中国股票市场收益率的可预测性研究  

Return predictability in the Chinese stock markets

作　　者：蒋志强 田婧雯 周炜星 JIANG Zhi-qiang;TIAN Jing-wen;ZHOU Wei-xing(School of Business,East China University of Science and Technology,Shanghai 200237,China)

机构地区：[1]华东理工大学商学院金融学系

出　　处：《管理科学学报》2019年第4期92-109,共18页Journal of Management Sciences in China

基　　金：国家自然科学基金资助项目(U1811462; 71532009; 71571121);中央高校基本科研业务费资助项目(222201718006)

摘　　要：构造了包括中国A股市场组合、行业组合、账面市值比组合和市值组合在内的31个投资组合,选取了8个预测因子(账面市值比、股利分配率、股息价格比、股息收益率、每股收益价格比、现金收益价格比、通货膨胀率、股票波动率),运用了可行拟广义最小二乘法对各因子对各投资组合收益率的可预测性进行了样本内和样本外检验.研究发现:1)中国股市收益率是可预测的,但是各投资组合收益率的可预测性在样本内、样本外、熊市和牛市均存在差异;2)绝大部分投资组合收益率的可预测性可由条件CAPM模型捕捉的时变系统风险溢价充分解释;3)行业投资组合收益率的可预测性与行业集中度显著负相关,可由信息摩擦理论解释.This paper investigates the return predictability to understand the theoretical foundation of quantitative investments in Chinese stock markets.Eight factors(book-to-market ratio,dividend-payout ratio,dividend-price ratio,dividend-yield ratio,earnings-price ratio,cash flow-price ratio,inflation rate,stock variance)are employed to carry out in-sample predictability tests by means of the feasible generalized least squares method(FQGLS).Out-of-sample predictability tests are conducted in different economic cycles.It is found that:1)The portfolios can be successfully predicted and the predictability is dependent on the cycle and the portfolio type;2)The predictability of most portfolios can be explained by the conditional CAPM model;and 3)The return predictability in industries is significantly negatively correlated with industry concentration.

关 键 词：可行拟广义最小二乘法 样本内预测 样本外预测 条件CAPM 行业集中度 

分 类 号：F830.91[经济管理—金融学]