Wages and Profits in Manufacturing Firms: Matched-Panel Evidence from Brazil Pedro S. Martins School of Business and Management, Queen Mary, University of London, United Kingdom Luiz A. Esteves Department of Economics & Graduate Program of Public Policy, UFPR, Brazil Abstract We provide evidence about the determinants of the wage structures of developing countries by examining the case of Brazil. Our specific question is whether Brazil’s dramatic income and wage differentials can be explained by the division of rents between firms and their employees, unlike in competitive labour markets. Using detailed individual-level matched panel data, covering a large share of manufacturing firms and more than 30 million workers between 1997 and 2002, we consider the endogeneity of profits, by adopting different measures of profits and different instruments and by controlling for spell fixed effects. Our results, robust to different specifications and tests, indicate no evidence of rent sharing. This conclusion contrasts with findings for most developed countries, even those with flexible labour markets. Possible explanations for the lack of rent sharing include the weakness of labour-market institutions, the high levels of worker turnover and the macroeconomic instability faced by the country. Keywords: Wage Bargaining, Instrumental Variables, Matched Employer-Employee Data, Developing Countries JEL Classification : J31, J41 Resumo O trabalho fornece evidência sobre os determinantes da estrutura salarial dos países em desenvolvimento, mais especificamente para o caso brasileiro. A pergunta endereçada no artigo é se a elevada desigualdade salarial observada no mercado de trabalho brasileiro pode ser, em alguma medida, atribuída à filiação industrial dos trabalhadores, ou seja, se a partilha de lucros econômicos (rents) entre firmas e empregados de setores oligopolizados, decorrentes de barganha, são responsáveis por alguma parcela da desigualdade. O trabalho utiliza um painel de microdados do tipo empregador-empregado, cobrindo grande parte das empresas industriais e mais de 30 milhões de trabalhadores entre 1997 e 2002. Nossas especificações econométricas incluem o controle da endogeneidade dos lucros por meio da adoção de diferentes instrumentos, além de controlar efeitos fixos de firmas e trabalhadores Revista EconomiA September/December 2012