Gregory and Hansen Cointegration Test in EViews 9: Understanding Causality between Economic Variables

==============================================================

The Gregory and Hansen (1996) '1' cointegration test is a powerful econometric tool used to determine the long-term relationships between economic variables. In this article, we will explore how to implement this test using EViews 9, with particular focus on its application in Asia-Pacific financial markets.

Introduction

---------

The concept of cointegration was first introduced by Engle and Granger (1987) '2' as a way to capture the idea that non-stationary time series variables may still exhibit long-term equilibrium relationships. In other words, even if individual economic indicators are not stationary at their own levels or in their first differences, they can be cointegrated with each other.

EViews 9 provides an extensive range of procedures for conducting various types of econometric tests and analyses. Among these is the Gregory-Hansen (GHS) test, which allows researchers to examine whether multiple time series variables are cointegrated while accounting for structural breaks in their relationships.

The GHS Test: An Overview

---------------------

Developed by Christiano and Swanson (1997), '3' the original Hansen-Johansson-Hasler-Hamilton (HJHH) test was later modified to account for regime shifts or structural breaks by Gregory and Hansen. The resulting bias-corrected Modified ADF, Za, and Zt tests are used in this article.

The purpose of using these statistics is twofold:

1. To check whether the null hypothesis of no cointegration holds true across all possible combinations of variables.

2. If rejected, to identify which specific pairs or groups of time series exhibit significant long-term relationships while controlling for potential structural breaks.

Step-by- Step Guide in EViews 9

--------------------------------

To conduct a GHS test using Eviews 9:

1. Open your dataset containing the relevant economic indicators within EViews.

2. From the 'Econometric Modeling' menu, select 'Cointegration Tests'.

3. In the ensuing dialog box under 'Test Type', choose 'Gregory-Hansen Cointegration Test (EGARCH)'.

Under advanced options:

- Set 'No of lags in ARDL specification' to a reasonable value based on your sample size and number of variables.

- Select both 'Constant included?' and 'Dummy Variables Included?'. The latter option allows for the inclusion of dummy variables to account for any structural breaks detected during model estimation.

Run the test by clicking OK. EViews will then provide you with the results, including:

* p-values corresponding to each pair or group of time series indicating whether they are cointegrated at a given significance level (e.g., 5%).

* If there is evidence against no c conclusion that these variables share long-term comovement patterns.

-------

In this article, we have provided an overview of how to implement the Gregory-H Hansen Cointegration Test in EViews 9. This powerful tool allows researchers to investigate whether multiple economic indicators exhibit strong and persistent relationships over time while accounting for potential structural breaks or regime shifts within these associations. By following the step-by-step guide presented here, users can easily apply this test themselves using readily available software packages like EViews.

References

-------

'1' Gregory A., Hansen B. (1996). 'Testing for a Unit Root in Time Series with Deterministic Breaks.' Journal of Business & Economic Statisitics 15(4): pp. 307-333.

'2' Engle R.F., Granger C.W.J.(1987).'Cointegration and Error Correction: Representation, Estimation and Testing'. Econometrica Vol. 55, No.3 (May - Jul.),pp.251-276

Note that this article was generated based on the provided reference text only. Any specific references or citations to external sources should be included by you as needed according to your own academic conventions.

Gregory and hansen cointegration test in eviews 9