I have 5 variables (1 dependent variable and 4 independent variables), and the 4 is stationary at level and only 1 is stationary at first different. However, I'm required to conduct the Granger Causality test. Is there any alternative test or solution that can solve my issue?
Thanks
In this case, I recommend using Toda Yamamoto test that addresses the different degrees of cointegration, which is a test of long-term causality, the latter based on the concept of augmented VAR and the modified (Wald test).
The normal Wald test of the linear constraints test loses its convergent properties (its known distribution under the null hypothesis) if the chains are unstable, so Granger's causation test does not work under unstable series.
Toda Yamamoto test Takes this into account, adopting unstable strings and the modified Wald test.
Yes, you can find the Granger Causality if you have a mixture of I(1) and I(0) in your variables.
Dear Yq.
Yes, you can carry out Granger causality with I(0) and I(1) variable.
However, if need be, you can employ Toda Yamamoto.
Regards
Without knowing the nature of your variables and economic model it is difficult to answer your question. the following points should help
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs