A REASON TO SMILE: The Effect of Twitter-Derived Sentiment on the Stock Market

Abstract

This paper studies the effects of multiple dimensions of investor sentiment on cross-sections of the stock market. Novel sentiment measures for happiness, sadness, surprise and hesitance derived from Twitter emoticons are assessed for their predictive abilities through statistical analysis. I predict that the emotions happiness and sadness will display the strongest predictive abilities for crosssections of the stock market that are predominantly held by retail investors: namely, small cap stocks and high growth stocks. Consistent with this hypothesis, empirical results in this study suggest that lagged values of happiness and sadness not only contain significant predictive information about future returns on the stock market, but also disproportionately affect small cap and high growth stocks. Although this study does not find significant effects of surprise and hesitance on stock market returns, the predictive abilities of Happiness and Sadness are a testament to social media as a tool for real-time sentiment measurement.