DEA Models with Undesirable Inputs, Intermediates, and Outputs

Abstract

In real applications involving the use of Data Envelopment Analysis (DEA) models, undesirable inputs and outputs have been frequently encountered and addressed, e.g., via data transformation. These studies were scattered in the literature and often confined to some particular applications. In this paper, we present a systematic investigation concerning the building of DEA models. First, we describe the desirability of inputs and outputs, as well as the disposability assumptions in the presence of undesirable inputs and outputs. Next we construct a number of DEA models with different disposability assumptions and performance measures for the case of single-stage DEA. Next, we try to systematically investigate two-stage DEA models with undesirable inputs, intermediates and outputs. Particularly, we utilize the free-disposal axioms to construct the production possibility sets and the corresponding DEA models with undesirable inputs, intermediates, and outputs.