The Value of Preserving Nature - Preference Uncertainty and Distributional Effects

This thesis deals with valuation of nonmarket goods using contingent valuation and consists of four papers and an introduction to the research area. Paper [I] examines the public benefits from preserving old-growth forest in the submountainous region in Sweden. Specifically, it analyzes a preservation program suggested by the Swedish Environmental Protection Agency. The results show that people value preservation for different reasons, not necessarily related to physical use. The paper finds that the estimated public benefits of the program exceed the estimated opportunity cost of forgone timber revenues and it should therefore be implemented. The paper also finds that there is no regional imbalance in the distribution of the benefits. Paper [II] examines the public benefits from preserving the four large predators in the Swedish fauna. Specifically, the paper focuses on the differences in attitudes and willingness to pay between people in wolf areas and other regions. We find that a clear majority of people in wolf areas are against preserving predators and that many of them need to be economically compensated in order to accept implementation of the predator policy package. The public in Sweden is, by a narrow margin, against implementation. The overall mean WTP is approximately SEK 300. It cannot be ruled out that the public benefits may be outweighed by the public costs following implementation. Paper [III] presents a new approach for treating preference uncertainty in contingent valuation. Specifically, it studies how data elicited from a multiple bounded question should be modelled. The new approach is compared to one of the conventional approaches and we find that: (1) it is more intuitive; (2) it better fits the data; (3) it gives more precise estimates of mean and median WTP; (4) it is less sensitive to distributional assumptions; and (5) it is better suited for policy analysis. Paper [IV] examines the income-effect in contingent valuation. Specifically three issues are analyzed: (1) the choice of income measure; (2) the choice of modelling assumptions; and (3) the social context. The results show that the estimated income-elasticity of WTP is fairly sensitive to different choices. The most statistically precise estimate is produced using household income and controlling for household characteristics. The third issue (social context) is approached by studying the answers to a WTP question conditioning respondents on a change in (1) their personal income and (2) the average income in Sweden. The results suggest that not only the income level per se influences WTP, but also its relation to the income of others.