Quality of Private Signals and Expectations About Returns

Abstract

This paper introduces a new measure of investor sentiment that is constructed using the information shared by traders on Twitter. The main advantage of the proposed index over existing measures of sentiment is the possibility of using the number of followers as a proxy for the quality of private signals. Moreover, the data allows for gauging sentiment directly with high frequency data. The index is used to test the implications of theories in asset pricing. The results show that (1) the follower-weighted sentiment index predicts the same day return of the stock market index, but the equal-weighted index has no predictive power for daily returns, (2) dispersion of expectations about future returns predicts volatility of the stock market returns, (3) information asymmetry is positively related to return volatility, and (4) the density of information arrival measured by the number of opinionated tweets is positively correlated with volatility and trading volume of the stock market index.