As a sensitivity analysis, I would like to replicate my FE and AB estimations with xtdpdml, or ML-SEM.

Is there a way to test whether the regressors are predetermined or exogeneity, or to put it more truly, is it possible to test the assumptions of exogeneity in xtdpdml? Incremental Hansen tests in xtabond2, for instance, gives a somewhat traceable path here.

Thank you.