Yale Economic Review PAECON Interview

  1. What do you see as the strongest point of the "post-autistic" critique?

Possibly the strongest point of the critique is word “autistic” itself, since it asserts that neoclassical economics has the characteristics of an autistic child. A dictionary definition of autism is:

A psychiatric disorder of childhood characterized by marked deficits in communication and social interaction, preoccupation with fantasy, language impairment, and abnormal behavior, such as repetitive acts and excessive attachment to certain objects. It is usually associated with intellectual impairment (The American Heritage Dictionary)

This is a strong but apt characterization of how neoclassical economics hangs on to its preconceptions, when serious analysis shows that they are untenable. The best way to illustrate this is with an example. Consider the “representative agent”, a concept that forms a large part of what passes for macroeconomics today: how did it come about?

Its genesis lies in a paper by Gorman in 1953 (Gorman, W.M., 1953 “Community preference fields”, Econometrica, 21: 63-80), in which he showed that the only conditions under which a market demand curve could have the same properties as an individual demand curve were (paraphrasing him somewhat) that (a) all consumers had identical tastes and (b) those tastes didn't alter with income.

Gorman's reaction to this discovery was itself autistic—in the sense of a “preoccupation with fantasy”. Clearly those conditions are absurd, so that the general rule must be that market demand curves aren't the simple downward-sloping functions of price shown in textbooks. However, Gorman commented that:

The necessary and sufficient condition quoted above is intuitively reasonable. It says, in effect, that an extra unit of purchasing power should be spent in the same way no matter to whom it is given. (Gorman 1953: 64)

There is nothing “intuitively reasonable” about that proposition: it instead amounts to ignoring the distribution of income, when analysis has just shown that it has a profound impact. Shafer & Sonnenschein, who rediscovered this problem after Gorman, made the far more realistic statement that:

market demand functions need not satisfy in any way the classical restrictions which characterize consumer demand functions… Only in special cases can an economy be expected to act as an ‘idealized consumer’ (Shafer & Sonnenschein in Arrow K. & Sonnenschein, H., (eds.) 1982: Handbook of Mathematical Economics, North Holland, New York, 671-2 )

Alan Kirman took the sensible approach to this issue (now known as the “Sonnenschein-Mantel-Debreu conditions”) when he stated that “If we are to progress further we may well be forced to theorise in terms of groups who have collectively coherent behaviour” (Kirman, A., 1989. “The Intrinsic Limits of Modern Economic Theory: The Emperor Has No Clothes”, The Economic Journal, 99: 126—139, 138). However economics chose the autistic path of ignoring the distribution of income by inventing the fantasy of the “representative agent”.

  1. How applicable is the critique to economics being conducted in America (if you can say) and in the rest of the world?

It's very applicable. Virtually every economics department contains one or more critics of neoclassical theory, but an uncritical neoclassical core is the norm of economics education throughout the world. From what I understand, the situation is worse in France—because of the imposition of a centralized curriculum—which possibly explains why PAECON originated there. But it struck a chord with skeptical students and dissident staff throughout the world, as can be seen by the student petitions at Cambridge UK, Harvard USA, etc.

I don't mind declaring myself horrified by what passes for macroeconomics in graduate education in most of the USA today. I've just read Ljungqvist and Sargent's Recursive Macroeconomic Theory, and in my opinion even “autistic” is too kind a word to describe that approach.

  1. How would you discard the argument that extreme simplification is necessary in order to model a situation and, cognizant of the model's imperfect representation of reality, gain various general insights from it?

There's a very big difference between a simplifying assumption and a counter-factual one. Neoclassical economics makes many of the latter, and then defends them as if they're the former.

An apposite example of this are the assumptions used by Sharpe to aggregate from his model of a return-maximizing and risk-minimizing individual to the market, when he invented the Capital Assets Pricing Model. Sharpe explicitly acknowledged that he assumed:

a common pure rate of interest, with all investors able to borrow or lend funds on equal terms [and] homogeneity of investor expectations: investors are assumed to agree on the prospects of various investments (Sharpe, W.F., 1964. “Capital Asset Prices: A Theory Of Market Equilibrium Under Conditions Of Risk”, The Journal Of Finance 29: 425- 442; 433-434)

Surveying the empirical wreckage of the CAPM some 40 years later, Fama and French noted that implicit in the second assumption was the proposition that “this distribution is the true one” (Fama E. F. and French K. R. (2004), “The Capital Asset Pricing Model: Theory and Evidence”, Journal of Economic Perspectives, 18, 25-46, 26): i.e., not only did investors agree on the future prospects of all shares, but their expectations were correct: they could predict the future.

I'm sorry, that's not a simplifying assumption: that's a counter-factual that ignores both uncertainty about the future, and the actual dynamic processes that drive the speculative pricing of assets.

A simple way to distinguish a simplifying assumption from a counter-factual one is to consider what happens when the assumption is removed. With a simplifying assumption, its removal should make the theory slightly less accurate in its predictions—it shouldn't cripple it entirely. Yet remove the assumptions that PAECON members object to from any neoclassical model, and it normally collapses.

Your phrase “extreme simplification” is also a guideline. Einstein coined the aphorism that “Theories should be as simple as possible, but no simpler”; the PAECON accusation is that neoclassical economics violates the second half of the aphorism, by making things far “simpler” than they are—by ignoring time and working in comparative statics (or using asinine devices like “overlapping generations” instead of time-based mathematics), by presuming homogeneity when heterogeneity is the obvious rule, and so on.

Ironically, that is often done to hang onto a mode of analysis that is inappropriate for the task—chiefly, using equilibrium methods to analyze dynamic processes—and the analysis would actually be a lot simpler if the fiction of equilibrium were dropped.

  1. What would a "post-autistic" economics look like -- what changes would need to be made?

It would certainly start by being based on dynamic modeling and systems theory. My major objection to neoclassical theory is its obsession with equilibrium, as if economic processes occur only in equilibrium. That's nonsense: economic processes, like those of all other dynamic and evolutionary systems, occur in time and far from equilibrium.

Technical training would start with differential equations rather than simultaneous ones, and models would necessarily include time, rather than ignoring it via equilibrium constructs. Its history of economic thought would eulogize Schumpeter rather than Walras, and praise parts of Keynes rather than Friedman.

Realism would be a methodological emphasis: we would take our guide on how to use assumptions from Musgrave (Musgrave, A., 1981, '‘Unreal assumptions’ in economic theory: the F-twist untwisted', Kyklos 34: 377-387, reprinted in Caldwell, B., 1984, Appraisal and Criticism in Economics: a Book of Readings, Allen & Unwin, London) rather than Friedman. Whether this goes as far as Tony Lawson's critical realism (Lawson, T., 2003, Reorienting Economics: Economics as Social Theory, Routledge, London) is a matter for debate.

Economic models would bear resemblance to those of meteorologists, though with the added difficulties of the absence of conservation laws, evolutionary change, and decisions being influenced by uncertain perceptions of the future. The knowledge that sciences now have of complex systems would affect how models were constructed and interpreted. The ignorance that economists currently have of this development—complex systems theory—would finally be overcome.

Models would explicitly reintroduce social class into economic analysis—the sensible response to the SMD conditions is the one Kirman made, that they legitimize the practice of Classical economists of reasoning in terms of social classes (primarily workers, capitalists, and bankers). They would also incorporate multi-sectoral dynamics, treat profit as a surplus rather than a marginal product, and give an endogenous interpretation of technical change along Schumpeterian lines.

The core mechanical analogy that still underlies neoclassical economics would go, to be replaced with the biological and evolutionary metaphor that Marshall paid no more than lip service to—and that modern neoclassicals sometimes invoke in a very amateur way.

With that change, a lot of the discipline's arrogance would disappear. Economists meddle with the economy in a way that ecologists do not meddle with an ecology. Neoclassical economists—and for that matter Keynesian economists before them—act as if they not only understand this most complex of systems, but also know how to make it function better: just make it look more like the textbook models. A student of mine once commented that the mechanical analogy encourages economists to tinker with the economy as if it were a car; but if the analogy were that of a rainforest, would economists blithely recommend that the forest would work better if we removed some species from it?

  1. What are some specific examples of a "post-autistic" approach?

Post-autistic economics isn't a school of thought in the same way that neoclassical economics is—or even Post Keynesian economics. Equally, there are some contributions to non-neoclassical economics that I'd regard as somewhat autistic for applying equilibrium and comparative static methods. And as you may know, I'm also highly critical of the “Labor Theory of Value” strands of Marxian economics. Generally, I favor work in nonlinear dynamical systems, evolutionary modeling, and quality data exploration work. As instances of post-autistic economics, I'd nominate

There are many others that I could note, but that should be a good starting list for an exploration of post-autistic approaches.

  1. If a new economics were created, would it be more, less, or equally math-intensive as the existing discipline? How might a new economic model, like a satisficing model, differ from existing models?

I think it would be as math intensive, though the nature of the mathematics would shift from the current arcane focus on simultaneous equations and econometrics to differential equations, nonlinear dynamics and systems theory.

It would add computing as a necessary skill too—while I don't see multi-agent modeling as essential, to do it you must know how to program in an object-oriented way. Simulation also requires knowledge of programming, and that would play a prominent role in any post-autistic economic theory.

We wouldn't necessarily start from the isolated individual, as Kirman noted. The SMD conditions are a classic case of “emergent properties” that are commonplace in complex systems—the aggregate has properties that aren't properties of each constituent item within it, and therefore reductionism is neither possible nor desirable. Thus we can have a model of individual behavior in which well-behaved individual demand curves can be derived, but where the aggregate can have any shape at all, because the relationships between individuals have more impact on the system than the isolated properties of each individual. In economics, emergent properties mean that: macro cannot and should not be reduced to a branch of micro—which is what neoclassical economics has attempted to do.

The computing focus would enable us to build “bottoms up” models where well-defined heterogeneous individual behaviors—such as Simon's satisficing model you mention—aggregate into very different macro behaviors. But they wouldn't be a fundamental part of the theory, because of emergent properties. That's a good thing, because evolutionary multi-agent models are notoriously difficult things to build, even for experienced programmers. I can see some potential for computer gaming technology helping out here, but it will take a long time. In the meantime, there is much to be done in building good, “tops down” nonlinear dynamic models.

But economics would also be multi-branched, and include what I call “barefoot economists”. We need knowledge of the intimate connections that exist in regional economies, as well as models of the macro-economy. Pamela Cawthorne gives an excellent instance of this type of work in a field study of Indian manufacturing.

  1. Is the "post-autistic" movement gaining traction among economists?

It's certainly emboldened critics of neoclassical economics, and maybe it's converted a few neoclassicals, but generally that's not the aim. PAECON is more of an appeal to the students of economics, rather than an attempt to convince existing economists to “change camps”.

PAECON economists, in general, don't believe that it is possible to have a meaningful dialog with staunch adherents of the neoclassical approach. Point out that you have to assume something patently absurd to reach a desired neoclassical conclusion, and ninety nine per cent of the time, ninety per cent of them will simply make that assumption. A fringe will be moved towards greater realism, but in general we feel we'd waste our time to try to engage committed neoclassicals in debate.

So instead, the movement has concentrated on getting critiques to the young, through the PAECON website that Edward Fullbrook maintains, and the books he has edited (Fullbrook, E. 2004, (ed.), Student's Guide to What's Wrong with Economics, Routledge, London). Books like my Debunking Economics (2001, Zed Books, London), and its associated website are also an appeal to the student, rather than an attempt to convert the neoclassical zealot.

In one sense, I'm not being critical of established neoclassical economists for continuing to cling to that manner of thinking. It's a difficult step for someone whose intellectual life has been in a given tradition to abandon it for the comparatively great unknown (though a minority of tenured economists do undergo an epiphany during their careers, which is why almost every economics department, even the most conservative, has a rebel or two on staff).

In this, economics is no different to physics: Max Planck famously commented “a new scientific truth” like quantum mechanics “does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it.” (Max Planck quoted in Kuhn, T., 1962, The Structure of Scientific Revolutions, University Of Chicago Press, Chicago, 151)

What makes economics different to sciences like physics, and therefore makes me pessimistic about PAECON ever gaining serious “traction” in economics, is the impossibility of confronting economic theories with experimental evidence. As a result, notions that would die an empirical death in an experimental discipline live on possibly forever in economics.

In this sense, I think Popper was wrong to downplay the importance of the experimental method in comparison to falsifiability. Even though, as Kuhn argued, a theory directs the choice of experiments by which it is tested, the essential “orthogonality” of experimental data to theoretical prediction means that scientific theories inevitably thrown up experimental anomalies that the science must in some sense confront. Economics is far less equipped to have that confrontation, and theoretical economists frequently ignore data that doesn't pander to their pet theories.

This side of economics—plus its inevitable ideological role—makes me doubt that economics will ever really escape from its autism. But that's not a reason to abandon the effort to move it forward.

  1. I reviewed a presentation in which you proposed that the face of economics would change completely if it were "wiped out by an asteroid and re-invented." What, then, has forced economics into its current mold and why has it not been corrected? (Or, why have economists chosen to disregard study of history of economic thought?)

I trace this back to the origins of economics as a theory of a market-based society as capitalism emerged out of English feudal society. Feudal society was rigidly ordered—though its militaristic and power-obsessed side meant that it was far from peaceful and stable (it wasn't called “feud”-alism for nothing!). But proponents of the feudal order could regale in horror at the apparent lack of order in a market society.

One way to champion the market order would have been to admit that it was less structured than feudalism, but argue it would be more creative. That, of course, would have been a pretty hard ideological barrow to push. Instead, Smith put forward the notion that the market established a “natural order”: in place of the rigid hierarchy of feudalism we would have the beneficent equilibrium of the market.

This has been the organizing vision of mainstream economists ever since—whether of Classical or Neoclassical bent. Only the Classical malcontents (Hobson, Marx, etc.), the Austrians, Schumpeter, and to some extent the Post Keynesians, have pushed the perspective that capitalist society is unstable; and only Schumpeter and the Austrians have seen this instability as a good thing. There has been a strong desire to prove preconceived notions of stability, optimality, equilibrium, welfare maximization, etc., and this has perverted the theory whenever it has transpired—as it almost always did—that theory, let alone reality, wasn't that straightforward.

Nowadays, researchers who are aware of complex systems theory know that any complex system is going to be cyclical, or “unstable” to some degree, and the only way to impose equilibrium on such a system is to kill it. If economics today was wiped out by my hypothetical asteroid, researchers from this modern tradition would have to fill the void, and they'd start with seeing the economy as a complex system which was necessarily cyclical.

They would then look for the endogenous sources of cyclical change, and something based both on inherent nonlinearities and endogenous technical change would emerge. It would look nothing like neoclassical economics, and to the extent that it evaluated capitalism qualitatively, it would portray this innate volatility as a strength rather than a weakness.

In this sense, it's well-worth reading Janos Kornai's work on the topic of why did capitalism succeed and socialism succeed. In it he sensibly describes capitalism as a demand-constrained economic system, and socialism as a supply-constrained one. The innovative forces that are needed to triumph in a competitive system subject to shortfalls in effective demand makes for faster growth than in a fully-employed, resource-constrained socialist system (Kornai, J., 1986, Contradictions and Dilemmas: Studies on the Socialist Economy and Society, MIT Press, Cambridge).

If economists could abandon that equilibrium-fixated mindset—seeing equilibrium where it doesn't exist, hypothesizing about efficiency in resource allocation when, as Kornai argues, excess resources are necessary for competitive survival in an uncertain environment, and so on—then we might just get a theory of economics that actually describes capitalism. But I think that I'm better off waiting for that asteroid to arrive instead!

As for why they have disregarded the history of economic thought in general, I blame our textbooks and the sausage-factory approach to education that comes out of the false scientism in economics. There's a completely misplaced belief that textbooks—especially those used in PhD programs, like Varian and Mas-Colell—encapsulate all that is worth knowing in economics, and that studying HET is simply raking over the coals of obsolete ideas. Instead, I think economics has tended to preserve the soot, and leave real coal—and some diamonds—lying untouched in the past.

In this sense, I'd urge any student today to read at the very least Schumpeter's Theory of Economic Development (Transaction Publishers), and Keynes's 1937 papers (downloadable from JSTOR). Then ask yourself why is it that the issues these authors treat so well don't even appear in modern economics.