Ambiguity in Dynamic Contracts

I study a continuous-time principal-agent model in which the principal is ambiguity averse about the agent's effort cost. The robust contract generates a seemingly excessive pay-performance sensitivity. The worst-case effort cost is high after good performance, but low after bad performance, which leads to overcompensation and undercompensation respectively and provides a new rationale for using performance-sensitive debt. I also characterize the agent's incentives when the contract is misspecified, i.e., when he is offered the robust contract, but his true effort cost is constant. I find that termination can induce shirking rather than incentivize effort. The strength of incentives is hump-shaped and agents close to firing prefer riskier projects, while those close to getting paid prefer safer ones. This feature resembles careers in organizations, most notably risk-shifting and the quiet life