Page 1 of 1

Structural Breakpoint Analysis

Posted: Thu Apr 04, 2019 5:10 am
by abby_d
Dear all,

I want to perform a structural breakpoint analysis of time series data and since I'm quite new to the topic, I have a few questions...

First: when I did the Bai, Perron test, that allows for multiple breaks, in my output I was told there is only 1 break. Does that mean, I'd be better off doing the Andrews single break test?

Second: when testing for an unit root, I'm not sure whether is Lee, Strazicich test is more appropriate or the Zivot-Andrews test. Anyone have some insights to share?

Third: when running my regression, I can either simply split the date into 2 and run 2 separate regressions OR I can run a breakpoint regression with all data. Again, any ideas on which approach gets more sound results?

Also if you have some undergraduate level literature recommendations, I'd be very grateful!! Thank you all in advance!
Greetings from Berkeley!