The news that this year's Nobel Prize equivalent for economics has gone to a couple of game theorists has triggered some discussion in the business press and the econ blogs over the suggestion that game theory is not a really a science because it does not yield testable predictions, so what good is it? See Marginal Revolution; Knowledge Problem; Daniel Drezner; Business Week; Slate.

From the point of view of a legal/policy analyst, this debate shortchanges some important points.

First, and most trivially, to say that game theory or any other economic theory is not a science is news only to economists. The rest of us knew it already. So, so what?

Second, Schelling never pretended that it was science. His genius is for the slantindicular analysis of the simple situation to show how the incentives at work will often lead to particular results. There is no pretense of comprehensiveness, scientism, or non-contingency.

Third, and most important, game theory problems, especially varieties of Prisoner's Dilemma, arise constantly in real life. Consumers coping with P2P file sharing services have a PD problem. So do toy makers or record companies coping with the rise of Wal-Mart. Many standards-setting situations present game theory issues that must be resolved. Setting prices in an investment-heavy economy presents serious game theory problems. There are many more -- including, I suspect, the Broadcast Flag issue.

Because the legal and regulatory systems have not incorporated Schelling's tool kit into their approaches, many official responses to these issues, particularly by antitrust lawyers, are absurd. (See above links.)

So, while game theory is far from a decision-making machine spitting out correct answers, it certainly helps its cognoscenti rule out many bad ones.

posted by James DeLong @ 8:48 AM | Economics, Game Theory & Public Choice

Link to this Entry | Printer-Friendly | Email a Comment | Post a Comment(0)