Sunday, September 16, 2007

What is wrong with econophysics?


Cosma Shalizi offers his take, which I have been looking forward to for quite sometime (over a year, actually). And it is Cosma at his best -- full of incisive arguments (with tons of links) as well as provocative insights. Here he is on the physicists' blind spot in econophysics:

If econophysics is dignified enough to have a tragic flaw, it is this. I have lost count of the number of times I have heard other statistical physicists insist, or explain, or just assume, that ecology, or evolution, or neuroscience, or, social networks, or, yes, economics, "is, after all, just another many-body problem", so of course it must yield to the insights of statistical mechanics. This is why our conquistador spirit leads us to make assaults on these disciplines, and not, say, classical philology. I don't even think that this is wrong. I think the problem is that we have a drastically impoverished notion of bodies, and how they might interact.

Here he is about neo-classical economics:

... [M]ainstream economics is clearly false. I don't say this just because perfectly competitive markets aren't the only economic institution in this world; the neo-classical framework now includes very sophisticated theories of imperfect competition, imperfect information and non-market institutions, and these developments are mainstream enough to result in Nobel Prizes (in, e.g., 1993, 1994 and 2001). The foundation on which the neo-classical framework is raised, though, is an idea about rational agents: rationality means maximizing expected utility, where expectations come from maintaining a coherent subjective probability distribution, updated through Bayes's rule; moreover, the utility function is strictly self-regarding. ... Alas, experimental psychology, and still more experimental economics, amply demonstrate that empirically it's just wrong. We are boundedly rational, and, for good or for ill, we give a damn about others. Moreover, there are very general reasons, having to do with the computational intractability of optimization problems, and the severe limitations on computable Bayesian learners, to think that no creature could ever be a "rational agent" in the neo-classical sense. Bounded rationality is the only kind we encounter, and the only kind we are going to encounter. ... But, as I said, all the rest of the neo-classical framework rests on this conception of individual decision-making; remove it and all the models are standing on air. So: neo-classical economics is false.

6 Comments:

  1. Ritwik said...

    Everything in the excerpt makes sense , except the bit about bounded rationality.

    rationality is definitely bounded, but that need not necessarily weaken the models of neoclassical economics. under MOST real world circumstances, the assumption of rationality works just fine because the rationality that needs to be assumed comes well within the bounds imposed by human nature and knowledge.

    saying that neoclassical mwechanics is false is similar to saying that newtonian mechanics is false. In the truest technical sense, yes it is- but for most real world phenomena, it works just fine.

  2. Ritwik said...

    *neoclassical economics*

  3. gaddeswarup said...

    Abi,
    I looked through the article quickly and am not sure whether he mentions the following development:
    http://www.sciencenews.org/articles/20070901/bob9.asp
    Excerpt:
    "A team of economists and physicists is now proposing a new way to look at development. The researchers have shown that a country's competitive edge can spread from one kind of product to another along a well-defined network of links, much as disease epidemics tend to spread among people who are socially connected."
    The links to the original papers are in the article and it looks like an interesting approach though may not be a big development.

  4. Anonymous said...

    I also believe in the concept that stuff like economics, neuroscience, etc can be reduced to many body problems. In his attempt to ensure the flow of his rant, he is missing the important point. The problem is not with many body physics or in physicists desire to apply many body theories in other fields (in his tone, poking their unusually long nose into other fields) but it is an issue of finding the correct model that will highlight the complex interactions in these highly complex systems. The difficulty in approximating the interactions with a suitable model cannot be construed as a difficulty with the many body approach itself. On paper, this many body approach will explain the various aspects of the fields like economics and neurosciences, as long as our model is capable of approximating these complex interactions well. Under such a scenario, there is nothing wrong in attempting to work towards a good model. He is expecting the results of exact solutions from the many body approach. It is evident that it is humanly and computationally (atleast at this point of time) impossible. We have to use statistical approximations. The quest is to find that approximation which could reasonably correlate with the complex interactions in these systems. I seriously don't understand why he thinks that such an approach is uncalled for. If this can help us understand the strongly correlated electron systems, it should, at least on paper, help us understand strongly correlated economic systems or neural systems, etc. It is just a matter of getting a correct model to approximate the interactions. I also want to emphasize that these many body approach can also tackle production apart from exchange. Obviously, I haven't read all his links. I am not sure if I am missing something. If I am not missing anything, I just feel that he is getting carried away with his rant.

  5. Abi said...

    Ritwik: It's interesting that you invoke Newtonian mechanics here. Cosma too invokes the same analogy, in pretty much the same way you do. He says, "... [P]hysicists know very well that classical physics is an extremely good theory in the right limits... In a similar way, I think, the neo-classical ideal is a tolerable approximation in certain limits. ..." He goes on to point out those limits as well. Do read his post!

    Swarup: Cosma doesn't get to the recent research on product clusters and export networks. He does, however, talk about interacting agent-based models. His grouse is that physicists don't bother with modeling the 'interactions' well.

    Krish: As a statistical physicist, Cosma is not against building more realistic models of interacting agents, based on psychological research (for example). By pointing out the "drastically impoverished notion of bodies and how they might interact", he's implying that such impoverished models aren't going to make a serious impact in economics.

  6. Anonymous said...

    Well these engineers-physicists turned financial analyst have been slightly on the receiving end after the recent market blow-up due to subprime mortgage crisis
    An interesting article in MIT Technology Review ( registration required) related to this -
    The Blow-Up
    https://www.technologyreview.com/article/19530/