Benefit Transfer from Multiple Contingent Experiments: A Flexible Two-Step Model Combining Individual Choice Data with Community Characteristics

This study proposes a new approach to utilize information from existing choice experiments to predict policy outcomes for a transfer setting. Recognizing the difficulties from pooling raw data from experiments with different designs and sub-populations we first re-estimate all underlying Random Utility Models individually, and then combine them in a second stage process to form a weighted mixture density for the generation of policy-relevant welfare estimates. Using data from recent choice experiments on farmland preservation we illustrate that our strategy is more robust to transfer inaccuracies than single-site approaches. The specification of "intelligent" mixture weights will be a fruitful ground for future research in the area of Benefit Transfer.