This award funds work on two separate projects. The first project develops new methods in dynamic game theory. The PI works to extend the definition of Markov perfect equilibrium to the case of imperfectly observable actions. He also uses the concept of evolutionary stability to examine whether it will prove to be a powerful method for sharpening the ability of the theory of repeated games to make firm predictions.
The second project addresses issues in the theory of political economy, again by using techniques and methods from game theory. The research characterizes what voting rules satisfy four key criteria (anonymity, neutrality, the Pareto property, and strategy proofness) for the widest possible set of voter preferences. This project shows what voting rules are most preferable in electoral systems. In addition, the PI builds and analyzes a model of legislative bargaining.
The broader impacts of this project include developing new methods for the many different communities in social science, computer science, and biology that use game theory as a fundamental method. The work in political economy will be helpful to those envisioning reform of electoral systems.
My work on intellectual property protection shows that patent protection may actually inhibit innovation. Specifically, such protection is counterproductive in industries that are highly sequential (sequentiality means that progress is made through a large number of small steps, each building on the previous ones). Roughly speaking, this is because if one of the steps is patentable, then the patent holder can effectively block (or at least slow down) subsequent progress by setting high license fees. Moreover, like any other monopolist, it has the incentive to set such fees. My work on elections shows that there is a precise sense in which true majority rule---the voting rule in which candidate A wins if, for each other candidate B, a majority of voters prefer A to B--- is better than all other possible voting rules for electing presidents or other leaders. Namely, it satisfies the most basic democratic principles---the consensus principle (the idea that if all voters prefer candidate A to candidate B, then B should not be elected); the equal-voters principle (the idea that all voters should count equally); the equal-candidates principle (the idea that all candidates should compete on an equal footing); the no-spoilers principle (the idea that a candidate with no chance of winning himself should not be able to change the outcome of an election by his decision to run); and decisiveness (the principle that an election method should always produce a clear-cut winner)---in more circumstances than any other voting rule. My work on evolution and repeated games shows that there are circumstances in which evolutionary pressure pushes a society toward more cooperative behavior. Cooperation entails that individuals should exert effort on behalf of some collective good. But cooperative behavior is tricky to sustain because cooperation by others presents an individual with the incentive to "free ride" and not exert effort himself. Long-term interaction can help solve the problem: A may cooperate today for fear that, if he doesn’t, B won’t cooperate tomorrow. However, such repeated interaction doesn’t guarantee cooperation--- we can get "stuck" in a situation in which individual A doesn’t cooperate because he anticipates B won’t cooperate, and vice versa. Even so, my work shows that if (i) there are occasional "mutations" (small groups of individuals occasionally try out other behaviors---including more cooperative ones) and (ii) an individual sometimes misperceives how others are behaving, then, in the long run, uncooperative behavior will not persist.
