Economics 201C, Part I: Introduction to Behavioral Game Theory Vincent P. Crawford Economics 319, 534-3452, vcrawfor@weber.ucsd.edu
Abstract: Part I (the first half) of Economics 201C will give an overview of behavioral game theory, an emerging blend of theory and empirical regularities whose goal is to move closer to the kind of understanding of strategic behavior needed to analyze economic, political, and social interactions. An important part of the theory deals with how individual preferences deviate from self-centered maximization in games and other social situations; but in this course I will focus on another part, which is unique to games: how players make decisions whose consequences depend on others' decisions they cannot observe, and must therefore predict. The answer to this question depends on the nature of players' mental models of others' decisions (including others' models of their own decisions). Game theory has approached this question in two very different ways, which coexist (a little too) peacefully in the literature. Traditional game theory assumes players form self-fulfilling beliefs about each other's decisions and make decisions that are optimal given their beliefs, and therefore in Nash equilibrium; in effect it assumes players have perfect mental models. Adaptive learning theories make assumptions directly about how players make and adjust their decisions over time in response to others' decisions; in effect they assume players have simplified mental models. These approaches differ on a cognitive dimension, the sophistication of players' mental models of others' decisions, which I will call strategic sophistication. Behavioral game theory can be viewed as a synthesis of the two approaches, which (among other things) treats the degree of sophistication as an empirical regularity, represented by behavioral parameters, and tries to measure it, usually in laboratory experiments. The course will discuss theories that are helpful in thinking about this issue and examine their performance in the light of experimental evidence. The course will emphasize theory and empirical regularities rather than experimental methods, though some discussion of the methods used in experiments will be needed to interpret their results.
[此贴子已经被作者于2004-10-12 7:27:54编辑过]