Intermediation, Information, and Diversity in Networks

Networks are ubiquitous in economic and social contexts. Individual decisions and welfare are affected by local connections, as well as by global network architecture. The research funded by this award investigates how the networks in which agents interact influence outcomes in a variety of market and group activities. There are three distinct projects. The first two investigate the impact of network structure on trade intermediation and information diffusion. The third project examines what network features are essential for the persistence of diversity in socially influenced choices.

The first project studies intermediation in networks. The theory suggests that the seller's profit depends not only on the number of middlemen along the path of trade, but also on the complete collection of competing paths of access to buyers offered by each middleman. The seller prefers to trade along paths that entail local competition among intermediaries with similar resale values. Trading paths do not necessarily minimize intermediation or maximize welfare. A decomposition of the network into layers of intermediation power characterizes equilibrium resale profits. Local competition is exploited within layers and hold-up problems arise in transactions across layers. This research has implications for combating organized crime and illegal trade, preventing corruption in hierarchical structures, and efficient production in supply chains.

The second project investigates how information (technological innovations, research ideas, news and gossip) is priced and diffused across the links of a network. In contrast to the research on intermediation, the information good is non-rival; any player who gets possession of the good may subsequently replicate and resell it to an arbitrary number of uninformed neighbors. The positions of the initial sources of information in the network affect the profits that actors accrue from acquiring and reselling the information and shape the diffusion process. Sellers compete for customers and establish individual spheres of influence in the network. The main objective is to prove that profits are confined within the blocks of a certain network partition. Links with endpoints within the same block constitute bottlenecks and play an important role in extracting profits. Links bridging across blocks are not critical in connecting the network and impose negative externalities on sellers.

The third project considers a setting where players situated in a network need to take actions or form opinions that conform with a majority of their neighbors. Unanimous coordination on any particular action constitutes an equilibrium. Many networks also feature equilibria with diverse action choices. The aim of this research is to characterize the networks for which diversity is robust to the introduction of stochastic mutations in the best response dynamic.

The tools and methods developed in this research will provide broader impact by allowing firms and policymakers to better understand how network relationships affect market outcomes.