Monday, September 26, 2011

Lucas and Krugman

I thought that this post by Paul Krugman required a response. It's related to a Lucas interview in the WSJ which, if nothing else, tells you why Krugman's assertions about the "moral failure" of non-Keynesian macroeconomists are absurd.

Krugman's blog post mainly tells us that his deficit in macro-knowledge persists. First, Krugman appears to think that Lucas's only contribution to 20th-century macroeconomics was "Expectations and the Neutrality of Money," in particular the signal-extraction money-surprise story about the non-neutrality of money. Then, Krugman states:
In the 1980s, the Lucas project failed — pure and simple. It became obvious that recessions last too long, and there are too many sources of information, for rational confusion to explain business cycles. Nice try, with a lot of clever modeling, but it just didn’t work.
Money surprises were a tiny part of the "Lucas project." While I think a widely-held view now is that the money-surprise mechanism is unimportant in explaining aggregate fluctuations, the lasting contribution from "Expectations and the Neutrality of Money" is a methodological one. And that methodological contribution is not only the use of rational expectations in a general equilibrium context - very useful in itself - but the use of explicit and consistent economic theory in a macroeconomic context. Further, the "Lucas project" also involved pathbreaking work in asset pricing, monetary economics, economic growth, and dynamic contracts under private information, among other things. Some failure!

Krugman seems to like New Keynesian economics, but:
I find NK economics useful, if only as a way to check my logic, although it’s not really clear if it’s any better than old-fashioned Keynesianism.
Judging from the output, I'm not sure there are any serious checks on Krugman's logic, but we'll give him the benefit of the doubt and assume that he actually grinds through Eggertsson and Krugman whenever he writes a blog post. However, if we take "old-fashioned Keynesianism" to be what is in the General Theory, or in an IS-LM model, then it is hard to argue that Woodford-style New Keynesian economics does not dominate. In its explicit form, New Keynesian economics in fact adheres to the "Lucas project." You can see exactly what is going on - there are optimizing agents, explicit preferences, endowments, and technology, and a well-defined equilibrium concept, in a dynamic context. The General Theory is close to impossible to decipher, and IS-LM is static, with a load of hidden assumptions.

Krugman does not like Prescott economics, and thinks that was a failure too. And he states:
But the math was impressive, and RBC became a self-contained, self-replicating intellectual world.
This is quite a mischaracterization, as RBC was ultimately neither self-contained, nor self-replicating. People built on the basic neoclassical growth model in work in monetary economics, economic growth, optimal taxation, commitment to economic policy, heterogeneous-agent incomplete markets models, etc. Indeed, New Keynesian economics, which is Krugman's logic-checker, evolved directly from RBC models - basic Woodford is RBC + Dixit-Stiglitz + Calvo pricing.

Krugman finishes with this:
But the descendants of the Lucas project know that Keynes was wrong — it’s what their teachers and their teachers’ teachers have been saying all these years. They cannot accept anything resembling a Keynesian explanation without devaluing everything they’ve done with their intellectual lives.
Actually, the descendants of the "Lucas project" are serious economic scientists, who would not say anything so outlandish as "Keynes was wrong." I think modern macroeconomists in general think of Keynesian economics as a small part of macroeconomics - just another friction. Whether Keynesian ideas are useful or not is an open question. There are many holes in Keynesian economics that need to be filled before serious macroeconomists can in fact take Keynesianism seriously. But I don't think anyone finds Keynes threatening. That's just silly.


  1. Dammit. After reading Krugman I have realized that the last 75 years of economic theory have been a waste. What have I done with my life?

    That is what he expects us to think, right?

  2. "The Fed's response was definitely needed and the fiscal stimulus was probably fine too and also I voted for Obama"


    There are plenty of things where Krugman is non-serious, but when it comes to talking about RBC or anything out of Chicago/Minnesota in the 70s/80s he's completely, entirely, insanely non-serious.

    That being said, the Lucas interview was rather odd. It was pretty random and incoherent, and i'm not sure if that's the fault of Lucas or if the journalist is simply bad at giving interviews and then translating them into a coherent story.

    From what I can piece together though I think his view of the recession/current situation is that the initial reaction to the recession by the Fed (definitely) and Obama (maybe) were proper and reasonable, but we're past the time where traditional stimulus can do anything and it's time to start looking at long-term growth strategy. Eg. boosting the labor factor and capital accumulation by tax incentives and reform, regulatory reform to boost efficiency/TFP and all that.

    Which seems like a fairly simple thing and is probably close to the "accepted" view among modern macroeconomists.

  3. And Quiggin has a response up to your paper on his book:

    It's... rather short.

  4. When Krugman tries to dissect modern macro, it always end in an embarrassing train wreck. I just don't understand how this man has won a Noble Prize in economics, but doesn't understand modern macroeconomic models. I understand these things better than him, and I'm a nobody. It's an embarrassment to the profession, especially since his prize makes people unfamiliar with the literature think he is saying something insightful. Any day now I'm just waiting for him to renounce his own work in international trade and economic geography, because it's not ad-hoc enough - maybe he'll return his Nobel too.

    I picked up on a few other interesting claims. First, he claims NK economics went back to "ad-hoc assumptions about wages and prices, with a bit of hand-waving about menu costs and bounded rationality." I don't see any of the NK literature putting much stock in menu costs, which aren't necessarily ad-hoc depending on how they are modeled. They use the even worse ad-hoc Calvo model, which doesn't even take a stance on why prices and / or wages are fixed. And what NK-type is using bounded rationality? Most are using that evil rational expectations stuff. Sometimes you'll see some people using econometric learning approaches in place of rational expectations, but I'd hardily call these "ad-hoc" - nor is this approach exclusively used by Keynesians. I think this just further confirms he knows nothing about the literature he praises. Secondly, he claims that RBC models are ridiculous because they imply that the unemployment was voluntary. This implies he think NK models (at least the few that have unemployment in them) have done better. All of the ones I've seen imply it's better off to be unemployed since your consumption is constant, but your leisure rises - the Great Recession becomes the Great Vacation. It goes without saying search-and-matching frictions aren't even considered. Thirdly, his claim that this recession is clearly not a "technology shock." Fair enough, but this doesn't seem like a compelling support for NK models - because they use equally unreasonable shocks (including technology shocks!). Usually when NK-types try to generate "demand shocks" they do it by introducing a temporary shock to the representative agents discount factor. If I thought about this literally, I'd say the recession had nothing to do with the financial crisis and the housing bubble, everybody just woke up one morning and had a temporary desire for increased savings for no apparent reason. I'm trying to figure out why this shock to patience explanation is somehow more plausible than a technology shock explanation. Now, most people would say that these preference shocks are stand-ins for some more realistic shock. But then why can't RBC economists claim that technology shocks are stand-ins for something else as well?

    I've basically come to the conclusion that Krugman has read only two papers in macroeconomics - Hicks (1937) and Blanchard-Kiyotaki (1987).

    As a side note, aren't the NK-types starting to bring back signal extraction models anyway? I'm thinking of the recent work by Chris Sims and Mike Woodford on signal extraction for pricing dynamics, and the somewhat related "sticky information" literature. While they look different that the Lucas approach, they are pretty clearly drawing on Lucas' idea (especially Sims). So, is Krugman completely sure that the idea of signal extraction was completely bankrupt since some of his Keynesian friends seem to think there might be something to the idea.

  5. Scuttlefish,

    1. Yes, I wasn't going to say much about the Lucas "interview" piece in the WSJ. That just seems badly written, and the interviewer did not know what questions to ask.

    2. I saw Quiggin's response. He would rather philosophize about rationality rather than discuss economics, apparently.


    Yes, I agree with most of that. Krugman is confusing the menu cost literature with NK - entirely different stuff. As you point out, Lucas's influence is all over the place. Ideas like the Lucas critique permeate all of macro post-1970s.

  6. "Any day now I'm just waiting for him to renounce his own work in international trade and economic geography, because it's not ad-hoc enough..."

    Coffee. Keyboard.

  7. "But I don't think anyone finds Keynes threatening."

    Libertarians and Republican politicians do because, like a lot of other economics, it says sometimes government economic action can be good, and that goes against their ideology.

  8. Robert Haugen, finance professor from the University of California, Irvine wrote:

    "Chaos aficionados sometimes use the example of smoke from a cigarette rising from an ashtray. The smoke rises in an orderly and predictable fashion in the first few inches. Then the individual particles, each unique, begin to interact. The interactions become important. Order turns to complexity. Complexity turns to chaotic turbulence...

    How then to understand and predict the behavior of an interactive system of traders and their agents?

    Not by taking a micro approach, where you focus on the behaviors of individual agents, assume uniformity in their behaviors, and mathematically calculate the collective outcome of these behaviors.

    Aggregation will take you nowhere.

    Instead take a macro approach. Observe the outcomes of the interaction – market-pricing behaviors. Search for tendencies after the dynamics of the interactions play themselves out.

    View, understand, and then predict the behavior of the macro environment, rather than attempting to go from assumptions about micro to predictions about macro..."

    -- The New Finance, 3rd edition, page 132.

    So what do you think of the idea that if you want to understand the smoke cloud (entire markets or the entire economy), how smoke clouds behave, how to control them well, how to predict their actions, then it is useful to just study the cloud as a whole and utilize stylized facts based on cloud observation in your model, instead of exclusively predicting its behavior from how individual molecules (complex individuals) work, and having to make some very strong simplifying assumptions on their behaviors so that the math is tractable?

  9. That WSJ interviewer was indeed terrible. I hope they release the full transcript so we can see all of what Lucas said and was responding to.

  10. The Minneapolis Fed Region publication has a good series of interviews. They are basically verbatim question-and-answer, probably with a light edit. You can find them here:

    There is a very old one with Lucas in there (1993):

  11. This is for Richard, from Lucas's 1993 interview:

    Region: The Chicago School was known for its orientation toward free market thinking. Is that still part of the Chicago School? Is that the essence of it?

    Lucas: I would have a hard time taking a given political issue going down the list of my colleagues and predicting who stands where on any given issue. It's not that political a group and a lot of people, particularly younger people, haven't really written or spoken much on political issues.

    With that said, I guess Chicago has a pro market bias or maybe a better way to put it is just a skepticism about the efficacy of government programs. The beauty of neoclassical economics is that it's not a revolutionary all-or-nothing kind of thing. It s a reformist line of thought so that you really have to take issues one at a time. I think most people here would agree that, for example, medical insurance in the United States is in bad shape and needs attention and they respect the fact that the new administration is facing up to that. Whether or not they're going to face up to it in a way that I like or some of my colleagues like, we'll have to wait and see. You've really got to take it issue by issue.

  12. So, Richard, modern macroeconomics is not ideological. Modern macroeconomists are serious scientists. The Keynesians take ideas from the non-Keynesians (as Woodford, Gali, etc., have) when they are good, and vice-versa. Politicians practice politics. Sometimes, economists aid politicians, and then they practice politics too. You just have to know when an economist is practicing economics and when he/she is practicing politics.

  13. Stephen, why do you keep advertizing Krugman's columns? You could have written about a very interesting interview with Lucas in the WSJ, but you chose to focus on a totally uninteresting blog post by Krugman.

  14. This is just part of my job. I know it's uninteresting in that we learn no economics from Krugman's column. But Krugman is out there every day speaking to an audience who think he is an expert in macroeconomics, and if I can disabuse some of those people of that notion, that would be a success. Further, I think we have agreed in the discussion above that Lucas's interview piece in the WSJ was actually crappy. I would have written more about that if it was any good.

  15. Steve,

    Did you see this interview with Peter Diamond by the WSJ?

    I had higher hopes for Diamond, but after listening to these statements, I am not sure it was such a bad thing that he didn't get confirmed for his appointment at the Fed. Among the things he says (and these are careful paraphrases):

    "Of course we need fiscal stimulus..."

    "Monetary policy should not narrowly focus on inflation. The labor market is important too."

    Why do we need fiscal stimulus? Why should monetary policy focus on the labor market? What framework is he working from?

  16. I watched a bit of the Diamond interview. What he said is consistent with my views here:

    Diamond sees what the Fed committed to at the last FOMC meeting as just more needed monetary stimulus. He clearly has no idea why it should have any effect. Further he seemed to think that the Fed was moving "one step at a time," which was certainly not the nature of its decision at the previous FOMC meeting, where they committed to a 0.25% interest rate on reserves for two years into the future. Yes, I think the world is better off with emeritus MIT professor Diamond than Fed Governor Diamond.

  17. Re: Krugman and Lucas
    I thought Krugman's compressed version of RBC was pretty good. You should not dismiss him as another NK crank. Just read David Laidler (old school monetarist) who post similar claims about thirty years with unrealistic GE assumptions (p. 14, The Monetary Economy and the Economic crisis). I know many of you will dismiss him as out of date, but there are many out there feeling the same way about GE models wo financial sector.
    True, the Woodford NK style economics does not dominate, but when was dominance a criteria for relevance. Even the IMF's independent Evaluation office report on the financial crisis criticizes the fund for "group think". Then it is a bad defense to argue that another school (NK) is doing no better. Neither shock to preference nor technology shocks can explain the current crisis. Read Perry Mehrling's History of Monetary theory in the US from 1920-70 of Alan Young and Edward Shaw, or Irving Fishers Econometrica article on debt deflation from 1933 and use that as your stepping stone for relevant research. They at least were cognizant about the inner working of an unstable monetary economy and try to do something about it.

  18. Its always the same from the defenders of Lucas - he wasn't quoted correctly, the reporter didn't understand. Please...

    As for all the guys claiming they understand macro better than Krugman, I'd love to see you in take him on in a debate. I think your bravado would evaporate pretty quickly.