Page 2 of 2
Re: 2 variables and 2 cointegrated eqns.
Posted: Sun Sep 11, 2011 12:35 pm
by startz
Good news. Testing for Granger causality works just fine on stationary series. Forget about cointegration. Search for Granger causality in the help system.
Re: 2 variables and 2 cointegrated eqns.
Posted: Sun Sep 11, 2011 1:39 pm
by dimitris
so if the series are I(0) what i have to do to go to the granger causaity?
Re: 2 variables and 2 cointegrated eqns.
Posted: Sun Sep 11, 2011 4:27 pm
by dimitris
thnx a lot for the valuable help!! and some last questions
if un is I(0) and inflation is I(1)??? can i do something or not???
also i saw an unemployment rate serie which was stationary in levels in 10% sigificance level by ADF test but in Phiplis Peron test wasn stationary in levels?? then what i do! i mean i follow ADF or PP!!
Re: 2 variables and 2 cointegrated eqns.
Posted: Wed Sep 16, 2015 4:44 am
by deleidimatteo
hi,
I am studying causality between 2 interest rate: central bank interest rate and lending commercial bank interest rate.
both series are I(1) at level and the probability of integration at the first difference is equal to zero.
when I estimate the Johansen cointegration test between the 2 variable the result is: 2 cointegrating equantions. what does it mean?
moreover, i can't estimate the VECM imposing 2 cointegrating eqns. because the number of variables must be greater than the number of cointegrating eqns.
even if the results of Johansen suggest 2 cointegrating eqns, can I estimate the VECM imposing just 1 cointegrating equation?
thank you in advance.
Matteo