Hello,
I am estimating a dynamic panel with xtabond. I suspect that one regressor
might be endogenous. In earlier postings, I have read that it is possible to
use a Hausman test to test for the endogeneity of regressors. However, I am
not quite sure about the syntax.
Assuming I have the semi-log model
logY = logY(t-1) X1 X2 TimeDummies
and I suspect X1 to be endogenous. I know that I can estimate the model
treating X1 first as endogenous (always consistent) and then a second model
treating X1 as exogenous (consistent only if X1 is exogenous) and do a
-hausman model1 model2-.
However, I am not sure about the syntax. With -xtabond-, would that look
like:
Model 1 (always consistent):
xi: xtabond logY X2 i.TimeDummies, lags(1) pre(X1, lags(1,.) end)
Model 2 (consistent only if X1 is exogenous):
xi: xtabond logY X1 l.X1 X2 i.TimeDummies, lags(1)
Is this the correct syntax? And with -xtabond2- would this be expressed as:
Model 1:
xtabond2 logY l.logY l(0/1).X1 X2 TimeDummies, gmm(l.logY, X1) ivstyle(X2
TimeDummies)
Model 2:
xtabond2 logY l.logY X1 l.X1 X2 TimeDummies, gmm(l.logY) ivstyle(X1 l.X1 X2
TimeDummies)
Thanks for your help,
Cordula
*
* 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/