Introduction

In economics, a principal–agent problem arises when there is a divergence of interest between one person – the principal – and the agent he or she employs (or otherwise relies on) to perform some costly task. Law-and-economics scholars should pay as much attention to the principal–agent problem in criminal law as they do to the problem in corporate law – perhaps more. Although the corporate division of ownership and control creates a substantial agency problem, it seems to pose a less substantial problem than the one in criminal law. Among other reasons, corporate agents are hierarchically arranged so that shareholders need only monitor and motivate the top corporate officer, who can then monitor and motivate lower-ranked agents. By contrast, the public's agents in the criminal justice system – legislators, police, prosecutors, juries, chief executives, and prison officials – are not hierarchically arranged; they enjoy separate discretionary domains. Moreover, corporate principals can usually find reliable output measures for what they care about – share price – in daily newspapers, whereas criminal law output data are inherently more complicated, contestable, and remote. Finally, corporations are subject to market competition, which will eliminate organizations that do too poor a job of controlling agency costs. Criminal justice agents are rarely subject to market competition and, by comparison, political competition allows more agency slack.

Despite these points, the field of law and economics pays not more but less attention to agency costs in criminal law than in corporate law. Agency costs are famously central to the economic analysis of corporate governance and have been since Jensen and Meckling's classic 1976 article on the subject. Canonical statements of the field begin with agency costs and model corporate law as a mechanism for minimizing those costs. By contrast, principal–agent models play at best a peripheral role in the economics of criminal law. Here, the analogue to Jensen and Meckling is Gary Becker's 1968 article on the economics of criminal law; the analogue to agency costs is optimal punishment. Becker first showed how an ideal social planner might achieve optimality by minimizing the sum of crime costs and crime-control costs. Since 1968, law-and-economics scholars have stayed largely on the path of optimal punishment, extending Becker's model in countless directions. Not everything is about optimal punishment, and there is some literature specifically on agency problems among criminal law enforcers, particularly about prosecutors and also the problem of corruption. There are even a few, though just a few, papers on the more general agency problem in criminal enforcement. However, compared to optimal punishment, there is no sustained or comprehensive analysis of the agency problem and no conventional model for criminal system agents. For instance, there is a large economic literature on how best to structure the pay of corporate CEOs, but virtually no literature on how best to structure the pay of police or prosecutors.