Non-homogeneity in Data Envelopment Analysis

Data Envelopment Analysis requires all the units that are being compared to be homogeneous. One of the homogeneity requirements is that they all share the same inputs

and outputs. In practice this is often not the case, as some units carry out more activities than others and may use more inputs to create more outputs than the rest. In an extreme case, some Decision Making Units (DMUs) may specialise in a single activity while others may aim for diversification. We present a new model that can estimate efficiencies for the various activities, even when some DMUs do not engage in some of the activities and do not share the same set of inputs/outputs with the remaining DMUS. The model is demonstrated with some data from the literature.