Abstract
This paper studies the problem faced by a monopolistic data seller when users are connected through an information-sharing network. When users’ prior information is sufficiently noisy, the seller’s optimal strategy is to target a maximum independent set, that is, the largest subset of users with no direct links between them. In this regime, the optimal precision of the data decreases as the network becomes denser. Nevertheless, using the Caro–Wei bound – a classical result in graph theory – we show that for most networks, the seller’s chosen precision remains strictly above the socially efficient level. Further, every core–periphery network is Pareto-efficient, and every Pareto-efficient network exhibits a quasi-core-periphery structure. When users can coordinate on network formation, the resulting equilibrium network takes this form as well.

This publication has 30 references indexed in Scilit: