This paper has studied the distributions of the log returns of the Chinese stock market, using the Generalized Hyperbolic Distributions (GH) as the reference distributions. The empirical results shows that in almost all the cases, both Normal Inverse Gaussian (NIG) distribution and the Hyperbolic distribution can fit the data very well. This suggests that using GH as building block, we may hopefully get some more realistic models in Value at Risk modelling and option pricing. Some empirical studies on this are also reported in this paper.
|