A quasi-concave DEA model with an application for bank branch performance evaluation

In this paper, we present a data envelopment analysis (DEA) model that replaces the standard DEA assumption of concavity for the production frontier with the less restrictive microeconomic assumption of quasi-concavity. The resulting quasi-concave model circumvents the potential specification error that comes with CCR and BCC models, if the production technology is non-concave, e.g. because of increasing marginal productivity. In addition, the model reduces the potential finite sample error associated with FDH models in small and medium-sized samples. A real-life application for the branch network of a large Dutch bank illustrates the operation of the quasi-concave model and its potential for improving performance evaluation. (C) 2001 Elsevier Science B.V. All rights reserved.