A Distributional Framework for Matched Employer Employee Data

Abstract

We propose a framework to estimate earnings distributions and worker and firm unobserved heterogeneity on matched panel data. We introduce two models: a static model that allows for interaction effects between workers and firms, and a dynamic model that allows in addition for Markovian earnings dynamics and endogenous mobility. We establish identification in short panels. We develop tractable two-step estimators where firms are classified into heterogeneous classes in a first step. We apply our method to Swedish matched employer employee panel data and report estimated earnings functions, sorting patterns, and variance decompositions.