Multitasking, two-part contracts, and bunching: an application to doctors' tasks

The optimal design of incentive contracts critically depends on whether the tasks performed by agents are complementary or substitutable, yet empirical evidence on this remains limited. This paper develops a novel empirical strategy to identify complementarities and substitutabilities in tasks, even in the absence of contract variation across agents, provided the incentive contract is piecewise linear. We apply this method to data on the management of chronic diseases by UK family physicians and find evidence that some tasks are complements, while none are substitutes. These complementarities may explain the widespread adoption of incentive contracts in healthcare. Furthermore, our findings suggest that healthcare systems centred around family physicians, rather than specialists, could achieve significant efficiency gains by consolidating complementary tasks under a single provider.