I'm using OLS Bai perron test to detect structural breakpoints in monthly data. The results show that sequential break dates and repartition break dates are different. Why does this happen? Do I have to include all break dates in my analysis or should I choose between sequential break dates and repartition dates?
1 2009M05 2005M08
2 2005M08 2012M02
BTW I've read Bai's paper(1997) and he said that "the sequential estimation method has a tendency to over- or underestimate the true location of a break point", "Repartition yields an estimator having the same asymptotic distribution as the simultaneous estimators". So I'm thinking maybe the repartition method is more accurate? However, Bai's statement is very vague so I'm not sure.
one of my variables has a linear trend, should I choose option 4 for my cointegration test? (i.e.: allow for linear deterministic trend for VAR+intercept and trend in CE) Also, after conducting the Johansen test, Eviews has already shown the cointegration equation. Then do I have to conduct VEC model next?
I'm new to Eviews. I'd really appreciate your help.
For econometric discussions not necessarily related to EViews.
1 post • Page 1 of 1
Who is online
Users browsing this forum: No registered users and 2 guests