The Good the Bad and the Trending: Microblogging Sentiment and Short Term Momentum
35 Pages Posted: 15 Mar 2019
Date Written: December 12, 2018
Over the past ten years microblogging services such as StockTwits and Twitter have become a popular way for users to express thoughts, opinions, and reactions in real time. This study explores the mechanism that connects the content of these posts to stock returns and theories of momentum at both the individual security level and in portfolios. Specifically I construct firm level measures of StockTwits sentiment using data from the StockTwits microblogging website to test theories of momentum from Daniel, Hirshleifer, and Subrahmanyam (1998) and Hong and Stein (1999). I find that trailing measures of sentiment have predictive power over future stock returns. Portfolios formed using StockTwits data have strong explanatory power over the daily momentum factor of Carhart (1997). These findings are consistent with Hong and Stein (1999) and also demonstrate that microblogging services provide an important new data set for testing asset pricing theories.
Keywords: Microblogging, Sentiment Analysis, StockTwits, Stock Returns, Momentum
JEL Classification: G11, G12
Suggested Citation: Suggested Citation