Leopoldo,
When both equations are used, I believe that xtabond2 reports the number
of observations used by the level equation, which is higher than the
number of observations used by the FD equation.
Jean Salvati
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of
> Leopoldo Avellan
> Sent: Monday, January 10, 2005 9:39 AM
> To: [email protected]
> Subject: st: Difference GMM and System GMM with xtabond2
>
> Dear Statalisters:
>
> I am estimating a dynamic panel using the difference and the
> system GMM estimators. However, the results tell that there
> are 50 more observations for the System GMM than for the
> Difference GMM, why is this happening? I cannot find the
> mistake that makes the number of observations increase for
> the system estimator, this should not occur.
>
> The following is the syntax of my estimation Difference GMM:
> xtabond2 depvar l.depvar l(0/1).( strictlyexogvars endogvars) if
> year>1974, robust gmm(depvar endogvars, lag(2 9)) iv(strictlyexogvars,
> eq(diff) passthru) iv( D.strictlyexogvars, eq(level)
> passthru) noleveleq
>
> system GMM:
> xtabond2 depvar l.depvar l(0/1).( strictlyexogvars endogvars) if
> year>1974, robust gmm(depvar endogvars, lag(2 9)) iv(strictlyexogvars,
> eq(diff) passthru) iv( D.strictlyexogvars, eq(level) passthru)
>
> Notice that the only difference between the two, is that the
> former has the noleveleq option so that the additional
> moments for the computation of the system GMM are ignored in
> order to just compute the difference GMM estimator.
>
> What is going wrong? Could anybody help me please.
>
> Thanks,
>
> Leopoldo
>
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
