The Role of Firm Heterogeneity in Determining Labor Market Outcomes

I construct and estimate an equilibrium search model with two-sided heterogeneity and on-the-job-search. Both \u85rms and workers are heterogenous in a productivity parameter, and solve their respective optimization problems in an environment with undirected search. Market equilibrium is achieved by the entry and exit of \u85rms run by pro t maximizing entrepreneurs. The structure of the model allows for a wide-range of wage-setting mechanisms; three of which, an exogenous output-sharing rule, a surplus-splitting rule and a Bertrand competition mechanism, the paper deals with explictly. The paper delivers a theory of equilibrium wage dispersion and labor turnover, and relates these to observable and unobservable worker and \u85rm characteristics. The structural model is estimated using matched a uniquely comprehensive employer and employee Austrian panel data. The full distribution of time-varying worker and \u85rm productivities and transition rates are estimated. Preliminary results suggest that rm heterogeneity explains one-quarter of residual (after accounting for observable worker characteristics) wage dispersion, and person-speci\u85c unobserved heterogeneity for the remaing three-quarters. Firm e¤ects explain perhaps 45 % of inter-industry wage di¤erentials, and up to 30 % of wage variation for low-skilled workers. Unobserved worker heterogeneity explain nearly all of gender wage di¤erentials and wage dispersion amongst high-skilled workers. Firm e¤ects account for three-quarters of the explained variation in wage growth rates. Over the lifecycle workers switch to \u85rms with higher productivity. Firm heterogeneity explains the \u85rmsize-wage e¤ect; turnover and the fraction of the workforce recruited from the pool of unemployed are decreasing in \u85rm productivity. I would like to thank Gary Becker, Steven Davis, Derek Neal, Robert Shimer, Robert Topel, Michael Wagner-Pinter and numerous colleagues and seminar participants at the University of Chicago for advice and helpful discussions. I am most grateful to Synthesis Forschung and Günter Kernbeiss for providing th