I am using MLE to estimate my IVProbit, which involves estimating a simultaneous system of equations. However, I am aware of the fact that, traditionally if we want to obtain statistics such as Weak Instrument Test or Underidentification Test, we would run 'ivreg2', which relies on the assumption that we run the model in two stages, especially OLS in the first stage. So, I am not sure whether I can still use the postestimation statistics obtained after running 'ivreg2' even if my MLE IVProbit model does not even use the two-step procedure.

Many thanks!