作者: Patrick Puhani
DOI:
关键词: Estimator 、 Collinearity 、 Step method 、 Statistics 、 Mathematics 、 Heckman correction 、 Sample selection 、 Monte carlo studies 、 Selection (genetic algorithm) 、 Econometrics 、 Maximum likelihood
摘要: not available in German This paper gives a short overview of Monte Carlo studies on the usefulness Heckman's (1976, 1979) two-step estimator for estimating selection model. It shows that exploratory work to check collinearity problems is strongly recommended before deciding which apply. In absence problems, full-information maximum likelihood preferable limited-information two- step method Heckman, although latter also reasonable results. If, however, prevail, subsample OLS (or Two- Part Model) most robust amongst simple-to-calculate estimators. Journal Economic Surveys