Hi
I would like to appy twostep system GMM with xtabond2 command. But I am not sure whether my model is true or not. I use the syntax below.

xtabond2 invest L.invest L2.invest L(0/2).(sales usercost cashflow), gmm(L(0/2).(invest sales usercost), collapse) iv(L(0/2) cashflow) small twostep orthogonal robust

Sales and user cost are endogenous variables here. Two lags of sales, usercost and casflow variables defined as independent variables to see the long run effect. In this case how should I decide gmmstyle subcommand. Should it also include two lags of all independent variables? Can gmmstyle command include only one lags of independent variabes? How should I decide that?

Thank you