Dual Industry Effects and Cross-Stock Predictability

This paper introduces the Peer Index (PI), a measure capturing dual industry-related effects in cross-stock predictability: the overall strength of a firm’s peer group and its relative position within the peer group. PI robustly predicts future stock re-turns, earnings surprises, and earnings growth at both the industry and stock levels across short and longer horizons. Its predictive power persists even after controlling for expected returns derived from machine-learning models applied to firm-own characteristics. We provide evidence that markets underreact to peer-related information, with the PI effect stronger when information uncertainty is higher and investor attention lower, driving cross-stock predictability.