Application of robust wild bootstrap estimation of linear model in econometric

Abstract

This paper investigates the use of robust wild bootstrap techniques on regression model as an estimator for economic indicators in a situation where heteroscedasticity and outliers are present. We introduced robust procedures, called robust weighted bootstrap least trimmed squares (RWBootWu) and robust weighted bootstrap least trimmed squares (RWBootLiu). The propose method uses the weighted residuals incorporating the Huber weighted function, least trimmed squares (LTS) estimator, bootstrap sampling procedure of Wu and Liu as well as the robust location and scale,. Numerical examples and simulation were carried out to evaluate the performance of the RWBootWu and RWBootLiu with the existing wild bootstrap BootWu, BootLiu, RBootWu, and RBootLiu method. The result of the study proved that the (RWBootWu) and (RWBootLiu) offer as a substantial improvement over the existing methods and proved to be good alternative estimators.