A comparison of semiparametric estimators for the ordered response model

Three semi-parametric estimators for ordered response models are compared: a “semi-nonparametric” series estimator, an estimator based on a polynomial transformation to normality, and a least squares estimator for a transformed dependent variable. The empirical estimates are found to differ markedly across estimators in a job satisfaction illustration presented, but much less so in the second illustration, a model of willingness to pay for environmental protection. Monte Carlo evidence comparing the three estimators is also presented.