Identification of congestion in two-stage data envelopment analysis

This paper aims at examining congestion in two-stage decision making units. Providing an example, it will be proved that presence or absence of congestion in the whole process of a two-stage decision making unit has nothing to do with presence or absence of it in each of stages. In other words, it is likely that the first stage to be weak efficient and the second one will have congestion, while the whole process lacks congestion. It is also possible that each stage has congestion, but it doesn’t mean that the whole process should have congestion. Then, to identify congestion in two-stage decision making units, modified Cooper model is developed.