The paper shows that several estimators for the panel probit model sug
gested in the literature belong to a common class of GMM estimators. T
hey are relatively easy to compute because they are based on condition
al moment restrictions involving univariate moments of the binary depe
ndent variable only. Applying nonparametric methods we discuss an esti
mator that is optimal in this class. A Monte Carlo study shows that a
particular variant of this estimator has good small sample properties
and that the efficiency loss compared to maximum likelihood is small.
An application to the product innovation decisions of German firms rev
eals the expected efficiency gains. (C) 1998 Elsevier Science S.A. All
rights reserved.