Date:
Wednesday, October 19, 2005   4:15 - 5:15p

Location:
301 College Avenue, Seminar Room

Abstract:
Newspapers, stock market commentators and, financial advisors, among many others, often talk about what the stock market knows or expects. In this commentary it is standard to assume that the market's beliefs about the value of assets are correct as is reflected in the oft stated proposition that you cannot beat the market. But what could this possibly mean? The stock market is merely a meeting place, now a virtual meeting place, for trade. It does not have expectations. However, market prices do serve as an aggregator of individual's beliefs about the value of the assets being traded. Is it in fact true that this aggregation produces prices that not only reflect some weighted average of individuals' beliefs, but that this averaging process yields correct expectations even if not all individuals in the market are correct?

There are two issues here. First, do prices of financial assets serve as an aggregator of individuals disparate information about the value of assets being traded? Second, does the aggregation process yield better predictions of the value of the assets than the individuals' predictions? There is an old answer to the first question. For any financial asset, the equilibrium price is such that the quantity that individuals want to sell equals the quantity that other individuals want to buy. This price thus reflects all individual's beliefs about the value of the asset. Exactly what type of averaging takes place depends on the details of the individual decision problems, but in any case there is an averaging process inherent in market prices. This reasoning of course assumes that the market price is an equilibrium price. This is an abstraction from reality, just as the idea of a market price is an abstraction, but it has proven to be a very useful approximation.

The answer to the second question is not so straightforward. If all individuals have correct beliefs, and use them correctly in their decision problems, then market prices reflect correct beliefs. The interesting question is what happens if some individuals make mistakes, misuse or ignore information, have bounded computational ability (which usually is not part of the model), have incomplete knowledge of their environment (which also is not part of the usual model), and so on. Averaging beliefs of these traders may yield better predictions than some, but these predictions may be worse than others, and there is no apparent reason for averaging to exactly smooth out errors.

However, there is another force at work in repeated asset markets that may account for the seeming accuracy of the predictions implicit in market prices. Simply put it is that the market selects for those whose behavior is, for whatever reason, most nearly consistent with the correct use of information. This is the idea I will explore. It has a long oral tradition in economics dating back at least to the work of Friedman and Fama. The argument is that smart traders drive out dumb traders, and for this reason the market eventually prices assets as if all traders have correct information and use it correctly.

In joint work Larry Blume and I have argued that under some interesting circumstances this intuition is correct: the market for financial assets selects among simple behavioral rules according to which is closest in relative entropy to a particularly simple rational rule. Further when traders are learning about their environment the market selects among them according to which is most nearly Bayesian. Finally, the market itself can be viewed as a Bayesian entity which aggregates the information residing in disparate individual rules according to a prior and which updates that prior according to Bayes rule.

Bio:
David Easley is the Henry Scarborough Professor of Social Science and a Professor of Economics at Cornell University. His research is in the fields of economics, finance and decision theory. In economics, he and Larry Blume work on learning and selection in stochastic intertemporal general equilibrium models. In finance, his work with Maureen O'Hara focuses on market microstructure and asset pricing. In decision theory, he works on modeling decision making in complex environments.

[handout]

If you would like to meet with David, or for more information, please contact Jeff Hancock.