Sources and effects of utility-theoretic inconsistency in stated-preference surveys

Abstract

Economists increasingly are turning to multiple-response stated-preference (SP) methods (sometimes called conjoint analysis) to value environmental and natural-resource commodities (Gan and Luzar; Opaluch et al.; Roe, Boyle, and Teisl; Adamowicz, Louviere, and Williams; Adamowicz et al.). These methods evolved independently of contingent-valuation (CV) approaches in the area of market research for consumer goods and services (Cattin and Wittink; Louviere). More recently, health economists have recognized SP as an appropriate technique for quantifying health-care preferences (Johnson, Banzhaf, and Desvousges; Jahannesson, Jönsson, and Karlsson; Ryan and Hughes). SP analysis encompasses several multiresponse, multiattribute, preference-elicitation techniques that allow respondents to systematically evaluate trade-offs among multiple environmental attributes or among environmental and nonenvironmental attributes. In this article, we describe and evaluate tests of consistency with welfare-theoretic principles for two surveys. The first survey was designed to elicit primarily nonuse values for preserving endangered salmon stocks. The second survey elicited use values from diabetes patients for a new insulin. We assess whether the difference in commodities affect response patterns and evaluate the effects of including or excluding nonwelfare-theoretic observations in estimating consumer surplus.