The identification of asymmetric conditional heteroscedasticity is often based on sample cross-correlations between past and squared observations. In this paper we analyse the effects of outliers on these cross-correlations and, consequently, on the identification of asymmetric volatilities.We...