許多實證研究指出股票投資報酬率並不一定服從常態分配，因此投資人可藉由極大化預測報酬率及偏態，建構一個投資組合，以規避下方 風險。本文以1997年金融風暴前後，臺灣股市資料為例，利用多項式目標規劃法，藉由極大化股票報酬率及偏態，來檢驗偏態投資組合能否獲 得較大盤為佳之報酬。實證結果發現，在金融風暴後，利用多項式目標規劃法所建構之偏態投資組合，普遍具有抗跌效果，當大盤下跌時，偏 態投資組合之報酬率下跌幅度較小；但當大盤上漲時，偏態投資組合之報酬率則低於買進與持有策略。此外實證結果亦發現投資組合之偏態效 應，不具備長期持續效果，亦即投資者於進行投資組合操作時，仍須不斷修正其偏態投資組合之投資比率。

Lots of empirical studies indicate that the distribution of equity returns is not necessarily a normal distribution. When the presence of skewness occurs, it is possible for the investors to construct an equity portfolio by maximizing equity's expected return and skewness simultaneously, consequently avoiding the market's downside risk. This paper uses the polynomial goal programming to maximize the expected return and skewness, and to test whether the optimal equity portfolio can earn an excess return than the stock index. The evidences conclude that, after the Asia financial crisis, our selected portfolios have better performance than the stock index in the bear market. However, in the bull market, the evidences show that the buying and holding strategy outperforms our equity portfolio. In addition, it is apparent that the performance persistence of the selected portfolio with skewness is limited. Rather, it appears that the investors or portfolio managers should regularly by periods modify their investment weights for the portfolio based on skewness.