The authors compare performance of the OLS, FGLS-SUR-IV (IV augmented regression with SUR that relies on feasible GLS), GMM and MEF approaches. The structural parameters for reduced form methods are estimated using minimum distance method. In order to solve the identification issue, the authors set δ = 0.05 equal to its true value for regression-based approaches and GMM. As can be seen from the table, FGLS-SUR-IV, GMM and MEF demonstrate very accurate estimates
of the true DGP parameters for γ,η and ρ but all of the have troubles correctly estimating κ. When taking the model to empirical data, the authors encountered some numerical optimization