On the Inference About Willingness to Pay Distribution Using Contingent Valuation Data

Abstract

Although contingent valuation (CV) is one of the main sources of estimates of non-market values of environmental goods, little guidance exists regarding parametric approaches for modelling CV data, which would reliably estimate willingness-to-pay (WTP) values based on binary choice, payment card or open-ended preference elicitation data, among others. CV studies often rely on relatively simple approaches to modeling stated preference responses. Lower-bound, non-parametric estimates seem to be preferred in legal cases, while studies that apply parametric approaches often select a specification among a limited set of commonly used distributions. To enhance the reliability of CV-based WTP estimates, we propose to adopt a more flexible approach to parametric modelling of a WTP distribution, by considering a wide range of parametric model specifications. We demonstrate the advantages of the proposed approach using databases from two large CV studies: the eutrophication reduction valuation for the Baltic Sea Action Plan and the Deepwater Horizon natural resource damage assessment. We find non-negligible differences in WTP value estimates across models with different assumed parametric distributions, and we observe the variation in the values to decrease when only better-fitting models are considered. This emphasizes the need for cautiously identifying the model best fitting to the data.

Keywords: contingent valuation, parametric modelling, stated preferences, willingness to pay, welfare estimates