Distribution Free Consistent Estimation of Mean WTP in Dichotomous Choice Contingent Valuation

Abstract

This paper develops a new estimation model for consistent estimation of mean WTP (willingness to pay) in dichotomous choice CV (contingent valuation). Consistent estimation of mean WTP in parametric models requires the correct specification of the probability distribution of WTP. However, it is not possible to foresee the probability distribution, and thus in practical applications there is always a danger of specification error, making it impossible to guarantee that mean WTP will be consistently estimated. This paper applies linear projection to develop an estimation model that enables consistent estimation of mean WTP without specification of the probability distribution. This linear projection model enables us to estimate mean WTP only by calculating a weighted mean of the indicator function defined as 1 (or 0) when a respondent says “yes” (or “no”) for a certain improvement in environmental quality. Moreover, when bids follow the uniform distribution, calculation is possible by multiplying “maximum bid” and “the rate of answering yes”. Therefore, a distinguishing feature of this model is that estimation can be accomplished very easily.