Referendum contingent valuation estimates: sensitivity to the assignment of offered values

Abstract

Contingent valuation methods (CVM) are becoming increasingly popular for assessing the value of nonmarket resources and public goods. In particular, CVM willingness to pay estimates are gaining currency for the assessment of damages in environmental litigation. Several studies have compared the value estimates resulting from alternative formats used for CVM survey questions and have speculated on the reasons for observed discrepancies. These reasons now include a whole taxonomy of possible biases. We take a closer look at one CVM format-the referendum-and demonstrate that simply the luck of the draw in assigning the referendum thresholds on individual questionnaires can produce a surprisingly wide variety of value estimates. We control for the behavioral biases that confound other comparison studies by using one sample of payment card CV data and simulating 200 samples of consistent referendum responses. Due to the inefficiency of the referendum format, we conclude that, where referendum questions have produced different value estimates than other formats, elaborate explanations for the apparent discrepancies may not be necessary.