本研究整理了有關於投資組合風險衡量模型的相關文獻,並在文獻中介紹四種不同風險衡量的投資組合模型,即均異模型、平均絕對偏差模型、下行風險模型、條件風險值模型,再從這四種模型中加入了再調整(rebalancing)以及放空(short selling)的動作,並以移動式視窗法進行多期交易測試。實驗結果發現,以三種績效指標綜合評估中,即市場價值、期望報酬率、投資風險,可以觀測出下行風險在績效綜合評估中是最差的,而條件風險值模型不論是在未允許放空或者是允許放空的模型當中,績效綜合評估都是最好的。對於允許放空的四種模型中,可以看出隨著放空在目標式權重的重要性增加,不但可以降低投資組合風險的波動性,也可以讓市場價值增高,因此本研究推論在台灣50成分股實際的歷史資料測試時,以條件風險值作為風險衡量的投資組合再調整模型,績效會是最好的。
With different risk measurement, the portfolio selection models including the Mean Variance model, the Mean Absolute Deviation model, the Downside Risk model, and Conditional Value at Risk model have been studied. Using the rolling window technique, multi-period trading simulation is performed while short selling is also taken into account for these four models. Three kinds of performance assessments which are market value, expect return, and standard deviation are used to compare the performances among these four models. The Conditional Value at Risk model exhibits superior performance no matter whether we take the factor of short selling into account or not. According to the four models which are taken the factor of short selling into account, we observe that with the increased importance of the objective of short selling, it’s able not only to reduce the volatility of the portfolio risk but also increase the market value. Therefore, according to this study, we can infer that conditional value at risk measure of portfolio rebalancing model can exhibit superior performance in simulating the historical data of stocks which consists in Taiwan 50 index.