中国股市收益率可预测性的实证检验

An Empirical Test of the Predictability of Chinese Stock Market Returns

摘要: 本文以2006年7月至2022年6月为研究区间，选取了8个预测因子，构造了市场组合、行业组合、账面市值比组合和市值组合在内的35个投资组合，运用广义最小二乘法探究了各个因子对各投资组合收益率的样本内和样本外预测性。研究发现：1) 中国股市不存在账面市值比效应，但在一定程度上符合市值效应。2) 现金收益价格比、股息价格比、每股收益价格比因子的样本内外预测能力均较强，通货膨胀率、股票波动率因子的样本内外预测能力均较差，大部分因子样本内外的预测结果具有一致性。

Abstract: This article takes July 2006 to June 2022 as the research interval, selects 8 predictive factors, constructs 35 investment portfolios including market portfolio, industry portfolio, book to market ratio portfolio, and size portfolio, and uses the generalized least squares method to explore the in-sample and out-of-sample predictability of each factor on the return rate of each investment portfolio. Research has found that: 1) There is no book to market effect in the Chinese stock market, but it is to some extent consistent with the size effect. 2) The cash flow to price ratio, dividend to price ratio, and earnings per share price ratio factors have strong predictive power both in-sample and out-of-sample, while the inflation rate and stock volatility factors have poor predictive power both in-sample and out-of-sample. Most of the factors have consistent predictive results both inside and outside the sample.