On Thu, Dec 10, 2009 at 2:44 PM, Edoardo Di Porto
<[email protected]> wrote:
> Dear Statalist,
>
> it is possible to develop spatial econometric models (as SAR, SEM, SAR
> fixed effect etc.) using STATA...?
>
> thanks
>
> Edoardo
Yes, it is possible. I have done all my spatial econometrics work
using Stata combined with Mata. I have developed some Stata routines
(not made publicly available yet) to facilitate the estimation of the
SAR model (a model with spatial lag autocorrelation) by spatial 2SLS,
ML, GMM; the estimation of the SEM model (a model with spatial error
autocorrelation) by ML, and the estimation of the SAC model (a model
with spatial lag and spatial error autocorrelation) by generalized
spatial 2SLS (GS2SLS). Those routines are being tested by some
students in my department and will be made publicly available
afterward. As Austin pointed out, David Drukker has written two
estimation commands, -gs2sls- and -sarml-, to estimate the SAC model
by GS2SLS and ML respectively. However, as far as I know, these tools
are not yet available to Stata users. There is also the -spatreg-
command written by Pisati to estimate the SAR and SEM models by ML
(type: findit spatreg). Since -spatreg- was written before the Mata
era, it has the limitation of the Stata's matrix commands such as the
-matsize- limit.
Edoardo, you can contact me off-list if you want to give my routines a try.
Regards,
--
P. Wilner Jeanty
Post-Doctoral Researcher
Dept. of Agricultural, Environmental, and Development Economics
The Ohio State University
2120 Fyffe Road
Columbus, Ohio 43210
(614) 292-6382 (Office)
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
