Incorporating Multiprocess Performance Standards into the DEA Framework

Data envelopment analysis (DEA) is a mathematical approach to measuring the relative efficiency of peer decision-making units (DMUs). It is particularly useful when no a priori information is available on the trade-offs or relationships among various performance measures. A shortcoming of the DEA model, however, is its inability to provide a measure of absolute performance for the DMUs under investigation. Traditionally, in the service sector, this has not been an issue that one could address, because performance standards in that sector have been difficult to establish. However, in those settings where it has become feasible to develop such standards, it is desirable to build these into DEA performance evaluation, thereby enhancing the capability of the tool. While there have been some attempts to incorporate standards into the DEA structure, these approaches have generally been indirect , in the sense that they have focused primarily on restricting the DEA dual multipliers. This paper introduces a new way of building performance standards into the model. Utilizing the conventional DEA framework and a set of activity matrices, a set of standard DMUs can be generated and incorporated directly into the analysis. We show that under normal circumstances, these generated DMUs are efficient relative to the normal ones, and therefore form a type of outer frontier against which regular units can be evaluated. The proposed approach is applied to a sample of 100 branches of a major Canadian bank, where time standards are used to generate a set of standard bank branches.