A mayor’s dilemma

Imagine that you are the mayor of a mid-sized American city. Over the past five years, your citizens have watched serious crime and disorder steadily rise, and now they are demanding action. Unfortunately, like the average municipality, your city has faced declining operating revenue as a share of GDP over the past decade. The additional funds provided by the American Rescue Plan in 2021 have long since dried up, and with Congress divided over spending, it’s risky at best to rely on new federal dollars to combat crime. This means you must do more with the resources you already have. 

Critically, you cannot simply transfer funds from other essential areas like schools and parks to bolster police and prisons. City budgets are too tight for such a shell game, and while there is some truth in Victor Hugo’s point that, “He who opens a school door closes a prison,” addressing so-called “root cause” social failures is not a viable solution to pressing crime problems, particularly when lives are at stake today. Instead, you must make strategic adjustments within your existing crime control budget. This presents what mathematicians call an optimization problem: how to allocate spending in a way that minimizes serious crime most effectively.

Broadly, crime-fighting expenditures fall into two categories that might be conceived as prevention and treatment. Prevention includes policies and practices that reduce the likelihood of crime occurring in the first place, while treatment includes after-the-fact responses to crime, such as punishment and rehabilitation. 

As a mayor committed to evidence-based policymaking, you turn to the relevant scholarly literature for guidance. The best empirical studies suggest that a 10 percent increase in incarceration reduces crime by a modest 1-2 percent. Meanwhile, a 10 percent increase in the number of cops on the street reduces crime by 5-10 percent, and an investment in neighborhood security infrastructure, such as street lighting and surveillance cameras, might yield returns 20 times greater than the initial outlay. Even though, according to your calculations, it costs twice as much to hire a police officer or install cameras as it does to put someone in jail, the expected net benefits of policing or infrastructure improvements still outweigh the net benefit of incapacitation. 

You are also a mayor who understands the value of real-world experience and is loath to rely solely on academic research. So, you examine how other state and local governments across the country have responded to similar challenges. What you find surprises you.

Other leaders have not approached the optimization problem in the same way. Despite the stronger returns from prevention—deterring crimes before they happen—most mayors and governors have relied on punishment instead. Since the mid-1970s, when crime began to trend upward at an alarming rate, city and state spending on police as a share of total budgets has remained flat. Investment in neighborhood infrastructure is harder to isolate, but no relevant category of such spending has risen as quickly as investments in punishment. New spending has largely accrued instead to prisons and jails, which now receive a share of budgets 50 percent larger than they did before the crime boom. 

As an experienced policymaker, you know the key question when it comes to choosing a policy is “compared to what?” Every policy option has a cost and an effect. Severity is expensive; prison cells are scarce, and the longer the average sentence assigned, the less frequently the existing resources for punishment can be used. Given constant resources, any increase in the severity of a punishment delivered must come at the expense of the certainty that other crimes will be punished.

Set these high costs beside their expected returns. Prisons do seem to reduce crime, but the historical overreliance on them is unjustified in light of other available options that have better return on investment. Crime is less severe than it was in the 1980s and 1990s, but could that result have been bought at a lower cost, and if so, could low crime rates be more sustainable? 

The ongoing crime surge in your city—mirroring trends across the country—suggests that policymakers have yet to solve the crime optimization puzzle. So, what should you do?

Certainty’s promise

Crime prevention and punishment are, at their core, problems of resource allocation. The criminal justice system has limited capacity: police can patrol only so many streets, courts can process only so many cases, and prisons can house only so many offenders. Preventing a crime means that the post hoc mechanisms of the criminal justice system do not have to be activated. 

Of course, not every crime can be prevented. Within these constraints, ensuring a high certainty of punishment allows the justice system to act as a tool of triage– distinguishing between would-be offenders who can be deterred and those who cannot. Certainty also reduces the chance that crime control devolves into what Mark Kleiman called “random draconianism” – harsh punishments handed out to whomever happens to get caught, regardless of the true danger they pose to the community. The goal is to reduce crime to a level where the most serious and persistent offenders—those for whom deterrence fails—can be incapacitated or treated with the scarce resources available, without overwhelming the system. 

This sorting process is crucial because both severity and leniency come at a cost. Prisons are expensive and cause long-term societal harms, while alternative responses–such as treatment programs–succeed only some of the time and require immense resources that could otherwise be used to improve enforcement. If the justice system fails to deter people with marginal incentives to commit crime, these more expensive interventions must be used on a much larger population, reducing the system’s ability to enforce punishment or deliver treatment consistently.

The concept of certainty as triage is not just theoretical—it underpins well-documented criminal justice interventions. Marquee programs such as the Boston Gun Project, the High Point Drug Market Intervention, HOPE probation, and 24/7 Sobriety all function on this principle: by making consequences swift and predictable, these initiatives remove deterrable offenders from the system, allowing policymakers to focus resources on those requiring more intensive interventions.

Making people think twice

For certainty-based interventions to be effective, the perception of risk must be managed. It is not enough for punishment to be certain in an administrative sense—it must be perceived as such by potential offenders. Behavioral economics suggests that people do not assess risk with perfect rationality. They often struggle to estimate probabilities, especially when the likelihood of punishment is unclear. If offenders believe the system will not respond, they may assume they will evade punishment. When consequences are clear and predictable, they adjust their behavior accordingly.

This principle guided the Boston Gun Project’s “Operation Ceasefire,” a response to rampant gun crimes in which law enforcement and community leaders sent a direct message to known gang members: violence would bring immediate and certain consequences. The result was a 63 percent decline in youth homicide rates—not because punishments became harsher, but because they became more certain. The same logic shaped the High Point Drug Market Intervention, where drug dealers were given explicit warnings and a clear choice: stop selling drugs or face prosecution. Many dealers chose to exit the drug trade, leading to a 57 percent drop in violent crime in the area. 

These strategies work because certainty influences decision-making more than severity does. A gunman or dealer who believes they will be caught quickly is more likely to change their behavior than one who faces a long but unlikely prison sentence.

Hot spots policing follows a similar principle but distributes scarce enforcement resources by place instead of people. Concentrating police presence in high-crime areas makes detection more predictable. While Operation Ceasefire and High Point succeeded by clearly communicating the threat of punishment to the small group of people who drove violent crime, hot spots and similar policing strategies succeed by communicating that threat in the places that are most attractive to criminals. A review of 65 studies of hot spots policing by Braga and colleagues found that focused patrols in crime-dense locations reduced criminal activity—not by making penalties harsher, but by making the risk of apprehension more immediate.

Increasing the perceived certainty of punishment does not always require active law enforcement. The design of public spaces also influences behavior, shaping the perception of risk even when police are not nearby. Crime Prevention Through Environmental Design (CPTED) applies this principle by modifying physical environments to increase the certainty of detection and discourage crime.

One consistent finding in CPTED research is the impact of improved lighting on crime rates. A study in New York City found that brighter street lighting in high-crime areas significantly reduced nighttime offenses. The reason was straightforward: increased visibility reduced offenders’ sense of anonymity, making detection feel more certain. Other environmental changes—such as clear sightlines, security cameras, and well-maintained public spaces—reinforce the perception that crime will be noticed and addressed, ultimately deterring potential offenders.

These successes show that certainty can be achieved not just through blanket enforcement, but through strategic deployment of limited resources.

Certainty as a means of triage

Certainty-based deterrence is not a universal solution to criminal behavior. A prolific minority of people who commit crimes, particularly those who struggle with impulse control, substance addictions, or deep social instability, do not respond to predictable punishment in the same way others do. Some evidence suggests even experienced professionals are poor judges of who is most likely to commit a crime. Because individuals assess risk differently—and some cannot be deterred by any level of threat—certainty serves as a powerful tool of triage, helping to distinguish those who can be influenced by deterrence from those who require more intensive interventions.

The limitations of certainty-based deterrence can be leveraged to distinguish those who are deterrable from those who are not, as seen in community supervision programs like HOPE probation and 24/7 Sobriety. These programs aim to prevent noncompliance through swift and certain consequences. 

The HOPE model, pioneered in Hawaii, subjects probationers with records of substance abuse to frequent drug tests with immediate but modest sanctions for failures. Most participants comply or fail only once, allowing officials to reserve more intensive interventions for those who continue using despite the certainty of punishment. Similarly, the 24/7 Sobriety program, pioneered in South Dakota, mandates twice-daily breathalyzer tests for DUI offenders, with an immediate but brief jail stay for positive results. Evaluations of both programs show that the vast majority of participants reduce their targeted behaviors.

Critics have argued that by focusing on punishment, programs like HOPE and 24/7 Sobriety overlook the underlying needs that drive criminal behavior. However, this supposed flaw is actually a key strength. Certain, non-severe punishments serve two purposes: they reduce problem behavior among the vast majority of deterrable individuals, and they identify the much smaller group who truly require compulsory treatment or incapacitation to address their issues. 

For certainty-based deterrence to work effectively, it must be understood as a mechanism of triage. Not all offenders require the same response. Certainty-based deterrence helps distinguish those who can be influenced by predictable punishment from those who cannot.

The backbone of justice

Mayors across America could take a lesson from parents with small children: Steep threats of punishment lose their effectiveness when not backed up by action, but there are only so many time-outs you can give in one day. Like parents, mayors are better off relying on vigilance rather than punishment. And, also like parents, mayors should recognize that when a clear expectation exists that undesirable behaviors will lead to swift but proportionate responses, persistent misbehavior signals the need for stronger measures.

Certainty of punishment is not only a tool of deterrence but a structural necessity for an effective justice system. When consequences are clear and predictable, many offenders avoid crime, while those who persist reveal themselves to need more intensive responses. This enables police, courts, and corrections to concentrate resources where they are most needed, striking a balance between deterrence, incapacitation, and rehabilitation.

Practices like focused deterrence, hot spots policing, CPTED, and swift-and-certain sanctions illustrate the power of certainty in shaping behavior. But they also reveal its limits. Some offenders require more than deterrence. The challenge is to refine this balance, ensuring that certainty is not just a principle of deterrence, but a foundation for a more rational and effective criminal justice system.