Dear Statalist users,
In the Two-step System GMM estimation using xtabond2, if all explanatory variables (x1, x2, .., xn) except for the time dummy variables (i.year) take 1-period lagged values, is it correct to treat all explanatory variables as predetermined variables and use them as gmm instruments together with the 1-period lagged value (L.y) of the dependent variable (y) as shown below?
xtabond2 y L.y L.x1 L.x2 ... L.xn i.year, gmm(L.y L.x1 L.x2 ... L.xn, gmm(2 .) collapse) iv(i.year) twostep collapse small
In the Two-step System GMM estimation using xtabond2, if all explanatory variables (x1, x2, .., xn) except for the time dummy variables (i.year) take 1-period lagged values, is it correct to treat all explanatory variables as predetermined variables and use them as gmm instruments together with the 1-period lagged value (L.y) of the dependent variable (y) as shown below?
xtabond2 y L.y L.x1 L.x2 ... L.xn i.year, gmm(L.y L.x1 L.x2 ... L.xn, gmm(2 .) collapse) iv(i.year) twostep collapse small
