Page 1 of 1

var

Posted: Thu Jun 12, 2014 1:08 am
by ANDROO1988
hi
i estimate a var model. i have 2 i(0) variables and 4 I(1) variables.
can i run this var model? Can i run this model by taking the first difference of first three variables and use 2 variable at level?
do I need cointegration test?
:(

Re: var

Posted: Thu Jun 12, 2014 9:20 am
by ANDROO1988
can one help me????

Re: var

Posted: Sat Jun 14, 2014 3:07 am
by ANDROO1988
please help me...

Re: var

Posted: Sun Jun 15, 2014 12:57 pm
by nishantvats12
Hi,

As some of your variables are I(1) it is better to apply ARDL rather than a normal VAR. You can read a lot about the theory of the ARDL and how to apply it on internet. Several posts on this forum also deal with the same.