Nonparametric and semiparametric estimation with discrete regressors

Preprint
Publisher Website
Google Scholar
    • preprint
      • Published in RePEc
Abstract
This paper presents and discusses procedures for estimating regression curves when regressors are discrete and applies them to semiparametric inference problems. We show that pointwise root-n-consistency and global consistency of regression curve estimates are achieved without employing any smoothing, even for discrete regressors with unbounded support. These results still hold when smoothers are used, under much weaker conditions than those required with continuous regressors. Such estimates are useful in semiparametric inference problems. We discuss in detail the partially linear regression model and shape-invariant modelling. We also provide some guidance on estimation in semiparametric models where continuous and discrete regressors are present. The paper also includes a Monte Carlo study.
All Related Versions

This publication has 0 references indexed in Scilit: