The Existence of Futures Markets, Noisy Rational Expectations and Informational Externalities

Abstract

It is a fact that futures markets exist in some commodities and not others. Similarly, contingent commodity contracts of the type described by Debreu do not exist for all commodities in all states of the world. Any explanation of this phenomenon must be intimately connected with a theory of what functions these markets serve. The KeynesHicks theory of commodity futures markets is that they provide a mechanism by which risk averse speculators insure other risk averse traders who hold (positive or negative) stocks of a commodity subject to price fluctuation.' We propose a new explanation of the role of futures markets as a place where information is exchanged, and where people who collect and analyse information about future states of the world can earn a return on their investment in information gathering. In particular, it is shown how the private and social incentives for the operation of a futures market depend on how much information spot prices alone can convey from to traders. (Firms which have information about future states of the world are called informed , while firms who do not are called uninformed .) In equilibrium, without a futures market, firms will use their information about next period's price to make spot market purchases. The commodity purchase is stored in anticipation of a capital gain. Therefore, the trading activity of firms in the present spot market makes the spot price a function of their information. Uninformed traders can use the spot price as a statistic which reveals some of the traders' information. When the spot price reveals all of the traders' information, both types of traders have the same beliefs about next period's price. In this case there will be no incentive to trade based upon differences in beliefs about next period's price. In general the spot price will not reveal all of the traders' information because there are many other factors ( noise ) which determine the price along with the traders' information. This implies that in equilibrium with only a spot market, and uninformed traders will have different beliefs about next period's price. The difference in beliefs creates an incentive for futures trading in addition to the usual hedging incentive. When a futures market is introduced uninformed firms will have the futures price as well as the spot price transmitting the firms' information to them. This is the informational role of futures markets. The model has the following testable implications. The degree of predictability of a future spot price from only a current spot price determines the private incentives for futures trading in a commodity which has no futures market. For commodities with futures markets the volume of futures trading is directly related to how poorly current and futures prices predict the future spot price, relative to how well various exogenous variables predict the future spot price. How well refers to mean square prediction error conditional on available information, not biasedness of the predictions.