Application of Sensitivity Analysis and Super-Efficiency DEA Models on Efficiency Evaluation of Public Sector Banks in India

The assessment of performance or efficiency is an essential step in the production process. Performance and efficiency evaluations can be done using parametric or non-parametric approaches. When analysing the effectiveness of similar organisations, also known as Decision Making Units (DMUs), with many inputs and outputs, one of the most widely used non-parametric methodology is Data Envelopment Analysis (DEA), which is based on a linear programming approach. Inefficient units are benchmarked (peers) and efficiency scores are provided by DEA for each DMU. Based on the number of peers, the Decision Making Units might then be ranked. A Tie between ranks, however, can happen. Andersen-Peterson’s (1993) model aids in breaking the tie. The super-efficiency model produces an infeasible result, which is confirmed by applying the prerequisites for infeasibility identification put forth by Seiford & Zhu (1999). In this paper, we present the results of our empirical investigation of Public Sector Banks in India during the 2019–20 period, focussing on three inputs and two outputs. We used variable return to scale (VRS) assumption based Banker Charnes and Cooper (BCC) and super-efficiency models, and also identified the infeasibility problem and performed sensitivity analysis. Finally, a stability region was constructed.