Estimation of Willingness to Pay Using Dichotomous Choice Data under Proportional Hazard Approach

Parametric and semi-nonparametric estimation techniques for the dichotomous choice contingent valuation model are receiving considerable interest. These methodologies originally have been developed to find the "willingness to pay (WTP)" or implicit valuation of non-market goods. The form of WTP distribution can be used to set up tariff of new service. Furthermore, it can be used for marketing strategy if we know the relations between WTP and covariate - salary, expenditure for telecommunication services, age, etc. The methodologies for finding WTP distribution can be classified three categories. The most popular approach is parametric approach. The second approach is fully nonparametric one. The most popular model is Turnbull's estimator and kernel method. The final approach is semi-nonparametric (SNP) approach which is the most preferable in some sense that take both advantage of parametric and nonparametric approach. In this paper, we apply a semi-parametric estimation procedure called proportional hazard regression model for binary discrete response data. We perform a series of simulation result under various model assumptions. We also give empirical studies of real situation for finding consumer's valuation of prescribed services.