システム情報系　社会工学域 ファカルティ・セミナー　　　　　　　　　　　　　　　　　　　　　　　    

人文社会科学研究科 社会的選択理論・厚生経済学研究会 共催

日　時　：　2013年6月6日(木)　　16:30-18:00

場　所　：　第三学群F棟　3F1102

講演者　：　室岡　健志　[　UC Berkeley経済学部   ]

司会者　：　秋山　英三

演　題　：　Deception under Competitive Intermediation

                           ＜ abstract ＞

I investigate conditions under which competing intermediaries motivated by commissions—such as financial advisors, mortgage brokers, and salespeople—promote products with unanticipated costs. Two types of firms sell products through common-agent intermediaries: one with a transparent product, and another with a deceptive product that has hidden attributes such as additional fees, hidden qualities, or unexpected risks. Each firm sets a product price and a commission, each intermediary chooses which product to promote and whether to disclose the hidden attribute, and each consumer visits multiple intermediaries to buy at most one item. If consumers are aware of the hidden attributes, then the profits of intermediaries are competed away and social welfare is always maximized. If consumers are inattentive to the hidden attributes, however, profitable deception occurs if and only if the amount of the hidden attributes is large and the number of intermediaries each consumer visits is small. In this case, intermediaries earn positive profits despite competition because the deceptive firms give high commissions to each intermediary in order to maintain deception. Such deception enables deceptive firms to sell socially-inferior or even socially-wasteful products. Furthermore, under profitable deception intermediaries as experts hurt consumers: because the high commissions to the intermediaries increase the total prices of the products, consumer welfare becomes lower when the intermediaries can educate about the hidden attribute consumers than when they cannot. The presence of sophisticated consumers may not limit profitable deception at all because intermediaries can sort consumers through shrouding. A policy regulating commissions can lead intermediaries to detect and unshroud any hidden attributes.