The theoretical structure of modern economics rests on the hypothesis that individual behavior is rational in the sense that choice can be described by utility-maximizing behavior. Most economists view this hypothesis not as an assertion of procedural rationality, but instead as a simple description of the outcomes of individual choice behavior. Furthermore, requiring that every individual behave according to the predictions of rationality is typically viewed as too stringent, even though many of our models assume it. We are most confident about the predictions of rationality for aggregate behavior rather than individual behavior. We believe that many, but of course not all, individuals behave approximately as predicted by models of rational behavior, and we hope that our models are robust to small departures from rationality. This belief is probably most firmly held for asset markets where many individuals interact repeatedly and the payoffs for good decisions can be large. But why should we believe that individual level rationality hypotheses yield good predictions for aggregate asset market behavior? There are two answers to this question. The first answer entails adaptive response: individuals who do not behave as predicted by rationality may modify their behavior over time. They could imitate the successful, they could experiment or they could learn from their past mistakes. The second answer entails natural selection: the market selects for those whose behavior is most nearly optimal. Those who, for whatever reason, behave as if they are rational take wealth away from those who behave otherwise. Ultimately the market is dominated by seemingly rational individuals and prices converge to their rational-expectations equilibrium values. This answer also requires some kind of innovation. If the initial set of decision rules contains no rational rule, natural selection alone will not drive the economy to rationality, but if randomly chosen new rules enter the market from time to time, rational or nearly rational rules will, sooner or later, enter the market and be selected for. In the course of our research we have addressed both adaptation and natural selection. Here we focus on natural selection. The appeal to natural selection for the defense of various rationality hypotheses has a long tradition in economics, but until recently there was little formal investigation of its validity. Alchian (1950) and Friedman (1953) both used it in defense of the hypothesis of profit maximization. Their argument was that firms that behave as if they maximize profits would drive out other, non-maximizing firms. Winter (1964, 1971, 1975) formalized and explored these ideas. But the link between profitability and fitness in his work was assumed.
[1]
Divergent rational expectations equilibrium in a dynamic model of a futures market
,
1978
.
[2]
Sidney G. Winter,et al.
Optimization and Evolution in the Theory of the Firm
,
1975
.
[3]
P. Samuelson.
The "fallacy" of maximizing the geometric mean in long sequences of investing or gambling.
,
1971,
Proceedings of the National Academy of Sciences of the United States of America.
[4]
Sidney G. Winter,et al.
Satisficing, Selection, and the Innovating Remnant
,
1971
.
[5]
David M. Kreps,et al.
Price Destabilizing Speculation
,
1986,
Journal of Political Economy.
[6]
L. Summers,et al.
The Survival of Noise Traders in Financial Markets
,
1988
.
[7]
A. Alchian.
Uncertainty, Evolution, and Economic Theory
,
1950,
Journal of Political Economy.
[8]
R. Zeckhauser,et al.
The Rationality Struggle: Illustrations from Financial Markets
,
1991
.
[9]
L. Summers,et al.
Noise Trader Risk in Financial Markets
,
1990,
Journal of Political Economy.
[10]
S. Winter.
Economic "Natural Selection" and the Theory of the Firm
,
1964
.
[11]
R. Rob,et al.
Learning, Mutation, and Long Run Equilibria in Games
,
1993
.
[12]
M. Friedman.
Essays in Positive Economics
,
1954
.
[13]
L. Blume.
The Statistical Mechanics of Strategic Interaction
,
1993
.
[14]
L. Blume,et al.
Evolution and market behavior
,
1992
.