Attrition in longitudinal panel data and the empirical analysis of dynamic labour market behaviour
分析失业和工作持续时间中，样本流失与劳动力市场行为的关系，通过联合建模估计两种持续时间的分布，解决因忽略相关性导致的估计不一致问题。
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.