Three essays on contingent valuation

Contingent valuation (CV), a survey-based method, is widely used by researchers and government agencies to assess the value of those goods or services whose market price is not well defined. This dissertation comprises three essays analyzing and extending the theoretical foundations, estimation methods and empirical applications of the CV method. The first and second essay focuses on producers' willingness to pay for novel inputs or technologies. The first essay analyzes the theoretical underpinnings of producer WTP for new inputs. In addition to conceptualizing the producer WTP function, I derived its comparative statics and show how these properties can be used to recover quantity demanded or supplied and, in some cases, price elasticities. I also discuss the implications of these relationships to specify empirical WTP models and survey design. The WTP model is developed within the context of neoclassical theories of utility and profit maximization. Producers' WTP function for novel inputs or technologies is derived using individual indirect utility function in combination with the firm's profit function. Comparative statics results show that producers' WTP is a decreasing function of the upgraded input price, its initial quality level, and an increasing function of output price and final quality level.;In the second essay, CV methods using online and mail surveys are employed to estimate the economic value that registered producers place on the services received from an Electronic Trade Platform (i.e., MarketMaker). Estimation of the WTP model was carried out using parametric maximum likelihood estimation procedures. Results indicate that producers, on average, are willing to pay ;The third essay proposes alternative distribution-free methods for the estimation of WTP models using nonparametric conditional imputation and local regression procedures. The proposed estimators involve iterated procedures that combine nonparametric kernel density estimation of the errors of the WTP function with parametric linear or nonparametric kernel regression of its conditional mean function. In contrast to other distribution-free procedures (i.e., Turnbull approach), the proposed estimation methodology allows the inclusion of covariates in the modeling of WTP estimates, as well as the thorough recovery of its underlying probability distribution. Monte Carlo simulations are employed to compare the performance of the proposed estimators with that of the Turnbull estimator. Simulation results show that the proposed estimators perform substantially better than the Turnbull approach, and that conditional mean and marginal effect estimates of these models are analogous to the ones obtained using the benchmark correctly specified parametric model. The performance of the procedures is also evaluated using a real data set.