Bai and Perron structural break test

For questions regarding the import, export and manipulation of data in EViews, including graphing and basic statistics.

Moderators: EViews Gareth, EViews Jason, EViews Steve, EViews Moderator

lakis
Posts: 13
Joined: Wed Oct 22, 2014 7:33 am

Bai and Perron structural break test

Postby lakis » Wed Oct 22, 2014 7:45 am

Hi guys,

I am trying to figure out how Eviews8 works with this Multi stuctural breaks tests.

Bai and Perron (2003) suggest 4 different test. So I am trying to figure out which eviews test corresponds to each test that authors suggest. So allow me to write down what I have understood and what I miss!!

SupF(l+1|l) test --------------> Sequential L+1 vs L (Eviews)
Double Maximum test (UD,WD) tests------------> Global Lvs none (Eviews)
LWZ and BIC tests ------------------------------> Global Info criteria (Eviews)
SupF test (fixed number of breaks vs 0) ---------> ??????????????????????????????

So thats what I miss. I cant find the test the Eviews uses for testing the "SupF test (fixed number of breaks vs 0)" .

It seems that http://www.eviews.com/EViews8/ev8ecmbreak_n.html link doesnt give me an answer.

Any help please??

Many thanks

EViews Glenn
EViews Developer
Posts: 2671
Joined: Wed Oct 15, 2008 9:17 am

Re: Bai and Perron structural break test

Postby EViews Glenn » Wed Oct 29, 2014 8:36 am

Those should be the individual rows of the Global L vs. none test.

pitsi
Posts: 19
Joined: Wed May 28, 2014 2:43 pm

Re: Bai and Perron structural break test

Postby pitsi » Sun Dec 14, 2014 12:00 pm

Dear Glenn,
i would like to see if cointegration exists among 2 series (prices of 2 countries). i find that cointegration exists but i want to test whether breaks exist.
Thus, i employ the least squares with breakpoints via evews 8. Do i have to choose first the method global L breaks vs. None(Udmax), then the Global information Criteria to get the rss and then the sequential l+1 breaks vs. L??
Or am i doing smthg wrong?
Thank you very much for your time

EViews Glenn
EViews Developer
Posts: 2671
Joined: Wed Oct 15, 2008 9:17 am

Re: Bai and Perron structural break test

Postby EViews Glenn » Wed Dec 17, 2014 4:48 pm

I don't understand the question.

pitsi
Posts: 19
Joined: Wed May 28, 2014 2:43 pm

Re: Bai and Perron structural break test

Postby pitsi » Thu Dec 18, 2014 2:18 am

my question is which test among the four of breaks should i take into consideration in order to find the breaks of my series.
Thank you very much
Last edited by pitsi on Sat Jan 10, 2015 8:53 am, edited 2 times in total.

EViews Glenn
EViews Developer
Posts: 2671
Joined: Wed Oct 15, 2008 9:17 am

Re: Bai and Perron structural break test

Postby EViews Glenn » Thu Dec 18, 2014 10:17 am

That's a preference question. There is no right answer. I will say that the consensus probably is that one should use one of the global methods as they evaluate all alternatives.

pitsi
Posts: 19
Joined: Wed May 28, 2014 2:43 pm

Re: Bai and Perron structural break test

Postby pitsi » Thu Dec 18, 2014 11:21 am

Thank you very much for your help! :D

pitsi
Posts: 19
Joined: Wed May 28, 2014 2:43 pm

Re: Bai and Perron structural break test

Postby pitsi » Sun Dec 21, 2014 10:58 am

Dear Glenn,
if i use the Sequential L+1 vs L with Eviews8, how do i get the rss ????
Thank you very much in advance!

pitsi
Posts: 19
Joined: Wed May 28, 2014 2:43 pm

Re: Bai and Perron structural break test

Postby pitsi » Tue Dec 23, 2014 4:26 pm

Hello again,
i have installed the add-in bai-perron breakpoints test.
Can i perform the Supf Test from the breakls menu of eviews and then do the bai-perron breakpoints test (add-in) so as to get the rss??
Because i have already done all the tests and results differ...
Thank you once again..

EViews Glenn
EViews Developer
Posts: 2671
Joined: Wed Oct 15, 2008 9:17 am

Re: Bai and Perron structural break test

Postby EViews Glenn » Mon Dec 29, 2014 12:11 pm

I can't speak to the differences since I'm not that familiar with the add-in.

I must admit to being a bit puzzled as to why you'd want the RSS for the sequential tests as the F-statistic is all that really matters, especially since you can do robust versions of the test. I'll look into an enhancement going forward, but I'm afraid that won't help you at the moment.

pitsi
Posts: 19
Joined: Wed May 28, 2014 2:43 pm

Re: Bai and Perron structural break test

Postby pitsi » Mon Dec 29, 2014 5:46 pm

Thank you very much Glenn,
I insist on the supf(l+1) test and the rss because my supervisor told me that i have to choose the breaks with the lowest rss.
I have done all the avaliable tests + the add-in breakpoint test and dates are similar.. 2 or 3 months is the difference among them .
Thank you very much once again :)))))))))))

EViews Glenn
EViews Developer
Posts: 2671
Joined: Wed Oct 15, 2008 9:17 am

Re: Bai and Perron structural break test

Postby EViews Glenn » Wed Dec 31, 2014 11:08 am

Choosing the break with the lowest RSS is really not the way that the SupF tests works. Moreover, specifying that as your objective doesn't really make sense since the lowest RSS is obtained by having (in essence) a break at every period.

pitsi
Posts: 19
Joined: Wed May 28, 2014 2:43 pm

Re: Bai and Perron structural break test

Postby pitsi » Sat Jan 03, 2015 6:49 pm

Thank you Glenn, i ll take that into consideration.
I have decided to choose the global test as you have recommended in a previous post.
Thank you very much really for your help


Return to “Data Manipulation”

Who is online

Users browsing this forum: No registered users and 13 guests