摘要
原有CAPM忽略了代表性投资者对高阶矩如偏度和峰度的偏好,也没有将上升风险和下跌风险进行有效分离(将高于目标受益的波动也认同为风险)。为了突破这两点假设,本文讨论了基于代表性投资者双侧矩和高阶矩偏好的资产定价方法,并用中国A股市场数据进行了实证检验,所得结论如下:(1)在影响股票收益的所有特征因子中,规模效应在国内股市存在,而账面-市值比不是一个独立的定价因子。(2)将市场做上下划分并纳入高阶矩偏好,对应的市场因子较特征因子能更好的解释股票的截面收益;通过加入规模因子和账面-市值比因子构建多因子模型进行检验,该结论照样成立。(3)投资者并没有将高于平均回报的"牛市"波动视为一种"收益"指标,而将其与下跌风险指标等同。
CAPM assumes that the representative investor doesn't care too much about higher moments of security return, such as skewness and kurtosis, and regards the upside risk and downside risk as the same. Asset pricing models based on the representative investor' preferences towards bilateral and higher order moments are put forwards to overcome these two shortcomings. Empirical test of Chinese stock market indicates that: ① of all these characteristic factors, ME factor can explain the cross sectional returns of stocks better but BE/ME is not an independent pricing factor; ② market factors of asset pricing models based on the preference towards bilateral and higher order moments show stronger explanatory power by contrast to characteristic factors such as ME and BE/ME; ③ the investor views upside volatility of return as a kind of risky factor.
出处
《管理评论》
CSSCI
2008年第3期3-7,共5页
Management Review
