A0767
Title: Forecasting time-varying conditional skewness with asymmetric Laplace distribution
Authors: Qian Chen - Peking University Shenzhen Campus (China) [presenting]
Abstract: Higher moment risk attracts increasing interest recently, partly due to its ability of accounting for those asset risk that can not be explained by the second moment of asset returns, e.g. some researchers found there could be critical loss even during low volatility period. As the conditional skewness distinguishes negative and positive returns, as well as describes the corresponding distributional features, it naturally carries important information about asset risk. Compared to an exhaustive list of time-evolving mean and volatility models for the asset returns, there are relatively limited and indirect models for time-varying conditional skewness. A GARCH skewness model with asymmetric Laplace conditional return distribution is proposed to capture the time-varying and persistent features lying in the conditional skewness of financial return series. The model is realized with a number of real stock index returns. The empirical results find there is less persistence in the conditional skewness compared to the persistence of conditional volatility. Furthermore, the high negative conditional skewness during low volatility period could be of potential risk as it is easily be neglected by the investors and decision makers. The improved comprehension of the time-varying features of the skewness is supposed to offer implication for the investors, financial institutions as well as the policy makers in risk control and decision making.
