Cointegration test

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

linluoau
Posts: 9
Joined: Fri Apr 26, 2019 4:06 am

Cointegration test

Postby linluoau » Mon Aug 29, 2022 10:31 pm

Hi
I have got this test result after testing for cointegration between inflation and import prices. My question is how come EViews says there are two cointegrations in these two variables? Does this mean there is no cointegration between the two? Thanks
Screenshot 2022-08-30 132658.png
Screenshot 2022-08-30 132658.png (29.69 KiB) Viewed 4070 times

virenrehal
Posts: 3
Joined: Thu Sep 22, 2022 3:07 am

Re: Cointegration test

Postby virenrehal » Thu Sep 22, 2022 11:58 pm

This means that your variables may be stationary (the cointegration test is at least reading them as stationary) because the Cointegration rank is equal to the number of variables. If the variables are stationary, there can't be any cointegration. Usually, cointegration exists in I(1) variables implying that the variables are not stationary at level.

If your variables are stationary, you don't need to apply VECM. You can use VAR in levels. But, remember that this cointegration test can be sensitive to trend specifications. You have to be careful about which trend or constant variables to include. Don't rely only on the Information Criteria, look at the graphs to check which specification would be accurate.

This link gives all the information related to this topic: https://spureconomics.com/vector-error-correction-vecm-theory/. Look under the heading "Choosing between VAR and VECM".


Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 11 guests