Consistent with recent findings that investors perceive performance by price changes, cross-sectional return extrapolation comes entirely from capital gains. Stocks that do not pay dividends (capital-gain stocks) experience robust return extrapolation, while dividend-paying stocks do not. Using a stock’s dividend-paying status to proxy for exposure to return extrapolation, I test cross-sectional predictions of return extrapolation models. Consistent with return extrapolation models, momentum and long-term reversal are stronger among capital-gain stocks. The value premium, however, is similar for both groups, suggesting that while return extrapolation may help generate momentum and long-term reversal, its role is limited for the value premium.

Online Appendix

A firm’s investment responds to the stock valuations of other firms headquartered nearby. This response is stronger among financially constrained firms, is robust to controlling for the investment of other firms in the region, and is driven by the valuations of large firms. These findings are difficult to reconcile with existing theories that link investment opportunities to firm valuations, but instead suggest that a firm’s access to credit rises and falls with the valuations of other firms located nearby. Consistent with this explanation, financially constrained firms issue more debt and receive lower loan spreads when neighboring firms have higher valuations.

We provide evidence that reference prices impact how investors respond to news. When current prices are farther from a reference price, investors react more strongly to news. We first document that individual investors are more (less) likely to sell a stock following bad (good) news when the stock's trading price is farther from the investor's purchase price. Motivated by this micro-level evidence, we construct a stock-level measure to capture the distance between a stock's trading price and its purchase price for the average investor. We provide evidence that this distance from purchase price produces a substantial amount of cross-sectional variation in the degree to which stocks over- or underreact to news. Stocks trading farthest from their purchase price react more strongly to news than stocks trading near their purchase price. Consistent with relative overreaction, stocks trading farthest from their purchase price also exhibit greater return reversals following news days. We document that a cross-sectional strategy exploiting these return patterns earns a monthly alpha of 0.93%. These findings are distinct from alternative explanations related to size, illiquidity, and volatility. Our evidence instead suggests that reference prices have a meaningful impact on how investors respond to news.


When Do Retail Investors Provide Liquidity? (with Hannes Mohrschladt)

Return Seasonalities: The Role of Institutional Investors (with Dongxu Li)

Anchoring and the Asset Growth Premium (with Ben Blau)