September 13, 2010

Brad DeLong Makes a Wishful Mistake

Attention conservation notice: Academics squabbling about abstruse points in social theory.

Chris Bertram, back from a conference where he heard Michael Tomasello talk about his interesting experiments on (in Bertram's words) "young children and other primates [supporting the view] that humans are hard-wired with certain pro-social dispositions to inform, help, share etc and to engage in norm-guided behaviour of various kinds", wonders about the implications of the fact that "work in empirical psychology and evolutionary anthropolgy (and related fields) doesn't — quelle surprise! — support anything like the Hobbesian picture of human nature that lurks at the foundations of microeconomics, rational choice theory and, indeed, in much contemporary and historical political philosophy."

Brad DeLong asserts that the microfoundations of economics point not to a Hobbesian vision of the war of all against all, but rather to Adam Smith's propensities for peaceful cooperation, especially through exchange. "The foundation of microeconomics is not the Hobbesian 'this is good for me' but rather the Smithian 'this trade is good for us,' and on the uses and abuses of markets built on top of the 'this trade is good for us' principle." Bertram objects that this isn't true, and others in DeLong's comments section further object that modern economics simply does not rest on this Smithian vision. DeLong replies: "Seems to me the normal education of an economist includes an awful lot about ultimatum games and rule of law these days..."

I have to call this one against DeLong — rather to my surprise, since I usually get more out of his writing than Bertram's. The fact is that the foundations of standard microeconomic models envisage people as hedonistic sociopaths [ETA: see below], and theorists prevent mayhem from breaking out in their models by the simple expedient of ignoring the possibility.

If you open up any good book on welfare economics or general equilibrium which has appeared since Debreu's Theory of Value (or indeed before), you will see a clear specification of what the economic agents care about: this is entirely a function of their own consumption of goods and services. Does any agent in any such model care at all about what any other agent gets to consume? No; it is a matter of purest indifference to them whether their fellows experience feast or famine; even whether they live or die. If one such agent has an unsatiated demand for potato chips, and the cost of one more chip will be to devastate innumerable millions, they simply are not equipped to care. (And the principle of Pareto optimality shrugs, saying "who are we to judge?") Arrow, Debreu and co. rule out by hypothesis any interaction between agents other than impersonal market exchange [ETA: or more exactly, their model does so], but the specification of the agents shows that they'd have no objection to pillage, or any preference for obtaining their consumption basket by peaceful truck, barter and commerce rather than fire, sword and fear.

Well, you might say, welfare economics and general equilibrium concern themselves with what happens once peaceful market systems have been established. Of course they don't need to put a "pillaging, not really my thing" term in the utility functions, since it would never come up. Surely things are better in game theory, which has long been seen to be the real microfoundations for economics?

In a word, no. If you ask why a von Neumann-Morgenstern agent refrains from pillaging, you get the answers that (1) the game is postulated not to have pillaging as an option, or (2) he is restrained by fear of some power stronger than himself, whether that power be an individual or an assembly. (Thus von Neumann: "It is just as foolish to complain that people are selfish and treacherous as it is to complain that the magnetic field does not increase unless the electric field has a curl.") Option (1) being obviously irrelevant to explaining why people obey the law, etc., we are left with option (2), which is the essence of all the leading attempts, within economics, to give microfoundations to such phenomena. This is very much in line with the thought of an eminent British moral philosopher — one can read the Folk Theorem as saying that Leviathan could be a distributed system — but that philosopher is not Dr. Smith.

One can defend the utility of the Hobbesian, game-theoretic vision, and though in my humble (and long-standing) opinion the empirical results on things like the ultimatum game mean that it can be no more than an approximation useful in certain circumstances, and ideas like those of Tomasello (and Smith) need to be taken very seriously. But of course those ideas are not part of the generally-accepted microfoundations of economics. This is why every graduate student in economics reads (something equivalent to) Varian's Microeconomic Analysis, but not Bowles's Microeconomics: Behavior, Institutions, and Evolution; would that they did. If you read Bowles, you will in fact learn a great deal about the ultimatum game, the rule of law, and so forth; in a standard microeconomics text you will not. I think the Hobbesian vision is wrong, but anyone who thinks that modern economics's micro-foundations aren't thoroughly Hobbesian is engaged in wishful thinking.

Update, 15 September: A reader observes, correctly, that actual sociopaths show much more other-regarding preferences than does Homo economicus (typically, forms of cruelty). I could quibble and gesture to dissocial personality disorder, but point taken.

Update, 24 December: In the comments at DeLong's, Robert Waldmann rightly chides me for conflating the actual social views of Arrow and Debreu with what they put into their model of general equilibrium. I have updated the text accordingly.

9 September 2012: A follow-up, of sorts.

Manual trackback: Stephen Kinsella; Marginal Utility; Marc Kaufmann; Brad DeLong; Contingencies; Krebscycle

*: Varian wrote a book, with Carl Shapiro, giving advice to businesses in industries with imperfect competition. The advice is to (1) extract as much as possible from the customer, to the point where they just barely prefer doing business with you to switching to a competitor or taking their marbles and going home, (2) disguise how much you will extract from your customers as much as possible, (3) participate in standards-setting and public policy formation, so as to ensure that the standards and policies will be to your commercial advantage as much as possible, and (4) generally engage in as much anti-competitive behavior as possible without risk of legal consequences. All this may in fact be sound advice for increasing the (more or less short-run) profits of such firms, but the premises are purely Hobbesian. Were there no risk of legal consequences, their arguments would extend straightforwardly to pillaging. The only reason Shapiro and Varian would counsel Apple against, say, running a phishing scam on everyone who bought a Macintosh would be that it was very likely they'd be caught, with adverse consequences; obviously if Apple made enough money from such a scam, Shaprio and Varian's arguments would say not only "go phish", but "lobby to make such phishing legal" (perhaps under the principle of caveat emptor).

The Dismal Science; Philosophy; The Natural Science of the Human Species

Posted at September 13, 2010 13:00 | permanent link

Three-Toed Sloth