Idiosyncratic Volatility Measures and Expected Return

We test whether expected idiosyncratic volatility is related to the cross section of asset returns. We find that, contrary to several recent papers, expected idiosyncratic volatility has no reliable relationship to expected returns. Further, realized contemporaneous idiosyncratic volatility does have a positive relationship with expected returns - this relationship is driven by unexpected idiosyncratic volatility. A look-ahead bias that has been present in recent papers has led to false conclusions about the relationship between expected idiosyncratic volatility and expected return. Our findings are robust to several choices of volatility forecasting models and systematic factor models.