Analysis on the individual efficiency prediction in the composed error frontier model. A Monte Carlo study

This study seeks to analyse some important questions related to the Stochastic Frontier Model, such as the method proposed by Jondrow et al (1982) to separate the error term into its two components, and the measure of efficiency given by Timmer (1971). To this purpose, a Monte Carlo experiment has b...

Descripción completa

Detalles Bibliográficos
Autores: Dios Palomares, Rafaela, Ramos Millán, Antonio, Roldán Casas, José Ángel
Tipo de recurso: artículo
Fecha de publicación:2002
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2099/4187
Acceso en línea:https://hdl.handle.net/2099/4187
Access Level:acceso abierto
Palabra clave:Probabilism
Inference
Bias
MSE
Point estimator
Production frointer
Jandrow formulae
Simulació (Matemàtica)
Inferència
Classificació AMS::62 Statistics::62F Parametric inference
Classificació AMS::65 Numerical analysis::65C Probabilistic methods, simulation and stochastic differential equations
Descripción
Sumario:This study seeks to analyse some important questions related to the Stochastic Frontier Model, such as the method proposed by Jondrow et al (1982) to separate the error term into its two components, and the measure of efficiency given by Timmer (1971). To this purpose, a Monte Carlo experiment has been carried out using the Half-Normal and Normal-Exponential specifications throughout the rank of the γ parameter. The estimation errors have been eliminated, so that the intrinsic variability of the conditional of u given ε can be evaluated. In addition, the behaviour of the mean and mode as point estimators of u is investigated. The results have yielded some interesting findings. We have observed that both the point estimates and the mean efficiency are more precise in cases of lower efficiency. This occurs when the variable that generates the inefficiency outweights the one that picks up the errors out of the control. The change in order found between the estimated efficiency and its estimated value is misleadingly high especially for low ε, which underlines the risk of estimating at these values.