AN EXAMINATION OF THE CLAIM FOR AN INFORMATION FUNCTION OF A FUTURES MARKET

The functions of a futures market and the question of why such markets exist for some commodities and not for others have long been of interest to economists. The discussion and debate has ranged from the Keynes-Hicks risk transfer hypothesis to Working's theory of storage costs and theory of anticipatory pricing. Recently Sanford Grossman and others have suggested that one function of a futures market may be to provide a place where people may earn a return on an investment in information collection. The purpose of this study has been to examine this claim by considering the following questions. (1) Do information differences among traders help to explain why futures markets develop for some commodities and not for others? (2) Does the development of a futures market reduce the informational differential among traders? (3) Can the information differential among traders help to explain the amount of trading activity within a futures market?;A futures market adds an additional price statistic to the current price system. This price increases the transfer of information from the informed who have invested an information collection to the uninformed who observe only current prices. This transfer is limited by the amount of noise in the price system. The second test of the study examines the predictive power of the price system before and after the initiation of futures trading. Using a vector of empirical proxies for available information, the predictability of the future spot price conditional on available information is compared with predictability conditional on information contained in the price system. The results for wool, pork bellies, and frozen concentrated orange juice indicate transfer of information through the futures price but suggest sufficient noise to maintain incentives for trading on information differences. For onions and Maine potatoes, the results indicate insufficient information differences among traders to warrant trading for a return to those differences.;The relationship between trading activity and information differences is examined using a measure of those differences from the Grossman model. The study relaxes an equal storage costs assumption and allows for varying ratios of informed and uninformed traders in an attempt to better characterize a market exhibiting an information function. The study indicates that an information function is valid for some, but not for all, commodities. Specifically, onion and Maine potato markets do not appear to support returns on information collection. Distinguishing those markets that will have increased activity from trading on information differences requires consideration of relative risk aversion among traders, storage costs differentials existing among traders, and relative numbers of traders who are well-informed.;The Grossman model is used as the theoretical base for the study. The first question is addressed by comparing how well current spot prices predict future spot prices in a non-futures market environment for a variety of commodities. "How well" is used in the Grossman sense of mean square prediction error conditional on available information. Those commodities with poor price predictors are most likely to develop a futures market with a return to information function. The results of this first test indicate a positive association between the inability of the current price system to predict the future spot price and the subsequent development of a futures market.