The Non-Parametric Identification of Generalized Accelerated Failure-Time Models

The author considers a class of models that generalizes the popular mixed proportional hazard model for duration data: the generalized accelerated failure-time model. He shows that the generalized accelerated failure-time model is nonparametrically identified (up to a normalization). He then reconsiders the nonparametric identification of the mixed proportional hazard model. He shows that the class of mixed proportional hazard models is not closed under normalization. This implies that a finite mean of the mixing distribution is a necessary condition for (nonparametric) identification of the mixed proportional hazard model. It is impossible to test this hypothesis without imposing arbitrary restrictions on the base-line hazard and/or the regression function.