Multi-stage financial index tracking model under GH distribution

To reflect the features of the risky assets' return distribution,we adopt the multivariate generalized hyperbolic (gh) distribution to describe the return data. then we establish a new active financial index tracking model which controls the downside risk through a probability constraint,and takes into account multiple market frictions. by using the normal variance-mean mixture definition of the gh distribution and the chapman-kolmogorov equation,the single-stage index tracking model is then extended to the multistage situation. empirical results based on the csi 300 index demonstrate the efficiency and practicality of our new tracking models and solution methods. © 2013 pakistan journal of statistics.