Contribuciones al análisis estocástico de la eficiencia técnica mediante métodos no paramétricos

Resumo

This thesis provides in the first chapter a brief review of efficiency and productivity methods. Secondly we examine how to analyze efficiency and productivity by DEA and we apply the method for the Spanish airports in the period 1992-94. Although the time period analyzed is fairly short, we study the impact of the crisis in the productivity of Spanish airports. We apply the Malmquist index since among its other advantages this ratio allows for the decomposition of total productivity changes into different sources of variation. We also use resampling methods to gain statistical precision and the bootstrap analysis yields further evidence given that for many airports efficiency and productivity is not statistically significant. In standard deterministic frontier analysis, either DEA or FDH techniques allowed to determinate inefficient units just but taking some measure between the estimated frontier and the related output. Unfortunately, when we assume that some symmetric noise is present in the data, then the previous task becomes much harder. The problem is that in this setting noise and efficiency are not identified and therefore, without an statistical tool it is impossible to decide whether a firm is efficient or not. In the third chapter we propose a test for efficiency in a stochastic nonparametric frontier analysis. Under weak conditions on the specification of the production frontier, and the null hypothesis of efficiency, we provide the asymptotic distribution of the test statistic. Furthermore we show the test is consistent against a broad set of alternatives of inefficiency. Evidence of the good properties of the test is given by a simulation study.