Idiosyncratic moments and the cross-section of stock returns in Brazil

Abstract

This online appendix reports additional robustness checks for our main results. Wepresent a set of tables with summary statistics for portfolios sorted on higher idiosyncraticmoments (expected skewness, realized skewness, and realized volatility) using two alternativetime horizons to estimate the measures of idiosyncratic skewness and idiosyncratic volatility:T = 24 months and T = 60 months. For both time horizons, consistent with the empiricalevidence presented in our paper for equally-weighted portfolios, the results indicate theexistence of a negative and statistically significant relation between expected idiosyncraticskewness and the cross-section of expected returns, as well as between idiosyncratic volatilityand the cross-section of future returns. Although we observe a negative relation betweenrealized idiosyncratic skewness and the cross-section of future returns, the results are notstatistically significant. To illustrate the time-varying nature of the idiosyncratic higher moments,we also present additional figures with the cross-sectional distribution of idiosyncraticskewness and idiosyncratic volatility for the alternative time horizons.