Industry Efficiency Analysis

Abstract

The characterization and estimation of productive efficiency of an industry have followed three stages of development over the past decade. One is the parametric theory, whereby a production or cost frontier is estimated by assuming a composed error model with two components of error: one measuring technical inefficiency, the other indicating purely random components. The method of nonlinear maximum likelihood (ML) is then applied. The second is the data envelopment analysis (DEA), which employs the basic notion of Pareto efficiency of economic theory by stipulating that a given firm (or decision-making unit (DMU)) is not efficient in producing its outputs from given inputs, if it can be shown that some other DMU or combination of DMUs can produce more of some outputs without utilizing more of any input. This DEA technique is sometimes called nonparametric or semiparametric, since it does not postulate any functional form of the production or cost frontier. In order to obtain reliable estimates of the production frontier, one may adopt smoothing methods and outlier rejection techniques for the observed data on inputs and outputs and then apply the DEA method to estimate the production or cost frontiers. The third approach to industry efficiency analysis is designed to improve the efficiency scores of the DEA model by incorporating various methods of error reduction, e.g. the bootstrap methods rescale the individual efficiency scores using average efficiencies calculated from different subsets of the data. Another approach is to apply the method of least sum of absolute errors (LAV) to the production or cost function and derive the estimates of the respective frontiers.