|Unit roots, cointegration
||[Jun. 29th, 2010|07:35 pm]
1. When testing for unit roots in eviews, if some of the series show trend coefficient (significant) and some only show significant intercept, would it effect my cointegration tests? |
2. For cointegration test what should I put the specification as?
3. How do I decide on the number of lags of for cointegration?
4. When I run the VECM model and if I have more than one cointegrating vector, how would I know which one to choose? When I specify the number of cointegrating vectors in eviews (for VECM), it gives me results for the number of cointegrating vectors specified. I am confused as to which one to choose? Can I just leave the number of cointegrating vectors as 1? Would it effect my results for speed of adjustment?