Performance measurement by means of Data Envelopment Analysis: a new perspective for undesiderable outputs

During the last years enormous attention has been given to the assessment and improvement of the performance of productive systems: all economic activities are affected by the world-wide trend for improved performance. On the other hand, the continuous economic recessions in the western world put enormous pressures on profit and not-profit organisations for the improvement of performance as a means to long run viability. In this context, two classes of methods were developed for evaluating the performance of organisational units: a non parametric one, typically represented by Data Envelopment Analysis (DEA), and a parametric one, represented by Stochastic Frontier Analysis (SFA). Both DEA and SFA approaches are derived from the methods of measuring efficiency introduced by Farrell (1957). Farrell’s argument provides an intellectual basis for redirecting attention from the production function specifically to the deviation from that function as a measure of efficiency. This empirical production frontier, also called “best practice frontier”, specifies for a unit the maximum quantities of outputs it can produce given any level of inputs and, for any levels of outputs, the minimum quantities of inputs needed for producing the outputs. According to Farrell, a technical efficient organisation would be one that produces the maximum possible outputs from a given set of inputs (output-orientated technical efficiency measure) or one that produces a certain level of outputs with the minimum amount of inputs (input-orientated technical efficiency measure). The aim of the present paper is to give an overall view of the non parametric approach to efficiency measurement and to propose a variant of DEA method to handle non standard applications, such as those which involve the presence of undesirable outputs. The paper is organised as follows. Section 2 reviews some of the theoretical background of DEA method and analyses its advantages and disadvantages with respect to SFA, Section 3 presents a modified DEA model which includes undesirable outputs, Section 4 shows three case studies which include the presence of bad outputs and Section 5 concludes.