Three essays in decision making

Abstract

This dissertation is composed of three essays about consumer judgment and decision making. In Essay 1, I develop a novel explanation for the well-known endowment effect - the tendency for owners to value goods more than non-owners do. According to a prominent explanation for this effect, the prospect of losing possessions creates psychological pain, inducing sellers to demand more than buyers will pay. My alternative account is based on transaction disutility: consumers are reluctant to trade on terms that are disadvantageous with respect to perceived market prices. The endowment effect appears to be caused by inflated selling prices because market prices usually exceed the value of ownership to consumers. But I show that reducing reference prices relieves sellers' transaction disutility, shrinking or eliminating the effect. Moreover, very low reference prices create disutility among buyers, resulting in disparities driven by a reluctance to buy, not to sell. Essay 2 explores the implications of transaction disutility for consumer preferences. Maximum buying and minimum selling prices are commonly believed to reveal preferences: a consumer who prefers one good over another presumably has a higher reservation price for it. But transaction disutility can distort reservation prices away from underlying values. If alternative measures of preference - such as binary choices between goods - are not regarded by consumers as transactions, they are not subject to such distortions. This difference can create preference reversals, that is, incoherence between explicit choices and the preferences implied by stated reservation prices. I find strong experimental evidence for this proposition. The "Bayesian Truth Serum" (BTS) is a survey scoring method designed to provide truth telling incentives for respondents answering multiple choice questions about intrinsically private matters: opinions, tastes, past behavior.

(cont.) My final essay discusses several tests of BTS. In one questionnaire, respondents indicated their familiarity with various items (e.g. electronics brands), one-third of which were nonexistent foils. BTS did in fact reward truth telling: the scoring method severely penalized "recognition" of foils. Also, survey takers viewed the BTS method as credible: people who were paid for achieving higher BTS scores claimed to recognize fewer foils, even when facing competing incentives to deceive.