Market Selection in Complete and Incomplete Markets

Cornell University, Ithaca NY

Investigators

Abstract

Markets bring together traders with widely divergent beliefs and motivations for the purpose of exchange. Economic tradition asserts the market selection hypothesis, that the market sorts traders. Those traders most directly concerned with profits and with the most accurate forecasts of the future will come to dominate the market. This hypothesis is used to justify a variety of modeling assumptions about markets, including rational expectations in macroeconomic models and informationally efficient markets in finance. However, it has never been carefully examined. This project has two goals: To study the validity of the market selection hypothesis and, more generally, the implications of trader heterogeneity for market behavior. The validity of the market selection hypothesis depends upon market structure. If market structure allows all potential gains from trade to be realized, the market selection hypothesis will be valid. If not, markets will sort traders in other ways which we hope to uncover. Market selection becomes a more delicate issue when traders have incorrect beliefs but are learning. In this case the question is how the market sorts across learning rules. Although general equilibrium analysis does not require similar traders, modeling practice in macroeconomics and finance eschews trader heterogeneity. We will show how market models with diverse traders exhibit very different (and more realistic) market behavior from the market models now in vogue. Our work opens a door for behavioral analysis in markets because the market selection hypothesis is not always correct, and even when it holds it has implications only for long-run market behavior. While it is possible to postulate the existence of traders with arbitrary behavior patterns, the market sorts among these traders and only a few behaviors will in fact determine market prices in the long run.