Attrition in longitudinal panel data and the empirical analysis of dynamic labour market behaviour

Abstract In the empirical analysis of unemployment durations and job durations, it is generally assumed that the stochastic processes underlying labour market behaviour and the behaviour concerning participation in a panel survey are independent. However, there are reasons to believe that the probability of dropping out of the panel is related to the rate at which a (different) job is found. If there is such a relation, and if it is ignored, then the estimator of the rate at which individuals become employed or change jobs will generally be inconsistent. In this paper we analyse the relation between the duration spent in a particular labour market state and the duration of panel survey participation, by explicitly modelling and estimating the joint distribution of both durations. The emphasis will be on models allowing for stochastically related unobserved determinants of both types of duration. We estimate models both for unemployment durations and for job durations.