Identifying managerial groups in a large Canadian bank branch network with a DEA approach

This paper develops a new grouping approach in using data envelopment analysis as a framework to identify management groups and group performance leaders. The management group relies on the fact that branches grouped together present similar managerial preferences over performance goals and resource deployments due to internal or external market forces. This grouping approach can help a firm to create continuous improvement opportunities with effectively promoting the best managerial practices within groups, given similar operating characteristics. This approach’s grouping power and rationality is examined in the context of a large Canadian bank with about 1000 branches. The advantages of this new grouping approach are further verified through comparisons with the results obtained from the traditional clustering algorithm and the collaborating Bank’s “community type and population size” grouping criteria.

► We compare the bank’s method of sorting branches for comparison with a DEA model.
► We establish a grouping process to match management’s needs for benchmarking.
► The method proposed helps the bank carry out continuous productivity improvements.
► The results obtained were verified against traditional clustering algorithms.
► We validated against the bank’s “community and population size” grouping criteria.