Instrumental Variable Estimation in Measurement Error Model Under Exact Restrictions

Abstract

Instrumental Variable Estimation in Measurement Error Model Under Exact Restrictions Shalabh Department of Mathematics and Statistics, Indian Institute of Technology, Kanpur, India ?: Most of the statistical analysis of data assumes that the data has been obtained correctly. In many practical situations, the measurement error creeps into the data and the true values of the variables cannot be observed. Instead, the data contaminated with measurement error is obtained. The presence of measurement error in the data disturbs the optimal properties of the estimators. In the context of measurement errors, the usual least squares estimators of the parameters become biased and inconsistent. The instrumental variable technique is one of the estimation techniques which yield consistent estimators of regression coefficient in measurement error models. In many practical situations, some prior information about the regression coefficient is available from some extraneous source. When such information is incorporated in the estimation procedure, the resulting estimators like restricted regression estimators are well known to have superior performance compared to the case when it is not used. How to use such prior information available in the form of exact linear restrictions in the instrumental variable estimation of regression coefficients in a measurement error model is the issue addressed in this talk.