Prewhitening-Based Estimation in Partial Linear Regression Models

a Comparative Study

Abstract

The problem of semiparametric modelling in time series is considered. For this, partial linear regression models are used, that is, regression models where the regression function is the sum of a linear and a nonparametric component. Two estimators for the nonparametric component are shown: one estimator takes into account the dependence structure in the errors of the regression function and the another estimator not. Both estimators are compared in practice for several real time series concerning economics and finance. In addition to the partial linear regression model, other regression models are included in the comparison.