Is Hypothetical Bias a Universal Phenomenon?: A Multinational Investigation

A concern with the contingent valuation method (CVM) is the finding that hypothetical and real statements of value often differ. We test whether hypothetical bias, broadly defined, is independent of location by comparing real and hypothetical votes on a dichotomous choice referendum in China, France, Indiana, Kansas, and Niger. We find significant differences in hypothetical bias across locations and reject the hypothesis that hypothetical bias is independent of location. As opposed to the typical finding reported in the literature, subjects in Niger significantly understated their willingness-topay in the hypothetical referendum. (JEL Q51) The contingent valuation method (CVM) is a popular tool to value changes in environmental amenities and is increasingly being used to value new technologies and private good attributes. The CVM, and other similar survey-based methods such as conjoint analysis, are often the only feasible modeling alternatives if the analyst is interested in obtaining the total value (i.e., both the use and non-use value) of a good or service. Recently, there has been a lively debate about whether and to what extent hypothetical bias permeates benefit estimation in contingent markets and whether the phenomenon complicates obtaining reliable measures of value (see, e.g., List et al. 2004). Simply defined, hypothetical bias is the difference between hypothetical and actual statements of value. A host of possible causes for the bias have been proposed, including strategic survey response (e.g., Carson, Groves, and Machina 2007), uncertainty about one’s value for the good (e.g., Champ and Bishop 2001; Johannesson et al. 1999), commitment costs or option values (e.g., Norwood, Lusk, and Boyer 2008), and social desirability bias (List et al. 2004; Lusk and Norwood 2008); however, there seems to be little consensus on the key mechanisms driving a wedge between real and hypothetical statements of value. While the finding of such bias is widespread (e.g., see List and Gallet 2001), there is a notable commonality amongst such studies—most, if not all, have been conducted in the United States or in western Europe. Whether hypothetical bias is identical across culturally distinct consumer groups remains an open empirical question. Indeed, we know of no such evidence in the literature. While hypothetical bias may appear to be an innocuous issue, it has implications for international benefit-transfer and consumer valuation. Contingent valuation is commonly used to carry out benefit-cost analysis, a technique increasingly used around the globe to improve the quality of government decisionmaking. Indeed, in the United States, every ‘‘economically significant’’ ruleThe authors are, respectively, assistant professor, Department of Agricultural and Applied Edonomics, University of Wyoming, professor and Willard Sparks Endowed Chair, Department of Agricultural Economics, Oklahoma State University, and professor, Department of Economics, University of Chicago. The authors wish to thank the following individuals for helpful comments on previous versions of the paper and/or providing assistance in conducting the experiments: an anonymous referee, Glenn Harrison, Myrdene Anderson, Tomas Nilsson, Cole Ehmke, Tony Torres, Glynn Tonsor, Sean Fox, Virginia Moxley, Carmel White, Min Wang, Jia Lv, Zuhui Huang, Tahirou Abdoulaye, Germain Ibro, Abdoulaye Mohammed, Moustpha Moussa, Bokar Moussa, Stephané Robin, and Bernard Ruffieux. This work is made possible through funding from the United States Department of Agriculture, Purdue Research Foundation, Kappa Kappa Gamma, and the American Association of University Women. Land Economics N August 2008 N 84 (3): 489–500 ISSN 0023-7639; E-ISSN 1543-8325 E 2008 by the Board of Regents of the University of Wisconsin System 1 ‘‘Contingent valuation’’ refers to a set of surveybased approaches for eliciting Hicksian compensating or equivalent surplus values for a hypothetical change in a