Consumers are willing to pay significant price premiums for goods with environmental attributes but these preferences often depend on an aspect of the good that need not relate to any fungible consumptive qualities of the good. Unlike product quality attributes (eg. appearance, flavor and durability) that are generally revealed either pre- or post-purchase, some environmental attributes (eg. sustainable, recycled, non-toxic, biodegradable, and cruelty-free) cannot be perceived by the consumer – goods with this feature are known as credence goods and they are pervasive in green markets, insurance markets, home repair services, medical markets, software services and taxi rides in unfamiliar cities. This project will construct a laboratory experiment to test how sensitive green credence good purchasing is to variations in the noise with which market purity is measured. In the experiment producers may choose to sell “brown” goods and “green” goods, but may also fraudulently label that their brown products are green. Buyers prefer greener products, but differ in their willingness to pay for them. However, green-ness is a credence good that buyers are unable to ascertain on their own at any point in the experiment - even after consumption. In each of 15 periods we create a market in which the market price of green goods (including any fraudulently labeled green goods) is endogenously determined using a double oral auction. After each period of production and purchase decisions, buyers are given information on the market purity for that round and they are told how purity is calculated (ie, they are told whether it is measured with or without noise). Market purity is calculated as a ratio of actual green goods to all goods sold as green in the market in that period (thus the denominator includes all fraudulently labeled green goods as well as all green goods). While a buyer is never told whether the goods he purchased are truly green or brown, he is informed about the purity of the entire market for the previous period by the experimenter. Our dependent variables of interest are production, labeling decisions (if the subject is a producer), purchase decisions (if the subject is a buyer) and beliefs about market purity. Our three treatments vary in how many green goods are sampled to calculate market purity, i.e. the treatments vary how noisy the purity measure reported to subjects is.
Team Members: Dr. Erin L. Krupka, School of Information; Dr. Thomas Lyon, Stephen M. Ross School of Business; Dr. Arnab Mitra, Portland State University; Student TBA, School of Natural Resources and Environment.