Can stock message board sentiment predict future returns? Local versus nonlocal posts
Journal
Journal of Behavioral and Experimental Finance
Journal Volume
34
Date Issued
2022
Author(s)
Abstract
Using textual analysis of stock message board posts, we find that investors’ sentiment expressed through messages can predict future one-day stock returns. For small stocks, message board sentiment can predict up to two-day cumulative future returns. This increased predictive power on small stocks is restricted mainly to local posts, which originate from the provinces in which stocks’ headquarters are located. Nonlocal posts exhibit a stronger trend-chasing sentiment than local posts. Furthermore, we find no evidence of the long-term predictive power of message board sentiment. Overall, our findings support the short-term information advantages of local investors. ? 2022 Elsevier B.V.
Subjects
Investor sentiment
Local versus nonlocal posts
Return predictability
Small stocks
Stock message board
SDGs
Type
journal article