Dear sir or Madam,
I'm doing my thesis, and i would like to read your suggestions about what i'm currently doing and some answers to a few questions i have.
Firts, i found that all my variables are stationary at firts difference.
In order to test for cointegration among them, is it correct to do it by using an "ARDL bound test" (normally if all variables are integrated of the same order, you test for cointegration through Johansen's test).
Also, given that all my variables are integrated of the same order and a cointegration relation exists, i would like to know if its correct to test for granger causality relations through a VAR model by following Toda - Yamamoto procedure?
Regards,
Joel
HI
Pls read this article. It's really good for Toda Yamamoto procedure.
http://shishirshakya.blogspot.com.au/2015/06/the-todayamamoto-approach-to-granger.html
Kind regards
Thushara
VECM seems more appropriate for the case when your variables are integrated at the same order and a co-integration relation exists.
You can use:
1) Johansen Methodology when variables are nearly co-integrated;
2) Engle's and Granger's;
3) VECM, when variables are integrated at the same order, and
4) etc (Augmented Dicky-Fuller assuming both have unit root)
You can check from many article which that solve many problem in the research methodology.
In this case, you can simply run VECM for causality. Becasue Toda - Yamamoto procedure is suggested with I(2) variables.
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series