Coordination in the Network Minimum Game
We study organizational design and its role in coordination failure using the network minimum game: a version of the minimum-effort game where dependencies between players are captured by a directed network. We show, theoretically and experimentally, that acyclic networks are most conducive to successful coordination. Indeed, introducing a single link to complete a cycle of dependencies may destroy coordination. Further, acyclic networks make coordination resilient: initial coordination failure is often overcome after repeated play in acyclic networks, but not in cyclic networks. Our findings provide a novel perspective on the near-ubiquity of acyclic (e.g., hierarchical) structures in organizations.
Additional material can be accessed here.
- Bandits in the Lab (John McMillan Prize for the Best Paper in Economics by a PhD Student)We test Keller, Rady, Cripps' (2005) game of strategic experimentation with exponential bandits in the laboratory. We find strong support for the prediction of free-riding because of strategic concerns. We also find strong evidence for behavior that is characteristic of their Markov perfect equilibria: non-cutoff behavior, lonely pioneers and frequent switches of action.
- Asset Ownership and the Hold-Up Problem We study a contract environment with an ex-ante investment stage and where ex-post bargaining takes place under one-sided asymmetric information. We offer a model where only the presence of an outside option allows for approximately ex-ante efficiency. Without an outside option, any static or sequential mechanism performs worse, which we view as a rationale for the role of ownership allocation in contracting environments with asymmetric information. We take these theoretical predictions to a laboratory setting and find that outside options as implemented through asset ownership are valuable, not because of efficient ex-ante investment but because they reduce ex-post frictions.