ABSTRACT

We develop a tractable econometric model in which agents receive rewards when achieving measurable goals across a variety of tasks. Different tasks can interact with others, and the principal (or social planner) would design suboptimal contracts if multitasking was not accounted for. Preferences and costs are identified from variation in contracts and in exposure to different tasks. We apply this framework to study England’s “Quality and Outcomes Framework” during the 2009-2019 period, under which thousands of groups of primary care physicians received payments rewarding different clinical out-comes among their patients. We estimate the distribution of physicians’ preferences and the degree to which different clinical outcomes are not independent, and then use
our model to quantify the impact of the program and to investigate the design of optimal incentives. Our results imply that the program led to a significant improvement in health as measured by increases in quality-adjusted-life-years, net of additional medical costs. We find that re-optimizing the design of the program’s incentives would increase-net-of-payments health benefits by an additional 3% from baseline.