Winner of the New Statesman SPERI Prize in Political Economy 2016


Thursday 27 August 2015

The day macroeconomics changed

It is of course ludicrous, but who cares. The day of the Boston Fed conference in 1978 is fast taking on a symbolic significance. It is the day that Lucas and Sargent changed how macroeconomics was done. Or, if you are Paul Romer, it is the day that the old guard spurned the ideas of the newcomers, and ensured we had a New Classical revolution in macro rather than a New Classical evolution. Or if you are Ray Fair (HT Mark Thoma), who was at the conference, it is the day that macroeconomics started to go wrong.

Ray Fair is a bit of a hero of mine. When I left the National Institute to become a formal academic, I had the goal (with the essential help of two excellent and courageous colleagues) of constructing a new econometric model of the UK economy, which would incorporate the latest theory: in essence, it would be New Keynesian, but with additional features like allowing variable credit conditions to influence consumption. Unlike a DSGE it would as far as possible involve econometric estimation. I had previously worked with the Treasury’s model, and then set up what is now NIGEM at the National Institute by adapting a global model used by the Treasury, and finally I had been in charge of developing the Institute’s domestic model. But creating a new model from scratch within two years was something else, and although the academics on the ESRC board gave me the money to do it, I could sense that some of them thought it could not be done. In believing (correctly) that it could, Ray Fair was one of the people who inspired me.

I agree with Ray Fair that what he calls Cowles Commission (CC) type models, and I call Structural Econometric Model (SEM) type models, together with the single equation econometric estimation that lies behind them, still have a lot to offer, and that academic macro should not have turned its back on them. Having spent the last fifteen years working with DSGE models, I am more positive about their role than Fair is. Unlike Fair, I wantmore bells and whistles on DSGE models”. I also disagree about rational expectations: the UK model I built had rational expectations in all the key relationships.

Three years ago, when Andy Haldane suggested that DSGE models were partly to blame for the financial crisis, I wrote a post that was critical of Haldane. What I thought then, and continue to believe, is that the Bank had the information and resources to know what was happening to bank leverage, and it should not be using DSGE models as an excuse for not being more public about their concerns at the time.

However, if we broaden this out from the Bank to the wider academic community, I think he has a legitimate point. I have talked before about the work that Carroll and Muellbauer have done which shows that you have to think about credit conditions if you want to explain the pre-crisis time series for UK or US consumption. DSGE models could avoid this problem, but more traditional structural econometric (aka CC) models would find it harder to do so. So perhaps if academic macro had given greater priority to explaining these time series, it would have been better prepared for understanding the impact of the financial crisis.

What about the claim that only internally consistent DSGE models can give reliable policy advice? For another project, I have been rereading an AEJ Macro paper written in 2008 by Chari et al, where they argue that New Keynesian models are not yet useful for policy analysis because they are not properly microfounded. They write “One tradition, which we prefer, is to keep the model very simple, keep the number of parameters small and well-motivated by micro facts, and put up with the reality that such a model neither can nor should fit most aspects of the data. Such a model can still be very useful in clarifying how to think about policy.” That is where you end up if you take a purist view about internal consistency, the Lucas critique and all that. It in essence amounts to the following approach: if I cannot understand something, it is best to assume it does not exist.


32 comments:

  1. "Or, if you are Paul Romer, it is the day that the old guard spurned the ideas of the newcomers, and ensured we had a New Classical revolution in macro rather than a New Classical evolution."

    I think Romer can sense that the days of RE are numbered and he is flaying about trying to find an explanation for why a promising new development in economics went off the rails. It's like he is mourning a still birth. He senses that last 40 years have been lost, more or less. But what he won't face up to is the fact that Lucas and Sargent purposefully set out to garotte Keynesianism. And more subtlely, he will not face up to the fact that what largely drove L & S was their underlying value system, which I suspect he subscribes to - this is the reason he has not dealt with this directly and has endeavoured to find a scapegoat in Solow.

    An extract from a letter written by Lucas after the FRB of Boston conference to Herschel Grossman appearing in a paper by Da Silva puts a different complexion on the situation. Lucas writes:

    "Tom and I were trying to jog people into an awareness that the cozy intellectual world of the neoclassical synthesis has gone the way of Queen Victoria's."

    (The Establishment of Robert E. Lucas Jr.'s Macroeconomics of Equilibrium in the 1970s - D. F. R. Da Silva - March 2014 - p.18)


    Lucas clearly admits he was out to provoke a reassessment by the adherents of the Neoclassical Synthesis.

    The agent provocateur here is Lucas, not Romer's Solow.


    Henry

    ReplyDelete
    Replies
    1. I think you are right about what Lucas was trying to do, but I do not think that is enough to show Romer is wrong. Lucas and Sargent's criticisms were major, and so in my view required a major change in what the then consensus was doing (on expectations for example). That could have led to evolution, but Romer argues traditional resistance led to revolution. I'm not saying Romer is correct, but its a difficult issue to resolve.

      Delete
    2. When I read "After Keynesian Macroeconomics" I am left with the impression L & S were out for blood, Keynesian blood. They were in a take no prisoners, uncompromising mode. They were driven by ideology and the powerful urge to make economics "scientific" by adopting a highly formalized mathematical approach. It was they that were seeking revolution. What Solow or anyone else might have said or done was irrelevant.

      Delete
    3. Sorry to keep banging the drum on this but if there is any doubt Lucas was baying for blood see the full extract from the letter on page 18:

      http://hope.econ.duke.edu/sites/default/files/The%20Establishment%20of%20Lucas%20Macro%20Danilo%20Silva%20CHOPE.pdf

      He said: "...it is not clear what value subtlety will bring...."


      Henry

      Delete
    4. Paul Romer has now devised a new paradigm for explaining why L & S went off the rails - they were doing "bench science" (and everybody else was doing clinical science). He should read Lucas' letter referenced above - L & S were forcing a palace coup, aiming to supplant Keynesianism - test tubes nowhere to be seen!

      http://paulromer.net/the-clinical-bench-science-distinction-in-macro/

      Henry

      Delete
    5. Henry: I think we are reading the same texts, but from different perspectives. Lucas and Sargent were correct about the need to use rational expectations as standard, and the need to start investigating microfounded models. They may have been wrong that their criticisms of traditional methods were fatal, but they were pretty serious nevertheless.

      If they genuinely felt that these points were being ignored by the establishment, then in my view they had every right to use the language they did. The point to focus on is that their criticism were not the fatal critique they believed it was, and to which I would guess 90% of academic macroeconomists under 50 still think it is!

      Delete
    6. "If they genuinely felt that these points were being ignored by the establishment, then in my view they had every right to use the language they did."

      When have you ever seen such a diatribe in a serious economic paper? Romer at one point blames Solow for L & S going off the rails. Clearly the evidence is that L & S set out to vanquish the opposition (the establishment) not just attract their attention. (And their hubris remains undiminished. :-) ) RE may have had a better hearing had they demonstrated its value and not set out to demonize Keynesianism. (I think they did this largely for ideological reasons.) I know you value RE and the microfoundations of macro - but I believe they did more damage to economics than good. And I would say so much for "the day that economics changed". Apologies for banging on about this.

      Henry.

      Delete
    7. No need to apologise. But is the debate over who was unkind too whom productive? What I find more interesting was why the new methodology won so comprehensively. There are so many reasons you can give (wanting to be like physics, the new generation wanting to start afresh, wanting to be like micro, inadequacies of then traditional models) it is difficult to know what was critical and what was not. Perhaps important if you want methodology to become less dominant.

      Delete
  2. Simon,

    A few questions.

    1) Did Ray Fair have a convincing response to Sims (1980)? I was under the impression that this fundamentally discredited the large scale SEM approach.

    2) Why would one use a SEM? Presumably for the purposes of pure forecasting it will invariably be outperformed by an atheoretical technique (e.g. I would guess ARIMAs and BVARs will both comfortably outperform SEMs). Since they are not Lucas Critique invulnerable, I assume they're not suitable for policy analysis. I know you've blogged about this many times, but I'm still unclear on what your argument is for why or when SEMs outperform.

    Cheers,
    Jonathon

    ReplyDelete
    Replies
    1. Why are the problems raised by the Lucas critique (which can be reduced by SEMs using rational expectations for example) worse than the problem of misspecification in every DSGE? SEMs, by being closer to the data, may be better than DSGEs for policy analysis - where has it ever been shown that this statement is false?

      Delete
  3. Without getting into the relative merits of SEMs or DSGE and whatnot (which I have no business discussing anyway), I think it's generally true in almost all modeling situations to try to avoid over fitting the model to past data with too many parameters because doing so tends to make the model great at "explaining" that past data while having less and less predictive power for future events. (Think about fitting a 99th order polynomial EXACTLY to 100 noisy data points, rather than fitting a straight line.... in many circumstances the straight line will be superior, both for interpolating results between past data points, and for accurately extrapolating out beyond the current set).

    ReplyDelete
    Replies
    1. Yes, but the question everyone is arguing over is essentially "What is the theoretically right order polynomial to use?" There is a LOT of room between first and 99th.

      Delete
    2. Agree with Jeff, but with this caveat. In SEMs there is a danger of over fitting, but this danger is almost non-existent in DSGEs. So SEM modellers have to be careful to avoid this danger.

      Delete
  4. Simon,

    The more I read your blogs the less I understand what you stand for. It's like grappling with a lively fish - one second you think you have him - one wiggle and he's gone.

    Do you think any kind of model could be so constructed to be able to predict episodes such as the GFC?

    Henry

    ReplyDelete
    Replies
    1. I prefer that to SWL being a dead fish.

      Delete
    2. I think I take that as a compliment, as long as I'm not inconsistent with myself. To answer you question, no. But the point I'm making here is not about predicting the crisis, but understanding its impact.

      Delete
  5. Much change, no progress
    Comment on ‘The day macroeconomics changed’

    ‘The Council of Nicaea repudiated Arianism and adopted the original Nicene Creed.’ Each belief system redefines itself from time to time and orthodox economics is no exception. Clearly, for others the specification of concepts that have been unacceptable ab initio is not such a memorable event.

    Keynes started macro with a repudiation of the classicals: “The classical theorists resemble Euclidean geometers in a non-Euclidean world ...” and pointed the way “... there is no remedy except ... to work out a non-Euclidean geometry.” (1973, p. 16)

    This Keynes did but his non-Euclidean axioms were not quite correct (2011a). The inconsistency, however, went unnoticed and Keynes’s foundational concepts became part of the neoclassical synthesis, which in turn lacked consistency because of the heterogeneity of the constituent parts.

    So, Keynes made a change, but no real progress, and the same happened with the neoclassical synthesis. Post-Keynesianism perpetuated the fatal logical mistake that was hidden in the formal groundwork of the General Theory (2011b). Lucas’s critique of the neoclassical synthesis was valid but his return to pre-Keynesianism, too, was change without real progress (2010).

    The fact that DSGE is applied in comprehensive econometric models does not prove too much and is analogous to the application of the faulty geocentric model: ‘The astronomical predictions of Ptolemy’s geocentric model were used to prepare astrological and astronomical charts for over 1500 years’ (Wikipedia). Ironically, just because of good empirical fit it took quite a time to find out that the underlying theory was false.

    To summarize the memorable day in 1978 from an objective viewpoint: macro has changed three times, but made no real progress. Macro will fail as long as it starts with these, or roughly equivalent, premises: “HC1 economic agents have preferences over outcomes; HC2 agents individually optimize subject to constraints; HC3 agent choice is manifest in interrelated markets; HC4 agents have full relevant knowledge; HC5 observable outcomes are coordinated, and must be discussed with reference to equilibrium states. (Weintraub, 1985, p. 147)

    No way leads from these behavior-centric premises to an understanding of how the actual economy works. After a long detour, macro is back where nothing really changed: ... there is no remedy except to make a genuine paradigm shift.

    Egmont Kakarot-Handtke

    References
    Kakarot-Handtke, E. (2011a). Keynes’s Missing Axioms. SSRN Working Paper
    Series, 1841408: 1–33. URL http://ssrn.com/abstract=1841408.
    Kakarot-Handtke, E. (2011b). Why Post Keynesianism is Not Yet a Science. SSRN
    Working Paper Series, 1966438: 1–20. URL http://ssrn.com/abstract=1966438.
    Keynes, J. M. (1973). The General Theory of Employment Interest and Money.
    The Collected Writings of John Maynard Keynes Vol. VII. London, Basingstoke:
    Macmillan.
    Quiggin, J. (2010). Zombie Economics. How Dead Ideas Still Walk Among Us.
    Princeton, NJ, Oxford: Princeton University Press.
    Weintraub, E. R. (1985). Joan Robinson’s Critique of Equilibrium: An Appraisal.
    American Economic Review, Papers and Proceedings, 75(2): 146–149. URL
    http://www.jstor.org/stable/1805586.

    ReplyDelete
  6. Spot on Simon - couldn't agree with you more. My background is similar to (though less distinguished) than yours. I used and built SEMs at the Treasury, OEF and the Bank (early '90s), but also promoted the use of DSGE in the Bank in the late '90s. Like you I think both approaches have much to commend them. I've learned huge amounts from both and really don't get the narrow-mindedness that says one is obviously superior to the other in all circumstances. Horses for courses and all that. My own preferred working-horse model is 5-equation, calibrated, IS-LM with flexible expectations formation and 2 asset prices. It isn't microfounded, but I know it could be and many of the 'structural' coefficients come from the SEMs I've worked with in the past. It would have done no better than any other approach in the run up to 2008, but using it always reminds you of its limitations. Also surely the criticism of Solow is over the top? It strikes me he's pretty open-minded in the sense I've described above - and the full text of his 1978 comment stands up very well today. He teased himself about his inability to take sides ideologically (can't remember the very amusing quote, but I think it's in his 1984 Hicks lecture). And anyone that annoyed both Milton Friedman and Joan Robinson in equal measure can't be all bad.

    ReplyDelete
    Replies
    1. It may also have been in the Hicks Lecture, but I think Solow's had a nice self-tease about his inability to consistently take ideological sides in his AEA Presidential Address. (Being retired, I no longer have access to the AER to check.) Almar.

      Delete
    2. ICYMI shamik

      You should have realized long ago that the IS-LM model is logically defective. See the blog post ‘Mental messies and loose losers’
      http://axecorg.blogspot.com/2015/07/mental-messies-and-loose-losers.html

      or the working paper ‘Mr. Keynes, Prof. Krugman, IS-LM, and the End of Economics as We Know It’
      http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2392856

      Delete
  7. Simon - you don't comment on the stock-flow consistent sectoral models by Godley & Lavoie. What is your view on these?

    ReplyDelete
    Replies
    1. Good point. I say if a model is not stock flow consistent it is a waste of time!

      Delete
    2. That is out of SW-L's paradigm so he can not comprehend it let alone discus it.

      Delete
    3. They are certainly within my paradigm, but just not very interesting I'm afraid. (You should remember I go back a long way!) DSGE models are stock flow consistent by the way. To be more positive, I think the key issue is how wealth stocks feed back on flow decisions: simple DSGE models take a somewhat extreme, but theory based position on this, but I'm not sure I can find in Godley a clear alternative view - but do correct me if I'm wrong.

      Delete
    4. Why should it matter if it is "interesting"? I find Godley and MMT very interesting
      http://www.bondeconomics.com/2013/12/theme-my-discontent-with-dsge-models.html?m=1
      Criticism with DSGE here.

      Delete
    5. For one thing, Godley and Lavoie models can cope with endogenous money, which I don't believe DGSE models do. Their post-Keynesian approach replaces utility maximisation and RatEx with stock-flow norms, it has a properly modelled banking sector, inventories, and households and firms which begin each accounting period with a consumption or investment decision, proceeding to a portfolio decision as a residual. I humbly submit that it's worth some of your time to engage with. I can email the last Godley book if you don't have a copy.

      Delete
    6. I hate to be so blunt, but this comment just displays your lack of knowledge about modern macroeconomics. The relevant criticism of the standard New Keynesian model is that money is absent, not that it is exogenous. People complaining about exogenous money have a (legitimate) dispute with the textbooks, not NK models.

      DSGE models are stock flow consistent, so your statement really just reads: post-Keynesians take out all the behaviour but keep in the accounting! The real difference between most DSGE models and Godley is how consumers and other agents react to wealth, and here post-Keynesians have a good point. But as long as they continue to misunderstand the mainstream, they will I'm afraid be ignored (which is a shame).

      Delete
  8. "That is where you end up if you take a purist view about internal consistency, the Lucas critique and all that. It in essence amounts to the following approach: if I cannot understand something, it is best to assume it does not exist."

    I think this is unfair. They are not stating that one should neglect what one cannot model. They are talking about what any SINGLE model should try to accomplish.

    I think that that paper is somewhat misnamed. Their main criticisms are not aimed at New Keynesian models in general, but particularly at large-scale NK DSGE models like Smets-Wouters. Their main critique is not the inclusion of ad-hoc nominal rigidities, but the inclusion of numerous dubious structural shocks to better fit the data.

    I interpret their argument as similar to Caballero's point in his paper on the "Pretense of Knowledge Syndrome". The distinction they are concerned with is between models that accept that they are incomplete, and attempt to simply and cleanly capture a few features of the data, and models that attempt to explain ALL the data.

    ReplyDelete
    Replies
    1. I agree that an important issue is the scope of a model, but I disagree that this is what Chari et al are talking about. Chari et al seem to include in their scope saying useful things for monetary policy, which is all that NK models - however elaborate - are designed to do.

      Delete
  9. > "What about the claim that only internally consistent DSGE models can give reliable policy advice?"
    Well..., economic reality is internally inconsistent, because it consists of economic actors with conflicting interests and wants.
    Even individual economic actors are internally inconsistent, because they (we) have various psychological drives and wishes that are regularly at odds with each other.
    How can internally consistent economic models possibly give reliable advice on how to deal with that reality?!?

    The problem I see with econometry is that it can't deal with final explanations (from economic actors' purposes) but only with causal explanation (to some extent).
    Even though human behaviour is 90something% explainable causally (because that part of what drives it is unconscious), the few percent that we manage to make conscious make us human and enable us to defy the (natural, social and paradigmatic) patterns of the past.

    ReplyDelete
    Replies
    1. DSGE is just another perpetual motion machine
      Comment on Wim Nusselder on ‘The day macroeconomics changed’

      Your fundamental error consists in the commonsensical intuition that economics is first and foremost about human behavior. The representative economist simply cannot get his head around the fact that economics is about the behavior of the economic system.

      No way leads from the understanding of human behavior to the understanding of how the actual economy works. Standard economics is a manifest failure and the ultimate methodological reason is that it is built upon false premises.

      All theories/models that take one or more of the following concepts into the premises are scientifically worthless: utility, expected utility, rationality/bounded rationality/animal spirits, equilibrium, constrained optimization, well-behaved production functions/fixation on decreasing returns, supply/demand functions, simultaneous adaptation, rational expectation, total income=value of output/I=S, real-number quantities/prices, and ergodicity. All these items are economic nonentities.

      From an economist who accepts one of these nonentities nothing of scientific value is to be expected. DSGE is caught in the proto-scientific cul-de-sac. To discuss its technical details is not any better than to discuss the mechanics of just another perpetual motion machine. Real scientists refuse to do this because they know for sure that this thing cannot work in principle.

      Egmont Kakarot-Handtke

      Delete
  10. Dear Egmont,

    "Your fundamental error consists in the commonsensical intuition that economics is first and foremost about human behavior. [...] economics is about the behavior of the economic system."

    What is 'the economic system' if not 'systematic human economic behaviour'?
    (I define 'economic' as 'organising what we want/need', but other definitions are possible too.)

    ReplyDelete

Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.