Johansen and Granger causality test

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

nsyenhong
Posts: 1
Joined: Wed May 03, 2017 6:26 am

Johansen and Granger causality test

Postby nsyenhong » Wed May 03, 2017 9:38 pm

Greetings everyone, I have a problem regarding the contradiction between the Johansen test result and granger causality ECT result.

Based on the Johansen test result, it show that there is 1 long run co-integration. But, when I proceed with granger causality based on VECM, the coefficient of ECT is insignificant, means that there is no long run co-integration.

May I know how can I justify this? Is Johansen test just a preliminary test?

Thanks in advance.

Frencholivier
Posts: 4
Joined: Sat May 20, 2017 10:25 am

Re: Johansen and Granger causality test

Postby Frencholivier » Sat May 20, 2017 11:51 am

What you need to know is that cointegration implies Granger causality.
However with cointegrated data in a VECM, it is difficult to interpret Granger causality as performed by Eviews. Eviews by default tests Granger causality on the first differences, excluding the error-correction term.
If you are interested in Granger causality in cointegrated VARs, I suggest the Toda-Yamamoto test. A post by Dave Giles will show you how.


Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 21 guests