Rank correlation estimators and their limiting distributions

Abstract

We examine a new rank correlation estimator, recently proposed by Bobrowski (Ranked modelling of risk on the basis of survival data. ICSMRA, Lisbon, 2007). It is obtained by minimization of a convex piece-wise linear criterion function. The main advantage of this estimator is the fact that it can be effectively computed by algorithms related to linear programming. We prove basic asymptotic theorems about the estimator: consistency and asymptotic normality.