xtabond2 command query

[Muhammad Ibrahim Shah]

I have 20 years of data from 2000-2019, so I created year dummies already. I want to estimate two-step system GMM for the following model

y=dependent variable, x1, x2,x3, x4 are independent variables. I suspect that x1 and x2 are endogeneous. So I want to use 2nd to fourth lags of L.y, x1 and x2 as instruments in gmm-style and current values of x3,x4 and year are to be used as iv-style instruments. Is my following model correct:

xtabond2 y l.y x1 x2 x3 x4 y1-y20, gmm(L.y x1 x2 , lag(2 4) collapse) iv(y1-y20 x3 x4 ) robust twostep orthogonal

Is it correct? gmm(L.y x1 x2, is it using first differences as instruments here as gmm type instruments?

[Jared Greathouse]

Please reread the FAQ on how to ask questions.

Also: In this business, very rarely is a model "correct"- aside from correctly specified via syntax, the appropriateness of your model depends heavily on your research question, your data, and other contextual requirements that we can't tell based on "My outcome is this from this year to that year, is this piece of syntax right", we need data and minimal worked examples to be able to effectively help you. This is why we ask for this in the FAQ

[Sebastian Kripfganz]

The following presentation might be of interest:

• Kripfganz, S. (2019). Generalized method of moments estimation of linear dynamic panel data models. Proceedings of the 2019 London Stata Conference.

[Muhammad Ibrahim Shah]

Thank you both.