Non-parametric cointegrating regression with NNH errors

Qiying Wang and Ying Xiang Rachel Wang

Abstract

This paper studies a non-linear cointegrating regression model with non-linear nonstationary heteroskedastic error processes. We establish uniform consistency for the conventional kernel estimate of the unknown regression function and develop a two-stage approach for the estimation of the heterogeneity generating function.

Keywords: Cointegration, non-parametric regression, kernel estimate, nonstationarity, nonlinearity, heterogeneity.

AMS Subject Classification: Primary C14; secondary C22.