Bibliografie

(eng): It is of great importance for those in charge of measuring and managing financial risk to analyse financial data by determining a certain probabilistic model. These data usually possess distribution with tails heavier than those of normal distribution. The class of alpha-stable distributions can be chosen for modelling financial data since this probabilistic model is able to capture asymmetry and heavy tails. In this paper, mixed alpha-stable model is applied for the analysis of return data of Lithuanian pension funds that usually contain a significant number of zero values. The distribution fitting and simulation algorithm are also described. Risk measures VaR (Value-at-Risk) and CVaR (Conditional Value-at-Risk) are chosen to evaluate the characteristics of return data, especially the degree of heavy tails. VaR and CVaR are estimated from return data, then computed from simulated data when using mixed alpha-stable law and finally compared to the measures obtained using alpha-stable model and Gaussian model.