Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: dynamic panel data, system GMM and fixed effect


From   "sistoand80" <sistoand80@libero.it>
To   "stata statta" <statalist@hsphsun2.harvard.edu>
Subject   st: dynamic panel data, system GMM and fixed effect
Date   Wed, 21 Sep 2005 10:55:33 +0200

Dear all,
I've a panel of 20 regions and 7 years (1998-2004)and I try to estimate a dynamic panel model as

yit = a*l.yit + b*(Xit) + uit.

My problem is that i've a heterogeneous panel. I tested for homogeneity constraints through a F-test but results showed that i could impose homogeneity on "a" and "b" only if fixed effect are included in the estimates. 

Standard approach for estimate DPD is to take the first difference to remove fixed effect (xtabond). However, this procedure imposes a high degree of homogeneity, so that with heterogeneous panel, this seems  to lead to inconsistent estimate as  a big portion of the variance of y is explained by fixed effect.

If a include fixed effect in xtabond, the command drop these dummy as it takes the first-difference. I've try to run a system of gmm (xtabond2) including fixed effect from the level equation.

However, I don't know if this procedure is correct.
Could anyone help me?
Is a system of GMM consistent in presence of fixed effect?

I've to run a dynamc fixed effect model (xtivreg)? 
Is a dynamicFE a consistent way to estimate such a dynamic panel (heterogeneity, N>T, etc.)? 

Best,
Andrea Sisto,
PHD student
University of Eastern Piedmont




*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index