Abstract

Journal of Actuarial Practice

Volume 13, 2006

Bivariate Archimedean Copula Models for Censored Data in Non-Life Insurance

Michel Denuit, Oana Purcaru, and Ingrid Van Keilegom

Abstract

We describe a methodology based on Archimedean copulas for analyzing nonlife insurance data with censoring present. Specifically, we propose a graphical selection procedure for the nonparametric estimation of the generator. An actual loss-ALAE data set is used for the numerical illustrations and for comparisons of our approach to a few others.

Key words and phrases: dependence, Archimedean copula, model selection, censored data

Corresponding Author:

Michel Denuit

Institut des Sciences Actuarielles

Université Catholique de Louvain,

6 rue des Wallons

B-1348 Louvain-la-Neuve

Belgium

E-mail: denuit@stat.ucl.ac.be