I would like to perform a granger causality test between two variables, A and B.
However, these two variables are of orders I(0) and I(1).
The variable A is stationary I(0) and the variable B is stationary in the first difference I(1).
However when I use VARSOC :
Code:
varsoc A difference_B
Please, I have two questions:
- Is it normal to have an optimal lag of 0 to perform a granger causality test? (I have read on the internet several times that it is not normal to have a granger causality test)
- What should I do when I have two variables of different orders?
- Should I use the variables at their stationarity level (i.e. variables A and difference_) as indicated here: https://stats.stackexchange.com/ques...-test-for-coin
- or should I use the first difference for each of them? Difference_A and Difference_B.
Pita
0 Response to Different order of integration - Var and granger causality
Post a Comment