Re: [Gretl-users] 2SLS
On Sat, 14 Mar 2009, Giovanni Madaro wrote:
I need another information. I'm struggling to make a two least
squares estimation from my panel data. Can you please give me
the text i need to write in the command window to tell gretl
that i want to estimate a system of two equations with
dependent, endogenous and instrument variables??
Please try the Help entry for tsls, it's quite explicit. The
command syntax is
tsls <depvar> <regressors> ; <instruments>
Exogenous regressors should be given in both <regressors>
and <instruments>.
I have one main equation:
RIT = VEGA + ASSET + TRAD + CIL + CONC + TIME DUMMIES
in this equation RIT and VEGA are endogenous, so i want the
second equation to be
VEGA = RIT + ASSET + TRAD + CIL + CONC + TIME DUMMIES
That's not going to work. Your instruments are ASSET, TRAD, CIL,
CONC and the TIME DUMMIES. If VEGA is endogenous you need at
least one more instrument.
Allin Cottrell