Small Sample Evidence of Regression Quantile Estimates for Structural Models: Estimation and Testing

Abstract

A small Monte Carlo study is conducted to investigate the small sample properties of regression quantiles estimates in structural econometric models with iid errors. Two versions of median regression (the 50 th quantile) estimates using instrumented regressors are used, with the fi rst stage estimated by least squares and by median regression. These estimators are compared to two stage least squares, ordinary least squares and simple median regression. It is found that the performance of the two stage regression quantile estimators is equal, if not better, than two stage least squares in terms of bias and that t tests have good size and power properties, under a variety of error distributions. Resumo Atraves de urn pequeno experimento de Monte Carlo, estudamos as pro priedades em pequenas amostras dos estimadores de regressao quantilica para uma equagao estrutural em urn modele de equag5es simultaneas com erros iid. Duas versoes de estimadores de regressao mediana (0 50Q quantil!percentil) us ando variaveis instrumetais sao empregados, estimando 0 primeiro estagio com minimos quadrados ou com regressao mediana. Comparando-os com estimadores de minimos quadrados em dois estagios(MQ2E) e ordinarios, e regressao mediana .. A previous version of this work was part of my dissertation The Effect of Personal Income Taxes on Labor Supply in Brazil: an application of quantile regression at the University of Illinois at Urbana�Champaign and was presented at the XVIII Encontro Brasileiro de Econometria (EBE) at Aguas de Lind6ia, SP 1 Brazil. I would like to thank the EBE session participants and the refer ees for comments and suggestions and Roger Koenker for very helpful comments. AJiy errors are my own. The financial support of CAPES�Brazil and UFRR at earlier stages of this work is ac knowledged. This paper was supported by CNPq-Brazil under project 301170/96�7. Joaquim Dias and Suzel L. Jensen provided research assistance.

Two‐stage quantile regression when the first stage is based on quantile regression

“…See alsoRibeiro (1998) for small sample simulations of two-stage LAD estimation with a first step of LS or LAD estimators. C Royal Economic Society 2004…”

Two-stage Quantile Regression when the First Stage is Based on Quantile Regression

“…See alsoRibeiro (1998) for small sample simulations of two-stage LAD estimation with a first step of LS or LAD estimators.6…”