The main aim of this dissertation is to present the most efficient evaluation impact methodologies for public policies, applying them to a data base of the Brazilian public education sector. We will analyse, in particular, the consequences of the municipalization process, which took place in the 90 s, where some municipality owned schools became stated administered. In our evaluation analysis we will be looking at the change of some indicators of education quality, constructing a comparison group based on the matching with difences in diferences. The central problem in the evaluation of any social program is the fact that individuals participating in the program cannot be simultaneously observed in the alternative state of no treatment. The solution to this problem is the estimation of a contrafactual effect, in which a response variable is measured in those participants, supposing that they had not participated on the program. The techniques presented here are able to deal with this problem and will allow us to make some conclusions about the public policy of municipalization. It is important to mention that this shift on the structure of the Brazilian public education has never been analysed before. Our conclusion was that the process had different impacts depending on the region it occurred. On the states of the Northeast included in the analysis and on the state of Santa Catarina we couldn t observe any significant impact on the quality of education. Nevertheless, on state of Minas Gerais, the impact was shown to be negative, reflecting a downgrade on the quality of basic public education.