Site Meter

Thursday, October 13, 2005

Mark Kleiman writes "Michael Mandel admires Schelling but seems dismayed that game theory isn't a perfectly predictive Theory of Everything. ([snip] If you doubt the utility of game theory, just ask an evolutionary biologist.)"

To be more specific ask Richard Lewontin who attempted to apply game theory to population biology and found it useless (at the time he was one of the worlds two leading population biologists). It is true that evolutionary biologists and game theorists have a mutual interest in so called "evolutionarily stable strategies."
However the concept was developed by an evolutionary biologist named Maynard Smith (the reason Lewontin was one of the two top population biologists). As far as I know, Smith did not draw on the work of game theorists. I see no reason to imagine he would in order to address an issue which is clearly germain to population biology and only arguable relevant to games.

Game theorists adopted this idea from evolutionary biology. Thus if you doubt evolutionary biology is useful, but are confident that game theory is useful, you are ignorant of the intellectual history of game theory. The reverse does not follow.

I'd say my objection to game theory is not that it isn't "a perfectly predictive Theory of Everything" but rather that predictions based on game theory about behavior in the simplest settings for which game theory was developed are consistently refuted by experimental evidence.

I challenge Kleiman to come up with an implication of game theory which is neither obvious nor false. This is not a satisfactory standard for a field of research like economics which aspires to be a science. I hasten to add that I don't have anything in particular against game theory which is, at least, more interesting than the rest of economic theory.

update: I'm sorry I was rude to Prof Kleiman. Based on the hints about it in his blog I think that I would like *his* research if I bothered to read it. Also, to be clear, I am pleased that Schelling won the prize. I just don't think that there should be a Nobel prize in economics until economists come up with some research worthy of the prize. Also I absulutely don't think that theoretical research in economics is as valuable as empirical research with identifying assumptions based on common sense. My rule is that if the average guy in the street doesn't immediately understand and agree with all the theory in the paper, then it's probably wrong. This does not eliminate all economic research or even all fun and interesting economic research. Suggested catch phrases "good instrument", "natural experiment" or catch word "experiment".

Oh I would guess that, if Kleiman were to notice my challenge, he would mention tipping points. I count that as true, related to game theory and obvious but maybe it wasn't always obvious. I learned about the idea in elementary school (ok maybe junior high) so it seems obvious to me.

No comments: