Asymptotics for sieve estimators of hazard rates: estimating hazard functionals

Working paper

This paper derives asymptotics for functionals of a hazard model with an exposure-time effect and time-varying covariates. A semi-nonparametric sieve maximum likelihood estimator of a competing risks model based on the Cox proportional hazard is considered. Consistency of the estimator and its rate of convergence in the Fisher norm are derived. These results are prerequisites for asymptotic normality of plug-in estimators of hazard functionals. This provides an inference procedure for a large class of functionals including the conditional probability of events and various asset pricing formulas for defaultable securities. Asset pricing formulas in this class include the value of mortgages, insurance contracts, bonds, swaps and other options.

Authors: Wolter, J

• Publication status: Published
• Publisher: University of Oxford
• Series: Department of Economics Discussion Paper Series
• Publication date: 2015-10-05
• Paper number: 760
• Keywords: conditional probabilities; hazard models; Sieve estimation