N: 2500
T: 3

For my purposes:
- Endogenous: x1, x2, x3
- Predetermined: x4
- Exogenous: x5 x6 x7 x8 x9 x10

Code:
xi: xtabond2 DV L.DV x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 i.waves, /*
*/ gmm(L.DV) gmm(L.x1) gmm(L.(x2 x3), collapse) gmm(x4) /*
*/ iv(x5 x6 x7 x8 x9 x10, eq(level)) iv(i.waves, eq(level)) robust two small
My questions are as follows:
1. Is this specification correct? I know that 3-waves produce lots of problems, but I am not so sure if I specify my code correctly.
2. I know that the coefficients should be between OLS and LSDV, but when I specify my model in this way, one of my coefficients become 1.50 instead of the 0.15-0.30 boundaries. What could be the reason for this?
3. Instead of System GMM, what can I do to supplement my FE analyses in terms of sensitivity analysis to mitigate reverse causality?
4. I also tried to replicate this code with xtdpdgmm but I couldn't be able to construct it. Can you help me?