NEOCLASSICAL: Stata module to estimate neoclassical education transitions model

neoclassical estimates a bivariate probit model with selection with coefficients re-scaled to facilitate cross-equation testing for equality. It is termed neoclassical for it is a variant of the neoclassical education transitions framework proposed in Lucas, Fucella, and Berends (2011). Note that the first transition appears in the second equation in the output. Yet, in specifying the model the temporally earlier transition is always regarded as the first transition. The estimates are the re-scaled coefficients and the appropriate standard errors. Any subsequent statistical tests will use this second set of results. A related command, retroclassical, estimates a neoclassical education transitions model while fixing the cross-equation error variance at zero. This provides the result one would obtain by calibrating the coefficients across transitions, yet not accounting for selection. Thus, it is a mid-way point between a neoclassical approach and the classical approach.