Investor Sentiment, Disagreement, and Breadth-Return Relationship

We extend the theory and empirics in Chen, Hong, and Stein (2002) by assuming that investors subject to market sentiment hold a biased belief in the aggregate. With a dynamic multi-asset model, we predict that the breadth-return relationship can be either positive or negative depending on the relative strength of two offsetting forces — disagreement and sentiment. Using the sentiment index developed in Baker and Wurgler (2006, 2007), we find evidence consistent with our predictions. The breadth-return relationship is positive when the sentiment effect is small. However, the relationship becomes negative when (i) the time-series variation of market-wide sentiment is high and (ii) the cross-sectional dispersion of firm-specific exposure to market-wide sentiment variation is large. Our unified framework reconciles a few seemingly inconsistent empirical studies in this literature and explains puzzling cross-sectional return patterns observed during the Internet bubble and the subprime crisis periods