Negative Twitter Sentiment and Analyst Behavior

Abstract

This study examines whether sell-side equity analysts use information in social media to improve their earnings forecasts. Using Bloomberg’s daily Twitter sentiment data for S&P500 firms, we show that negative Twitter sentiment contains useful information about future firm-level earnings and earnings surprises. This effect is distinct from the predictive power of traditional news sources and macro-economic indicators. Analysts who cover firms with greater sensitivity to negative Twitter sentiment, issue relatively less optimistic and more accurate earnings forecasts, especially when the information environment is more opaque. Using an exogenous event that changed the information content of individual tweets and a difference-in-differences design, we establish a causal relation between Twitter sentiment and analyst forecasts.

Keywords: Social media, Twitter sentiment, earnings forecasts, sell-side analysts, firm-specific information.

JEL Classification: G14, G24.