Frisch seminar: Victor Podinovski, Loughborough University
What if we know more than we can model: tailor-made models of data envelopment analysis
In many applications of DEA our knowledge of the production process may exceed the generic assumptions incorporated in the conventional models. Not using this knowledge leads to on an overestimation bias in the efficiency evaluation. We consider several applications, from education to agriculture, in which additional information is worth incorporating in a DEA model. For example, this may include our knowledge about specific (e.g., proportional) relationship between some inputs and outputs, and also the information about substitutions between different resources or products. Based on our experience, we demonstrate existing modelling tools at our disposal that can be used to construct a better-informed model for a specific application.