High Dimensional Modelling and Simulation with Asymmetric Normal Mixtures
18 Pages Posted: 11 Aug 2007
Date Written: July 20, 2007
Abstract
A family of multivariate distributions, based on asymmetric normal mixtures, is introduced in order to model the dependence among insurance and financial risks. The model allows for straight-forward parameterisation via a correlation matrix and enables the modelling of radially asymmetric dependence structures, which are often of interest in risk management applications. Dependence is characterised by showing that increases in correlation values produce models which are ordered in the supermodular order sense. Explicit expressions for the Spearman and Kendall rank correlation coefficients are derived to enable calibration in a copula framework. The model is adapted to simulation in very high dimensions by using Kronecker products, enabling specification of a correlation matrix and an increase in computational speed.
Keywords: Dependence, copula, normal mixtures, Kronecker product,
Suggested Citation: Suggested Citation