投资者情绪指标与股票市场——基于扩展卡尔曼滤波方法的研究  

A Study on Investor Sentiment Index and Chinese Stock Market Based on the Extended Kalman Filter Method

作　　者：池丽旭[1] 张广胜[1] 庄新田[2] 宋大雷[3] 

机构地区：[1]沈阳农业大学经济管理学院,辽宁沈阳110866 [2]东北大学工商管理学院,辽宁沈阳110004 [3]中科院沈阳自动化研究所机器人学国家重点实验室,辽宁沈阳110016

出　　处：《管理工程学报》2012年第3期122-128,165,共8页Journal of Industrial Engineering and Engineering Management

基　　金：国家自然科学基金资助项目(70871022);教育部社科研究青年基金资助项目(12YJC790018)

摘　　要：基于扩展卡尔曼滤波(EKF)方法,首次构造出过滤市场噪声的投资者情绪指标,并在此基础上应用向量自回归模型分析我国投资者情绪指标与股票收益、股本规模等因素的经验关系,实证结果表明:(1)扩展卡尔曼滤波方法可以获得一个更加清晰反映投资者情绪的状态变量;(2)情绪的变化量比情绪指标本身具有更强的市场收益预测能力;(3)大规模公司股票的收益对投资者情绪的影响程度高于小规模公司股票,而投资者情绪对小规模公司股票的影响显著高于大规模公司的股票,并且情绪波动能够预测小规模股票的短期收益惯性和跨期收益反转的特征,证明情绪波动是影响资产定价的重要主观因素。Researchers in the behavioral finance field have been trying to explain the market anomalies. Investor sentiment is a belief about future cash flows and investment risks that cannot be rationally explained by factual data. This paper focuses on investor sentiment and its relationship to stock returns in Chinese stock market. The objeetive of this paper is to explain the relationship between sentiment and stock returns in Chinese stock market from January 2004 to December 2009. We conduct our analysis in two main steps.First we measure investor sentiment based on the EKF （Extended Kalman Filter） method. Second, we test the relationship between investor sentiment and stock returns using the VAR （Vector Auto-Regressive） model. The way to measure investor sentiment becomes the key point because the main purpose of this paper is to test investor sentiment in Chinese stock market. The commonly used sentiment measures in China include closed-fund discount rates, IPO scales and consumer confidence index, etc. However, most of these indices are noise proxies for sentiment. These noise proxies may confound the relationship between investor sentiment and stock returns. Therefore, we use the EKF methodology to filter the market noise in these proxies and extract a clear sentiment variable in the first part. In the second part, we adopt the VAR model to investigate the investor sentiment and its relationship to stock returns. We notice that the level of investor sentiment shows strong self-correlation and weak connection with stock returns in both Shanghai and Shenzhen stock markets. Furthermore, we test the relationship between sentiment fluctuations and stock returns and find that sentiment changes have significantly effect on stock returns. In addition, we compare the effect of investor sentiment on small firms and large firms, and find that large companies＇ stocks has stronger impact on investor sentiment than small companies＇ stocks do. On the other hand, investor sentiment has stronger effect on stock retu

关 键 词：投资者情绪 扩展卡尔曼滤波 向量自回归模型 股票收益 

分 类 号：F830.91[经济管理—金融学]