st: xtabond2 subsamples dummies

[Sabina Kummer <sabina.noo@postmail.ch>]

Dear all,

I have a sample of 56 countries and I perfom xtabond2 (system GMM) in order to estimate GDP growth. As my sample of countries is heterogeneous, I would like to perform the same regressions for lower income countries and higher income countries. I would like to avoid to split the sample into 2 sub-samples in order to keep the same number of observations (and of degrees of freedom).
So, in order to get the results for higher income countries, I create a dummy (d=1) for this category of countries. Then, I use the same specification as for the same sample but I multiply each variable by the dummy. I would like to know whether this way of doing is correct.

Example - My Stata commands are the following:
whole sample: xtabond2 fdiratio l(1/1).fdiratio l(1/1).gdp i.year, gmm(fdiratio, laglimits(2 2)) gmm(fdiratio, laglimits(2 11) collapse) twostep robust small
high income countries: xtabond2 fdiratio l(1/1).(d*fdiratio) l(1/1).(d*gdp) i.year, gmm(fdiratio, laglimits(2 2)) gmm(fdiratio, laglimits(2 11) collapse) twostep robust small

Many thanks in advance for your help.

Kind regards.

Sabina



*
*   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/