基于新闻文本数据的我国股市波动性测度及其应用

Measurement and Application of News-based Equity Market Volatility in China

本文基于对我国报纸新闻的文本大数据分析,构建了一个能较好测量中国股票市场波动性的新闻指数(CEMV),并进一步利用我国2005年1月—2020年10月的宏观经济数据,通过建立结构向量自回归(SVAR)模型,实证研究了我国政策性新闻驱动下的证券市场波动对宏观经济及资产价格的影响。研究发现:(1)CEMV指数峰值的出现与我国证券市场历次较大的波动相关,同时CEMV对股票市场指数已实现波动率有显著的解释能力。(2)本文进一步构建了货币政策、财政政策、贸易政策、汇率政策4类特定政策类别的CEMV指数,发现对于不同金融事件的冲击,各指数的反应有所差异。(3)政策性新闻驱动下的证券市场波动冲击会对产出增速、价格水平、股票市场收益率产生短期负面影响,对货币供应增速产生总体正向影响,并且加剧了投资者情绪的短期波动。

This paper constructs a newspaper-based Chinese Equity Market Volatility(CEMV)tracker that moves with the realized volatility of the Chinese stock market.Using the macroeconomic data from January 2005 to October 2020,we further estimate a structural vector autoregressive(SVAR)model to assess the impact of CEMV on the Chinese economy and equity prices.We find that:(1)The CEMV index peaks are closely related to the large market fluctuations.Regressing the realized volatility of the stock market index on contemporaneous EMV values yields a significant slope coefficient.(2)Parsing the underlying text,we construct the CEMV indexes of four policy categories that respond differently to specific market shocks.(3)The stock market volatility driven by policy news foreshadows declines in output growth,price levels and equity returns.Moreover,CEMV has an overall positive impact on money supply growth and exacerbates the short-term fluctuations in investor sentiment.