Statistical models and methods for dependence in insurance data

Abstract

Copulas are becoming a quite flexible tool in modeling dependence among the components of a multivariate vector. In order to predict extreme losses in insurance and finance, extreme value copulas and tail copulas play a more important role than copulas. In this paper, we review some estimation and testing procedures for both, extreme value copulas and tail copulas, which received much less attention in the literature than corresponding studies of copulas.