Re: [Gretl-devel] adf and restricted samples
Patricio Cuarón schrieb:
Between the two /smpl/ I changed the sample via GUI.
About the rest, I understand what you're saying. Should it use the
pre-sample periods?
I don't think there's a canonical answer to that. It would not be wrong,
just a matter of convention.
(If the underlying assumption about pre-sample values is correct, it may
be worth checking if there is an issue with the samples used for
lag-length information criteria. For comparability the same sample
should be used. I'm sure in general that's what gretl does, but maybe
there are some corner cases like existing pre-sample values in the
datafile, but not enough to cover the max lag, or somesuch. But this is
just a wild paranoid thought, I remember nothing like that. And I don't
have time right now to test, despite the impression I'm giving with all
these emails...)
cheers,
sven