Local Constant Kernel Estimation of a Partially Linear Varying Coefficient Cointegration Model

In this paper, we consider a partially linear varying coefficient cointegration model. We focus on the estimation of constant coeffients. We derive the saymptotic result for the local constant kernel estimator, which complements the results in Li, Li, Liang and Hsiao (2013) where the local polynomia...

Full description

Saved in:
Bibliographic Details
Published in:Annals of economics and finance 2015-11, Vol.16 (2), p.353-369
Main Authors: Wang, Luya, Liang, Zhongwen, Lin, Juan, Li, Qi
Format: Article
Language:English
Subjects:
Cointegration analysis
Regression analysis
Simulation
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we consider a partially linear varying coefficient cointegration model. We focus on the estimation of constant coeffients. We derive the saymptotic result for the local constant kernel estimator, which complements the results in Li, Li, Liang and Hsiao (2013) where the local polynomial estimation methods are studied. However, Li et al. (2013) impose stronger conditions to rule out the local constant estimation due to technical diffiulties. We give the full treatment of the local constant method in this paper based on a novel proof. From the simulation results reported in the paper, we show that the local constant and local linear estimators perform similarly, but the local constant method requires less data. Also, in fnite sample applications the local linear estimation could suffer from the matrix singularity problem.
ISSN:1529-7373