Page 1 of 1

Eviews 5 var and vecm

Posted: Tue May 11, 2010 7:16 am
by henrikstudent
I am currently testing two time series to see if one leads the other in finding a new price, both series are related by arbitrage relations in theory.
Previous studies haves used the johansen cointegration test and then gone to use a vecm. However i have found through the use of the augumented df test that the residuals after running an ols on the time series are stationary, so my question is can i then run a simple var or do i have to use the johansen method followed by a vecm.

Thanks