Using daily news tone data between 2017 and 2020 we examine whether news tones can predict stock returns in Chinese A-share market. We first document that the news tones significantly and positively predict the cross-sectional stock returns over next day and over the next 12-weeks. When we separate the news into online news and paper news the online news exhibit strong predictive power for future returns while the printed news only displays marginal predictive power. We hypothesize that the online news is more related to firm fundamentals while the paper news is more linked to political aspects of firm information. Our results using earnings surprises and SOE subsamples provide supportive evidence for the hypothesis.
G10 G11 G12 G14.
Media tone text-analysis cross-sectional prediction.