中国股票市场三因素定价模型的实证研究 摘要 本文利用510家沪深两地A股上市公司1998年7月到2009年1月的数据验证了Fama-French(1993)三因素定价模型在中国A股市场的有效性。我们发现一个包含市场因素、总市值因素和账面价值/总市值比率因素的三因素模型可以很好的解释样本股票的收益。只含有市场因素作为唯一变量的传统的资本资产定价模型(CAPM)可以在一定程度上解释样本股票收益,而在传统的CAPM中加入SMB和HML两个变量会显著的提高模型对样本股票收益的解释能力。与美国股票市场不同,变量SMB对样本股票收益的解释能力强于变量HML。变量SMB和HML的平均值均为正,说明样本中小市值公司股票的平均收益高于大市值公司股票收益,高账面价值/总市值比率公司的股票收益高于低账面价值/总市值比率公司的股票收益。对大多数线性回归的结果来说,截距为零的假设不能被拒绝,说明三个变量RM-RF、SMB、和HML能够解释样本股票的收益,模型设定合理。用真实的市场平均收益代替样本平均收益、按可流通比例把样本股票分成高流通比例和低流通比例子样本、把样本区间分为股改前和股改后子区间分别做模型稳健性检验,三因素模型均依然有效。 关键词: 三因素模型、资本资产定价模型、中国股票市场 An empirical study of the three-factor model on Chinese stock returns CUI Duolun (Major in Economics) Directed by Professor Minsoo Lee Abstract This paper uses 510 A-market listed firms’ data from July of 1998 to January of 2009 to test the validity of Fama-French (1993) three-factor model on Chinese stock market. We find strong evidence that the sample stock returns can be explained by the three-factor model which includes a market factor, size and BE/ME ratio. The traditional CAPM model with market factor as the sole risk factor explains sample stock returns to a satisfactory level, and adding variables SMB and HML to the traditional CAPM helps to improve the explanatory power of CAPM effectively. However, variable SMB has stronger explanatory power than HML. The average values of variables SMB and HML are all positive, which means small stock returns are statistically higher than big stock returns, and high BE/ME stock returns are statistically higher than small BE/ME stock returns. The hypothesis that the intercepts of the regressions are zero can not be rejected for most of the regressions, which means that these three variables RM-RF, SMB, HML used together petent to explain excess average returns and the model specification is reasonable. The three-factor model is still valid when we test it with real market return instead of sample market return, divide our samples stocks into high tradable ratio and low tradable ratio
