Conference Agenda

Benefit Transfer (BT) has evolved as the dominant non-market valuation method for large-scale environmental benefit-cost analyses, including those required of U.S. federal agencies. Yet, even best-practice approaches for BT based on Meta-Regression Models (MRMs) typically exhibit poor predictive fit and out-of-sample efficiency. This article introduces Random Forests (RFs) for nonparametric estimation of MRMs and construction of BT predictions. We compare the performance of a variety of RF models to current best practice approaches for BT, including a globally-linear MRM and Locally-Weighted MRM (LWR). We find that forest-based models substantially improve the within-sample accuracy of welfare predictions and tighten confidence intervals of predicted benefits for out-of-sample transfers. The best-performers reside within the family of Local Linear Forests (LLFs), essentially a hybrid approach that combines elements of RFs and LWR. We also examine the utility-theoretic properties of each specification. Results suggest that this new approach has the potential to substantially improve BT accuracy for environmental policymaking without sacrificing theoretic properties, while simultaneously reducing econometric and computational difficulties relative to leading alternatives.