Efficiency vs Bias of Willingness-to-Pay Estimates: Bivariate and Interval-Data Models

Dichotomous choice contingent valuation (CV) surveys with follow-ups attempt to obtain more precise information about the respondents′ willingness to pay for environmental quality while preserving the desirable properties of the referendum CV method. This paper compares the associated interval-data and the recently proposed bivariate models of willingness to pay (WTP). Monte Carlo simulations show that the interval-data estimates of mean or median WTP can be surprisingly robust to departures from the true, bivariate model, and that the interval-data model is often superior to the bivariate model in terms of the mean square error of the estimates. Two empirical examples that confirm and disprove this finding, respectively, are reported.