On Wed, Jul 27, 2011 at 12:32 AM, Eloy Arturo Fisher Hogan <eloyfisher(a)hotmail.com> wrote:
> I am pretty sure it is improper, yet I will ask nonetheless -- is it still > feasible to use tsls for the panel estimation? I am pretty sure you can't, > since it doesn't demean the data and the like, but I want to be sure since I > am more familiar with the test outputs reported there. You can use plain tsls with panel data only if the data are poolable, that is, there are no significant differences in the conditional means across the groups. If that condition is not satisfied, you can use the ivpanel() function. To get help on the script version, open gretl's function package listing (the "fx" button or under the File menu), right click on the ivpanel line, and select Info. Or in the console, do include ivpanel.gfn help ivpanel The "dpanel" command is another alternative, but that is for dynamic models. Allin Cottrell
