Inference for Aggregate Efficiency: Theory and Guidelines for Practitioners

We expand the recently developed framework for the inference for aggregate efficiency, by extending the existing theory and providing guidelines for practitioners. In particular, we develop the central limit theorems (CLTs) for aggregate input-oriented efficiency, analogous to the output-oriented framework established by Simar and Zelenyuk (2018). To further improve the finite sample performance of the developed CLTs, we propose a simple yet easy to implement method through using the bias- corrected individual efficiency estimate to improve the variance estimator. The extensive Monte-Carlo experiments confirmed the developed CLTs for aggregate input- oriented efficiency and also confirmed the better performance of our proposed method in the finite sample sizes. Finally, we use two well-known empirical data sets to illustrate the differences between all the existing methods to facilitate the use by practitioners.