Thanks a lot.
José Perles

2012/7/16 Summers, Peter <psummers@highpoint.edu>
Well this isn't the Lee & Strazicich test, but I've recently translated gauss code for a Bayesian unit root test with multiple breaks, and I'm happy to share it. So far only breaks in the mean are allowed, but hopefully a version with variance breaks will follow shortly.

The main program is "Bayes_adf_break.inp", which calls the other two utility scripts. You specify the max number of breaks you want to allow for, then the program loops over models with 0 to that many breaks. Currently it's set up to store the posterior distribution of rho (the coefficient on the lagged dependent variable) for each model. It also gives you summary output files including the marginal likelihood for each model. All that could be changed, of course. More details are in my paper "Bayesian evidence on the structure of unemployment, Economics Letters v. 89, 299-306, (2004).

Questions, comments, etc all welcome!

PS

PS -- I also think Sven's idea is a good one.

-----Original Message-----
From: gretl-users-bounces@lists.wfu.edu [mailto:gretl-users-bounces@lists.wfu.edu] On Behalf Of Allin Cottrell
Sent: Monday, July 16, 2012 10:37 AM
To: Gretl list
Subject: Re: [Gretl-users] suggestion for function package management [was Re: Unit root and structural breaks]

On Mon, 16 Jul 2012, Sven Schreiber wrote:

> Am 16.07.2012 12:31, schrieb JOSE FRANCISCO PERLES RIBES:
>> Hi:
>>
>> Has anyone written any code to perform the Lumsdaine and Papell
>> (1997) or Lee and Strazicich (2003) unit root test with two structural breaks?
>
> Sorry for hijacking this thread, but IMHO this shows that it would be
> useful to add categories (=tags/keywords) to the function package
> repository. In the future when hopefully there are gazillions of
> user-contributed packages it would be nice to be able to apply such a
> filter to the list to find what you are looking for more easily.

Yes, good idea.

On the original question I think it's safe to say there's no such code in the gfn repository at present -- although somebody, somewhere may have coded such tests for their own use.

Allin
_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users

_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users