[Limdep Nlogit List] Hausman-Taylor estimator for panel data

[William Greene]

Bill.  If you have no time invariant observed data, then you can just use
2SLS or a random effects version, 2SRE.  H&T is written for models that
contain time invariant effects (observed and unobserved).
Regards
Bill Greene

On Tue, Jan 14, 2020 at 4:01 PM William Spitz <bills at gra-inc.com> wrote:

> Hi. I'm trying to the Hausman-Taylor estimator. There are four sets of
> variables in the model:
> x1 = time-varying uncorrelated with u[i]
> x2 = time-varying correlated with u[i]
> f1 = time-invariant uncorrelated with u[i]
> f2 = time-invariant correlated with u[i]
>
> The manual claims that only the first of these required. For my purposes
> I have only x1 and x2 defined; I have no time-invariant variables, but I
> still want to use H-T estimator to account for the x2 variables that (I
> suspect) are highly correlated with the group effects.
> But when I issue the final command to get the H-T estimates, I get an
> error message stating:
> No time invariant vars uncorrelated with u(i)
> No time invariant vars correlated with u(i)
>
> How can I get H-T estimates without time-invariant variables?
>
> --
> William Spitz
>
> _______________________________________________
> Limdep site list
> Limdep at mailman.sydney.edu.au
> http://limdep.itls.usyd.edu.au
>
>

-- 
William Greene
Department of Economics, emeritus
Stern School of Business, New York University
44 West 4 St.
New York, NY, 10012
URL: https://protect-au.mimecast.com/s/M_x5C1WZXriQjVA9TLYNG1?domain=people.stern.nyu.edu
Email: wgreene at stern.nyu.edu
Ph. +1.646.596.3296
Editor in Chief: Journal of Productivity Analysis
Editor in Chief: Foundations and Trends in Econometrics
Associate Editor: Economics Letters
Associate Editor: Journal of Business and Economic Statistics