EFICIÊNCIA TÉCNICA, ECONOMIAS DE ESCALA, ESTRUTURA DA PROPRIEDADE E TIPO DE GESTÃO NO SISTEMA HOSPITALAR BRASILEIRO

We estimate DEA (Data Envelopment Analysis) technical efficiency scores for 1170 Brazilian hospitals included in the SUS (Central Health System) using a recently proposed method that combines bootstrap and jackknife resampling to eliminate the influence of outliers and possible measurement and recording errors in the data. We use the variable returns to scale variants of the DEA method. After computing the efficiency scores, we use econometric methods, especially quantile regression, to investigate the determinants of those scores. Our results confirm the importance of the scale effect (measured by the total number of services) over the technical efficiency of this sector, controlled by the average permanence in the hospital, its average costs and human capital variables. Regarded to hospital's management, we highlight the negative effects of non-profit organizations over the efficiency, which differs from the empirical literature due the usage of quantile regression in this paper. Excessive specialization also has a negative effect over the efficiency scores, pointing out the existence of an optimal mix between specialization and generalization hospitals' characteristics. As for property, private hospitals have its performance negatively affected for those units over the 30th efficiency quantile, suggesting that market failure effects are important in this sector.