Evaluating density forecasts of the model with a conditional skewed-tdistribution for China's stock markets

Abstract

This study sets up a model which assumes a conditional skewed-t distribution for returns on four of China's stock price indexes (Shanghai A, Shanghai B, Shenzhen A and Shenzhen B). We employ Chen and Fan's (2004) pseudo-Wald test via the copula approach to evaluate both in- and out-of-sample density forecasts of the model. The results show that our model characterized by modelling conditional skewness and conditional kurtosis has a good in-sample fit as well as a good out-of-sample performance of density forecasting.