ost empirical work and empirically oriented illustrations dealing with M the hedging effectiveness of futures contracts utilize either one of two approaches, namely: Risk minimization or payoff maximization. In the first approach, hedging is perceived as a combination of a futures position with an existing cash position that yields a desired exposure to risk on the overall portfolio. Hedging effectiveness in this framework is measured therefore in terms of risk reduction.' In contrast, the second approach assumes that for a given level of risk, a hedged position could provide an improvement in expected return compared with the initial position. Effectiveness, in this second sense, is measured by examining both the risk and return consequences of investing in futures. Up till now, the first approach has been more popular in professional as well as in academic circles, as evidenced by publications emanating from brokerage houses, exchanges, and professional and academic journals. The second measure of effectiveness, although more realistic, has been confined so far to academic publications.
Without probing into the details of these studies, it is useful to re-examine these issues in a different and broader perspective. The purpose of this paper is therefore to present a more general approach to the problem of evaluating the hedging effectiveness of financial futures. Several elements that have been so far overlooked in the literature on hedging will thus be brought to the fore and analyzed. First, we will use a traditional tool of risk theory, namely: The utility function, thus making risk aversion together with expected return, a determinant feature of the solution. Second, the analysis will be presented in The authors are grateful to P. Yourougou, R. Nadeau, J.-C. Cosset, and two anonymous referees of this journal for useful comments.
'To be more precise, this first approach to measuring hedging effectiveness is compatible with traditional theories of hedging as well as with one-period portfolio models. In both cases, the focus is placed on reducing price risk. Indeed, the only difference between the traditional measure of effectiveness and that emanating from the portfolio model is that, in the first case, risk is defined in ex-post terms (e.g. ex-post dollars), whereas in the portfolio model it is evaluated ex-ante (e.g. through the correlation coefficient). See for example Fortin and Khoury, 1984).