Re: [Gretl-users] Tobit Regression for Panel Data in Gretl
Am 03.05.2018 um 14:33 schrieb 3J LEMA:
Dear members,
How to perform Tobit Regression with Panel Data in Gretl?
The Dependent Variable of my balanced panel data is bounded between 0
and 1. Six years across 15 companies.
Your comment is highly appreciated.
Hi, I'm not aware of a function (built-in or packaged) to do that in gretl.
I presume you are asking about random-effects tobit. Because quoting
from Stata's documentation: "Unconditional fixed-effects tobit models
may be fit with the tobit command with indicator variables for the
panels; [...] . However, unconditional fixed-effects estimates are biased."
Running this (biased) estimator is also easy in gretl: With your panel
dataset, do "genr unitdum" and you get group dummies du_1 ...du_x, where
x is the number of groups. You can collect them in a list like this:
"list groupDs = du_*". Then add this list to your regressors when
running "tobit". But again, this is probably not what you had in mind.
cheers,
sven