期貨與選擇權之起源，主要在於配合經濟社會規避價格風險之需求。期貨合約之避險理論，大多採用Markowitz的投資組合選擇理論所導出。該理論之隱含假設，乃資產報酬率為常態分配，或投資人的效用函數為二次式。事實上避險後資產報酬率分配未必為常態，而投資人的效用函數更鮮為二次式，因此以變異數降低百分比作為避險效果指標未必恰當。尤其當避險工具為選擇權時，更易產生不對稱 ( asymmetric ) 之報酬分配，因此平均數－變異數分析，便未必能提供公平的比較基礎。 本文之目的，在於探討當避險後報酬率機率分配並非常態，以及投資人效用函數未知之情況下，如何提供更具一般性之避險效果評估指標，以作為選擇避險策略之參考。所考量之指標包括：報酬率偏態、β風險、以及就整個分配進行比較之隨機優勢指標 ( Stochastic dominance ) 與擴充式吉尼係數指標 ( Extended mean-Gini coefficient ) 等。並以1990年11月至1992年10月股價指數期貨及選擇權市場之資料，驗證各策略在本文所建議之各種指標下之相對表現。

The major economic benefit provided by futures and options markets is risk man-agement through hedging. Extensive literature in futures hedge focused on the applica-tion of mean-variance analysis of Markowitz portfolio selection theory. As is well known, mean-variance analysis is based on the assumption that either returns are nor-mally distributed or decision makers have quadratic utility functions. Unfortunately, the assumptions of the mean-variance model are subject to serious criticisms in the em-pirical studies. If the distributions of hedged returns are asymmetric or the utility func-tion of the decision maker is unknown, the traditional mean-variance criteria to evaluate hedging effectiveness will not be suffice. In this article, we propose that, in addition to the first three moments of return distributions, the stochastic dominance rules and the extended mean-Gini coefficient be employed to analyze the performance of alternative futures and options hedging strategies.