Abstract

Using a decomposition approach, this paper provides new cross-sectional evidence in supporting the pricing role of idiosyncratic risk, and reconciles the idiosyncratic risk puzzle documented in the literature. When the total idiosyncratic risk consists of both the long-run and the short-run components, our empirical tests document that the long-run idiosyncratic volatility is positively related to future returns, while the short-run idiosyncratic volatility is negatively associated with future returns. These results are robust to model specifications, sample periods, different samples of stocks, and the possible January effect. Moreover, our evidence suggests that the conflicting results offered by existing studies on the pricing of idiosyncratic risk is a result of different weightings toward the long- or the short-run component in a particular measure of idiosyncratic risk.