When Less Liability Leads to More Care: Adverse Effects of Liability Regimes in Multitask Principal Agent Settings

The problem of multitasking concerns the dilemma of designing incentives for an agent among an array of tasks when some tasks are more difficult to observe than others. So far the multitasking approach, as introduced by Holmstrom and Milgrom, has only been used in the context of incentive schemes based on positive rewards resulting in the use of lower-powered incentives in comparison with single-task settings. The main argument of the paper posits that these results can also be applied to incentive schemes based on negative rewards such as a liability system. The underlying idea claims that the additional utility from an increase in the first task due to a positive deterrence effect is lower than the disutility from a decrease of the second less observable or unverifiable task due to rising opportunity cost for the agent. The idea is discussed in the context of the optimal design of medical malpractice regimes.