The Influence of Misspecification in Stochastic Frontier Analysis Models: A Simulation Approach

Abstract

This paper examines the influence of misspecification in stochastic frontier analysis (SFA) models. It explores the impact of misspecification on frontier and efficiency estimates as well as efficiency scores. Using a simulation approach with several data generating processes the general findings are that variables for which is assumed that they influence technology/heterogeneity should also be included as variables that influence efficiency and the other way around.