causal machine learning

Using causal machine learning, we examine the influence of sentiment from earnings-related news articles on firms’ return, volatility, and trade volume. Our analysis considers price and volume reactions to different sentiments within varying economic, financial, and aggregated investor mood conditions. The findings indicate significant differences in the effects of sentiment types, larger reactions to negative sentiment, and investors’ general underreaction to news, especially in adverse macroeconomic conditions or high stocks liquidity.