投资者情绪与股市政策失灵：基于互联网大数据的分析与实证

Both the conclusion of behavioral finance theory and the extreme expression of the stock market in our country show that the stock market policy adjustment needs to consider the role of investor sentiment，otherwise, it is difficult to achieve the expected objectives. However, put the investor sentiment into policy decision-making system faces the following problems: (1) How the Investor sentiment interferes the policy is not clear. (2) It is difficult to make scientific policy evaluation beforehand. (3) The measure of investor sentiment also faces the lack of representation, objectivity and timeliness. In view of the above problems, this program takes the behavioral finance as the theoretical foundation, combined with the advantages of network big data in the analysis of behavior and emotion, and carry out the following research: (1) The policy transmission mechanism and theoretical model of based on both economic factors and investor sentiment. (2) the quantification of Investor sentiment based on big data. (3) The intervention and prevention of stock market policy failure, mainly including the key path identification in the policy transmission mechanism, the evaluation of monetary policy impact based on the investor sentiment and the pre-evaluation of the stock market policy. (4) Evaluation and policy suggestions on the hotspot policy of the stock market. The research of this program is a beneficial supplement to the behavioral finance and policy evaluation methods, and also.It helps to achieve goal of macro prudential supervision in the stock market.

行为金融理论的研究结论和近期我国A股市场的极端表现，都表明股市政策调控需要考虑对冲投资者情绪的作用，否则难以达到预期目标。将投资者情绪纳入政策决策体系面临以下问题：（1）投资者情绪对政策的扰动机制不明；（2）难以实现事前科学评估；（3）投资者情绪的测度还存在代表性不足、客观性不足、时效性不足等缺陷。针对上述问题本课题以行为金融学为理论基础，结合网络大数据在行为和情感分析方面的优势，开展以下研究：（1）基于经济因素和投资者情绪的双路径政策传导机制和理论模型的构建；（2）基于网络大数据的投资者情绪量化；（3）股市政策失灵的干预与防范措施，主要包括基于双路径传导机制的关键路径识别、考虑投资者情绪的货币政策影响和股市政策事前评估的实证研究，（4）对股市热点政策问题的评估及政策建议。本课题研究内容是对行为金融学和政策评估方法的有益补充，对实现股票市场的“宏观审慎”监管具有重要意义。

本研究基于投资者的微观情绪与金融市场的宏观趋势之间的关系，围绕“投资者的微观行为机理——投资者行为对股票市场的影响传导机制——投资者行为与股票市场波动”三项核心工作，取得以下进展： 一是发现了人际互动有助于纠正某些投资者的非理性行为。以处置效应为例，基于网络金融社区的多源异构数据的建立了投资者微观行为实验机制，利用在线社交及模拟交易数据，构建了社交中心性、社交发布内容、社交参与度和社交关注四个社会互动维度，分析了不同维度社会互动对投资者处置效应的影响。结果表明，社会互动会显著降低投资者的处置效应，更高的社交参与度、尤其是在交易时间内的社会互动ꎬ可以通过提高投资者卖出亏损股票的比例来降低处置效应。二是发现了外部事件冲击下、投资者微观行为和股票市场宏观趋势之间的影响机理。我们选择雾霾作为公共事件的代理变量，考察雾霾冲击下投资者行为与股市波动之间的关系。发现雾霾对股市的影响是依靠投资者情绪的中介作用产生的，这一影响对股市整体表现出时变特征；而对于不同的行业股票，表现出异质性影响，投资者情绪对总营收和市值较小、市盈率较高的行业受到的影响更加显著。三是基于投资者对股市影响的机理、构建了TOP-SVR风险管理模型。基于热最优路径方法识别投资者的时变影响、选出投资者情绪领先于价格变动的股票，并在此基础上利用SVR模型对此类股票价格进行预测，取得了非常好的预测效果。.基于上述系列工作，本项目构建了股票市场微观行为与宏观趋势的关联分析框架，基于相关理论和方法方面的研究，在国际期刊发表JCR一区论文3篇、JCR二区论文4篇，在国内期刊发表基金委管理学部认可的A类期刊论文3篇。