Studying the reach of deterrence: Can deterrence theory help explain police misconduct?

This article reports the first perceptual deterrence study of a sample of police officers. The study investigated the influence of traditional deterrence considerations, extralegal sanctions, and impulsivity on the intention to commit several hypothesized acts of police misconduct. The results were largely consistent with perceptual deterrence findings from samples of college students, experienced offenders, and corporate managers. In particular, this study found that both legal and extralegal sanction threats potentially deter police misconduct. Further, it found that impulsivity diminished the deterrent influence of both sanction forms. The study also found that some of the effects of the explanatory variables depended on whether officers had prior punishment experience. The article discusses the implications of its findings for combating police misconduct and for deterrence research generally.     

The challenge of cyber attack deterrence

National security planners have begun to look beyond reactive, tactical cyber defense to proactive, strategic cyber defense, which may include international military deterrence. The incredible power of nuclear weapons gave birth to deterrence, a military strategy in which the purpose of armies shifted from winning wars to preventing them. Although cyber attacks per se do not compare to a nuclear explosion, they do pose a serious and increasing threat to international security. Real-world examples suggest that cyber warfare will play a lead role in future international conflicts. This article examines the two deterrence strategies available to nation-states (denial and punishment) and their three basic requirements (capability, communication, and credibility) in the light of cyber warfare. It also explores whether the two most challenging aspects of cyber attacks – attribution and asymmetry – will make cyber attack deterrence an impossible task.     

Studying deterrence with active residential burglars

Research designed to test deterrence/rational choice theory has contributed much to the knowledge of sanction threats, and the extent to which they are used in individuals' decision-making processes. The accumulation of knowledge about such perceptions among active and incarcerated offenders has surfaced only in recent years. We identified and interviewed active residential burglars to examine how they take into consideration the perceptions of risks and rewards before committing a burglary. Employing a series of estimation techniques which have not yet been used to study this research question, we find that offenders are influenced by the perceptions of both risk and rewards, though the latter are a stronger predictor of the decision to engage in a residential burglary. Theoretical implications and directions for future research are discussed.     

Error minimization and deterrence in agency control

In the presented oversight model, in which a regulatory agency may collude with regulatees, a watchdog organization may scrutinize the agency’s decision-making and find evidence speaking for collusive behavior. Found evidence is of a specific, stochastic quality. Courts will overturn the administrative decision when the evidence presented in court exceeds a minimum quality standard set by the political principal. Lowering the quality standard increases the odds of finding evidence of sufficient quality and, hence, leads to increasing collusion deterrence and to a lower probability of acquitting collusive administrators (type I error), but also to a higher probability of convicting an innocent administrator (type II error). It is shown that, when welfare-maximization gives rise to an interior solution, the welfare-maximizing standard of evidence is lower than the one that merely minimizes the costs of legal errors without taking deterrence costs into account, but will imply incomplete deterrence. However, conditions can and will be identified under which both error cost minimization and complete deterrence coincide with welfare-maximization.     

Police patrol and the deterrence of crime

A nonrecursive model of the deterrent effect of police presence was formulated and tested for 26 cities. Victimization data were employed as measures of crime, unpublished FBI data on the number of police patrol units as the measure of levels of police presence, and data for 11 exogenous variables were derived from FBI and census reports. It was found that per capita, police are a positive function of rates of violent crime and that clearance rates are a positive function of police presence. However, some offenses that have traditionally been thought to be deterrable are not inversely related to clearance rates. The equations were re-estimated using official measures of crime and found to be more consistent with the deterrent hypothesis. It is suggested that official data may generate a spurious correlation. Finally, caution is suggested in consideration of alternative forms of police presence.     

Plea bargaining and deterrence: An institutional approach

Previous economic analyses of plea bargaining have largely ignored its impact on the deterrence of crime. Instead, they have focused on the bargaining between a defendant and a prosecutor once a crime has been committed. This article remedies this deficiency by asking how the practice of plea bargaining influences the determination of criminal punishment and thereby the supply of crime by rational offenders. The key question examined is, how do the ex post objectives of prosecutors affect the ability of legislatures to implement criminal punishments aimed at achieving optimal deterrence? Various prosecutorial objectives are considered in answering this question.     

Correlates of deterrence: Problems of theory and method

A central methodological problem in ascertaining the correlates of deterrence is that there is no unambiguous method by which deterred persons can be identified. Assuming that the identification of a legal threat by persons who indicated that they have not committed a crime (the use of marijuana) is evidence of deterrence, this study differentiates deterred from undeterred persons based on a number of demographic variables. The results show that persons who were deterred from marijuana use by legal threats were indistinguishable on the basis of age, sex, race, marital status, and occupation from those who refrained from using marijuana for other reasons.     

Perceptional foundations of deterrence: The case of prostitution

The authors conducted a survey to determine community perceptions about the certainty, celerity, and severity of punishment with regards to the crime of prostitution. A representative random sample of Mecklenburg County, which includes Charlotte, North Carolina, was taken and 850 individuals responded. The results revealed that (1) citizens perceive sanctions against prostitution violations as being applied only infrequently; (2) most respondents perceive the time from violation to arrest as being relatively short but slower as the offenders proceed through the system, and (3) the severity of punishment is not perceived as being harsh. The authors argue that the low likelihood of apprehension, both actual and perceived, plus the mild sanctions may partially account for prostitution’s continued survival.     

Statistical measurement of the deterrence effect: A regression framework

When the dependent variable is a ratio, and its numerator also appears as the denominator of an explanatory variable, doubts exist with respect to the bias of the estimated effect. Although the problem arises in many contexts, it is posed here in the context of estimating the deterrence effect of punishment on crime. This is the context in which Logan (1978) suggests the method of “part-correlation.” By expressing this method in a regression framework, the results of the analysis of specification error are drawn upon to simplify Logan's procedure, proper measures are derived, and results are generalized. Finally, the effects of measurement errors are introduced and the bias examined.     

Saturated enforcement: The efficacy of deterrence and drunk driving

Deterrence theory generally finds support among studies focusing on changes in apprehension certainty, in punishment celerity, or in punishment severity. However, scientific evidence establishing a clear deterrent effect of criminal justice institutions on criminal behavior remains problematic partly because of measurement, design and/or experimental control difficulties. This study attempts to overcome some of these methodological difficulties by measuring deterrence in a drunk driving enforcement program in a small Wisconsin city. It demonstrates a statistically significant relationship between a carefully designed enforcement program and driver response measured by accident rates. This finding is contrary to some recent studies challenging the efficacy of enforcement efforts aimed at traffic accident reductions. A general discussion of deterrence and enforcement amplifies the findings and conclusions. A word of caution is added regarding the findings.     

Further findings on the deterrence question and strategies for future research

Findings for recent data (circa 1970) are consistent with previous purported tests of the deterrence doctrine. For each of seven types of crime there is an inverse relation among the states of the U.S. between the rate and the certainty of imprisonment, but there is no significant relation between the rate and severity of imprisonment (length of sentence served). Although the findings justify reconsideration of the deterrence doctrine, the utility of further tests is questionable without first restating the doctrine as a systematic theory, one that incorporates (inter alia) perceptual variables pertaining to punishment (e.g. perceived certainty of punishment). However, several empirical question should be contemplated prior to stating a deterrence theory; and conventional data (e.g. published statistics on imprisonment) can be used to answer some of those questions.     

Pulling levers focused deterrence strategies and the prevention of gun homicide

A number of jurisdictions have been experimenting with new problem-oriented policing frameworks to understand and respond to gun violence among gang-involved offenders. These interventions are based on the “pulling levers” deterrence strategy that focuses criminal justice and social service attention on a small number of chronically offending gang members responsible for the bulk of urban gun violence problems. Unfortunately, there is relatively little rigorous evaluation evidence on the effectiveness of these approaches to violence prevention. In Stockton, California, an interagency task force implemented a pulling levers strategy to prevent gun homicide among gang-involved offenders. A U.S. Department of Justice-sponsored impact evaluation suggests that the pulling levers strategy was associated with a statistically significant decrease in the monthly number of gun homicide incidents in Stockton. A comparative analysis of gun homicide trends in Stockton relative to other midsize California cities also supports a unique program effect associated with the pulling levers intervention.     

Situational deterrence and claim padding: results from a randomized field experiment

While it is generally argued that threats of legal sanctions are more effective with offenders involved in economic crimes than with delinquents committing more conventional crimes, such an affirmation rests on weak empirical evidence. Also, most studies supporting this proposition were conducted with non-experimental designs, thus, undermining the interval validity of the results. On the other hand, studies base their predictions on individual factors and do not incorporate contextual factors. To overcome those limitations, a randomized field experiment was implemented in four insurance companies to incorporate contextual factors into the dynamics of deterrence. This study assessed the effect of a written threat (a deterrent letter reminding insured persons of the punishment for insurance fraud) on claim padding behaviours of insured persons filing claims for residential theft. A deterrent-letter project was implemented in four insurance companies, with claims randomly assigned to the experimental or the control group. Cases belonging to the control group were managed as usual, while individuals in the experimental group received the written threat. The experimental design made sure that the deterrent stimulus was exactly delivered to the insured persons when they had the opportunity to exaggerate the value of their claims. Findings demonstrate that claimants in the experimental group were less likely to pad their claims than were those in the control group. The letter was effective, regardless of the means of delivery. In conclusion, the administration of a written threat at the moment of criminal opportunity appears to be an effective strategy for preventing economic crimes.

A note on exogenous changes in incentives for and deterrence of corruption

In this note we amend the model elaborated by Acemoglu and Verdier (2000, The American Economic Review, 90(1),194–211), to examine how agents respond to an exogenous change in incentives. In particular, we apply the theoretical framework to the consequences on the corruption level in Italy of a famous sentence of the Italian Corte di Cassazione in plenary session: in this sentence, no. 500 of 1999, a revolutionary interpretation of the law has increased the area of civil liability for both the public administration and bureaucrats. This is one of the few cases in which the judge substantially legislates in a system of civil law, and the modification in incentives towards or away from corruption comes from an authority that is not part of the game.