基于分位数Granger因果的网络情绪与股市收益关系研究

发布时间：2018-01-19 02:32

本文关键词： 分位数回归 Granger因果检验网络情绪股市收益微博数据　出处：《管理科学》2017年03期 　论文类型：期刊论文

【摘要】：行为金融学理论认为,股票市场的价格变动除受宏观基本因素影响外,还在很大程度上受众多个体投资者或噪音交易者行为左右。中国股票市场拥有庞大的个人投资者群体,且股民群体与网民群体之间具有高度耦合性,使用网络情绪等信息能够探索中国股市收益变动基本规律。为揭示个体投资者行为对股市收益的影响,以个体投资者情绪为视角,以网络环境中个体投资者的情绪信息为切入点,检验投资者情绪与股市收益的关联关系,评估网络情绪信息价值。使用中文文本情感分析方法,从新浪微博文本中提取出网络情绪时间序列;分别运用均值Granger因果和分位数Granger因果检验方法,探讨网络情绪波动与股市收益之间是否存在因果关系;将股票市场发展阶段进行细致划分,研究不同市场阶段下网络情绪波动与股市收益之间的因果关系。对沪深300指数收益进行实证研究,结果表明,尽管在均值框架下网络情绪波动与股市收益之间因果关系并不明显,但基于分位数Granger因果分析却发现两者在极端分位点区间处存在广泛且显著的因果关系。数据显示,在40个因果关系检验中,分位数Granger因果检验的因果关系发现了23个显著的因果关系,发现率为57.5%,远高于均值Granger因果检验的7.5%。此外,股市收益受到网络情绪波动影响的程度和方式在不同市场阶段下有所不同。研究结果具有一定的理论意义和应用价值。在一些特定分位点区间网络情绪波动对股市收益存在显著因果关系影响,这为在特定条件下股市收益的可预测性提供了佐证。网络情绪能够预测股市收益的尾部(上尾或下尾)行为特征,可以为金融风险防范提供决策参考。研究结果为股票市场的定价、收益预测和波动率估计等相关研究提供了新的研究思路,也为网络情绪信息使用提供了新的方向。
[Abstract]:The theory of behavioral finance holds that the price change of stock market is not only influenced by macroscopic basic factors. The Chinese stock market has a large group of individual investors, and there is a high degree of coupling between the investors and Internet users. In order to reveal the influence of individual investors' behavior on stock market returns, we can explore the basic rules of stock market returns by using information such as online sentiment. Based on the emotional information of individual investors in the network environment, this paper examines the relationship between investor sentiment and stock market returns, and evaluates the value of online emotional information. The online emotional time series is extracted from the text of Sina Weibo. Using the mean Granger causality test and the quantile Granger causality test, the paper discusses whether there is a causal relationship between the network emotion fluctuation and the stock market return. The development stage of stock market is divided into two parts, and the causality between the network emotion fluctuation and the stock market return is studied in different market stages. The empirical research on the Shanghai and Shenzhen 300 index returns shows that. Although the causal relationship between online emotional volatility and stock market returns is not obvious under the mean value framework. However, based on the quantile Granger causality analysis, we found that there was a wide and significant causal relationship between the two in the extreme locus interval. The data showed that there were 40 causality tests. In the quantile Granger causality test, 23 significant causal relationships were found, and the discovery rate was 57.5, which was much higher than that of the mean Granger causality test. The degree and mode of the stock market income affected by the network emotion fluctuation are different in different market stages. The research results have certain theoretical significance and application value. There is a significant causal relationship between market returns. This provides evidence for the predictability of stock market returns under certain conditions. Internet emotions can predict the tail (upper tail or lower tail) behavior characteristics of stock market returns. The research results provide a new approach for the stock market pricing, income forecasting and volatility estimation. It also provides a new direction for the use of network emotional information.
【作者单位】：合肥工业大学管理学院;合肥工业大学过程优化与智能决策教育部重点实验室;
【基金】：国家自然科学基金(71671056) 国家社会科学基金(15BJY008) 教育部人文社会科学研究规划基金(14YJA790015)~~
【分类号】：C912.6;C913.4;F832.51
【正文快照】： 为揭示个体投资者行为对股市收益的影响,以个体投资者情绪为视角,以网络环境中个体投资者的情绪信息为切入点,检验投资者情绪与股市收益的关联关系,评估网络情绪信息价值。使用中文文本情感分析方法,从新浪微博文本中提取出网络情绪时间序列;分别运用均值Granger因果和分位数G

【相似文献】