Managing automation in teams

In this paper, we study a principal's decision to introduce automation into a production process governed by a team of employees. When introduced, automation displaces an employee with a machine. This displacement increases efficiency as the machine carries out the tasks of the employee at a lower cost, and reduces the scope of moral hazard as the machine does not make unobserved effort choices. We show that, despite the direct benefits, a principal may prefer not to adopt automation due to its indirect costs. Before automation is introduced, the principal is able to take advantage of her ability to shape the interactions between the team members to manage the agency problem. Automation eliminates this ability and removes an incentive device at the principal's discretion, resulting in an indirect cost. On the one hand, adopting automation is always optimal when the principal incentivizes employees independently, abstaining from creating a team interaction. On the other hand, automation may be suboptimal when the principal incentivizes employees by encouraging them to compete via a “relative performance evaluation” contract or to cooperate via a “joint performance evaluation” contract. We offer two extensions to test the robustness of these findings qualitatively. First, the findings carry through if we consider alternative effects of automation, where it impacts employees symmetrically without displacing any employee. Second, the findings also remain consistent when there are synergies between the efforts of team members.