Identification and Inference of Duration Models Driven By Stochastic Processes

Thesis (Ph.D.)--University of Washington, 2015I study a new class of duration models driven by stochastic processes. In contrast with the standard hazard-based models, the duration outcome or survival time is de fined to be the the fi rst time a Lévy subordinator--a stochastic process with stationary, independent and non-negative increments--crosses a random threshold. Such a model is of substantial inter- est because not only is it related to optimal stopping time models where agents optimally time their discrete actions, but it also applies to the scenario in which the termination of duration is caused by some gradual and irreversible accumulation of damage