以隨機優越法評估美國股票市場元月效應之實證研究

Abstract

本論文旨在使用幾乎隨機優越與廣義幾乎隨機優越這兩種新方法，檢視一月效應的存在。1927至2014年美國NYSE、AMEX與NASDAQ公司的月報酬資料顯示最小規模等分(最高淨值市價比等分)至少幾乎二階隨機優越及廣義二階幾乎隨機優越於較大規模等分(較小淨值市價比等分)。也可看到所有公司規模等分的一月報酬皆至少二階幾乎隨機優越及廣義二階幾乎隨機優越於其他月份(除了最大規模等分的十一月、第六至十規模等分的十二月)。證據顯示一月效應確實存在。

This paper aims to examine the existence of January effect by using almost stochastic dominance and generalized almost stochastic dominance. Monthly return data of NYSE, AMEX, and NASDAQ firms from 1927 to 2014 show January returns in the smallest size decile (highest book-to-market decile) dominate larger (smaller) deciles by at least almost second-degree stochastic dominance and generalized second-degree stochastic dominance. Similarly, January returns in all size deciles dominate other months (except for November in the largest size decile and December in sixth to tenth size decile) by at least almost second-degree stochastic dominance and generalized second-degree stochastic dominance. The evidence show that January effect does exist.