I was suspicious of a lot of behaviorial economics: laboratories are not markets. But having read a couple of passionate defenses of the methods and findings of behaviorial economics, I had come to think that my suspicions reflected merely my own serious intellectual failings.

Well, maybe not. Here's Gary Becker:

But I would have some major differences with behavioral economics as it is usually defined. Let me say two things: First, there is a heck of a difference between demonstrating something in a laboratory, in experiments, even highly sophisticated experiments, and showing that they are important in the marketplace.

Economists have a theory of behavior in markets, not in labs, and the relevant theories can be very different. One reason is the division of labor in markets. Consider, for example, the claim by behavioral economists that most people cannot calculate probabilities accurately. I agree to some extent but am confident that people who work at blackjack tables know the relevant probabilities very well. Otherwise they don't work at these jobs. People go into activities when they are either good at those activities, or when they learn to be reasonably good. So while the average person may not calculate certain probabilities very well, they do not end up in jobs where they need to make those calculations.

A market economy is a group of specialists who are integrated by exchange. It may be that each of these specialists is terrible at other activities, but the whole aggregate can be highly efficient. The aggregate may make few mistakes. One of the things some behavioralists have missed is that a specialized economy eliminates many mistakes because vulnerable people don't get put into positions where they can make these mistakes.

The second related criticism I have is that some of the defects in behavior claimed by behaviorists tend for a different reason to be eliminated in an exchange economy. It is sometimes claimed, for example, that people's preferences are not transitive. Transitivity means that if I like a group of goods donated by A over another group, B, and B over C, then I like A over C.

Now suppose a case that violates transitivity. Suppose I prefer one apple to one orange and one orange to one pear, but I prefer one pear to an apple. That is intransitive preferences. What would happen? Someone would come to me and say since I prefer an orange to one pear, I should give him a pear plus money in exchange for an orange. Then he gives me an apple for my orange plus money. OK again. But since by assumption I prefer a pear to an apple, he gives me back my pear in exchange for the apple plus more money. I end up with the pear I started with, but I lost money in three transactions. Why? Because in each case I have given him some money plus a piece of fruit.

This is called the Dutch book argument in economic theory. I do not know if you have heard that before. I don't know where the expression comes from, but it is illustrated here by a series of exchanges that can take advantage of these intransitive preferences. Many other of the behaviorists' claims are subject to similar Dutch book arguments that either make behavior more rational, or a person goes broke. Since this is mainly an implication of exchange, it is hard for me to believe that such inconsistent behavior is important in modern exchange-based economics.

Barnum said there's a sucker born every minute of the day. Well, suckers lose their money. Another example: I have fair dice, but you believe that a 12 is going to come up half the time. I would love to play craps against you. You will continue to lose until either you change your beliefs, or you lose your shirt. Exchange and the division of labor do not eliminate all the issues brought up by behavioral economics but I believe “behavioral” economics has a different place in modern economies than is often claimed.