Rethinking the puzzle of escalating penalties for repeat offenders.

• Author: Dana, David A.

1. INTRODUCTION

   The legal system punishes repeat offenders more severely than nonrepeat offenders. Second-time offenders receive more severe punishment than first-time offenders; repeat offenders with many previous offenses receive more severe punishment than repeat offenders with a few previous offenses. Indeed, the general principle of escalating penalties based on offense history is so widely accepted that it strikes most people as simple common sense. The principle is embedded in formal federal, state, and administrative codes and in the enforcement norms of prosecutors and government officials at all levels of government.(1)

   Penalty escalation based on offense history, moreover, is not limited to any particular kinds of offenses. Escalation certainly characterizes the treatment of "traditional" criminal offenses, such as theft, assault, and murder.(2) But escalation also characterizes the treatment of regulatory or "public welfare" offenses, such as violations of environmental, worker safety, health, immigration, and labor regulations.(3) Thus, the Inspector General of the federal Environmental Protection Agency recently explained that

   [O]ne generally agreed upon enforcement concept is that of escalating enforcement actions for repeat violations. For instance, a violator may initially be required to comply with an administrative order or be assessed a relatively small monetary penalty. If these actions do not bring about compliance, the enforcement actions may be escalated to civil or criminal judicial actions and progressively higher monetary penalties.(4) For economists and law-and-economics scholars, however, the principle of escalating penalties based on offense history is puzzling. In the standard economic model, the purpose of penalties is to deter conduct that creates greater social costs than benefits; where the system of penalties is calibrated to produce "optimal deterrence," offenses that produce net social costs will be deterred while offenses that produce net social benefits will go undeterred. Within this paradigm, the key factor in assessing the optimal penalty for a given offense is the social harm that will result from the offense. But the social harm from a given offense would seem to have nothing to do with the offense history of the offender; the illegal discharge of waste into the ocean causes as much social harm when the discharging company is a first-time offender as when it has a long history of such offenses. Thus, standard economic analysis would seem to suggest that, contrary to actual practice, penalties should not escalate based on offense history.

   A central claim of this Article is that the phenomenon of escalating penalties based on offense history presents an even greater puzzle from an economic perspective than has been previously recognized. The economic model of optimal deterrence actually supports declining penalties based on offense history for some categories of offenses, rather than nonescalating or escalating penalties. Contrary to the assumptions of the existing literature, the probability that a person or corporation will be detected committing a violation correlates with offense history: In other words, holding all other variables constant, people and entities with "records" have a higher probability of having their offenses detected than people and entities without records. Moreover, people and entities with records assume higher risks of audits for previously undetected past offenses when they commit new offenses. Because previous violators face higher probabilities of detection, conventional economic theory dictates lower penalties(5) for previous offenders than for those with clean records. The case for declining penalties based on offense history, however, is not uniformly strong. The case is strongest in the context of regulatory offenses, such as violations of environmental, safety, and labor regulations, and weakest in the context of the core traditional criminal offenses, such as assault and murder.(6)

   The remainder of the Article :is a three-part exploration of the question of whether, by broadening the analytic focus beyond conventional economics, penalty escalation might be explained on the basis of offense history. First, the Article explores whether behavioral economics and cognitive psychology can help reconcile the theory of optimal deterrence with the principle of escalating penalties based on offense history. The opposite turns out to be true. A central contribution of the behavioral economics literature has been the identification of certain heuristic biases that help to explain why people act in ways that seem to deviate from the predictions of economic models that assume complete rationality in decisionmaking. The literature regarding salience bias--the bias toward exclusive or over-consideration of personal, recent, vivid experience--suggests that individuals who have recently been detected and punished for a violation should face reduced penalties because their recent detection and punishment experience will bolster their fears of future detection and punishment. The behavioral economics literature also suggests that individuals will be overoptimistic about their chances of avoiding detection of wrongdoing until they are actually caught. This analysis suggests, once again, that, in order to achieve optimal deterrence, our legal system should provide for less severe penalties for repeat offenders than for first-time offenders instead of, as it currently does, reserving the least severe penalties for the latter group.

   Second, the Article explores the possible significance of social, extralegal sanctions in explaining the phenomenon of escalating legal sanctions based on offense history. The punishment suffered by an individual or entity for any offense is the sum of both social and legal sanctions. For a variety of reasons, social sanctions probably decrease with offense history: The marginal reputational loss from a first conviction for securities fraud is much greater than the marginal reputational loss from a fourth conviction. To the extent this is true, and to the extent that the decline in social sanctions based on offense history merely counteracts the escalation in legal sanctions based on offense history, there may, in fact, be no escalation in aggregate sanctions based on offense history. Escalation in legal sanctions thus may be consistent with economic theory and hence not puzzling because it is part of a regime of nonescalating aggregate sanctions.

   The Article also explores whether penalty escalation might be justified as a means for the government to express moral condemnation of legally prohibited conduct. Penalty escalation, particularly the escalation from noncriminal to criminal sanctions, probably helps to create and reinforce popular perceptions that legally prohibited conduct is not merely legally wrong but also morally wrong and should not be engaged in for that very reason (as opposed to reasons based on perceived probabilities of magnitude and certainty of punishment). These perceptions of moral wrongfulness matter, because people sometimes will make compliance decisions based on moral considerations and because it is far more costly for the government to achieve compliance based on fear of punishment (as that necessitates investments in enforcement resources such as police, inspectors, courts, etc.). In other words, penalty escalation, through its expressive power to shape popular perceptions of moral wrongfulness, allows society to conserve its resources for investments other than law enforcement.

   The answer to the question whether economics can explain penalty escalation, then, is quite complex. Two previously ignored factors--escalating probabilities of detection for repeat offenders and salience and optimism biases--suggest that an optimal penalty regime should employ declining, rather than escalating, penalties, at least in some contexts. These factors indicate that, for example, a landowner who repeatedly violates prohibitions against filling wetlands should receive a lower sanction than a first-time violator. At the same time, two other previously ignored factors--the interplay of legal and social sanctions and the power of escalation to encourage compliance by expressing moral condemnation--support penalty escalation notwithstanding the presence of escalating probabilities of detection and cognitive biases. This Article thus does not definitively solve the puzzle of escalating penalties for repeat offenders, but it does enrich an understanding of its complexity and point to new avenues for theoretical and empirical research.

   Part II presents the standard account of optimal deterrence and explains why, given escalating probabilities of detection, optimal deterrence would seem to dictate declining penalties for repeat offenders. Parts III and IV explore the relevance of cognitive biases and extralegal sanctions, respectively. Finally, Part V develops the argument that escalating penalties serve an expressive function as a means of labeling certain conduct as not merely legally, but also morally, wrong.

2. DECLINING PENALTIES FOR REPEAT OFFENDERS AS "OPTIMAL" DETERRENTS: THE SIGNIFICANCE OF ESCALATING PROBABILITIES OF DETECTION

   This Part first explains the concept of optimal deterrence and its central role in the economics literature regarding public enforcement and sanctions. The Part then explains why, on average, probabilities of detection(7) increase with an individual or entity's number of previously detected offenses. This fact, which is not addressed in the current literature,(8) calls into question previous efforts to explain penalty escalation based on offense history as a practice consistent with the optimal-deterrence model employed in conventional economics. As explained below, escalation in probabilities of detection is likely to be most significant in the context of regulatory...