The Heckman Correction for Sample Selection and Its Critique

TLDR

Such models occur frequently in empirical work, especially in microeconometrics when estimating wage equations or consumer expenditures. This paper gives a short overview of Monte Carlo studies on the usefulness of Heckman’s (1976, 1979) two‑step estimator for estimating selection models. The authors assess the estimator’s performance through simulation studies. Exploratory collinearity checks are strongly recommended; when collinearity is absent, full‑information maximum likelihood outperforms Heckman’s two‑step, but if collinearity is present, subsample OLS or the Two‑Part Model is most robust.

Abstract

This paper gives a short overview of Monte Carlo studies on the usefulness of Heckman’s (1976, 1979) two‐step estimator for estimating selection models. Such models occur frequently in empirical work, especially in microeconometrics when estimating wage equations or consumer expenditures. It is shown that exploratory work to check for collinearity problems is strongly recommended before deciding on which estimator to apply. In the absence of collinearity problems, the full‐information maximum likelihood estimator is preferable to the limited‐information two‐step method of Heckman, although the latter also gives reasonable results. If, however, collinearity problems prevail, subsample OLS (or the Two‐Part Model) is the most robust amongst the simple‐to‐calculate estimators.