Friday, August 17, 2007

Family Resemblances Among Games

Do mainstream economists have a theory of prices? General Equilibrium theory is obviously one candidate for such a theory. But economic theorists have found many problems with it, some internally generated. After examining the failure of General Equilibrium Theory, in a paper in the Cambridge Journal of Economics, Rizvi turned to look at game theory, another candidate:
"It is important to understand the reasons why general equilibrium theory had to be replaced. This has to do with the emergence of arbitrariness results in general equilibrium theory during the 1970s and early 1980s. Because of these results, it became obvious that individual rationality postulates gave no structure to aggregate theorizing: the aggregate realm was undetermined and arbitrary. This vacuum provided an environment in which rational choice game theory, with all of its problems, was able to prosper. Here, too, the conception of rationality was inadequate to the task at hand. Nash play required entirely unbelievable assumptions, and created a plethora of equilibria; the simplest form of subgame perfection was bedeviled by experimental nonconfirmation. These problems gave impetus in turn to the development of boundedly rational and evolutionary approaches to game theory. To the extent that these approaches have had success in giving basis to some notions from rational choice game theory, they do this by, to a large extent, doing away with rationality. Yet since many of these results were obtained in an oversimple setting, it does not seem likely that evolutionary approaches will find widespread economic applications. Another response to the arbitrariness results has been the work on market demand. Explicitly, authors in this tradition do away with individual rationality to the extent possible. What remains, then, of optimization and its relation to best choice? Since best choice has been the basis of normative economics, evaluative concerns are sacrificed in the effort to obtain aggregate-level results of the most straightforward sort. Many of these approaches are to some extent allied to the emergence of experimental economics. But there is controversy concerning the role of experiments in economics. Are they just a source of paradoxes to be resolved in the usual way in theoretical economics, or should experiments form the centerpiece of an inductive approach to economics? If the latter way is found to be attractive, it is doubtful that there is much chance that the view of economic behavior or satisfaction that emerges will be anything like the universal and individually rational agent.

In the course of examining these transformations in economic theory, we can see quite clearly a loss of ambition in terms of generality, and an increasingly tenuous grasp on the concept of rationality. In consequence, there is a loss in the ability to say anything much about normative considerations. To some, the current period of rapid theoretical change and even to some extent pluralism will be attractive; but durable results are hard to find and many of the fundamental concerns of economics have fallen by the wayside." -- S. Abu Turab Rizvi (1999?). "Rationality, Evolution and Games"

No comments: