Abstract:
Despite several decades of research developments, the survey-based contingent
valuation (CV) method remains one of the most controversial topics in environmental
economics. As the procedure matures and more applied studies are completed, there
is increasing pressure to make the results policy relevant. This research is composed
of three distinct yet related manuscripts and investigates several important issues in
the application of CV for measuring the economic value of nonmarket goods.
Specifically, this dissertation explores the accuracy, applicability and transferability of
CV measures, with a focus on the dichotomous choice (DC) elicitation format.
The first study asks whether individuals directly questioned about their
willingness-to-pay (WTP) for a public good respond differently, in a setting of
apparent social pressure, than those provided with additional anonymity. This
question is explored in a quasi-experimental setting using a split sample design to
compare a randomized response questioning format with a direct questioning format.
Econometric models assess the robustness of explanatory effects across survey types.
Redacted for Privacy
The results support the hypothesis of compliance bias.
The second manuscript investigates an experimental fee hunt pheasant stocking
program that developed subsequent to a previous CV. This case study evaluates the
performance of the original study, and addresses the question of how CV may be used
in converting WTP into public revenues. The results are a mix of good news and bad
news; while WTP appears to be a real value, the findings indicate the difficulty of
conducting any external validation against a real market.
The third study explores several issues in the derivation of demand curves
from DC-CV models. Application of an available specification test provides a
technique for addressing endogeneity questions on measure-of-use variables. Absent
endogeneity, there is still the opportunity for considerable bias in demand curves. A
typology is presented and used to distinguish how CV models incorporate such
variables. The empirical results from applying the specification test to a previous
study provide initial confirmation of the proposed typology. Caution is urged in
reconstructing demand curves from "off-the-shelf" DC-CV models for benefits transfer
purposes.
