Zhao (2008) presents an interesting “all-or-nothing monitoring” result for a multitask moral hazard agency problem with partial effort observation. We argue that the optimal contract based on the non-verifiable observation of the agent's effort in Zhao (2008) can be regarded as a limitation on the incentive schemes available to the principal. I then propose some arguably more appropriate approaches for analyzing such agency problems. (D82, D86, M54)