Sensitivity of technical efficiency estimates to estimation methods: an empirical comparison of parametric and non-parametric approaches

Abstract

This paper highlights the sensitivity of technical efficiency estimates to estimation approaches using empirical data. Firm specific technical efficiency and mean technical efficiency are estimated using the non parametric Data Envelope Analysis (DEA) and the parametric Corrected Ordinary Least Squares (COLS) and Stochastic Frontier Analysis (SFA) approaches. Mean technical efficiency is found to be sensitive to the choice of estimation technique. Analysis of variance and Tukey’s test suggests significant differences in means between efficiency scores from different methods. In general the DEA and SFA frontiers resulted in higher mean technical efficiency estimates than the COLS production frontier. The efficiency estimates of the DEA have the smallest variability when compared with the SFA and COLS. There exists a strong positive correlation between the efficiency estimates based on the three methods.