A Reasonable Benchmarking Frontier Using DEA: An Incentive Scheme for Managers

This paper constructs on principal-agent theory and its extensions to production theory to propose an incentive system for the agent. The incentive scheme suggested here rewards the agent (hospital manager) on the basis of the distance that separates observed output, the result of past decisions, and the potential output that would maximise principal’s welfare (health authority). But setting up such a scheme is challenging, since we have to estimate potential output and to do this we need to know the production function faced by each hospital, something we ignore. To overcome this difficulty we resort to Data Envelopment Analysis (DEA). This way of using DEA identifies a frontier that sets first-best effort levels; i.e., the level of effort attributed to those hospitals that are successful at delivering the greatest level of outputs for a given consumption of resources. A very delicate issue in this context is the selection of the hospitals that define the frontier. If the frontier contains hospitals whose performance cannot be matched (i.e., outliers), efficiency targets will be distorted for the remaining hospitals, and this may create disincentives rather than incentives. We propose an iterative procedure which allows us to identify a “reasonable frontier” that should set the objectives to be met by agents. The empirical part of the paper is based on data from Spanish public hospitals. From a theoretically point of view, this paper addresses the problem of performance assessment and incentives management in delegated decision making. From an applied point of view, our paper contributes to understand how opportunistic agents will strategically respond to productivity studies, and to know how desired responses can be motivated most efficiently.