Parametric and Non-parametric Models for Efficiency Measurement

Abstract

The first part of this chapter outlines 12 parametric panel data models classified into four groups in terms of the assumptions made on the temporal behavior of inefficiency. The models are grouped by their underlying assumptions: models assuming the inefficiency effects to be time-invariant and individual-specific; allowing inefficiency to be individual-specific but time-varying; models separating inefficiency effects from unobserved individual effects; and models separating persistent inefficiency and time-varying inefficiency from unobservable individual effects. In general, all performance measurement methods are expected to generate individual-specific effects. The chapter sheds light on whether particular panel data stochastic frontier models are better suited to different datasets.The second part of this chapter includes four non-parametric models. Two of these models are full frontier models (DEA and FDH) and the other two are partial frontier models (Order-m and Order-α).

Some parametric models of this chapter are discussed in Rashidghalam et al. (2016).