Cointegrated but a not statistically significant error ter in VECM.
Posted: Fri Feb 02, 2018 12:03 am
Hello there,
All of my variables are I(1) and Johansen test says that my multivariate model has one cointegration. However, my VECM is not statistically significant nor has a right sign. What should I do next? Should I do general to specific approach or VAR with first difference?
Thank you for consideration.
All of my variables are I(1) and Johansen test says that my multivariate model has one cointegration. However, my VECM is not statistically significant nor has a right sign. What should I do next? Should I do general to specific approach or VAR with first difference?
Thank you for consideration.