Benefit Transfer: A Hierarchical Bayesian Approach

Benefit transfer is a collection of methods widely used in costbenefit analysis and policy analysis to assess environmental values in contexts where original valuation work is deemed too expensive and/or too time-consuming. The essence of benefit transfer is conditional or unconditional prediction of environmental values in a new context. This paper proposes a hierarchical Bayesian method for benefit transfer. The model specifies one prior for the parameters governing the distribution of values in a given context. These parameters are in turn given a second prior (or hyperprior) capturing the distribution of these parameters across different contexts. The Bayesian framework provides the posterior distribution of all the model parameters. The marginal posterior distribution for the hyperparameters can be utilized to construct a posterior predictive distribution of environmental values across new contexts. This distribution summarizes the uncertain prediction of values in new contexts and is the proper tool for use in benefit transfer. The proposed method for benefit transfer is illustrated with an application to improved water quality.