The use of futures contracts has grown enormously in recent years. From 1979 to 1985 the number of futures contracts traded literally doubled. Most of the growth can be attributed to the development of recent contract innovations. Trading in financial futures, alone, increased sixteen fold over this period. This remarkable rise in futures usage and the importance of innovation highlights the constant struggle by exchanges to develop and initiate successful contracts. However, there is no known process for actually identifying potentially successful contracts. lt is this general question of what leads to a successful contract that forms the initiative behind this work. Formally, this study is a theoretical and empirical analysis of futures usage. The purpose of the theoretical section is to develop a model of contract usage that leads to a set of testable hypotheses about the determinants of contract use. Usage is defined in this study as being measured by the number of contracts in existence at a certain time. The theoretical work is general in the sense that it is not directed at behavior in one specific contract; but rather, it rests on the belief that certain underlying fundamental economic factors will affect, in general, usage in all futures contracts. The theoretical model is based upon firm behavior in an uncertain world with the firm having the ability to enter a portfolio of futures contracts. The purpose of the empirical section is to provide support for the theoretical section by determining, through time series analysis, which fundamental variables affect futures usage and how these effects are transmitted. The exogenous variables center upon the variance-covariance matrix of actual price series, transactions costs, and production levels. The empirical results yield strong support for the theoretical section developed in this work and the overall portfolio approach. Additionally, the results draw into question the importance of several variables which have classically been considered essential in determining usage. While the results strongly support the model and the portfolio perspective, they did not suggest a specific set of variables that uniquely determine contract usage across a wide set of different contracts. / Ph. D. / incomplete_metadata