Experiments on Lotteries for Shrouded and Bundled Goods: Investigating The Economics of Fukubukuro

Fukubukuro (or lucky bag) is a familiar institution in Japan and many other countries used by retailers for disposing of unwanted stock during the New Year sales. Two features of the institution are important: First, in fukubukuro, stores bundle goods of related items into sealed bags rather than selling items separately. Secondly, while general information about the contents is provided, details of brands and specifications are concealed creating a lottery for the purchaser. Motivated by the fukubukuro example and the lack of evidence on risk attitudes in lotteries involving goods, we conduct a laboratory experiment and follow-up survey to investigate preferences for lotteries in which the outcomes are bundled or unbundled goods. In general, we find that risk has a negative effect on subjects’ WTP for a product lottery. Nevertheless, a minority of subjects are risk-seeking and value the lottery more highly than the highest valued individual product. Conversely, we do not find much evidence of an uncertainty effect. Although subjects’ WTP responses to bundled product lotteries are less heterogeneous than their responses to single product lotteries, there is no significant advantage of selling bundled product lotteries over single product lotteries in relation to subjects’ risk preferences. We follow up the experiment with a hypothetical choice questionnaire in which we confront subjects with three options for a variety of goods: a certain product, its substitute, and a product lottery. We find that subjects who are riskseeking or have less product knowledge and familiarity are more likely to choose a product lottery. Furthermore, subjects are more likely to choose a product lottery when the choice task consists of complex products rather than simple products. We speculate that risk seeking and less-informed subjects may find a lottery between products to be a direct and simple way to solve their buying decision tasks.