The purpose of this paper is to assess the technical efficiency of credit unions by taking into consideration not only desirable but also undesirable outputs and measure the contribution of individual inputs and outputs to efficiency scores.
Data envelopment analysis is used to measure the relative performance of credit unions under different models, both with and without undesirable outputs, i.e. Banker–Charnes–Cooper, slacks-based measure and slacks-based measure extended to the case of undesirable outputs. Efficiency scores are compared between models using non-parametric statistical analyses. Implications are derived for credit union performance enhancement.
Credit unions should focus on increasing the quantum of loans disbursed as well as the number of members served. The omission of undesirable outputs, i.e. NPLs from consideration, leads to a negative bias in efficiency scores due to the failure to reward credit unions for efforts to reduce NPLs. Although NPLs are high in Indian credit unions, they are not the main contributor to inefficiency. The major contributions of this study are to fill the void in studies on credit union efficiency in the presence of undesirable outputs and identify the specific problem areas for improvement in credit union operations by separately measuring the contribution of individual inputs and outputs to efficiency scores.
The main limitation of this study is the use of cross-sectional data. In future studies, data may be collected from multiple periods so that trends in efficiency as well as changes in contributors to efficiency over time may be identified.
The paper identifies specific problem areas for improvement in Indian credit union operations, namely the quantum of loans disbursed as well as the number of members served.
This study fills the void in studies on credit union efficiency in the presence of undesirable outputs and measures the contribution of individual inputs and outputs to efficiency.
