Dynamic Nonmonetary Incentives *

We consider a dynamic principal-agent environment in which investments and rewards (compensation opportunities) arrive stochastically over time, and are either taken immediately or foregone. The agent privately observes whether an action is currently available, but he needs the principal’s consent to take it. We show that there exists a unique optimal mechanism and analyze its qualitative properties. As the agent’s promised compensation increases, the principal gradually becomes more selective about incentivizing investments and less selective about the rewards that she allows. Interestingly, the unique optimal use of each reward turns out to be via “time allowances”; that is, the principal allows the agent to enjoy all rewards that arrive before a certain point in time.