| 摘要: | 我們利用大規模的全球個股對風險因子模型與個股特徵模型進行檢定。我們採用Brennan, Chordia and Subrahmanyam (1998)建議的個股橫斷面迴歸模式檢定下列兩個因子模型: Connor and Korajczyk (CK, 1986) 的主成分分析因子模型和Fama and French (FF, 1993)的特徵因子模型。我們發現規模效果和淨值市價比效果經過FF模型調整後依舊存在，而之於CK模型並不存在。此外，價格效果、短期動量效果和長期反轉效果在因子調整前存在,但在CK和FF因子調整後均不存在。最後，CK調整後的截距項和迴歸係數皆不顯著，而FF調整後卻非如此。這些證據顯示CK模型提供對於全球股票報酬一個理性架構的解釋。 We use a large-scale global data to test risk-based factor models against specific non-risk characteristics on individual stocks. We employ the individual cross-sectional regression analysis, suggested by Brennan, Chordia and Subrahmanyam (1998), to test two different factor model, the asymptotic principal components approach of Connor and Korajczyk (CK, 1986) and the characteristic-factor based approach of Fama and French (FF, 1993). Generally, the size and book-to-market ratio effects are still strong in the FF method, but the CK factors attenuate the magnitude and statistical significance of these effects. Moreover, there are strong evidence of price effect, short-run return momentum and long-run return contrarian before FF risk-adjustment, but there is no evidence after FF risk-adjustment. Finally, after adjusting risk based on CK factors, the intercept term and most coefficients are insignificantly different from zero as predicted by the null hypothesis, but not for the case of using FF factors. It is apparent that the CK factor model provides a more complete description of global equilibrium returns and a continuing validity of the rational pricing paradigm. |
