Incentives in Networked Teams

Abstract

This paper studies incentive provision and optimal network design in team environments. We develop a two-stage moral hazard framework in which agents first invest effort in preparation and then decide on continuation effort after observing their peers' first-stage actions. Peer information is represented by an observability network, modeled as an undirected graph in which connected agents mutually observe each other's actions. A central insight of the paper is that the role of the network depends critically on whether agents' efforts are substitutes or complements. Under substitute technology, incentives depend only on the number of active agents, and network structure is irrelevant. Under complementary technology, by contrast, agents condition continuation effort on observed behavior, so that network topology directly shapes incentive provision. We then study the optimal design of the observability network. We show that a key determinant is how total compensation—comprising incentive rewards and fixed salaries—varies with agents' network position. A horizontal team structure is optimal when total compensation increases with degree at the margin, whereas a hierarchical structure is optimal when it decreases with degree.