Revealed preference (19 June 2006)

In the latest issue of The Economist is an article about immigrants' social mobility in America, in which appears this statement: "In absolute terms, Mexican have grown much richer by coming to the United States. If they had not, they would go home." I don't necessarily disagree with this statement, but the logic seems a bit specious. Economists in general often fall into this kind of de facto reasoning that presumes that there is always a perfectly logical set of incentives that explain any situation, and that since individuals are responding to incentives of their own free will, no one need bother interfere -- certainly not the government. But people don't always respond rationally to incentives, or to put that differently, incentives don't always correspond to maximizing utility at the margin, as economists assume. Imperfect information, stubbornness, altruism or sheer stupidity can wreak havoc on such models that ascribe to humans a uniform rational competence. In fact, the advertising industry exists to create information asymmetries, to produce irrational behavior, to provoke marginal stupidity. But since no one has forced these people to respond to ads,and no one can judge objectively what another person subjectively needs, their behavior is said to constitute revealed preference.

Revealed preference is an attractive (if unforgiving) notion because it assesses people's interests by what they do and not what they say. (People say they want smart programming on TV, but they actually watch Big Brother. People say they are going to quit smoking, but they buy cigarettes by the carton online.) And it eschews making value judgements on the interests so revealed or assigning standard one-size-fits-all human preferences to everyone. Thus it seems to admit the greatest possible scope for individual freedom. But it seems to exempt economists (which should be read to mean conservative economists of The Economist stripe) from making any useful prescriptions whatsoever and consign them to mere description and modeling. Of course, that's not so mere in practice, but it leads to, for example, The Economist's reticence in its article in the same issue about American inequality, which is very frustrating. The article points out a huge and ever-widening gap between the rich and poor in America, but is strangely passive about assigning blame for it and certainly advocates no actions to ameliorate it -- instead there is faint speculation about what might happen, as though we have no possible say in the matter and are left to the whims of the economic gods, as if this unfair division of the society's spoils is not the result of political machination.

I guess my resistance to revealed preference arguments ultimately stems from the obvious empirical fact that many people (myself included) don't always no what they are doing or why, and that my consumption choices reveal simple confusion or inertia or rather than any actual preferences -- not a preference for confusion or inertia or decision-avoidance. I may just object to the word preference in this context, because I think markets can be coercive -- they can constrain choices and lead to self-misrecognition -- the "this is not my beautiful house" moment so vividly described in the Talking Heads song. Of course, I may be resisting the fact that preferences change constantly, and "revealed preference" reveals most of all that mercurialness.