A Monte Carlo simulation comparing DEA, SFA and two simple approaches to combine efficiency estimates

In certain circumstances, both researchers and policy makers are faced with the challenge of determining individual efficiency scores for each decision making unit (DMU) under consideration. In this study, we use a Monte Carlo experimentation to analyze the optimal approach to determining individual efficiency scores. Our first research objective is a systematic comparison of the two most popular estimation methods, data envelopment (DEA) and stochastic frontier analysis (SFA). Accordingly we extend the existing comparisons in several ways. We are thus able to identify the factors which influence the performance of the methods and give additional information about the reasons for performance variation. Furthermore, we indicate specific situations in which an estimation technique proves superior. As none of the methods is in all respects superior, in real word applications, such as energy incentive regulation systems, it is regarded as best-practice to combine the estimates obtained from DEA and SFA. Hence in a second step, we compare the approaches to transforming the estimates into efficiency scores, with the elementary estimates of the two methods. Our results demonstrate that combination approaches can actually constitute best-practice for estimating precise efficiency scores.