Semiparametric Estimation of the Binary Choice Model for Contingent Valuation
This paper is concerned with the estimation of the binary choice model without imposing any parametric structure on the distribution of the stochastic term. We adapt the distribution-free maximum likelihood method developed by Cosslett (1983) for deriving the contingent value function with respect to observable exogenous variables. We present Monte Carlo comparisons with the probit estimates and discuss the asymptotic consistency and relative efficiency of the approach. Data from a forest environment valuation survey are used for empirical estimations.