SEMI-MARKOV AND COMPETING RISKS MODELS WITH APPLICATIONS TO OCCUPATIONAL MOBILITY

Abstract

This chapter presents two approaches of stochastic modeling of social processes, namely, semi Markov processes and competing risks models, with applications to occupational mobility. Models of competing risks are developed by biostatisticians and can be interpreted as mechanisms governing special semi Markov processes. However, most of the competing risks literature uses the concept of latent sojour times, each associated with a specific cause of termination of the actual, observable sojourn time. A semi Markov process is completely characterized either by the matrix of transition probabilities and the vector of its initial distribution or by the conditional distributions of the sojourn times, the probabilities of transitions. The chapter discusses a few generalizations of concepts used in the statistical analysis of single sojourn times. The dependent and the independent model are indistinguishable by means of the observable quantities.