Measuring the efficiency of banks: the bootstrapped I-distance GAR DEA approach

The efficiency of the banking sector, particularly in developing countries, has captivated the attention of various researchers. Contributing to this issue, we present the results of in-depth analysis of the efficiency of Serbian banks during the period 2005-2016. Unlike previous papers evaluating the efficiency of South-Eastern European banks, we emphasize the importance of applying weight restrictions in Data Envelopment Analysis (DEA). The aim is to incorporate every aspect of a decision-making unit's performance to avoid misevaluation of a bank's efficiency. As a possible remedy to the issue, a bootstrapped I-distance is suggested as a statistically sound framework for determining weight bounds in the Global Assurance Region (GAR) DEA model. In terms of average efficiency, the banking sector of Serbia exhibits an improving trend over the period analyzed. The results show how banks can be evaluated when the impact of all the operating inputs and outputs are properly factored into the study.