Measurement error in quantile regression models with an application to returns to education

Abstract

This project seeks to develop estimation and inference of quantile regression with measurement error models. We shall propose a semiparametric easy to implement two-step estimator under the assumption of repeated measures. The asymptotic properties of the estimator will be derived, namely, consistency and weak convergence. We shall also develop practical statistical inference procedures. Monte Carlo simulations will assess the nite sample performance of the proposed methods. These should be particularly useful for applications in which there exists of measurement error in the dependent variable. Education is notoriously measured with error. Thus, based on a sample of 428 female monozygotic twins from the Twins Research Unit, St. Thomas'Hospital United Kingdom, we will revisit and analyze the important example of returns to education, and measure the marginal impact of education on earnings. (AU)