Dear Statalist,

I have estimated the 3SLS model below.

reg3 (eq1: y1 = x1 x2 x3 x4 x5 dum1 dum2 dum3 i.year) (eq2: x1 = x3 x4 a1 a2 a3) (eq3: x2 = x1 x3 x4 a1 a2 b1 b2)

I have read several articles and I have found that the best technique for panel data is system-dynamic GMM in the case of large cross-sections (N) and small-time dimensions (29 countries and 8 years of unbalanced data).

Does anyone know how to turn this model into a system-dynamic GMM using the xtabond2 command?

thank you in advance,

Regards,

Gamze,