China | Measuring news media sentiment using Big Data for Chinese stock markets

We construct five sentiment measures based on the GDELT database, representing the Tone, Optimism, Attention, Tone Dispersion, and Emotional Polarity of Chinese stock markets. All these news media sentiment measures are shown to have significant predictive power for Chinese stock market returns and volatilities.

Key points

• Key points:
• Using the big database of news reports from GDELT, we construct four sentiment measures, namely, the general Tone (daily average tone change), Optimism (proportion of news reports with positive tones), Attention (number of news reports), and Tone dispersion (standard deviation of article tones) for the Chinese stock markets.
• Our results suggest that news media sentiment plays a significant role in predicting future returns and volatilities in China. Higher news media sentiment (Tone and Optimism), fewer news media coverage (Attention), and a smaller news media tone dispersion imply higher next-day market returns.
• For market volatilities, lower news media sentiment (measured by Tone and Optimism), more news media attention, and a larger news media tone dispersion, may indicate much larger market volatilities.
• We also document the existence of asymmetric sentiment effects on stock market returns and volatilities in China. Market returns and conditional volatilities overreact to negative shocks to the news media sentiment, and these asymmetric sentiment effects are more profound for the Shenzhen stock market.

Authors

Geographies

Topics

Tags