Semiparametric Estimation Of The Binary Choice Model For Contingent Valuation PDF Download

Semiparametric estimation of monotonic single index models is studied. In this class of models, the unknown components are a monotonic link function along with finite-dimensional parameters including the coefficient of the single index. The proposed method optimizes objective functionals with respect to both finite and infinite dimensional parameters. In the first chapter, proofs of consistency, rates of convergence, asymptotic normality and semiparametric efficiency are offered. The main result is applied to the semiparametric Least Squares (LS) estimation, semiparametric Least Absolute Deviation (LAD) estimation, and the semiparametric Maximum Likelihood (ML) estimation for various types of single index models. The second chapter focuses on an iteration-based method proposed by Wang and Zhou (1995) for the standard binary choice model. The algorithm is kernel free, very fast and easy-to-implement. The estimator is consistent and nearly efficient. These desirable large sample properties of the estimator, however, have not been rigorously proven so far. In this chapter, a set of sufficient conditions for consistency and asymptotic normality of the WZ estimator will be given. In the third chapter, the estimation methods developed in the previous two chapters are applied to dichotomous choice contingent valuation, which has been one of the most popular methods to estimate Willingness-To-Pay (WTP) for non-market goods, such as environmental resources. The proposed method is a two-step estimatior. In the first step, the underlying binary response model is estimated by the method studied in the previous chapters. In the second step, the distribution of the WTP is computed based on the first estimates. Consistency, asymptotic normality, and semiparametric efficiency of the estimator are studied.