Estimation of the conditional tail moment risk measure under random right censoring
Résumé
Estimators of the conditional tail moment risk measure based on extreme Kaplan-Meier integral constructions are proposed. The situation when observations are heavy-tailed and subject to right-censoring is considered, which arises often in non-life insurance. Weak convergence is established for both standard and bias-reduced versions of the estimator, and the finite-sample performance is studied through simulations. A real data application to a theft guarantee from a Danish non-life insurer is considered.
Origine | Fichiers produits par l'(les) auteur(s) |
---|