Do idiosyncratic skewness and kurtosis really matter?

This paper empirically examines the relationship between idiosyncratic volatility, skewness, and kurtosis and future stock returns. We find evidence of significant time-series variation in the average idiosyncratic moments. However, the single and double sorted portfolio-based tests show that only the expected idiosyncratic skewness and volatility have some pricing effects reflected in significant differences in the returns of extreme portfolios. Cross-sectional tests using portfolios of stocks sorted on expected idiosyncratic kurtosis, but not individual stocks, reveal that only the expected idiosyncratic skewness is consistently priced with a significant and negative relationship with expected returns. Furthermore, there is strong evidence that the maximum daily return (MAXRET) is highly significant in all tested models suggesting that stocks with extreme positive returns tend to decline in price in subsequent month, leading to negative returns.