Environmental variables and relative efficiency of bank branches: a DEA-bootstrap approach

We analyse the efficiency of the branches of a local bank in Cologne, Germany, using a novel two-stage Data Envelopment Analysis (DEA) approach. Traditional DEA studies frequently found that most branches of a bank were efficient, which led researchers to conclude that banks are usually "well managed" and that there would be little scope for improvement. However, due to the nature of the DEA estimator, analyses based on a limited number of observations using a multitude of parameters will lead to results that overstate the efficiency of the observations. The two-stage approach makes it possible to determine which so-called environmental parameters have a significant impact on efficiency. A traditional DEA model that includes several environmental parameters seems to imply that 25 out of 31 branches in our data are efficient. Using the two-stage model, we are able to demonstrate that eight of the 25 branches in our data were wrongly classified as efficient by the traditional model. Therefore, the conclusion that bank branches are usually well managed and that there is little scope for improvement may be somewhat premature.