ABSTRACT

To measure productive efficiency of competing firms a linear programming technique called Data Envelopment Analysis (DEA) is widely used. It requires first to adopt a production possibility set, that is axiomatic based and inefficient plans are projected on to the boundary of pp set, for the purpose of which radial and non-radial measures are used. These measures emanate from varying philosophies. The boundary of the pp set is sensitive to the extremely efficient firms, that are potentially influential observations. This study proposed a methodology to arrange the influential observations according to their influence on inefficient firms. The method goes beyond the super efficiency method introduced by Andersen and Petersen (1993), the purpose of which was to rank the efficient observations all of which posses unitary efficiency as revealed by Charnes, Cooper and Rhodes (1978) DEA formulation.