DEA with non-monotonic variables. Application to EU governments’ macroeconomic efficiency

Abstract

Conventional Data Envelopment Analysis (DEA) considers monotonic variables, i.e. the lower the inputs and the larger the outputs, the better. There are, however, occasions when the monotonicity of a variable with respect to efficiency depends on the value of the variable, i.e.in a certain range of values an increase in the variable is desirable, while in another range it is a decrease of the variable that is desirable. In this paper, a DEA model that solves problems considering non-monotonic variables is proposed. An application to assess the macroeconomic efficiency of European Union (EU) Member States, as regards taxation, gross debt, GDP growth and employment is presented.

Dominance Network Analysis: Hybridizing Dea and Complex Networks for Data Analytics

“…Another feature of DEA is that, although approaches with non-monotonic variables (Villa & Lozano, 2016) and dual-role factors (Toloo et al, 2018) exist, in principle the attributes must be either positive or negative. Positive attributes are those that the larger, the better, while for negative attributes it is the opposite.…”