社会杂志 ›› 2017, Vol. 37 ›› Issue (2): 51-73.

• 专题一:大数据与社会 • 上一篇    下一篇

网络舆情是否影响股市行情?基于新浪微博大数据的ARDL模型边限分析

陈云松1, 严飞2   

  1. 1. 南京大学社会学系;
    2. 清华大学社会学系、斯坦福大学亚太研究中心
  • 出版日期:2017-03-20 发布日期:2017-03-20
  • 通讯作者: 陈云松 E-mail:yunsong.chen@nju.edu.cn
  • 基金资助:

    本文得到国家社会科学基金一般项目“中国民众主观阶层的结构特征和动力机制”（16BSH011）和清华大学自主科研计划的支持。

Does Online Sentiment Predict Stock Market Indices? The ARDL Bounds Tests Based on Sina-Microblog Data

CHEN Yunsong1, YAN Fei2   

  1. 1. Department of Sociology, Nanjing University;
    2. Department of Sociology, Tsinghua University;The Walter H. Shorenstein Asia-Pacific Research Center, Stanford University
  • Online:2017-03-20 Published:2017-03-20
  • Supported by:

    This study is supported by the National Social Science Funding Project "Determinants and Dynamics of Social Class Identification among the Chinese in the Transitioning Period"(16BSH011) and the Tsinghua University Fundings for Self-Selected Topic Project.

PDF

1221

摘要：

本文基于新浪微博大数据，分析互联网上的股市舆情是否影响真实世界中的股市行情。在梳理网络舆情，特别是微博影响股市的机制的基础上，我们利用具有“利好”和“利空”含义的股市术语的微博出现词频（“热词指数”），生成股市的“微博信心指数”。“格兰杰因果检验”和“自回归分布滞后模型”（ARDL）边限检验表明：在股市震荡期，早前三天内的“微博信心指数”有助于预测上证指数；“微博信心指数”和“上证指数”存在正向相关的均衡关系；在股市行情平稳期，以上的统计关联并不存在；网络舆情通过影响入市资金流进一步影响股市行情。

Abstract:

This study focuses on the relationship between online sentiment and the stock market as both play a significant role in modern life. More specifically, based on a time series analysis of Sina-Microblog data and SSE composite index data, this paper explores whether and how Sina-Microblog feeds affect stock market trends in China. We extracted the frequency of occurrence of stock market-related terms on Sina-Microblog to construct an indicator of online market sentiment. Considering the complexity of the effects of online sentiment on the process of decision-making, we divided the stock market into low volatility and high volatility to analyze the effects. Granger tests using the T-Y process and co-integration analyses based on bound tests using the ARDL model show that:1) the indicator of online market sentiment in previous 1-3 days is a statistically significant predictor of the daily SSE composite index; 2) there is long-term relationship between online market sentiment and the SSE composite index; 3) the relationship is insignificant when the market volatility is low, and 4) the relationship is mediated by the capital flow into the stock market. As an empirical study of the online sentiment influences real economic and social consequences, this paper contributes to our understanding of the vital role that social media plays in the economic and social process.