Comparing Fuzzy and Probabilistic Approaches to Preference Uncertainty in Non-Market Valuation

Abstract

In seeking to value environmental amenities and public goods, individuals often have trouble trading off the (vague) good or amenity against a monetary measure. Valuation in these circumstances can best be described as fuzzy in terms of the amenity being valued, perceptions of property rights, and the numbers chosen to reflect values. In this paper, we apply fuzzy logic to contingent valuation, employing a fuzzy clustering approach for incorporating preference uncertainty obtained from a follow-up certainty confidence question. We develop a fuzzy random utility maximization (FRUM) framework where the perceived utility of each individual is fuzzy in the sense that an individual’s utility belongs to each cluster to some degree. The model is then applied to a Swedish survey that elicited residents’ willingness to pay for enhanced forest conservation and to a Canadian survey of agricultural landowners that elicited their willingness to accept compensation for a tree planting program. Both the WTP and WTA measures we obtain using the fuzzy approach are well below those obtained using standard probability methods. Based on goodness of fit measures and Monte Carlo experimentation, a case can be made for using a fuzzy preference approach for modeling preference uncertainty as opposed to incorporating respondent uncertainty within the random utility maximization framework.