Page 1 of 1

Rank of cointegration using ARDL - EViews 9

Posted: Tue Aug 18, 2015 12:11 pm
by benjwin
Hello,

because of my small sample size and different orders of integration among my 5 variables, I am using an ARDL framework to test for cointegration.
Eviews 9 does a great job for this purpose!
The result suggest that my variables are cointegrated and hence I proceed by estimating a VECM framework.
However, estimating a VECM requires the specification of the "Number of cointegrating". Where do I find the rank of cointegration within my ARDL estimation?
Furthermore, the optimal lag length as given by ARDL estimation result is defined as ARDL(2,0,1,2,2), but what lag length should I use for the VECM framework?

Hoping for valuable feedback!

Benjwin