I need to estimate a non-linear equation using intrumental variables for
some of the covariates involved. Does anyone have suggestions on how
to combine the nl proceedure in STATA with 2sls to ensure that the
non-linear term is orthogonal to the error term?