable of correcting. Some people look at the state of the world now and think that the difference between real GDP today and what it would have been if it had continued to grow at 3% per year since the end of World War II is all inefficiency. You can find other people who think that where real GDP is today is about the best we can do. There are a lot of other people who aren't sure one way or the other.

So, how would either Larry Summers, or Janet Yellen, fit in as leader of this group? Brad DeLong tells us that other people think that Summers is a "right-wing hyena," but he thinks Summers is a good guy. DeLong fits nicely in the Old Keynesian camp, so maybe Summers is an Old Keynesian. But clearly Paul Krugman does not like Summers. He seems to like Yellen better. Krugman is certainly an Old Keynesian too, and maybe DeLong is just sticking up for his friend and coauthor Summers, so maybe Summers is not so Old Keynesian after all.

Janet Yellen is firmly Old Keynesian. Her training happened before the revolution in macroeconomics took place, and it's quite clear that she thinks about the world in a conventional IS-LM, Phillips curve, demand-management style.

If we were to choose a Fed chair based on academic records, then this would be no contest. Summers, in spite of being occupied as a policy wonk for a considerable period of time, still has a REPEC ranking of 25. That's a #25 ranking in the world among economists, based on publications in academic journals, quality of publications, citations to those publications, etc. Yellen's ranking is 805. Summers was tenured at Harvard at age 28, while Yellen never made it past the assistant professor rank there. Of course we know that central banking requires some different skills from what it takes to publish papers. Some of our colleagues should definitely not be let out in public.

A good Fed Chair should be collegial. That means making FOMC members feel comfortable in expressing their views, so that diverse information can be forged into some kind of coherent synthesis, as part of the policy process. Bernanke has been very good at exploiting the strengths of the unusually-decentralized Federal Reserve System. The regional Presidents have their own staffs, their own views, and Bernanke by all reports is willing to listen, and has used ideas from the regional Feds as part of the monetary policy program. As well, the regional Presidents seem free to speak their minds in public, and they often do. I think all of that is healthy.

Some clues to how Summers thinks are in the transcript and video for this forum on "new economics." The worst it gets is when Summers says this:
When I was in the government, I got a lot of papers in the mail. To the first approximation, I attempted to read all the ones that used the words ‘leverage,’ ‘liquidity,’ ‘deflation’ or ‘depression.’ And I attempted to read none of the ones that used the words ‘neoclassical,’ ‘choice theoretic,’ ‘real business cycle,’ or ‘optimizing model of.’ (laughter) There were more in the second category than there were in the first. But there were a reasonable number in the first, and they told you a lot.

There is a lot in Badgett [sic] that is about the crisis we just went through. There’s more in Minksy and perhaps more still in Kindleberger.2 There are enormous amounts that are essentially distracting, confusing, and problem denying in the stuff that is the substance of the first year courses in most PhD programs.

So I think economics knows a fair amount. I think economics has forgotten a fair amount that’s relevant. And it has been distracted by an enormous amount.

You can find other things in that forum, and in what Summers has said in other contexts, that put him in a more favorable light, but I think that quote is telling. It's interesting to read Bagehot, or Minsky, or Kindleberger, but we're going to find a lot more serious ammunition to bring to bear on thinking about financial problems in work on money, banking, information economics, mechanism design, contracts, and incentive problems, that people have been working on in the last 30 years or so. Summers has revealed himself here to be shockingly closed-minded - he wants to ignore a large segment of research that the mainstream of the economics profession takes very seriously. A leader of central bankers needs to dig into all the alternatives and understand them, even if he or she doesn't agree with everything. Summers seems to impress people as always wanting to demonstrate that he is the smartest person in the room. He is quick to come to a conclusion, and can be an intellectual bully. Not exactly Mr. Collegial.

If you have never seen Janet Yellen in action, here's a speech she gave at the Haas Business School at Berkeley. She's clear, articulate, and thoughtful, I think. Yellen has a reputation in the Fed system as being a good listener - she's certainly collegial. There is no reason to expect that she would not follow Bernanke's example in encouraging independent thinking in the hinterland - places like St. Louis. Yellen may be overly inclined to see everything going on in the U.S. economy as a demand management problem, but maybe that's not the worst flaw we could have in a central banker.

Summers has the additional problem of being too closely allied with various power structures. He is part of the Cambrige, MA clique. He has worked for hedge funds and too-big-to-fail banks. Not good for central bank independence.

I think I'm with Krugman on this one. Though there are potentially better choices around, Janet Yellen would be fine, and I think Larry Summers is just too risky a bet, though he might surprise us.


  1. I think in some contexts you might want to be relatively "hawkish" and in other circumstances "dovish".

    To switch to a different set of animals, what do you think of the turkey vs hummingbird distinction?

  2. Nice post. On the one hand, Summers in quick and sharp, while on the other he is opinionated and shallow.

    1. That is precious, thanks for the laugh.

    2. Those things sometimes go together. Particularly in academic settings, people can score points by thinking quickly, deciding what they should think, and then going with it. Having a big voice, and training in high school debate helps. The problem is that the quick and sharp type often has a short attention span. More often the quiet person in the back of the room who actually spends some time pondering the problem gets it right.

  3. I don't think it's fair to say that Krugman "doesn't like" Summers. It seems pretty clear to me that Krugman has a lot of respect for Summers as a thinker-- it's hard not to; Summers is smart as hell. I also seriously doubt that Summers is as ignorant of the last thirty years of research as this post seems to suggest-- I think he just recognizes that straightforward IS-LM has done a pretty admirable job of explaining the state of the world today; a lot better than the "new era" RBC work of the past few decades, which has pretty much gone down in flames.

    But I also think the distinction between "old Keynesians" and "New Keynesians" isn't as neat as this post wants to make it seem. What I hear people like Krugman and Summers (and, even more strongly, Richard Koo) saying is that Mike Woodford-style NGDP targeting is worth pursuing, but that the mechanisms through which it would be effective (influencing expectations) are inherently difficult, and that there's no guarantee of success. I think that point is almost certainly true-- if influencing market expectations were easy, Ben Bernanke would have done it by now. And I think Yellen is even less skeptical than Krugman on this point; she does head the Fed's communication committee, whose communication strategy is substantially substantive rather than just communicative.

    But my takeaway from reading Krugman on this issue (while Summers has been opaque on it, I'd guess he's pretty much in the same place) is that he thinks the "New Keynesian" approach is worthwhile, but that success is far from guaranteed.

    1. "I don't think it's fair to say that Krugman "doesn't like" Summers."

      Actually, I think it's clear he's anti-Summers and pro-Yellen. You can't speak out against a fellow Cambridge insider, though.

      "I also seriously doubt that Summers is as ignorant of the last thirty years of research as this post seems to suggest..."

      He's willfully ignorant of some things. He thinks he got it early in the 1980s, and has stopped paying attention.

      "straightforward IS-LM has done a pretty admirable job of explaining the state of the world today"

      I know that's Krugman's line, and it's just silly. IS-LM is just not useful, given the alternatives.

      "a lot better than the "new era" RBC work of the past few decades, which has pretty much gone down in flames."

      Here you're being inconsistent. Obviously it hasn't gone down in flames, as neoclassical growth theory is central to what economists who work on growth do, and the RBC model is central to New Keynesianism, which obviously you think is alive and well.

      "Mike Woodford-style NGDP targeting"

      What Woodford does with his models is not NGDP targeting. He may have said something about that (and Evans may have mentioned it), but the New Keynesian work that backs up forward guidance has nothing to do with NGDP targeting, and as far as I know NGDP targeting is not collectively taken seriously on the FOMC.

    2. "straightforward IS-LM has done a pretty admirable job of explaining the state of the world today"

      I know that's Krugman's line, and it's just silly. IS-LM is just not useful, given the alternatives.

      Looks like IS-LM matches the data better than your model. So much about silliness and usefulness.

    3. "Looks like IS-LM matches the data better than your model."

      That statement makes no sense. IS-LM is not a dynamic model, and can't even make a start in fitting aggregate time series. My model is not quantitative, and therefore not set up to even address anything in the time series. You could do that if you wanted to, though.

  4. Contemporary research economics is dominated by New Keynesian economics. New Keynesian economics has adapted from the New Classical school the RBC hypotheses including rational expectations and associated methods, combining it with nominal rigidities (price stickiness) and other market imperfections. It has led to the development of dynamic stochastic general equilibrium (DSGE) models.

    Brock and Mirman, and Kydland-Prescott are key aspects of RBC hypothesis and hence came to New Keynesian economics via the New Classical school. It's hard to find anyone who calls themselves New Classical these days, although the phrase pops up as a Freudian slip in papers from time to time (e.g. Lee Ohanian). At the time of their founding in the mid-1970s New Classical economists did not share the monetarist belief that monetary policy could systematically impact the economy and broke with Keynesian economic theory completely, while the monetarists had built on Keynesian ideas. New Classical economics had pointed out the inherent contradiction of the neoclassical synthesis, that is, Walrasian microeconomics with market-clearing and general equilibrium could not lead to Keynesian macroeconomics where markets failed to clear. New Keynesians recognized this paradox, but, while the New Classicals abandoned Keynes, New Keynesians abandoned Walras and market-clearing.

    So when you deride IS-LM, AD-AS and the Phillips curve as Old Keynesian ideas found only in some undergraduate textbooks, they are in fact important parts of the New Keynesian toolkit, and a substantial portion of the best selling introductory graduate macroeconomics textbooks, such as David Romer's "Advanced Macroeconomics," is in fact devoted to a discussion of these models.

    In your discussion of New Keynesian economics you manage to leave out all of the defining characteristics that in fact separate New Keynesian economics from New Classical economics, namely, in addition to Old Keynesian ideas, nominal and real rigidities, coordination failure, efficiency wages, and insider-outsider models. Interestingly, Yellen's earliest and best cited papers make fundamental contributions to the study of the concepts of nominal rigidities and efficiency wages.

    So Yellen's familiarity with IS-LM, AD-AS and the Phillips curve is characteristic of all New Keynesian economists. (Otherwise why call them *Keynesian* at all?) And her important contributions to the distinguishing concepts of New Keynesian economics make her a New Keynesian economist in the best sense. It is no exageration to say that Janet Yellen is a founding mother of New Keynesian economics.

    In contrast, your confused description of the economic schools reinforces my opinion that you are really a New Classical economist who simply doesn't know, or refuses to admit, that is indeed what you are.

    P.S. Charles Plosser is not an Old Monetarist but rather a founding father of Real Business Cycle Theory (RBC), a school of economics for which money and prices don't even exist, and that tells you pretty much all you need to know about his thinking on monetary policy. Esther George and Dick Fisher are difficult to classify simply because they aren't even economists much less macroeconomists.

    1. "It's hard to find anyone who calls themselves New Classical these days..."

      I don't think anyone you would call a "new classical" ever actually applied that name to his or herself.

      "At the time of their founding in the mid-1970s New Classical economists did not share the monetarist belief that monetary policy could systematically impact the economy..."

      Old monetarism was not about "beliefs." It was mainly based on empirical work, e.g. Friedman and Schwartz. Further, Friedman wasn't really about systematic effects of monetary policy. You could think of Lucas's Expectations and the Neutrality of Money as a formalization of some of what Friedman was thinking. As well, Friedman liked talking about "long and variable lags," i.e. the effects of monetary policy were not "systematic" in his mind.

      "New Keynesians abandoned Walras and market-clearing."

      Not quite right. It's not Walrasian alright, but there's a well-defined equilibrium concept.

      "they are in fact important parts of the New Keynesian toolkit"

      Wrong. There are reduced-form versions of New Keynesian models, e.g. the three-equation model with the "IS curve," "Phillips curve" and the Taylor rule. If you think that's the same thing as AS-AD etc., then you don't understand it properly.

      "It is no exageration to say that Janet Yellen is a founding mother of New Keynesian economics."

      Oh boy. Now you're really getting carried away.

      "you are really a New Classical economist who simply doesn't know, or refuses to admit, that is indeed what you are."

      You'll have to tell me what you think a new classical economist is, exactly. When I hear those words, I'm imagining someone who uses straight competitive equilibrium models - preferences, endowments, and technology, complete markets, perfect information, no problems of enforcement, commitment, etc. Relative to that frictionless paradigm, if you read my work you'll find that my models are loaded with frictions - private information, limited commitment, for example. In my model environments, things get very screwed up - people need to use banks and monetary arrangements to trade, and there are important problems that the government needs to solve. There can be multiple equilibria, credit can rationed, etc. Relative to Woodford-style New Keynesian economics - that's complete markets, full information, and one friction, sticky prices - it's well at the other end of the spectrum. Basically, you don't know what you're talking about.

      With regard to the P.S.: Here's another sense in which you don't know what you're talking about. Plosser was on the Shadow Open Market Committee before he came to the Fed. That's a well-known group of Old Monetarists that has existed since the 1970s.