本文藉由二階隨機優越 (Second-Degree Stochastic Dominance, SSD) 與1.5階隨機優越 (1.5-Degree Stochastic Dominance, 1.5SD) 法則篩選出SD贏家與SD輸家之股票，並且檢測買進SD贏家與賣出SD輸家的套利投資組合是否有顯著超額報酬。隨機優越不像平均數-變異數 (mean-variance)需要假設特定之報酬分配或者效用函數，僅以過去報酬率排序的方式判斷風險性資產之優越性，並且不同於動能策略，隨機優越有著理論的支持。 實證結果發現，藉由過去6個月形成期為基礎，不論利用二階隨機優越法則或1.5階隨機優越法則選出的SD贏家投資組合皆有顯著的額外報酬，並且買進SD贏家放空SD輸家之套利投資組合無論持有期為3個月、6個月、9個月或是12個月，超額報酬皆為正值。除此之外，此超額報酬使用CAPM單因子、Fama-French三因子以及Carhart四因子迴歸模型皆無法圓滿解釋。 ;In this paper, we construct SD-winner and SD-loser portfolios based on 1.5 and second degree stochastic dominance rules and examine the performance of arbitrage portfolio which by longing SD-winner stocks and short selling SD-loser stocks. Compared with the most widely accepted Mean-Variance framework, SD rules require neither a specific return distribution nor a specific utility function. SD rules just need to determine SD-winner or SD-loser on ranking previous 6-month returns. In empirical test, we form portfolios through the previous 6-month ranking period and hold them up to 12 months. The results show that SD-winner portfolios produce statistically, significant abnormal returns and arbitrage portfolio has positive excess return. Moreover, these returns are robust with respect to the single index CAPM, the Fama-French three-factor model, the Carhart four-factor model under various criteria of the SD investment strategy.
