This project investigates the question of why local governments cooperate with one another for service provision and coordinated policies. It proposes that the selection of interlocal cooperation among local leaders in the Unites States can be best understood as a diffusion process by which local elites learn from the cooperative experiments of neighboring jurisdictions and reproduce them in order to realize similar gains when it makes sense to do so. This process, I argue, is driven by the mechanisms of learning, development of networks of trust, and interlocal competition. The project presents theory, methods, and results in three manuscripts. The first uses a newly constructed longitudinal dataset of financial transfers by local governments to show that localities are more likely to cooperate when larger shares of their neighbors were cooperating in the past. This process is amplified in regions with more intense interlocal competition. The second manuscript demonstrates that the diffusion of cooperation is most intense within particular types of local service provision, namely those that involve capital-intensive and system-maintenance functions of government, such as highways, sewers, and water delivery. Finally, the third paper presents results from an original, national survey of mayors and councilors that involved embedded experiments to tease out the hypothesized mechanisms of diffusion. Findings provide strong support for the role of development of trust and learning in the spread of interlocal cooperation.