How wrong can you be? Implications of incorrect utility function specification for welfare measurement in choice experiments

Abstract

Despite the vital role oftheutilityfunctioninwelfaremeasurement,theimplicationsof workingwithincorrectutilityspecificationshavebeenlargelyneglectedinthechoice experimentsliterature.Thispaperaddressestheimportanceofspecificationwitha specialemphasisontheeffectsofmistakenassumptionsaboutthemarginalutilityof income.MonteCarloexperimentswereconductedusingdifferentfunctionalformsof utility togeneratesimulatedchoices.Multi-NomialLogitandMixedLogitmodelswere then estimatedonthesechoicesundercorrectandincorrectassumptionsaboutthe true,underlyingutilityfunction.Estimatedwillingnesstopaymeasuresfromthese choice modelingresultsarethencomparedwiththeequivalentmeasuresdirectly calculatedfromthetrueutilityspecifications.Resultsshowthatfortheparameter values andfunctionalformsconsidered,acontinuous-quadraticoradiscrete-linear attributespecificationisagoodoptionregardlessofthetrueeffectstheattributehason utility. Wealsofindthatmistakenassumptionsaboutpreferencesovercostsmagnify attributemis-specificationeffects