Brad DeLong Blows a Gasket

In an all-too-common gesture of blog-civility Brad DeLong has decided that I’m an idiot because he thinks I’m asking for some sort of balance in viewpoints on economics regardless of the soundness or lack thereof of the views.  Put differently, it seems that Brad thinks that my earlier post was saying that we should grant some credibility to Prescott’s outlandish claims because Paul Krugman has said some outlandish things before. 

That is not at all what the earlier post was saying. The only substantive economic statement in the post dealt with the effects of monetary policy. Here’s exactly what I said regarding that statement: “Prescott is wrong.” Now, I realize that in any statement there is always some room for interpretation but there doesn’t seem to be a lot of room in that statement.  

The main point of my earlier post is that we should think a bit more before immediately turning to Nobel laureates for policy analysis. We should ask ourselves why we are soliciting their opinions. It could be that we turn to them because “hey, they won a Nobel Prize; let’s see what they have to say.” It could be that we turn to them because we think “oh, they have a Nobel Prize and therefore they are going to know better than most other economists about economic issues.”  I worry that some people think that the Laureates sknow better about all economic issues than people who haven’t won a Nobel Prize, or people who just work in the industry or in the private sector…. 

Now, unquestionably, when you consider the Nobel laureates, you are looking at some of the smartest people in our field.  If you want to get an opinion, talking to a smart person is definitely not a bad thing to do.  But it doesn’t mean that it is the best thing to do. Again, let me use Paul Krugman as an example.  I’ll try to be very careful to avoid offending Brad’s delicate sensibilities.  I’m not using Krugman as an example because I think his analysis is particularly flawed (or flawed at all). I’m using him as an example because he is very prominent in the popular press and because he finds himself in this position a lot. Krugman is asked to weigh in on all sorts of areas. For instance, Paul Krugman has said many things about health care policy. That’s fine. Paul is really smart and he knows a lot about economics.  Moreover, I think he tries hard to keep up to date on the details of current policies. You could do a lot worse than asking Paul Krugman for his opinion. That said, Krugman doesn’t really have any particular expertise about the economics of health policy.  (At least I don’t think he does.) There is nothing wrong with getting the opinions of a smart informed economist but that’s really all you are getting when Krugman starts talking about health policy. The same thing is true when he talks about the minimum wage. Krugman knows about the literature on the minimum wage; I’m sure he is very interested, as most economists are, in the research on the minimum wage but he’s not an expert on that area. In any case, he is the one who is asked to write about the minimum wage. Why? Because he has a Nobel Prize. 

Interestingly, in Brad’s post he inadvertently gives a very good example of what I’m talking about.  He presents an example of Robert Lucas talking to the Council on Foreign Relations on issues surrounding bailouts and the financial crisis (and also testifying to Congress about fiscal policy). Like Krugman, Lucas is also very, very smart.  In fact you could argue that Lucas has looked deeper into the field than any other economist in the latter half of the twentieth century. (There is really no way around it – Lucas’s work is incredibly significant to the development of modern economics.) In any case, there he is talking to the Council on Foreign Relations about bailouts. Why? Does he know anything about bailout policies? Not really. But then why is Lucas being asked to give his opinion?  Because he has a Nobel Prize.


18 thoughts on “Brad DeLong Blows a Gasket

  1. I would say that if are not able to provide one single example of a nutty statement from Paul Krugman, then your original post was not a balanced opinion.
    However, let me ask you one thing. Krugman defended that the government printed a trillion dollar coin to benefit from a legal loophole to circumvent the debt ceiling limit.
    Wouldn’t you agree that that is a rather nutty thing to argue?

    • LA-C views as nutty Krugman’s suggestion that the government print a trillion dollar coin to benefit from a legal loophole to circumvent the debt ceiling limit. I’m sure that if we lived in a rational world Krugman would agree that the idea is nutty. But we don’t live in that world. The reason Krugman proposed the idea is that the Congressional right uses the debt ceiling to extort concessions from the left that Krugman feels are very damaging to the economy. In that context, I don’t think the idea is nutty.

      • Did Krugman defend the idea? Reread what he wrote. He said that the government seriously considered the option, not that it was a good option.

      • I am replying to M.
        Here is one quote, there are others if you look for them:
        “Should President Obama be willing to print a $1 trillion platinum coin if Republicans try to force America into default? Yes, absolutely. He will, after all, be faced with a choice between two alternatives: one that’s silly but benign, the other that’s equally silly but both vile and disastrous. The decision should be obvious.”

      • This is a reply to Al
        Al, if that idea gains traction, that would imply that the president would have control over the monetary base. And, by the way, that would also imply that the next Republican president culd also manipulate the monetary base. That could be the beginning of the end of any idea of an independent Central Bank.
        Basically, the American government would be emulating the worst monetary examples in the world.
        This is an incredibly nutty idea and I find it hard to believe that Chris House disagrees with me.

  2. It’s extremely bizarre you’ve gotten so much criticism (or at least far more than I would expect) for your last few posts. I don’t think there’s been anything that would be regarded as cotroversial among the majority of economists

  3. DeLong tends to throw insults when he’s not either link-spamming or doing block-citations of other blog posts (the main reason I stopped following his blog), so I’d just take criticism from him with aplomb. It doesn’t mean much, compared to criticism from Krugman or somebody with knowledge on a particular issue.

  4. The nuttiness you have in mind seems partly to be the willingness to make decisive pronouncements on areas well away from one’s expertise. This has less to do with any specific utterance and more in (i) a delusional confidence in one’s pronouncements to the point where (ii) one accuses any who disagree of bad faith or (in your case)… “idiocy.” Krugman passes this test. And his horde, deep down inside, when it briefly pauses between attacks on his critics–real or imagined–knows it.

  5. I don’t think DeLong’s use of “idiot” is productive, but at the same time, I’m reluctant to support your ideas about Krugman, even as somewhat softened in this post.

    Krugman is not a “credentialed expert” on health care policy, but he has been following the issue for decades and seems to have a good sense of whom to listen to on the issue. His columns and his blog posts are not meant to be “wonkish” except in his specific areas of expertise (and he explicitly labels those posts as such to segregate them). Regarding health care, he has always been an aggregator and distiller, and that role is important. I can’t believe you don’t think it is critical to have people who can communicate in clear prose the issues that are at stake, especially because wonks and credentialed experts do not always have such skills.

    I care not that he lacks credentials in health care economics. His credentials (to me) are that he continues to write about the issue over and over and over again without embarrassing himself or saying anything “nutty”. That is the evidence that is important to me. Credentials and someone else’s legitimization can be left for those without the ability to think for themselves. I need them not.

    So Krugman is not a wonk in the areas you cite, and I’ll never be a wonk in those areas either. We still need to be able to find a way to efficiently navigate the world despite the inability to focus on a particular issue like a wonk does.

    I just think you picked a very, very poor example. An extremely poor, extremely unfortunate example. I do not always agree with Krugman. In fact, at times, he has struck me as stubborn and myopic (like everyone else can be), but you have basically created the mother of all false equivalences and further focus on it is detrimental to you (in my opinion). The thing is, you obviously have a point. A really good point. So focus on that instead and bury the false equivalence where it cannot be found again.

    • This comment basically nails it.

      I would add that although Krugmans nobel price added quite a bit to his fame, it has taken on a life of it’s own at this point. These days, people read what he writes because he is a very well-known economist who is good at explaining complex economics to non-economists and not because he won a noble prize.

    • Your point is well-taken. Communication is also one of Krugman’s strong suits. But there is no false equivalence in this post. There is no equating of anyone in this post. Banging on about false equivalence is just off-reservation and harkening back to a post three days ago. I’d hate to be one of your kids, you’d never let them ever move on from anything they ever said or did.

  6. Here’s the problem: most of the time some people are right and others are wrong.

    There exists a method to tell the difference that works in all but the most extreme cases, a method that is well known and accessable to all educated people willing to devote enough time and interest. The set of claims that are either genuinely subject to debate or impenetrable without special training are easily identified.

    Unfortunately, academics and journalists are committed to engaging in lengthy discourses that ostensibly have truth as a goal but have zero connection to the aforementioned well known method. This discourse is excruciating to behold and generates reactions like Brad’s. The Krugman is nutty claim is a textbook example of this discourse.

  7. Three points. First, your point about Nobel prize laureates not being renaissance scholars who know everything about economics and focusing on arguments instead of stupid prizes was crystal clear. Second, Brad’s point about your dubious attempt to appear balanced is also crystal clear.
    Krugman was simply empirically right during the last few years. Prescott or Lucas or any other Chicago economist who is in demand denial are simply wrong.

    “In fact you could argue that Lucas has looked deeper into the field than any other economist in the latter half of the twentieth century.”
    I guess field means recessions, short-run macro. Perhaps I am totally ignorant but as far as I know his only contributions to the field are rational expectations and the Lucas critique. The Lucas critique is ridiculous, why should technology or preferences be “deep parameters” that are stable over time? About rational expectations, to get something like the housing bubble (or any other bubble) which is financed via Ponzi scheme some people have to be fools.

    I know that this is basically a critique of mainstream New Keynesian DSGE macro and not merely Lucas. But you guys made a big mistake when you tried to appease the Chicago radicals via using their models and adding sticky prices to them. There is more useful (useful meaning being roughly in synch with what is happening at the moment) stuff in Hicks 1937 or Minsky 1992 than in any Woodford paper. The reason is simple, Old Keynesian or Post Keynesian writers are Keynesian, New Keynesian economist on the other hand are not. Keynes wrote explicitly in Ch.19 of his General Theory that more wage flexibility would worsen and not alleviate the problem.

    So mainstream economics has to either ignore the microeconomic causes of recessions like Solow, Tobin and Samuleson did or focus on the right ones: Stiglitz and Greenwald wrote a bunch of brilliant papers in the nineties which focused on money as credit (no liquidity preferences, no money magically appearing in the utility function) plus asymmetric information and the ensuing incentive problems in credit markets as the key inefficiency.

    • Macro is an evolving field. And one of the reasons why it takes so long to evolve is that every turn in the business cycle is just one data observation.

      The models prior to 1970 were Keynesian. But they couldn’t predict the supply side shocks that flowed from the increase in oil prices nor the collapse of the Phillips Curve. Lucas, Sargent and Co explained that with a combination of Rational Expectations and the Lucas Critique. But that was only one data observation. Now, we have this Lesser Depression. And it spurs another research agenda (financial frictions, bank runs), but again it is only one data observation.

      Minsky got the trigger of this Lesser Depression right. But his model is completely inapplicable to so many other recessions. Even he, with all his vision, only explains one data observation.

      As macroeconomists, we inch forward. And we do so by learning from experience. And the single greatest lesson Lucas gave all economists is the way to incoporate what we have learned, what we observe into a standard discipline; a consolidated method. Into a way of thinking. The framework: model the optimal actions of agents under constraints and subject to some market clearing conditions.

      Now, this is where we start. It is the benchmark. And we depart from Rational Expectations (Learning, Bounded Rationality, Adaptive Expectations) and the Lucas Critique (ad hoc innovations such as Calvo Pricing) and we include frictions (real and nominal) and more and more agents to see whether we can get closer to the way the world is.

      That is why economist have massive reverence for Lucas. Not for policy pronouncements. Because he was the beginning. He gave us the toolkit and the instruction manual. And, yeah, we have a long way to go.

  8. Lucas wasn’t the beginning of macro, Hume, Wicksell and Keynes were. Treating a guy / the work of a guy who thought ten years ago that the “central problem of depression-prevention has been solved”, who wanted us to unlearn everything about short-run macro that is relevant, with reverence is ridiculous.

    “The framework: model the optimal actions of agents under constraints and subject to some market clearing conditions.”
    In a recession output and the labour markets do not clear and during the housing bubble not all agents behaved rationally, some were stupid (there is no Ponzi finance unless some folks make mistakes).

    I do not mind DSGE with all the changes you mentioned. But I mind the attitude that DSGE should be the only model. Why should we only use a framework which, in its basic form, cannot describe any short-run macro phenomenon of interest? What is wrong with some model variety? “One model to rule them all” seems hardly in the economic spirit of competition.

    Besides the empirical problem of the basic DSGE model there are theoretical problems.
    The Arrow-Debreu general equilibrium model is anything but robust. Leijonhufvud has shown that without the magical Walrasian auctioneer trade at disequilibrium prices can lead to a collapse of demand, if there are increasing returns to scale the model does not work and Stiglitz and Greenwald have shown that the two welfare theorems do not hold when there are incentive problems due to asymmetric information (which is basically always the case).

    Basic DSGE only incorporates one market failure, Calvo pricing as a short form of the Blanchard/Kiyotaki style demand externality due to monopolistic competition which leads to price rigidity.
    Again, I am not arguing against this approach but for more openness of mainstream macroeconomists towards models which do not start with DSGE and then add frictions (not to mention that it is technically incredibly difficult to model several frictions at the same time; basic DSGE does not even one friction, model monopolistic competition. directly but uses the short-cut of Calvo pricing).

    • Keynes Still Matters:

      When I wrote Lucas is the beginning, I meant he was the beginning of a common language. Not of macro. I agree with you. Macro has its origins much earlier in all the guys you stated. But it had less formal structure. Less of a consistent structure. Reading Keynes is hard because of it. There is definte value there, but some aspects of the General Theory are debated and contested and it took Hicks to put some structure around it. Lucas was the beginning of the common language.

      You cherry pick a Lucas quote. And he was wrong. But everyone makes mistakes. Krugman has been wrong here and there (the Euro), and has at times contradicted in newspaper articles what he writes in his own textbook (the minimum wage). But I am far less willing to write off people of massive intellect and value like both Lucas and Krugman for an error of judgement now and then, or a slip of the tongue, or a moment of misplaced exuberance, or for the sharp rebukes they may dole out to other economists.

      Paul Krugman and Brad DeLong don’t seem to like Bob Lucas. I don’t know why. But I think that he deserves a great amount of respect for the way he revolutionised the profession. I also think that Krugman and DeLong deserve respect; Krugman for revolutionising trade theory and Delong for his papers on noise traders.

      Now, most importantly, we are talking at cross purposes. I want to be very clear here. The modeling framework is a baseline. A benchmark. By making changes we can study how the model reacts when the real world is not consistent with the features of the model. Basically, what happens when we relax assumptions.

      So, DSGE isn’t A model. It is a class of models. A framework. I think you may be confusing DSGE (which is broad – so many models are dynamic, stochastic general equilibrium) with RBC, which was one of the original versions of the model. But everything has moved on from there. So there is no one model to rule them all. Far from it. DSGE models are a class of models which give us insight into the mechanisms at play in the real economy. To what may occur when shocks hit the economy. But, a caveat, I would never profess to them ‘ruling’.

      One version of this class of models matched the data from the recent Lesser Depression pretty well (Del Negro, Giononi and Schorfeide – written up accessibly below by Noah Smith. So it is not correct to say that they ‘cannot describe any short-run macro phenomenon of interest’.

      Almost everything you wrote in the second paragraph of your reply paragraph is covered by my fifth paragraph. Note the different ways of forming expectations (Learning, Bounded Rationality, Adaptive Expectations etc) that permit departures from Rational Expectations. A way of forming expectations is necessary in a forward looking model because it permits agents to look forward in time and make calculations based on what they infer. I should have added innovations that do not rely on market clearing have also been incorporated – search and matching theory. This is a tractable departure from the market clearing assumption and one that has helped versions of DSGE models match markets, such as the labour market, much better.

      I don’t know anything about Leijonhufvud and his criticism of the Walrasian auctioneer, but I will look it up and I thank you for the tip.

      When the welfare theorems fail, it means two things. One, the competitive market outcome is not pareto optimal and two, a pareto optimal outcome does not correspond with a competitive market outcome. This is not a failure of the framework. This is an important insight from the framework. It basically means that there is a role for government in obtaining a pareto optimal outcome. In this model, government can step in and make everyone better off. This is only a problem for people who want to argue that there is no role for government. So you are confusing a result for a criticism.

      I dispute that ‘the Arrow Debreu Model is anything but robust’. Many economists think that the complete markets frework is too unrealistic assumption, but that the real world resembles something like the ADM model minus a few markets (ie incomplete market structure). For a reference on this, see the new book by Athreya. There are also DSGE models with incomplete markets that incorporate this (I forget the reference a tthe moment, but I think one example was a modelk by Aiyagari in 1994.

      There are a multitude of market failures in DSGE models aside from price stickiness. There are real and nominal frictions (including wage stickiness), financial frictions, bank runes, zero lower bounds and others. So this class of models is capable of capturing market inefficiencies. Again, it is not correct to say the only friction is Calvo pricing and monopolistic competition. Sure, this is the case in the baseline New Keynesian model (as propounded by Gali), but models are much richer now. They capture much more.

      I agree with you that openness to new techniques, or even the older techniques, is important. I myself use both Keynesian macro and microfounded macro.

      I guess my main point is that DSGE captures so much. It is far more sophisticated than what is represented in your reply. And there is a long way to go. Many improvements and adjustments to be made. Many wrong paths to take and then backtrack up and many things to learn. But DSGE has already traversed so much, and that all started with Lucas.

  9. Thanks for your long and well-argued response. I am in agreement with much of what you wrote.

    Due to my lack of familiarity with all the varieties of DSG models I have a question, how good are they ar incorporating several frictions at once? All I know are some basic DSGE models and they often use Calvo pricing as a short-cut for monopolistic competition and the resulting aggegate demand externality.

    Suppose it is indeed difficult, labour-intense or not feasible to include several market failures in a DSGE model.
    Why not do something less elaborate in the meantime like focusing on one market failure? The benefit are that you are able to actually do something, the disadvantage is that you cannot analyze the interplay of this particular market failure with other markets.

    Suppose I am wrong and it is actually possible to add all kind of frictions to RBC. First there is the triade of imperfect competition, externalities and incentive issues due to asymmetric information. Then one has to relax the homogenity assumption and undo the representative consumer and the representative firm (and yet it would still not be guaranteed that aggregration problems vanish!). Then one has to undo rational expectations with the stuff you mentioned, learning, bounded rationality and so on (I don’t know the literature but given that pre-Lucas adapative expectations meant that you can fool all the people all the time and that rational expectations mean that you cannot fool anybody my gut feeling is that this “third way” is promising; intuitively it is straightforward that during the housing bubble some people were stupid but afterwards they learned that housing prices cannot rise forever, that Ponzi financing isn’t stable).
    The resulting GE model would not resemble RBC anymore so why start with Arrow-Debreu in the first place?

    I have to strongly disagree with your claim that Arrow-Debreu is theoretically robust. Imperfection competition, non-convexities or a little bit of asymmetric information (Stiglitz’s seminal papers made this point over and over again) and it collapses. About empirical robustness, I guess we aggree that one cannot explain real world phenomena like involunatary unemployment, advertisement, the existence of financial intermediaries and so on with the ADM.

    “Macro has its origins much earlier in all the guys you stated. But it had less formal structure. Less of a consistent structure. Reading Keynes is hard because of it. There is definte value there, but some aspects of the General Theory are debated and contested and it took Hicks to put some structure around it. Lucas was the beginning of the common language.”
    My last methodological point is that seemingly less elaborate theoretical work is often more illuminating. You are totally right that pre-Samuelson macro writings are often vague and imprecise and I am all for formalizing an argument via the help of mathematics in order to be able to see easier whether it is sound or not. But one can exaggerate formalism. When a subject is difficult it might be (so far) impossible to approach it via the “common language”, when your understanding of something is vague you cannot pretend that you have a clear solution of it.

    I am not aware of any modern paper which provides a better understanding of the problems of public debt and deflation than Fisher’s. You mentioned Hick’s formalization of Keynes’ GE so let’s take a look at its evolution. If you take IS-LM and endogenize prices you get AD-AS. Rising prices have a positive effect upon supply because they decrease the real wage and falling prices have a positive effect upon demand because they increase real balances.
    The model works perfectly fine but it is totally wrong. And to understand why the Pigou effect and why deflation is dangerous you must have read your Fisher. The model alone does not suffice.

    To sum up this long post, I am for micofoundations, DSGE and GE in general. But I am also for “quick and dirty back of the envelope” style models, prosaic economics and constant awareness of the limitations of models (which means constant awareness of all the implicit assumptions of a model). And last but not least I value external consistency higher than internal consistency. A Woodford paper is beautiful but the world is a mess and a Richard Koo paper might make more sense of this mess.

    • Correction, Fisher wrote about private debt.
      While we are at it, I only saw DeLong made the obvious point that QE reduces the riskiness of private asset and increase the riskiness of public assets. This is the kind of “quick and dirty” thinking I advocate. If you cannot model these balance sheet issues at least think about it (Koo isn’t doing theory and I know too little about Post Keynesian literature to tell whether it includes theoretical modeling of debt / balance sheet issues).

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s