Theorising in science: theorising in economics

Robert Waldman has a fantastic critique of Paul Romer’s recent missives on economic science. He’s commenting ultimately on why Lucas’s work isn’t such a breakthrough. In it he highlights something of immense importance. It’s hard to think of many developments in economic theory in recent decades to which his observations do not apply. His point is that ‘advances’ in theory tend to be advances in conventions of professional practice, rather than new insights into the world.

General Relativity explained an anomaly … the precession of the perihelion of Mercury… predict[ed]… how much gravity caused light to curve… has yielded a huge number of predictions which fit the data exactly… was easily modified to correspond to an expanding universe…. Physicists are quite sure general relativity is not the truth (because it is inconsistent with quantum mechanics and therefore a lot of data). But it is a very empirically successful theory.

In contrast, the Dixit-Stiglitz example did not attempt to explain anomalies. … The aim was to make models with imperfect competition tractable… an example…. They made a modelling choice… neither would guess that people might actually have Dixit-Stiglitz preferences…. It meant there was a standard way to handle imperfect competition…. [But] there are no general results…. Together the assumptions of imperfect competition and Nash equilibrium imply almost nothing….[Dixit-Stiglitz] made it possible to have the illusion that economic theorists understood imperfect competition, but this was discovery by assuming we have a can opener. The example was fruitful because, once a lot of people decided to explore the same special case, they could discuss its interesting behavior…. Theory can grow if people agree on core assumptions. This is progress if the assumptions are useful approximations. Once a field of economic theory has developed, its core assumptions are no longer vulnerable to data. I do not think the the development of a new branch of theory is necessarily scientific progress.

I think economic theory was massively improved by the Dixit-Stiglitz example because it made economists outside of industrial organization willing to consider imperfect competition. But… it gave the impression that there were simple elegant results based on assuming imperfect competition similar to those based on assuming perfect competition. There aren’t….

A radically new hypothesis which turns out to be totally false was not a contribution to scientific progress. Finding out that it was false was and such dead ends are inevitable in science. But in science development of new theory is not necessarily progress. OK, what about the Solow growth model and the Lucas supply function?… Solow… made a large number of extreme assumptions yielding a tractable model… that… fit the data surprisingly well… [in spite of] excellent arguments for why one shouldn’t be able to treat capital and labor as scalors…. There was no reason to think that the concept of disembodied technology would be useful. But it helps economists fit the data.

Lucas formalized an argument about price level misperception and fluctuations which had been made many times (for example by Keynes in The General Theory as Tobin explained to Lucas in 1971)…. Lucas definitely did not identify a previous conceptual error of treating the expectations unaugmented Phillips curve as a stable relationship–this is a myth. He added a focus on expectational errors alone and the insistence that economists assume rational expectations….

Unlike Dixit-Stiglitz… in 1972 the Lucas model was obviously grossly false – it implies that output is a white noise and it was well known that economic fluctuations aren’t. It requires that agents have very limited information on the price level when, in fact, they have a lot of information. I think it was clear that Lucas’s new research program would be sterile. I think it was entirely sterile…. I certainly don’t think this of the Solow growth model or imperfect competition with a Dixit-Stiglitz preferences or a Dixit-Stiglitz aggregator. Yes, Lucas 1972 was path breaking, but the new path Lucas blazed lead to a dead end.

It seemed to me that ‘New’ or ‘Strategic Trade’ theory was similarly a dead end. Indeed it was a dead end that John Hicks had warned that assumptions of imperfect competition would be in Value and CapitalHere’s Hicks:

[In the presence of scale economies,] [t]here must be something to stop the indefinite expansion of the firm, but it can just as well be stopped by the limitation of the market as by rising marginal costs. . . . It is, I believe, only possible to salvage anything from this wreck – and it must be remembered that the threatened wreckage is that of the greater part of economic theory – if we can assume that the markets confronting most . . . firms . . . do not differ very greatly from perfectly competitive markets. . . . At least this get-away seems well worth trying. We must be aware, however, that we are taking a dangerous step, and probably limiting to a serious extent the problems with which our subsequent analysis will be fitted to deal. Personally, however, I doubt if most of the problems we shall have to exclude for this reason are capable of much useful analysis by the methods of economic theory.

Given the canonical state of this passage and the book they come from, and the fact that it was these concerns that hung over trade theory and justified the almost ubiquitous assumption of perfect competition in formal models (along with the rider that scale economies were important and should not be ignored, but couldn’t be easily handled using formal theoretic methods) you’d think that the new agenda of strategic trade theory might have addressed Hicks’ concerns. That is you’d think that, in constructing the new field, its pioneers would have addressed this question – would have explicitly argued that what they were doing was not, now a waste of time. Certainly there were new techniques, but you’d hope they’d give some thought to their likely utility. I know this is very verbal and all, not conveyed in an equation, but that’s what you’d expect if their energies were to be deployed rigorously. The alternative is what I call the mountaineer’s approach to the utilisation of theory.

Q: Why did you rebuild trade theory around the Dixit-Stiglitz model?

A: Because it was there.


This entry was posted in Economics and public policy, Philosophy. Bookmark the permalink.

14 Responses to Theorising in science: theorising in economics

  1. derrida derider says:

    “You’d think that in constructing the new [Strategic Trade] field, its pioneers would have addressed this question – would have explicitly argued that what they were doing was not, now a waste of time”
    As I’ve noted here previously, at least one – Krugman – did. See

    • Nicholas Gruen says:

      DD, you’ve cited a long piece. Can you please point to the bits of it you think bear out your claim.

      • derrida derider says:

        The piece uses the history of development economics to talk about “the evolution of ignorance” – that as a field of study is formalised we lose insight through the incredibly narrow focus formalism gives, but ultimately this is worth it for the improved quality of insight we get. The relevance here is that the canonic example he gives – the (mathematical) incorporation of scale economies into formal models of development based on inter-industry intra-national trade – is exactly the same one as those he used in his strategic trade theory of intra-industry inter-national trade. And he makes the claim about method explicitly about both fields (oh, and meteorology) too – formal but “silly” models “that do violence to the rich complexity of the world” were extremely useful in advancing knowledge, but there were definitely losses as well as gains along the way..

        • Nicholas Gruen says:


          That doesn’t address the point I was trying to make. The Krugman piece is a retrospect. The operative expression in my comment was a question as to anyone had raised the methodological question of whether it would all end without anything much of use being accomplished “in constructing the new field”.

          To find a counterexample, you need to show me an article by Krugman or someone else who is doing the work in New Trade Theory in the 1980s which engages with Hicks’ critique of using imperfect competition in trade (not necessarily referring to Hicks) and explains why “this time it’s different”.

          On the Krugman piece, though I’m a fan of almost all his economic commentary, I’m afraid I find his methodological reflections deeply complacent.

          Hicks’ formula is to use simple, indeed to use Krugman’s word, “silly” models to create a kind of road map of the territory and then to be vigilant in the use of commonsense to try to avoid obvious mistakes. Krugman treats the making of obvious, counterintuitive mistakes as business as usual for economists. The remedy, not to try to build some resilience against obvious blockheadedness into economics, but rather to start a fashion in which the thing that is being ignored, is modelled. The actual modelling leaves you pretty much where you were before, because modelling scale economies ends up producing intractable models. But hey, you’ve started a fashion so then it’s OK to talk about scale economies. Then you can use your commonsense on them.

  2. GrueBleen says:

    Just a small quibble with Waldman’s presentation: one might also say that quantum physics “is not the truth” because it is incompatible with general relativity. After all, it wasn’t relativity that had to invent (and then “normalise”) renormalisation to eliminate integral infinities.

    So, relativity “is not the truth” because it is incompatible with quantum, and quantum “is not the truth” because it is incompatible with relativity.

    So, two theories that are not the truth are incompatible with each other. That’s a great leap forward for science (and economics) isn’t it.

    PS. Should we now all join Brad DeLong in rejecting “reality in science” ?

    • john Walker says:

      Grue Hi
      Funnily enough , New Scientist has just reported :

      IT’S official: the universe is weird. Our everyday experience tells us that distant objects cannot influence each other, and don’t disappear just because no one is looking. Even Albert Einstein was dead against such ideas because they clashed so badly with our view of the real world. But it turns out we are wrong – the quantum nature of reality means, on some level, these things can and do happen. A groundbreaking experiment puts the final nail in the coffin of our ordinary view of the universe, settling an argument that has raged for nearly a century.

      • GrueBleen says:

        Hi John,

        Yeah, it’s all good fun, isn’t it. Like most people, I guess, I’m just a little bemused by entanglement and its ‘demonstration’ of non-locality. But I have just this one small problem: if General Relativity “is not the truth’ then why should we accept lightspeed as an absolute limit on the speed of inter-particle communication ?

        GenRel was, and is, the last great creation of the classical – ie realistic, deterministic – universe in which entanglement loopholes can be closed by taking the particle detectors etc sufficiently far apart that a lightspeed limited signal cannot cross between them in time.

        But, to repeat myself for clarity and emphasis, if GenRel “is not the truth” then it is not necessarily so that signal speed is limited to the speed of light.

        Now what I propose is that we all take a very deep breath and we look up what (Nathaniel) David Mermin had to say about the difference between ‘description’ and ‘explanation’ (and we might just remind ourselves what “ding an sich” supposedly means and ask how we are ever actually going to detect it).

        The trouble with Quantum – and it will likely ever remain the trouble with Quantum – is that our instruments and detectors are inseparably entangled with our theory. After all, if photons and electrons didn’t exist in ding an sich ‘reality’ then what would our sensors actually be detecting ?

        • john Walker says:

          “our instruments and detectors are inseparably entangled with our theory.”
          Even more so with economics, were we are both the subject and the observer, no?
          In any system of representation that is powerful enough to make statements -theories- about itself, statements that are true but also mutually exclusive, must pop up from time to time.

        • GrueBleen says:

          Hmmm. Well, John, if what you’re saying is that GenRel and Quantum, though mutually exclusive, could both be true, then have I got a name for you: Teilhard de Chardin :-)

          Anyway, all I can do is quote myself as follows:
          “Economists have always had a lot of trouble defining their ontology, their phenomenology and their epistemology. Other than that, there’s no problem whatsoever with the ‘science’ of economics.”

          And I most sincerely mean that.

          Now I’m going to have to find time to read the Krugman link provided by Derrida Derider (above) to see just how good Krugman was.

          On the other hand, maybe Nicholas will read it and provide a commentary thus saving me from much most likely fruitless labour.

          PS: Not fruitless because I don’t think Krugman lived up to your analysis, DD, but fruitless because I probably can’t live up to Krugman.

        • john Walker says:

          It is much more likely that both theories are, incomplete.

        • GrueBleen says:

          But John, how can any theory ever be complete ?

        • John walker says:

          Expect that when it comes to things as big as solar systems or things like the behaviour of protons that completeness, or near enough in practice , is possible .
          As for the idea that ‘it’ only exists if somebody is looking, tend to agree with Dr Johnston’s , rock kicking,response to Bishop Berkleley.

  3. Mike Pepperday says:

    Other social scientists can only view economics with awe and envy. Economists have a common language and they run the world.

    In my view, one reason for the success of economics is that it theorises extreme concepts: perfect information, perfect competition, market clearing, homo economicus, and probably more. This all-or-nothing, go/no-go approach to concepts has worked. Other social scientists try to theorise the typical, or averages, and it fails. Economists think their concepts simplify but that misses the mark: they are extremes.

    Natural science theory also deals with extremes: gravity theorises point masses in a vacuum, not landslides; aerodynamics theorises thin layers flowing over mathematically perfect surfaces, not fluttering leaves; gas theory is of gases in tanks, not in volcanoes. Yet the theories apply to landslides, leaves and volcanoes. Those applications of theory are, as in economics, not too accurate but the theories are coherent and everyone agrees on what they are talking about. As your quote of Romer says: “Theory can grow if people agree on core assumptions.”

    No one thinks that homo economicus is a typical person. However, he is theoretically coherent and insensitive to waffly ideas of what theorists might think a typical person is. He is insensitive because he is extreme (not because he is a simplification). Solow’s “disembodied technology” sounds like another extreme and apparently it works, too.

    Hicks confirms these ruminations by saying that attempts to theorise imperfect competition have failed. “It is… only possible to salvage anything … if we can assume that the markets confronting most . . . firms . . . do not differ very greatly from perfectly competitive markets.”

    Could it be that, at bottom, the problem is that people just cannot agree on what imperfect competition means? Given that there are no units of measure (no kilograms, no parts per million) of competition (or of any other social science concept), perhaps it is simply impossible to make coherent soc sci theory that does not deal with extreme concepts, concepts that are either present or absent and can have no theoretical in-between state.

    If theorising extremes in science is necessary to ensure mutual comprehension, to obviate discussion of the meanings of terms, and to rescue the theory from sensitivity to differences in the meanings of concepts, then the only possibility in social science would be with concepts restricted to the all-or-nothing.

  4. paul frijters says:

    Economics has many faces and perhaps nowhere more so than with competition and markets, where the tribe adopts a different face depending on the needs of the questioner and the type of economist. The face of economic regulators and government economists is the one that probably matters most, and there the mainstream economic story is not that markets are close to perfect but that there are simple rules of thumb to make them more perfect. As such, our main theories are about what could be, rather than what is (and in that sense the Stiglitz models of imperfect competition are largely irrelevant. Indeed, any description of ‘what is’ is not that relevant: what is relevant is ‘what could be’).

    For instance: the market cross, perhaps the most recogniseable symbol of economics, is a theory of imperfect competition, of dynamic adjustment of demand and supply due to restrictions in prices and quantities, none of which makes much sense in perfect competition general equilibrium. There is no producer surplus in perfect competition! Hence the main theory taught the most students is one of imperfect competition.

    There are many many more examples. It is for instance clear that taxation and compliance with legislation in a formal labour market only can occur because markets have significant returns to scale in them. Indeed, the whole capitalist system hinges on the interplay between limited liability and limited appropriability, both forms of imperfect markets. There wouldn’t be a stock market if markets were perfect! Etc. Etc.

    So the above quote by Hicks is hence about as relevant for the treatment of competition in modern economics as a manual on fishing radars. A curiosum.

Leave a Reply

Your email address will not be published. Required fields are marked *

Notify me of followup comments via e-mail. You can also subscribe without commenting.