Title: Misspecification of data envelopment analysis

Authors: Kekoura Sakouvogui

Addresses: US Department of Commerce, US Census Bureau, Decennial Statistical Studies Division, Washington, District of Columbia, USA

Abstract: Data envelopment analysis (DEA), a non-parametric efficiency estimator uses linear programming technique for the computation of estimates of decision-making units, such as, universities, schools, hospitals, banks, or mutual funds. There has been an ongoing debate about the application of the DEA model for model misspecification and in particular the inefficiency error of production for input and output variables. This paper contributes to this debate by examining several misspecifications of the DEA model in Monte Carlo (MC) simulations. MC simulations are conducted to examine the performance of the DEA model under two different data generating processes, stochastic and deterministic, and across five different misspecification scenarios, inefficiency distributions (traditional and proposed approaches), sample sizes, production functions, input distributions, and curse of dimensionality.

Keywords: data envelopment analysis; DEA; inefficiency distributions; Monte Carlo simulations.

DOI: 10.1504/IJOR.2022.128395

International Journal of Operational Research, 2022 Vol.45 No.4, pp.467 - 491

Received: 20 Apr 2019
Accepted: 09 Jan 2020
Published online: 20 Jan 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article