Abstract | This paper presents a family of distributions based on what we name GeneralMoments Expansions (GME).We show that the GME preserves the flexibility and good performance of the densities based on Edgeworth and Gram-Charlier series for modelling asset returns. Nevertheless the GME is simpler and more general, since it can be straightforwardly applied to expand any distribution with finite moments up to the truncation order. The goodness-of-fit of a GME using the Normal as basis is tested and compared to both the Gaussian and Student’s t by means of an empirical application for forecasting financial risk. |
---|