II. Semiparametric estimation: the maximum score Many semiparametric models impose assumptions of symmetry or mono tonicity, which can be used to derive consistent estimators. Manski's maxi mum score procedure is one of them. Consider a binary choice model with y = 1 (Note: not 0-1! using 1 is a convenient choice for what fol lows). We observe iid observations (yi,xi)n i=1. The maximum score estima tor chooses to maximize Sn() = 1 n n i=1 yi 1(xi 0