Page 1 of 1

Estimation of VAR - I(0) and I(1) processes

Posted: Tue Feb 14, 2012 8:21 am
by shepps89
Hi All,

I am completing an econometrics topic, estimating a VAR. I am looking for some advice/confirmation on my method.

I have an information set combining of four I(1) process and one I(0). I am estimating a VAR model, and when first testing for cointegration (Johansen LR Test) I enter all I(1) process but exclude the I(0), as clearly cointegration cannot occur.

No cointegration is found so I wish to proceed with a VAR estimation. Now, I am unsure, but clearly all my I(1) processes need to be integrated before entering the system? But can the I(0) process go straight in aswell? (or) Do I include the I(0) as an exogenous variable?

Any feedback would be greatly appreciated.

Many Thanks, : )

Re: Estimation of VAR - I(0) and I(1) processes

Posted: Tue Feb 14, 2012 11:23 am
by variance
I have an information set combining of four I(1) process and one I(0). I am estimating a VAR model, and when first testing for cointegration (Johansen LR Test) I enter all I(1) process but exclude the I(0), as clearly cointegration cannot occur.
Who told you that? this is not true

I am pretty confident that the following article will help you:

http://www.springerlink.com/content/2460p239wr828885/

regards