On Mon, 21 May 2012, artur tarassow wrote: > I tried to access the LR-test test statistics and p-value for (G)ARCH terms > after estimating a GARCH like this: > > <hansl> > open b-g.gdt > dY = diff(Y) > garch 1 1 ; dY const --stdresid > <hansl> > > but I could not find any accessors. I guess we could use the $chisq accessor for that purpose. > I was just thinking whether it makes sense to activate the > 'modtest' option for this kind of models as well? Presumably only when using the --stdresid option, which would be a bit awkward. Easy enough to save the standardized residuals and run tests on them? Allin Cottrell _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users

On Thu, 24 May 2012, artur tarassow wrote: > definitely it is simple to run the test by hand. But I was just thinking > that in the case gretl already runs a test itself, why not to allow to > access the test results... ;-) You can actually get the result of the LR test for (G)ARCH terms using $["garch_LR"]. We haven't attempted to document this yet but any named data attached to the last model can be got using the bundle-type syntax, $["<key>"]. Right now you'd have to peek into the C sources to see what named things are attached to models of different kinds. Allin Cottrell _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users

find . -name "*.c" | xargs grep garch_LR