Cross-efficiency measurement is an extension of Data Envelopment Analysis that allows for tie-breaking ranking of the Decision Making Units (DMUs) using all the peer evaluations. In this article we examine the theory of cross-efficiency measurement by comparing a selection of methods popular in the literature. These methods are applied to performance measurement of European warehouses. We develop a cross-efficiency method based on a rank-order DEA model to accommodate the ordinal nature of some key variables characterizing warehouse performance. This is one of the first comparisons of methods on a real-life dataset and the first time that a model allowing for qualitative variables is included in such a comparison. Our results show that the choice of model matters, as one obtains statistically different rankings from each one of them. This holds in particular for the multiplicative and game-theoretic methods whose results diverge from the classic method. From a managerial perspective, focused on the applicability of the methods, we evaluate them through a multidimensional metric which considers their capability to rank DMUs, their ease of implementation, and their robustness to sensitivity analyses. We conclude that standard weight-restriction methods, as initiated by Sexton et al. , perform as well as recently introduced, more sophisticated alternatives.
Bibliographical noteFunding Information:
José L. Zofío thanks the financial support from the Spanish Ministry of Science and Innovation (Ministerio de Ciencia e Innovación), the State Research Agency (Agencia Estatal de Investigación) and the European Regional Development Fund (Fondo Europeo de Desarrollo Regional) under grants EIN2020-112260 and PID2019-105952GB-I00 (AEI/FEDER, UE).
Jos? L. Zof?o thanks the financial support from the Spanish Ministry of Science and Innovation (Ministerio de Ciencia e Innovaci?n), the State Research Agency (Agencia Estatal de Investigaci?n) and the European Regional Development Fund (Fondo Europeo de Desarrollo Regional) under grants EIN2020-112260 and PID2019-105952GB-I00 (AEI/FEDER, UE).
© 2021 The Author(s)