「風險管理」是一門投資者都關注的課題，但是風險的發生無法準確地被預 測，本文希望將風險控制在可忍受程度之下，找到一組權重能使得投資組合有最 大期望報酬，傳統衡量風險方法中，變異數和平均絕對離差法皆只考量整體的平 均離散程度，卻忽略極端値的發生；風險値、條件風險値和最小順序統計量等方 法，其著重於尾部損失，卻忽略整體的風險。本文提出一個新的風險衡量指標-風險加權法，希望此新的風險衡量指標兼具兩種不同衡量風險的特性，一方面考慮整體的風險另一方面考慮尾部損失發生的可能性，避免金融市場又有大波動時造成的過度損失。本文建立了一個比較的基準，來探討新的風險衡量指標與傳統 風險衡量指標應用於投資組合表現的差異性，根據實證結果顯示：本文所提出新的風險衡量指標-風險加權法在不同的風險忍受程度下，不管在平穩時期或是金融海嘯時期，實際持有投資組合時，其平均報酬的表現與其他傳統風險衡量指標比較起來，並不會是最好的，但也不會是最差的能維持在一定的水準，不同的傳 統風險衡量指標會隨著景氣的變化而變動很大，但風險加權法屬於一個穩健的風險。

“Risk Management” has become a vital topic. Since the occurrences of risks cannot be respectively pinpointed, we would like to construct robust risk measurement methods to keep risks within tolerable levels, while yielding maximum expected returns. Using traditional methods of risks measurements, variance and mean absolute deviation only take the overall discrete degree into consideration, but does not take occurrences of outlier and extreme values into account. Methods such as Value-at-Risk, Conditional Value-at-Risk and Minimum stress the importance of expected tail loss, at the expensive of the overall risks. The purpose of this paper is to propose a new risk management index, Risk-weighted method. This method will incorporate two orientation for assessing risks: to evaluate overall risks while taking the possibility of expected tail loss into consideration and to avoid losses caused by major disruption in financial market, guaranteeing certain degree of expected returns at identical risk conditions.