Social Media Toxicity and Capital Markets
Abstract
This paper asks about the existence, determinants, and implications of toxic content in financial social media. Using machine learning algorithms to measure toxicity in Seeking Alpha articles and comments, we find measurable toxicity concentrated primarily in user-generated comments rather than articles. Comment toxicity is greater for firms with more investor attention or uncertainty. We document three key results about toxicity in capital markets. First, toxicity displays a feedback loop in platform participation: past toxicity is associated with more future toxic contributors for a given firm. Second, firms receiving more toxic comments have greater retail trading volume but less informative retail trades, especially when comments disagree. Third, toxicity is associated with slower price discovery around earnings announcements. Our findings suggest that financial social media toxicity relates to both user behavior and market outcomes and raise important considerations for platform governance in financial markets. Data Availability: The Seeking Alpha articles and comments used in this study were obtained directly from Seeking Alpha under a data-sharing arrangement and are proprietary; we are not permitted to redistribute them. These data may be requested from Seeking Alpha. Compustat, Execucomp, CRSP, and New York Stock Exchange (NYSE) Trade and Quote (TAQ) data are available through Wharton Research Data Services (WRDS) to subscribing institutions.
- DOI
- 10.2308/tar-2024-0017
- Pages
- 1-31
- Language
- en
- Sources
- crossref