Wednesday, February 27, 2008

Thanks Ezra.

Ezra Klein defends the idea that behavioral economics is policy relevant from Will Wilkinson

Over at Free Exchange, Will Wilkinson is taking some shots at behavioral economics and its occasional implications for tighter regulation of markets. "These alleged irrationalities," writes Will, "are general tendencies of the species. So they must afflict every voter, every politician, every bureaucrat, every power hungry general. How exactly is 'a larger role' for the government supposed to improve on the coordinating function of the price mechanism?"


Klein reponds convincingly, but I would like to add an observation.

It is known that people have difficulty dealing with probabilities and updating subjective probabilities with new information. Fortunately, with large data set and assumptions which everyone finds extremely plausible, we can sometimes estimate probabilities accurately by calculating frequencies. In this case, the fact that the person making decisions based on freequencies doesn't understand probability is not a problem.

Clearly individuals can't do this on their own -- collecting and analyzing the data is a large fixed cost. Even the evaluation of such studies requires the investment of a lot of time listening to or participating in debate. This effort is completely different from individuals effort to learn from their experience and conversation with friends.

I guess I am mainly repeating what Ezra said but making a hash of it. My point, if any, is that we can (collectively) overcome our mental handicaps by conducting experiments and analyzing natural experiments. This happens when we look at frequencies rather than trying to calculate probabilities.

update: Actually Klein's argument is much much better than mine. He shifts the debate to a debate about policy proposals, in particular replacing opt in 401(k)'s without opt out 401(k)'s. I don't think that Wilkinson has a plausible rejoinder. If the debate is about policy proposals (as you know the policy debate should be) then the argument "politicians and their advisers are irrational too" doesn't get Wilkinson very far.

If it is agreed that economists must assume that people are rational, one can argue "that policy can't be good because no one can come up with a plausible model in which people are rational and it is good." If it is agreed that people are not perfectly rational Wilkinson's argument becomes "You are not perfectly rational so your policy proposal can not be good." There are three replies to this 1) You aren't rational either and laissez faire is a policy proposal 2) I may not be perfectly rational but I'm not perfectly reliably wrong either and 3) even a blind squirrel sometimes finds an acorn.

Behavioral economists are not asking you to assume that we are right but just to consider a proposal based on status quo bias the same way you would consider a proposal based on say the intertemporal elasticity of substitution of consumption (you can even ignore the fact that evidence for status quo bias is very strong and that it is not proven that the intertemporal elasticity of substitution of consumption is positive).

No comments: