Hi All,

I want to run the regression for my research using xtabond2 command. I have read Roodman (2009). since xtabond2 is very sensitive and there are high possibilities of manipulation when running the regression, I just want to make sure that I am applying the regression in the correct way.

I have unbalanced panel data
Code:
xtset code year
where code 1 to 85 and year 2010-2018

the dependent variable is roa
the explanatory variables are as follows:
bsiz div edu bus mtg size growth opexp lev, which identified as endogenous variables based on the literature
age wgi inf gdp, which identified as strictly exogenous based on the literature.
I used i.country and i.year to control for country and year effect

I've run the following command on Stata

Code:
xtabond2 roa l.roa bsiz div edu bus mtg size growth opexp lev age wgi inf gdp i.country i.year, ///
gmm(l.roa bsiz div edu bus mtg size growth opexp lev, lag (1 2) collapse) iv(age wgi inf gdp i.country i.year) robust
sometimes Hansen test result does not support the validity of instruments, so I adjust the instruments by excluding some of the country dummies until they become valid instruments, as follows:

Code:
xtabond2 roa l.roa div edu bus mtg size growth opexp lev age wgi ing gdp i.country i.year, ///
gmm(l.roa div edu bus mtg size growth opexp lev, lag (1 2) collapse) iv(age wgi ing gdp i(1 2 3 4 5 6 7 8 9 10 11 12 13).country i.year) robust
country take the values 1-26


Am I on the right track?

Any help is highly appreciated