Abstract
This paper investigates the origins of the collinearity problems encountered in the two-step estimation method for sample selection models. The analysis reveals several critical misconceptions and deficiencies in the literature. Remedies to the collinearity problems are proposed and evaluated. Monte Carlo experiments as well as an empirical example based on Mroz's (1987) data are used to illustrate the practical relevance and importance of the analysis. © 2000 Kluwer Academic Publishers.