BUYERS, SELLERS, AND MIDDLEMEN: VARIATIONS ON SEARCH-THEORETIC THEMES
International Economic Review
We study bilateral exchange, both direct trade and indirect trade that happens through chains of intermediaries or middlemen. We develop a model of this activity and present applications. This illustrates how, and how many, intermediaries get involved, and how the terms of trade are determined. We show how bargaining with one intermediary depends on upcoming negotiations with downstream intermediaries, leading to holdup problems. We discuss the roles of buyers and sellers in bilateral
... and how to interpret prices. We develop a particular bargaining solution and relate it to other solutions. In addition to contrasting our framework with other models of middlemen, we discuss the connection to di¤erent branches of search theory. We also illustrate how bubbles can emerge in intermediation. Although we think there is more to be done, several subsequent studies have attempted to rectify the situation by analyzing the roles of middlemen, and how they a¤ect the quality of matches, the time required to conduct transactions, the variety of goods on the market, bid-ask spreads, and other phenomena. Rubinstein and Wolinsky (1987) themselves focus on search frictions, and for them, middlemen are agents who have an advantage over the original suppliers in the rate at which they meet buyers (a di¤erent but related recent analysis can be found in Ennis 2009). Focusing instead on information frictions, Biglaiser (1993) and Li (1998 Li ( ,1999 ) present models where middlemen are agents with expertise that allows them to distinguish high-from low-quality goods, and show how the presence of informed intermediaries helps to ameliorate lemons problems. ,b), middlemen hold inventories of either more, or more types of, commodities that help buyers obtain their preferred goods more easily. See In general, middlemen may either hold inventories, or act as market makers whose role is to get traders together, without buying and selling themselves ((2010a,b), de...ned by an endeavor to explicitly model the exchange process, and institutions that facilitate this process, like money, intermediaries, etc., often using search theory. We say more later about the relationship between intermediation and money; for now, we mention that early search-based models of monetary exchange like Kiyotaki and Wright (1989) not only make predictions about which objects might emerge as media of exchange, as a function of the objects' properties and agents' beliefs, they also make perhaps-less-well-known predictions about which agents emerge as middlemen.