xtabond2 doubt

[Sudhanshu Gupta]

I am working on effect of leverage on firm performance. I have 1220 firms data for 17 years, i am using xtabond2 two-step first difference-GMM Arellano and Bond methodology. Profitability is dependent variable, I am using lag of dependent variable as explanatory variable, leverage is endogenous variable, other control variables I am using salegrowth, tangibility, size and selling and distribution expenses. As leverage is endogenous variable so i am thing of using lag of leverage. so please tell me that is my following equation is correct or not:

xtabond2 profitability L.profitability L.Leverage size salegrowth tangibility S&Dexp yr2000-yr2016, gmmstyle(L.profitability L.Leverage) ivstyle(size salegrowth tangibility S&Dexp yr2000-yr2016) noleveleq robust twostep h(2)

please suggest me alternative or more appropriate equation if it there

[Kaleem Khan]

Dear Gupta

I need to ask how did you determine that Leverage is endogenous "only"?
Another thing is that using industry dummy as instrumental variable is out of my understanding also.

Equation seems correct. test statistics of AR1, AR2, Sargan/Jensen etc can tell you validity of equations and instruments more clearly. I would be very much interested to know the test results of this equation.

[Sudhanshu Gupta]

Dear Khan,
I have not used industry dummy as instrumental variable, I dont know how you get this impression.
Secondly, if you see the capital structure theories and empirical studies than you will find that they used profitability as independent variable to find the capital structure. I having simultaneity problem. I used leverage as endogenous variable.

[Toussaint Houeninvo]

Dear Sudhanshu

I am using stata 2013 which did not have xtabond2 command.It has xtabond but not xtabond2. Did you downloaded it from internet? If it is the case could you please advise on how did you proceed?

[Sudhanshu Gupta]

Dear Toussaint can you please tell me which stata version you are using. there is no stata 2013 because all stata after 9 version have xtabond2.
Please go to statistics--> longitudinal/ Panel data ----> Dynamic Panel data -----> Arellano-Bover/Blundell-Bond estimation

[Toussaint Houeninvo]

Dear all
I finally found how to download Xtabond2
. Thanks

[Toussaint Houeninvo]

Dear Sudhanshu
Thanks for your advice. You are right when you say there is no stata 2013. In fact I am using stata 13 with GMM system. I regressed Log of GDP per capita (LGDPPC) at constant price on the lag of the LGGDPPC(LGDPPC.L1) Log of Inglation (LINF) and log of Government expenditure (LGEXP). The command for second order autocorrelation test is "estat abond" in stata13 . When I tried to perform it I got a message saying "varlist not allowed" r (101) . Therefore I was not able to performe AR (2) test. Gratful if anyone can help to solve it . Thanks