We present a model where the value of the life insurance benefit is random. The policy is at each point in time assumed to be in one of a finite number of states and the evolution of the policy through time is modelled by a time-continuous, non-homogeneous Markov chain. The insurance period of a...