Unemployment duration models with non-stationary inflow and unobserved heterogeneity

This paper is concerned with econometric problems and methods involved when estimating duration models using data on uncompleted unemployment spells provided by standard labour force surveys. In particular, it considers how the model estimates are affected by the commonly applied assumption of stationary inflow rates when the true inflow rates are non-stationary and when different assumptions about unobserved heterogeneity are maintained.