A threshold cointegration test with increased power

Authors:
Steven Cook
Affiliations:
Department of Economics, University of Wales Swansea, Singleton Park, Swansea SA2 8PP, United Kingdom
Venue:
Mathematics and Computers in Simulation
Year:
2007

Citing 2
Cited 1

Time Series Analysis, Forecasting and Control

Time Series Analysis, Forecasting and Control
The behaviour of US stock prices: Evidence from a threshold autoregressive model

Mathematics and Computers in Simulation

Small sample improvements in the threshold cointegration test using residual-based moving block bootstrap

Mathematics and Computers in Simulation

Quantified Score

Hi-index	0.00

Visualization

Abstract

The low power of threshold, or asymmetric, cointegration tests is addressed. A new test is developed which combines momentum-threshold autoregression (MTAR) and local-to-unity detrending via generalised least squares (GLS). Critical values for the newly proposed GLS-MTAR threshold cointegration test are provided under alternative decisions regarding the deterministic terms employed when implementing the test. Simulation analysis of the test shows it to provide a substantial increase in power relative to the previously proposed MTAR threshold cointegration test of Enders and Siklos [W. Enders, P. Siklos, Cointegration and threshold adjustment, J. Business Econ. Statist. 19 (2001) 166-176].