A Tutorial on Using Dynamic Network DEA to Benchmark Organizational Performance

In investigating organizational performance or efficiency, two perennial problems confront us, namely, loss of information in aggregated data, and determining the influence of lower level variables when they are not explicitly identified in a performance model. It can be quite complicated to quantify various interactions using parametric methods that make distributional assumptions. An additional complication is how to capture the effect of undesirable outputs on organizational performance across time. The primary purpose of this tutorial is to illustrate step-by-step how non-parametric, dynamic network data envelopment analysis (DN-DEA) can be used to overcome the above problems and provide a refined estimate of relative efficiency. The tutorial begins with a discussion of the rationale for studying relative efficiency and a non-technical introduction to standard DEA. It then continues to illustrate how interactions across time among various inputs used in providing different levels of outputs can be captured and main potential improvements identified. DN-DEA enables an analysis where lower level variables in an organization are accounted for by disaggregating data, while simultaneously maintaining the focus on performance of the organization as well as its main divisions. Efficiency analysis of divisions across time can be an integral part of organizational learning and a source of competitive advantage