Benchmarking with data envelopment analysis: An agency perspective

Abstract

Most studies on benchmarking in data envelopment analysis (DEA) focus on setting targets on the “best practice frontier”, few of them pay attention to motivating decision-making units (DMUs) to realize these benchmarks. Moreover, such “DEA-based best practice” may not be the actual “best practice” of evaluated DMUs. In this paper, subordinates’ strategy behaviors are first considered to realize the actual “best practice” during the benchmarking process, and agency theory is employed to reveal subordinates’ strategy behaviors. A novel incentive game with yardstick competition theory is established. Together with the ex-post targets (“DEA-based best practice”) and the actual production, we propose a reimbursement scheme to motivate DMUs to realize their “best practice”. In addition, we prove that DMUs’ best responses to our incentive game are just to realize their “best practice” when the reimbursement scheme satisfies strong monotonicity in outputs, and these responses constitute the strong Nash equilibrium of our incentive game. Finally, several DEA models are provided to hold the reimbursement scheme's strong monotonicity in outputs and minimize the compensations paid to all DMUs.