Measuring Banking Performance in a Network DEA context: A General Weight Assurance Region Model

This paper adopts the General Weight Assurance Region (GWAR) model of Kourtzidis et al. (2020) to the case of banking institutions with a two-stage network structure. The proposed model can handle multiple input, output and intermediate variables at both stages, including leakage variables in the first stage and inputs that enter the system in the second stage. The general structure of the model gives greater flexibility to the weights assigned to the two stages relative to the Weight Assurance Region (WAR) and the conventional additive network DEA model. Consequently, the weights of the GWAR model are not strictly non-increasing but they can take any possible value in the range of 0 to 1. As a result, they do not suffer from the criticism of Ang and Chen (2016). Furthermore, the GWAR model can be considered as the general case of all additive network DEA models. Regarding the modelling of banking efficiency, the network structure preserves the dual role of deposits by treating them as an intermediate variable, thus providing a solution to the deposits’ dilemma. Finally, incorporating nonperforming loans is essential to our suggested solution for the GWAR weights.