Price-Path Convexity, Extrapolation, and Short-Horizon Return Predictability

Abstract

The curvature of intramonth stock price paths, which is distinct from cumulative return over the same period, contains significant additional return predictive power. In the cross section, stocks with the least convex price paths subsequently outperform stocks with the most convex price paths. This effect ranges from 1.34% to 1.53% per month and is not driven by small stocks, the bid-ask bounce, or other short-term return predictors. We find similar results in different time periods, in non-US G7 countries, and even at the aggregate market level. We argue that price-path convexity uniquely captures investor over-extrapolation of recent price changes. Therefore, our results provide broad evidence that extrapolative expectations are an important contributor to short-horizon return predictability.