投资者情绪指数构建及其与超额收益率关系研究

A Study of the Construction of Investor Sentiment Comprehensive Index and Its Relationship with Excess Returns

有效度量市场投资者情绪是研究投资者情绪对资本市场影响的关键。选取2008—2022年证券投资者信心指数、消费者信心指数、市盈率、成交量、换手率、新增投资者开户数的月度数据,用主成分分析法构建市场投资者情绪复合指数(CIMIS_(t)),并通过“孪生股票”模型和定性分析对CIMIS_(t)进行有效性检验。采用ARMA-GARCH模型消除CIMIS_(t)和超额收益率(R_(a))的自相关和异方差性,对其残差进行回归分析和Granger因果检验。实证发现,CIMIS_(t)能较准确表征中国股市投资者情绪,去除自相关和异方差性的投资者情绪与R_(a)呈显著正相关,对R_(a)拟合效果更好;短期内超额收益率明显影响投资者情绪,而投资者情绪不能明显影响我国股市的表现。研究结果为投资者在投资决策以及监管者在制定市场政策时提供参考。

Effective measurement of investor sentiment is the key to study the impact of investor sentiment on the capital market.After selecting the monthly data of securities investor confidence index,consumer confidence index,price-earnings ratio,trading volume,turnover and newly opened accounts of investors during the period from 2008 to 2022,this paper constructs a comprehensive index of market investor sentiment(CIMIS_(t))by principal component analysis,and conducts a validation test with Siamese twins and qualitative analysis.The ARMA-GARCH model is used to remove the autocorrelation and heteroscedasticity of CIMIS_(t) and excess return(R_(a)).Regression and Granger causality tests are performed for the residuals of the model.The results indicate that CIMIS_(t) can represent the investor sentiment of Chinese stock market and removed the impact of autocorrelation and heteroscedasticity,the investor sentiment is significantly positively related to R_(a).Besides,the excess return can predict the investor sentiment in short terms,while the investor sentiment cannot significantly affect the Chinese stock market.This paper provides a reference for stock investors in investment decisions and regulators in formulating corresponding market policies.