This paper estimates the relative technical efficiency of Brazilian municipalities in the public education using the DEA model along with tobit regressions with the aid of re sampling methods. The jackstrap method enabled the estimation of robust efficiency scores, while the bootstrap allowed the estimation of the bias in those indexes. In effect, kernel density estimates portray remarkable changes in the distribution of efficiency. The econometric analysis relied upon the algorithm which corrects the serial correlation in the scores. It is noteworthy how accounting for environmental factors alters the dispersion of efficiency in Brazil. In a final step, we implement group comparisons tests based on information relative to certain management practices in the area of public education.
technical efficiency primary public education two-stage DEA environmental factors management practices