Alternative Estimators of Cointegrating Parameters in Models With Non-Stationary Data: An Application to US Export Demand

20 Pages Posted: 8 Aug 2009

See all articles by James J. Forest

James J. Forest

University of Massachusetts - Amherst

Paul Turner

Loughborough University - Department of Economics

Date Written: August 3, 2009

Abstract

This paper presents Monte Carlo simulations which compare the empirical performance of two alternative single equation estimators of the equilibrium parameters in a dynamic relationship. The estimators considered are Stock and Watson’s dynamic ordinary least squares (DOLS) estimator and Bewley’s transformation of the general autoregressive distributed lag model. The results indicate that the Bewley transformation produces a lower mean-square error as well as superior serial correlation properties even with lower truncation lags for the lagged variables included in the estimation equation. An application is then provided which examines the nature of the equilibrium relationship between aggregate US exports, world trade and the US real exchange rate. This confirms that estimation of the equilibrium parameters of this relationship by the Bewley transformation produces results which are superior to estimation by DOLS.

Keywords: DOLS estimator, Bewley transformation, US Export Demand

JEL Classification: C12, C15

Suggested Citation

Forest, James J. and Turner, Paul, Alternative Estimators of Cointegrating Parameters in Models With Non-Stationary Data: An Application to US Export Demand (August 3, 2009). Available at SSRN: https://ssrn.com/abstract=1443102 or http://dx.doi.org/10.2139/ssrn.1443102

James J. Forest

University of Massachusetts - Amherst

Department of Finance
Amherst, MA 01003
United States

HOME PAGE: http://https://www.researchgate.net/profile/James_Forest2

Paul Turner (Contact Author)

Loughborough University - Department of Economics ( email )

York House
Loughborough LE11 3TU
Great Britain

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
195
Abstract Views
999
rank
169,950
PlumX Metrics