Bai and Perron Sequential test

For requesting general information about EViews, sharing your own tips and tricks, and information on EViews training or guides.

Moderators: EViews Gareth, EViews Moderator

testeviews123
Posts: 8
Joined: Tue Mar 01, 2022 6:49 am

Bai and Perron Sequential test

Postby testeviews123 » Tue Mar 01, 2022 11:42 am

I am having trouble to find out where the coefficient estimation for the 5 breaks are in EVIEWS when i apply the Bai and Perron Sequential Test in EVIEWS. I cannot find anything on their website about this. They only represent the output given for the breaks that are significant by the sequential test. Yet i wanna see the full output (coefficients, standard deviations p values etc) of the 5 breaks.

Thanks in advance!

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13305
Joined: Tue Sep 16, 2008 5:38 pm

Re: Bai and Perron Sequential test

Postby EViews Gareth » Tue Mar 01, 2022 12:02 pm

Did you try using the breakls estimation method?
Follow us on Twitter @IHSEViews

testeviews123
Posts: 8
Joined: Tue Mar 01, 2022 6:49 am

Re: Bai and Perron Sequential test

Postby testeviews123 » Tue Mar 01, 2022 12:23 pm

EViews Gareth wrote:Did you try using the breakls estimation method?


yes but i cannot see the output of coefficients and standard errors T test results for the 5 breaks. I put m=5 and i want to see the results of those 5 coefficients and standard errors T test results.

To be more clear I want to see the values of coefficients and standard errors T test results.

Like follows:

break 1 (value of coefficient) (value of standard error) (value of T stat) (P-value) etc...

Because now EVIEWS only report the values when the Sequential Method found the breaks to be significant. But i just wanna see all the values for the 5 breaks


any idea how to retrieve those?

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13305
Joined: Tue Sep 16, 2008 5:38 pm

Re: Bai and Perron Sequential test

Postby EViews Gareth » Tue Mar 01, 2022 2:07 pm

You cannot, other than using fixed number of breaks.
Follow us on Twitter @IHSEViews

testeviews123
Posts: 8
Joined: Tue Mar 01, 2022 6:49 am

Re: Bai and Perron Sequential test

Postby testeviews123 » Tue Mar 01, 2022 2:12 pm

EViews Gareth wrote:You cannot.


I dont understand the creators of EVIEWS or any other programming language. Why give output but not provide details of the estimation results??? it really doesnt make any sense right? So in order to obtain those estimate what should i do now??

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13305
Joined: Tue Sep 16, 2008 5:38 pm

Re: Bai and Perron Sequential test

Postby EViews Gareth » Tue Mar 01, 2022 2:25 pm

Because those results aren't calculated.
Follow us on Twitter @IHSEViews

testeviews123
Posts: 8
Joined: Tue Mar 01, 2022 6:49 am

Re: Bai and Perron Sequential test

Postby testeviews123 » Tue Mar 01, 2022 2:37 pm

EViews Gareth wrote:Because those results aren't calculated.


huh how come ? When you retrieve an RSS automatically you should get an coefficient and standard deviation T test p value etc right? Am i missing here something? and where do i go wrong in here?

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13305
Joined: Tue Sep 16, 2008 5:38 pm

Re: Bai and Perron Sequential test

Postby EViews Gareth » Tue Mar 01, 2022 3:29 pm

No, when you calculate the RSS, you calculate the RSS. You could then, from the RSS, compute other statistics, but that's totally inefficient if you don't need those statistics.
Follow us on Twitter @IHSEViews

testeviews123
Posts: 8
Joined: Tue Mar 01, 2022 6:49 am

Re: Bai and Perron Sequential test

Postby testeviews123 » Tue Mar 01, 2022 3:38 pm

EViews Gareth wrote:No, when you calculate the RSS, you calculate the RSS. You could then, from the RSS, compute other statistics, but that's totally inefficient if you don't need those statistics.


In the Bai and Perron paper (2003) They choose 2 breaks while the sequential method returned 0. They choose the two breaks based upon the Sup(2|1) and Sup(2) 10% level, So now if you wanna write a paper you should also return the coefficients and the respective standard errrors right.

So why leave those results out of the estimation. If you calculate the RSS then you also have the betas and the standard errors t values p value etc.

because the formula is (y-xb)^2=e^2 , where we minimze b , so we would for each segment have the betas and thus the standard errors etc .. so why return the RSS but not those??? you understand me now?

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13305
Joined: Tue Sep 16, 2008 5:38 pm

Re: Bai and Perron Sequential test

Postby EViews Gareth » Tue Mar 01, 2022 4:17 pm

There are many mathematical tricks you can use so that to calculate the series of RSS you don't need to compute individual coefficients, and certainly not standard errors.

You can always use the fixed number of breaks option if you want to view the coefficients and standard errors for two breaks (or any other).
Follow us on Twitter @IHSEViews

testeviews123
Posts: 8
Joined: Tue Mar 01, 2022 6:49 am

Re: Bai and Perron Sequential test

Postby testeviews123 » Wed Mar 02, 2022 7:31 am

EViews Gareth wrote:There are many mathematical tricks you can use so that to calculate the series of RSS you don't need to compute individual coefficients, and certainly not standard errors.

You can always use the fixed number of breaks option if you want to view the coefficients and standard errors for two breaks (or any other).



ok so now, I can imagine the tricks being able to conduct in order to obtain coefficients and t stats etc... but why so much hassle for those just provide them because you already have the RSS its just printing the other values. It still doesnt make any sense. I hope the creators will focus on these issues and not come up with recommended tricks but just provide things which are already calculated.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13305
Joined: Tue Sep 16, 2008 5:38 pm

Re: Bai and Perron Sequential test

Postby EViews Gareth » Wed Mar 02, 2022 7:44 am

That's the point - coefficients and t-stats are not calculated. If you want to calculate them, it takes extra machine power, which slows down the whole routine.
Follow us on Twitter @IHSEViews

testeviews123
Posts: 8
Joined: Tue Mar 01, 2022 6:49 am

Re: Bai and Perron Sequential test

Postby testeviews123 » Wed Mar 02, 2022 9:33 am

EViews Gareth wrote:That's the point - coefficients and t-stats are not calculated. If you want to calculate them, it takes extra machine power, which slows down the whole routine.


Not calculated?...The Sum of Squared Residuals i am talking about ... The Sum of Squared Residuals in the Bai and Perron (2003) paper calculates per segment the RSS. So now , it minimizes the coefficients B and Delta right.... (Y-xb-Zdelta)^2 ??? What else does it minimize then ? So now if it minimizes the coefficients B and Delta then just provide their value and related stats such as standard deviations etc...

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13305
Joined: Tue Sep 16, 2008 5:38 pm

Re: Bai and Perron Sequential test

Postby EViews Gareth » Wed Mar 02, 2022 9:59 am

I feel like we are going round in circles here.

To compute the RSS at each date, you do not need to calculate the coefficients, and you certainly do not need to calculate the standard errors, t-stats.
Follow us on Twitter @IHSEViews

testeviews123
Posts: 8
Joined: Tue Mar 01, 2022 6:49 am

Re: Bai and Perron Sequential test

Postby testeviews123 » Wed Mar 02, 2022 10:07 am

EViews Gareth wrote:I feel like we are going round in circles here.

To compute the RSS at each date, you do not need to calculate the coefficients, and you certainly do not need to calculate the standard errors, t-stats.


You make me confused. I am clearly looking at the RSS formula which is RSS=SUM (y-xb)^2=e^2 , so now b has to be estimated first by minimizing (y-xb)^2 with respect to b which turns out to be b=SUM(xy)/SUM(xx) so now then we sum RSS right?

What am i missing in this ?? Your answers does not clearify the issue enough. If i am wrong in my understanding then explain where i go wrong.


Return to “General Information and Tips and Tricks”

Who is online

Users browsing this forum: No registered users and 14 guests