Page 1 of 1

Johansen cointegration tests and Granger Causality on VECM

Posted: Sun Jul 07, 2013 3:16 pm
by sotia_econ
Hi all !

I am undertaking Johansen Cointegration tests in four variables pair-wise (2 variables in each test). For those that I detect cointegration relationships I have the following questions:

1) How do I form the VECM through a Johansen estimation window (I did Proc-VECM-fullfilled the number of cointegrated relationships found with Johansen in Cointegration Tab but need to know how do I apply the Wald criterion on the VECM to specify the appropriate number of lags)

2) How do I apply a Granger Causality test from a VECM estimation output window and again how do I use the Wald criterion for the specification of the lags.

Help would be precious!!!