Impact of sentiments on stock returns, volatility and liquidity

Abstract

This study aims to predict stock characteristics such as returns, volatility, and liquidity of companies involved in sustainable investments, using sentiment analysis. The paper uses lexicon-based sentiment analysis and develops regression-based predictive models for testing the impact of Twitter sentiments on stock market movements using four sentiment scores (positive, negative, total, and directional). It is found that positive Twitter sentiments are better able to predict stock returns and volatility than liquidity. While negative sentiment scores can better predict volatility and liquidity than stock returns. All stock market variables are found to be significantly affected by the total number of tweets, i.e., posting volumes. Our study suggests that investors investing in companies with sustainability interests tend to keep track on their corporate social platforms, and thus, companies involved in sustainable investments should remain active on social networking platforms such as Twitter to maintain their corporate image and enhance social value.