Performance-Based Incentives in a Dynamic PrincipalAgent Model