GLS-based Unit Root Tests with Multiple Structural Break

If you are seeking RATS code for implementing a particular technique or replicating results from a paper, post your request here. Be sure to include complete citations for any papers or books.
ege_man
Posts: 85
Joined: Sat Jul 07, 2012 2:39 pm

GLS-based Unit Root Tests with Multiple Structural Break

Unread post by ege_man »

Dear Tom,

I am looking for the RATS code of the following paper if it is available.

Carrion-i-Silvestre, J.L., D. Kim and P. Perron (2009), GLS-based Unit Root Tests with Multiple Structural Breaks both Under the Null and the Alternative Hypotheses, Econometric Theory 25, 1754-1792.

You can also find Gauss code for the paper in the following link

http://people.bu.edu/perron/code/Replic ... T-2009.zip

Thanks
TomDoan
Posts: 7814
Joined: Wed Nov 01, 2006 4:36 pm

Re: GLS-based Unit Root Tests with Multiple Structural Break

Unread post by TomDoan »

Are there any applications for that? The paper and the Gauss code are only for simulated data.
ege_man
Posts: 85
Joined: Sat Jul 07, 2012 2:39 pm

Re: GLS-based Unit Root Tests with Multiple Structural Break

Unread post by ege_man »

Unfortunately not, as far as I understand the paper is just based on generated numbers.
TomDoan
Posts: 7814
Joined: Wed Nov 01, 2006 4:36 pm

Re: GLS-based Unit Root Tests with Multiple Structural Break

Unread post by TomDoan »

The GLS detrending with multiple breaks is not much different from GLS detrending with one. (The @GLSDETREND procedure handles one break). The only other thing that's novel is the use of some variant of the Bai-Perron algorithm for making the calculation more efficient. But since anything beyond two breaks is likely pointless, the efficiency gain isn't that important.

The Lee-Strazicich test (@LSUNIT) also allows for multiple structural breaks both under the null and the alternative; it does OLS rather than GLS detrending. There's a rather obvious question of why the authors of the new technique wouldn't have done a comparison to show that their approach is "better".
Post Reply