Sunday 8 June 2014

E. Glen Weyl on Price Theory

One of the young hot-shots of economics in recent years is E. Glen Weyl of the University of Chicago. He and I are terribly unlike each other in the way that he has a stellar career whereas my eminence is yet to be recognized, but on the other hand, he and I are very much alike in the sense that we were both incredibly lucky to have several opportunities to speak to the late Gary Becker, whose untimely death about a month ago, I am sure, left a deep hole in the hearts of everyone fortunate enough to have met him. Professor Weyl shows his appreciation of Professor Becker in a new paper attempting a summary of price theory, which, from the abstract, he defines as follows:
[...] a methodological approach that derives low-dimensional "prices" sufficient to characterize simple allocative problems in complex economies.
What Professor Weyl does in the paper is to describe the development of price theory as defined above. The paper is a good one, but I will nevertheless only discuss one disagreement I have with it. My disagreement is that Professor Weyl has too structural a vision for the future of price theory, arguing that Price Theory can adopt parts of the structuralist method. Me, I do not believe its future lies in many theoretical developments, but rather in new applications of it, theoretical as well as empirical.
 
"Structure" means that the economist imposes substantial and fairly detailed theory on the issue under analysis. If he does empirical work, he might assume that, say, the Median Voter Theorem is true and in consequence impose the one-dimensional restriction on public opinion surveys. This requires some sort of "scoring system" for locating individuals on the left-right axis, so that being for more provisions to the navy has a certain score and being anti-gay is associated with another one. This example will obviously become intractable and unwieldy; most structuralists would shy away from this sort of work (though note that the resultant model may turn out to be "true"), but it is, I believe, what a strucuralist hell-bent on really getting at a problem involving public opinions and political economy must resort to doing.
 
Not all structure is this scary. For instance, one may safely assume that firms maximize profits and use this assumption in one's theoretical or empirical work. However, I believe most economists have something much more involved in mind when they think of "structure". The appeal of "structure" can, I believe, be traced back to the following proposition. Economists, structuralists or not, face one ugly fact: good ideas are really scarce. What Gary Becker could do theoretically required next to no messy assumptions (only "clean" "structure" in agreement with very basic propositions in economic theory), but no-one has come up with nearly as great theoretical ideas as did Professor Becker. To work on more established ideas then requires some semblance of progress, which leads the structuralists to impose progressively messier requirements in order to get as progressively less valuable "insights". I believe this is one reason behind the crazy mathematics frequently demanded at doctoral programmes in economics in the past five decades or so. If one does not work this way, one must have novel and original ideas, or write fewer papers. Hence the appeal of "structure".
 
By contrast, Price Theory the way I think of it simply means, and should continue to mean, that people respond to incentives in a consistent way, both synchronically and diachronically. I agree completely with Professor Weyl's remark that price theory is "low-dimentional". Price Theory and General Equilibrium Theory are not really "enemies", but Price Theory is focused on understanding the real world and thereby lends itself to partial-equilibrium analyses. There are many things about the details of how people respond to incentives which we do not know, but rather than trying to impose messy assumptions I believe properly-practised Price Theory should rest on developing clean empirical approaches (in addition to continued theoretical treatments of new issues), often in the form of so-called field experiments, which highlight simple variations plausibly resultant from simple differences in treatment.
 
To flesh out this last remark, consider the most important breakthroughs in theoretical and empirical economics in the past few decades. Did the empirical advances rely on imposing heavy theory? The foremost practitioners are, I estimate, mainly people who come up with clever ways of obtaining insights from easily-interptretable real-world phenomena (e.g. Levitt and Donohue's abortion and crime paper). The same thing may be said for theoretical advances. Profit-maximization is rarely questioned these days, but all kinds of rationales as to why it is false were offered before Armen Alchian showed by evolutionary analogy that firms survive better if they fail less at maximizing profit, assuming only a fairly stable environment ("market"). A Price Theory which continues to rely on simple mechanisms and clever ways of obtaining evidence will probably produce fewer papers than one which (partly) approaches structuralism, but its papers will be more impactful and last longer. Or are the most important "structural" breakthroughs really even nearly as celebrated as the more classic price theory papers?

(HT: MarginalRevolution's Tyler Cowen)

No comments:

Post a Comment