Amit Goyal, Narasimhan Jegadeesh, Avanidhar Subrahmanyam
Review of Finance, Volume 29, Issue 1, January 2025, Pages 241–273, https://doi.org/10.1093/rof/rfae038
Ever since the publication of Jegadeesh and Titman (1993), the profession has been puzzling over the causes of stock return momentum (the tendency for the relative performance of stocks over the past 3-12 months to repeat itself). Because this evidence is not explained by well-known asset pricing models, the literature proposes and empirically tests several other explanations. Although the empirical work testing the explanations mostly uses U.S. data, momentum strategies are profitable in many international markets as well. Because the pattern of momentum is similar in both the U.S. and internationally, a natural question that arises is which, if any, of the explanations apply internationally. This is the issue addressed in the paper.
One strand of momentum explanations proposes behavioral theories. One of them proposes that momentum is due to investor overconfidence. The idea here is that overconfidence builds as investors receive public signals that confirm their initial beliefs but does not subside equally when they receive contradictory ones, which causes investors to continue to overreact to an initial belief. Another prominent explanation, labeled “frog-in-the-pan” hypothesis, posits that because of their limited attention span, investors underreact to information that arrives in small bits but correctly react when information arrives in large chunks. Another explanation in the same category is that stocks less followed by analysts have slower news diffusion, and this leads to overall momentum. A further explanation posits that investors anchor their valuation to 52-week high prices and hence stocks that are close to this high tend to be perceived as expensive and therefore underreact to good news (and vice versa). A risk-based explanation posits that momentum is due to time-varying value of real options. Under this explanation, winning stocks have more growth options, implying higher risk, and therefore higher required returns.
We test the above explanations with international data using regression based and portfolio approaches. We test them with the full international sample and within subsamples stratified into developed and emerging markets. Across all samples, we find consistently strong support for the frog-in-the-pan explanation. This explanation is proxied by a quantity that represents the degree to which the signs of daily returns match the sign of the past momentum return. To a lesser extent, we find support for overconfidence, captured by the market-to-book ratio. We also obtain confirmation of another U.S.-based result across all samples: that momentum profits are higher in up-market and low-volatility states; these are states hypothesized in the literature as those in which investor confidence is greater.
Overall, our paper provides strong international support for the notion that momentum arises from slow news diffusion, and to a more modest extent, investor overconfidence.