Einladung zur Vortragsreihe aus Finanz- und Versicherungsmathematik
Prof. Dr. Alexander McNeil
Department of Mathematics, ETH Zürich
Some New Copulas for Risk Modelling
The idea of using copulas to describe the dependence of financial risk factors or insurance loss types has enjoyed a lot of attention in recent years. Data of this kind mostly show strong evidence against a Gaussian distribution and copulas provide useful tools for talking about dependence in a non-Gaussian context and formulating ideas such as that of "dependence in the tail".
In this talk we survey the current state of knowledge in risk modelling using copulas. We comment on the families that seem most useful in the market and credit risk contexts and describe some useful new models. In particular we look at copulas that arise in multivariate normal mixture models and provide extensions to the popular t copula; we also look at copulas arising in the theory of multivariate extremes. Examples illustrating the use of these new copulas will be given.
Direktor Helmut Holzer
Gen.-Dir. Dr. Siegfried Sellitsch
o.Univ.-Prof. Dr. W. Schachermayer
(top of page)
© by Financial and Actuarial Mathematics, TU Wien, 2002-2020 |