The objective of this paper is to investigate the returns to schooling issue for Brazil based on distinct methodologies in order to reach a reliable estimate for this type of measurement. The first one suggests that bias is caused by the job search strategy adopted by the agent. In this case, wages do not depend solely on the wage supply but also on the job-search strategy (Gronau, 1974). Second, we use the methodology developed by Garen (1984) that allows dealing with the optimal choice of years of schooling due to the fact that this variable is continuous and ordered. Lastly, a pseudo-panel analysis (Deaton, 1985) is conducted in an attempt to deal with the relevant variable omission related to ''ability bias'' (Griliches, 1977) in the model. All results are compared to returns to schooling equation generated by OLS. Among main conclusions we have evidences about comparative advanced from Willis e Rosen (1979), endogeneity on schooling' decision and no relevant evidence on the existence of ''ability bias''.
