A Note on Stein’s Overreaction Puzzle

Recently, Christoffersen et al. (2013) argue that the overreaction puzzle of Stein (1989) can be explained by a variance-dependent pricing kernel. In this note, we challenge this view. Our theoretical results are in line with their argument that the variance under risk-neutral measure is more persistent than the variance under physical measure due to a negative variance risk premium. But our results do not support their argument that the more persistent variance is able to qualitatively explain Stein’s findings. We show theoretically that the persistence of the volatility cannot amplify the movements of long-term variance to short-term fluctuations in variance, and, therefore, conclude that Stein’s overreaction puzzle is still unsolved.