Rank Estimation of Duration Models: Dissertation Summary

The articles contained in this PhD thesis give one of the first attempts to formalise Instrumental Variable (IV) methods for duration models. The methods proposed can handle the typical features of duration data such as, censoring and time-varying covariates. It also gives a new solution to well--known inference problems of the, commonly used, Mixed Proportional Hazard models. The estimation methods in this thesis are all based on extensions of the (inverse of) log--rank statistic. The studies in this thesis focus in particular on the evaluation of a re-employment bonus experiment that is contaminated by selective compliance to the assigned treatment group. The estimation of the effect of the bonus on the re-employment probability corrected for the possible selection bias illustrates the methods developed in the first and third study. The method in the third study is a generalisation of the method in the first study. In the second study only the data on the control group is used to illustrate the use of a very general semi-parametric estimation procedure for duration models