[Gretl-users] teste Hansen-Johansen

Sven Schreiber svetosch at gmx.net
Sun Nov 29 06:34:44 EST 2015


Am 28.11.2015 um 17:38 schrieb Augusta Pelinski:
> Fiz o  coint2rec.gfn
> Podes me explicar PQ ESTA VINDO ASSIM O RESULTADOS (EM ANEXO)?.
> mas nao ta dando certo.
>   Não deveria vir assim neh? o que estou fazendo errado?
>

Hi Augusta,
as Oleh said, please use English, and it doesn't matter if there are 
language mistakes.

I'm not completely sure whether "fazer o teste de cointegração com 
quebra" means you want to *allow* for a break in the cointegration 
relations, or if you want to check whether there is such a break.

Our function package coint2rec.gfn basically just does the latter, 
checking the stability of the cointegration relations. As far as I 
remember it does not perform an extended cointegration test where breaks 
would be allowed.

With respect to the error message "system roots not well behaved", I 
would have to check the code to remember the details, but that could 
mean that you have explosive roots in the system at some stage, which 
would violate the assumptions of the stability test. This could be 
either bad luck, or it could mean that the fundamental specification of 
the system in terms of variable choice and necessary data 
transformations has to be revisited.

good luck,
sven



More information about the Gretl-users mailing list