Re: [Gretl-users] VAR roots plot
john w schrieb:
You are wrigt Jack.
At the end, it's better to work on completely new features rather than
spending time on cosmetics. Let it stay.
don't give up so quickly ;-)
> >
> > In VECM module the terminology is quite clear, but in VAR module a
> little
> > bit cunfuses me.
> > My suggestion for VAR module is:
> > 1. To include one more check option for including a trend (like for a
> > constant),
>
> The way the source code is now, that would be a little inconvenient,
> unless we're willing to add a "--trend" option to rhe "var"
command. I
> wouldn't be against it, but adding a trend to a VAR is also very easy
> now,
> and such an option would be a tad gratuitous. Hence, I'm not either
> strongly pro nor against such a change. What do other list members think?
I would (mildly) favor consistency for the interface of VARs and VECMs, but it's not
terribly important.
>
> > 2. To replace "Deterministic variables" with "Exogenous
variables".
>
> Any adjective you use in this context is rather arbitrary. The word
> "exogenous" may mean at least 4 different things in the econometrics
> literature, so my very personal preference is for using it as
> sparingly as
> possible, unless the context makes what I mean very clear. But others may
> disagree.
In this context "deterministic" seems wrong, because you can add any variable
you like. "Exogenous"
may not be precise, but at least it's correct. So I think john w is right and it
should be changed
(that way it's also consistent with the vecm formulation).
cheers,
sven