I want to check either the cross-sectional dependency between units is homogeneous or heterogeneous. This is to decide whether to use the standard GMM estimator or a consistent GMM estimator under heterogeneous error cross section dependence. Your response is highly demeaned. Sarafidis et al. (2008) argued that the standard GMM with time dummies is consistent if the there is homogeneous cross-sectional dependency. Hence, I badly want to know the Stata command for their test(Sargan's difference tests based on DIF and SYS).