Zivot Andrews and VAR
Posted: Tue Jan 23, 2018 2:12 am
After performing Zivot Andrews test, I find structural breaks in the data. The test ofcourse takes care of it while employing the unit root test, but how should we incorporate that break in our VAR?
Should we create a dummy or is there an alternative to reflect that break in our VAR model.
If creating a dummy is the option, how would that be?
Suppose x has a structural break at Q2:2009 and Q3:2014, and y has a structural break Q4:2014, how could this be solved?
After running the UR test, VAR, should we worry about this dummy while Granger Causality and Impulse Response Function?
Or should we go for SVAR, now?
Regards,
Deven N Valecha
Should we create a dummy or is there an alternative to reflect that break in our VAR model.
If creating a dummy is the option, how would that be?
Suppose x has a structural break at Q2:2009 and Q3:2014, and y has a structural break Q4:2014, how could this be solved?
After running the UR test, VAR, should we worry about this dummy while Granger Causality and Impulse Response Function?
Or should we go for SVAR, now?
Regards,
Deven N Valecha