Panel data: r(3000) with ivreg2 struct ms_vcvorthog undefined

[Grace Mzungu]

Hello,
I have a panel data of macro variables.

" xtset CountryCode year
panel variable: CountryCode (strongly balanced)
time variable: year, 1990 to 2012
delta: 1 year"

When I run xtreg, my regression works fine.

I would like to instrument one of my variable which is endogenous with a few lagged values of the variables that I used in the main regression with ivreg2 and I get this error struct ms_vcvorthog undefined.

I looked up in previous posts and tried this:

"mata: mata mlib query
.mlib libraries to be searched are
lmatabase;lmataado;lmataopt;lmatasem"

Do you know what is the problem? I am not yet familiar with xtabond2 and I am not sure that command would fix my problem.
Many thanks,
Grace

[Mark Schaffer]

It looks like the Mata library livreg2.mlib hasn't been installed properly. This should install automatically along with the other ivreg2 files. Try uninstalling ivreg2 (however you installed it) and then reinstalling from within Stata using

ssc install ivreg2, replace

[nihel]

hi I WORK ON A SUBJECT treating a model of dynamic panel (balanced) and I wonder if the use of commands: xtabond ..... year, vce (robust) xtdpdsys ....... year, vce (robust) estat Abond, artests (4) is sufficient to have cohérants and meaningful results? thank you

[Grace Mzungu]

Thanks Mark. It worked.
Nihel, I am not sure I am understanding your question. I am new to xtabond and xtabond2 actually but I believe dynamic panel means you include lagged values and you use that in xtabond.
What are your other possible instrumental variables you can use?

[nihel]

When lagged dependent variables are included as explanatory variables, both the ‘within groups’ and ‘random effects’ estimators are biased and inconsistent, except when the number of time periods is large.To solve this problem, the dependent variable lags (y_i,t-1) are used as instruments for independent variable.

[nihel]

In my paper I use Yit as a dependent variable and Yit-1 as an explanatory variables; that's why I'm using the GMM (dynamic panel)
when searching through the help of STATA I found this link:
http://www.stata.com/stata10/dpd.html

the problem is that each time the dependent variable shown in the output STATA
I do not know how to set the outputs to my paper and how to analyze them later?

thanks a lot Grace Mzungu for your Answer
nihel halouani

[Steve Samuels]

Please don't post a question like this in the forum, titled "Using the Forum", which is designed for questions about how to use the forum, not how to use Stata. Post in the General forum instead