Deadlines and infrequent monitoring in the dynamic provision of public goods

We consider a dynamic game of private provision of a discrete public good. In our model, a group of agents contributes to a project over time, which is completed once the cumulative contributions reach a threshold. Provided that this occurs prior to a prespecified deadline, each agent receives a lump-sum payoff. We show that a shorter deadline can induce the agents to raise their efforts, but no matter the length of the deadline, effort provision is inefficient due to the agents' frontloading incentives. Only if the agents do not monitor progress until the deadline are their frontloading incentives eliminated, so by committing to a deadline equal to the first-best completion time, it is possible to restore efficiency. Recognizing that deadlines are not renegotiation proof, we show that by committing to monitor progress to date at the first-best completion time, and then again at a sufficiently later date, efficiency can be attained. In this case, that monitoring date acts as a self-enforcing deadline.