Asymmetric and Long Memory Volatility Process in the Chinese Stock Markets: A FIAPARCH Skewed Student-t VaR Approach

It is well known that the distributional properties of financial asset returns exhibit fatter‐tails and skewer‐mean than the assumption of normal distribution. The correct assumption of return distribution might improve the estimated performance of the Value‐at‐Risk (VaR) models in financial markets. In this paper, we investigate the relevance of the skewed Student’s t distribution innovation in capturing long‐memory and asymmetry features in the volatility of three Chinese stock markets including the Hong Kong, Shanghai and Shenzhen stock markets. In this perspective, we also examine the performance of in‐sample and out‐of‐sample value‐at‐risk (VaR) analyses using the FIAPARCH model with the normal, Student’s t, and skewed Student’s t distribution innovations. The results from the FIAPARCH model estimation suggest that returns of the Chinese stock markets exhibit long‐memory and asymmetry features in volatility. In the in‐sample and out‐of‐sample analyses, the FIAPARCH VaR models with the skewed Student’s t innovation predicted critical loss more accurately than did the models with the normal and Student’s t innovations for both long and short positions. In addition, the expected shortfall analysis indicates that the non‐normality distribution models fail less than normal distribution models, but when they fail, it happens for large (in absolute) returns: the average of these returns is correspondingly large. Therefore, risk managers and portfolio investors can estimate VaR and optimal margin levels most accurately by using the skewed Student’s t FIAPARCH VaR models of long and short trading positions in the Chinese stock market.