Book excerpt: This paper concerns moral hazard problems in multi-agent situations where cooperation is an issue. Each agent chooses his own effort, which improves stochastically the outcome of his own job. Agents also choose the amount of 'help' to extend to other agents, which improves the performance of other agents. By selecting appropriate compensation schemes, the principal can design the task structure of the firm: The principal may prefer an unambiguous division labor, where each agent is inclined not to help other agents and specializes in his own job. Or the principal may prefer teamwork where each agent is motivated to help other agents. The analysis identifies two important determinants in choosing the optimal task structure; the effect of 'interpersonal interaction' and the attitude of the agents in providing 'small' amounts of help.