MORE@DIAG Seminar, Thursday 4 July 2013, Aula Magna at 12:00 a.m.
CondEff: A Toolbox for Conditional Efficiency Measurement Luiza BadinDepartment of Applied MathematicsBucharest University of Economic Studies,& Gh. Mihoc - C. Iacob Institute of Mathematical Statisticsand Applied Mathematics, Bucharest, Romania A simple method for the determination of scale economies in non-parametric models Giovanni CesaroniDipartimento Sviluppo Economie Territoriali,Presidenza del Consiglio dei Ministri, Roma CondEff: A Toolbox for Conditional Efficiency Measurement By Luiza Badin, Cinzia Daraio and Leopold Simar AbstractMotto: “Science is what we understand well enough to explain to a computer. Art is everything else we do" Donald KnuthRecent theoretical developments and empirical applications (e.g. Daraio and Simar, 2005, 2007; Badin and Daraio, 2011) have shown the usefulness of conditional efficiency models for comparing the performance of Decision Making Units (DMUs) taking into account different external environmental conditions. Indeed, introducing heterogeneity in the comparison of performance is a relevant issue, often neglected in the literature or addressed by relying on strong parametric or semi-parametric assumptions, many times unrealistic. Nonparametric conditional efficiency models allow for a complete and general handling of heterogeneity and of the impact of external-environmental factors on the performance of DMUs. Latest contributions in this area (Badin, Daraio and Simar, 2010; 2012a,b) propose a general two-stage approach, which does not rely on the separability condition (meaning that the input-output space is independent from the external environmental factors that can influence only the distance of DMUs towards the efficient frontier and not the efficient frontier itself). Nevertheless, the computations require technical and programming skills in order to correctly implement the algorithms and run the applications. Up to now none of the existing software available has implemented conditional efficiency scores. We fill this gap by proposing in this paper a toolbox which implements the most updated research on conditional efficiency analysis. It is suitable for applied researchers in order to perform the computations and obtain rigorous empirical evidence. Some illustrations are provided on simulated and real datasets. The comprehensive CondEff Toolbox will be developed in Matlab and will be freely available.Keywords: FDH, DEA, conditional efficiency, robust measures, bootstrap, bias correction, nonparametric regression, optimal bandwidth. A simple method for the determination of scale economies in non-parametric modelsBy Giovanni Cesaroni and Daniele GiovannolaAbstractWe propose a procedure for the determination of scale economies (cost-based returns to scale analysis), which is based on the minimization of the ray average cost and allows a more straightforward approach when compared to existing methods. In particular, for the frontier points under evaluation, we can avoid both the need to solve the NIRS problem, required in Fare and Grosskopf (1985), and that of solving the dual problem set by Sueyoshi (1999) to compute the intercept of the supporting hyperplane and the corresponding value of the cost-based scale elasticity. Therefore, our procedure allows to dispose of the indeterminacy implied by the multiplicity of solutions in the intercept method. Moreover, we conjecture that multiple solutions to the CRS cost program (i.e. multiple optimal scale sizes) are not a problem, because a result analogous to corollary 2 in Banker and Thrall (1992) holds. Another advantage of our approach is that of dispensing from assumptions regarding the differentiability of the transformation function and of the cost function. The theoretical properties of the proposed methodology are derived under the only assumption of strong disposability, so that - contrary to Sueyoshi’s method - it applies to FDH models. Key words: Returns to scale, Scale efficiency, DEA, FDH