When I was young and irresponsible, I met with the domestic policy team of a Democratic presidential nominee. To my suggestion that the candidate run on smarter crime control—and the gross failure of mere toughness as a strategy—the head of the team said, “No, crime isn’t our issue; it’s their issue.”

A generation later, that remains more or less the Democratic office-seeker’s approach to crime: try to talk about something else. On this theory, the voters aren’t buying what the good guys are selling: less punishment and more social services in the name of “crime prevention.” Conservatives, who want to fight crime by hurting people who commit it, have a natural rhetorical advantage over liberals, whose tendency is to fight crime by helping those who have committed it or might do so. No campaign speech, on either side, is likely to mention the prospect that more strategic use of law enforcement might reduce both crime and punishment at once.

Many progressives not seeking office cling to the views that the politicians have had to abandon: that crime is mostly an imaginary problem, vastly overstated by conservatives to put racial fears in socially acceptable code; that the root of criminal activity is inequality and injustice, and that only basic social reform can control it; that the criminal-justice system functions primarily as an engine of racial and social-class oppression, as evidenced by the huge disproportion between the racial makeup of the prison population and the racial makeup of the country as a whole; and that enforcement should focus on white-collar crime. Strong opposition to the death penalty, or to disproportionate incarceration, establishes one’s progressive bona fides; proposals for reducing the homicide rate (other than via gun control) or disproportionate victimization, not so much. The progressive tendency is to fixate on the plight of those punished rather than the plight of those victimized, though of course these are often the same persons under different labels or at different moments.

The corresponding conservative tendency is to regard “victims” and “perpetrators” as distinct groups—as if most criminals hadn’t first been victims—and punishment as good, and more punishment as better, conditional only on actual guilt and some sort of due process. That leads them to dismiss concerns about having 1 percent of all adults, 4 percent of all African-American men, and 37 percent of young African-American men without high-school diplomas behind bars at any given time as without moral significance.

Thus the debate over criminal-justice policy often seems to take place between the disciples of Michel Foucault and the disciples of the Marquis de Sade, with the Foucauldians winning the academic debate even as the sadists mostly get their way in the real political world. The resulting policies manage to combine enormous cruelty with unsatisfactory crime-control results: The United States leads the developed world in both homicide and incarceration, and both of those evils land most heavily on poor African Americans.

We can and should do better. But “doing better” doesn’t mean simply focusing on social services and systemic reforms and ignoring the need for punishment. It means using punishment intelligently, which means using it as sparingly as possible but also as much as necessary. As Machiavelli warned his fellow opponents of tyranny, a reluctance to punish comes naturally with good-heartedness, but those unable to overcome that reluctance are as unfit to rule as those who have no such reluctance to begin with.

I argue that (blue-collar) crime—theft and assault, in all their varieties—is still a real and major problem; that its economic and social costs are vastly under-appreciated; that its primary victims are disadvantaged minorities and poor people; that the current criminal-justice system wrongs them by under-enforcing the law against those who victimize them (who are, of course, mostly people like them in racial and class terms); that better criminal-justice policy could give us less crime and less incarceration; and that better and more equal law enforcement ought therefore to be as central a progressive political goal as better and more equal education or health care.

After all, the Fourteenth Amendment forbids any state to deny any person equal protection of the laws. As Randall Kennedy pointed out two decades ago, current actual policy in every state routinely violates that proscription, by punishing less consistently those who victimize poor people and members of socially disadvantaged ethnic groups, most of all African Americans. That system helps sustain intolerably high rates of victimization—especially violent victimization—in poor black neighborhoods, with disastrous consequences. Progressives ought to demand action.

The Costs of Crime

That crime is a terrible problem is among the received pieties of American politics, but at first glance it isn’t obviously true, especially after the great crime decline of the past 15 years. Homicide is no longer among the 15 leading causes of death; accidents are eight times as common, suicides more than twice as common. The average burglary loss per household is about $50 per year. Such calculations seem to support the suspicion that “crime” is a largely invented issue, created for the purpose of justifying repression.

And yet those calculations also seem to miss the point. Crime isn’t just something that happens to people. Crime is something someone does to someone else. As Justice Holmes remarked, even a dog knows the difference between being stumbled over and being kicked.

That difference matters because being hurt deliberately rather than inadvertently is degrading as well as damaging: It adds insult to injury. By asserting lawless power over the victim, the perpetrator not only puts the victim in fear of the event happening again (with or without a change of victimizer) but also acts out the message that the victim is not someone whose person or property others need to respect. Impunity drives the message home; there is evidence from the social-psychology lab that people tend to think less of the victims of unpunished crimes. This helps make sense of the demand of victims and their families that someone—even, sometimes, the wrong person—be punished, to give the victims “closure.”

Nor do all the losses from criminal activity accrue to crime victims. To avoid being victimized, people lock their doors, stay at home, and avoid dangerous neighborhoods. All of those forms of crime avoidance are costly, and not only to those who engage in them. When those people and businesses that can abandon high-crime neighborhoods do so, they leave those neighborhoods poorer—in social connections, in services, and in job opportunities—and even more dangerous than they already were. It would be impossible to understand the residential patterns in any metropolitan area without reference to the crime map. A crime decline that partly stems from the depopulation of high-crime neighborhoods is nothing to be proud of, and even today’s reduced crime rates are still at least 50 percent above those of the 1950s and early 1960s. Homicide has fallen further, but that seems to stem in part from better medical care: The rate of intentional gunshot injury has not been falling.

Geographically concentrated crime also corrodes social ties at the individual and neighborhood levels; it is both a cause and an effect of distrust. And it feeds on itself in other ways as well. Since enforcement tends to be less concentrated than criminal activity, criminal behavior is safer where there is a lot of it, and both the opportunities for gainful licit activity are fewer and the gains from such activity less secure. Living in chaos makes people more present-oriented and less averse to risk, two characteristics that make crime, with its immediate and certain gains and its deferred and uncertain losses, appear more attractive. Little wonder crime is so heavily concentrated, that those concentrations are so stable over time, and that it concentrates among those otherwise worst off. And concentrated it is: African Americans face more than six times the murder risk of the rest of the population, and there are neighborhoods where homicide is the leading cause of death among males between 15 and 30.

But crime has an equally evil twin: punishment. The United States, with 5 percent of the world’s population, has 25 percent of its prisoners. We have more people behind bars than does China: not merely more per capita, but more total prisoners. Our current incarceration rate is five times our own historical norm, and five times the level of any other economically and politically advanced country. And incarceration is just about as concentrated by race as is homicide victimization. While the national incarceration rate is about 750 per 100,000 population, the rate among African Americans is more than 3,000 per 100,000. (The late William Stuntz of Harvard Law School calculated that African Americans are locked up at a higher rate than Russians were under Stalin.) Add social class to race and the problem looks even worse: An African-American male who does not finish high school is more likely than not to spend time in prison.

Mass incarceration can be self-defeating. Converting imprisonment from a shocking disgrace to a routine incident of early manhood greatly reduces the stigma that carries much of its deterrent power when it is sparingly used. Purdue sociologist Bert Useem and Rutgers economist Anne Morrison Piehl estimate that, under current conditions in the median U.S. state, adding a prisoner creates as much crime as it prevents.

Many critics of mass incarceration focus their attacks on the “war on drugs.” In fact, the 500,000 people behind bars at any one time for drug-law violations constitute only about 20 percent of those locked up. That’s still a shockingly large number of people—the United States has a higher rate of incarceration for drug-law violations alone than any Western European country has for all offenses combined—but “ending the drug war” would not put much of a dent in the mass-incarceration problem, even if we assume no additional crime due to additional drug abuse. (Alcohol, our one licit addictive intoxicant, is involved in something like half of all violent crime.)

Since victimization is overwhelmingly intraracial, doing a better job of vindicating African-American victims would be hard to do without further increasing the disproportion in incarceration rates. Is it any surprise that progressives would rather talk about root causes, prevention programs, and procedural justice?

The Costs of Crime Control

Policing takes a big share of municipal budgets. Incarceration swallows up a noticeable chunk of state budgets, and rising prison costs have helped squeeze state-university funding.

Still, the overall costs of the criminal-justice system do not loom especially large among forms of public expenditure. Counting everything—police, courts, prosecution, prisons, jails, probation, parole, and publicly paid defense counsel—and combining all levels of government, we spend about $225 billion a year on law enforcement, less than half the cost of public K-12 education, about a third of what we spend on defense, or a fifth of the public-sector contribution to health care. More than half of that $225 billion goes to police; prisons and jails combined account for about a quarter, or $55 billion a year. Thus the vast “prison-industrial complex” whose budget could be “reinvested” in massive social reforms is largely imaginary; shutting down all the prisons and jails tomorrow wouldn’t finance a 10 percent increase in education spending.

A focus on the budget leads to solving the wrong problem. It is not the dollar costs of criminal-justice operations that should most concern us. Those are modest, not merely compared to any reasonable computation of the costs of crime, but also compared to the damage the system does to those it acts on: the suffering inflicted by arrest, prosecution, conviction, and incarceration, including all of the residual disabilities that go with the label “ex-convict,” and the fear created by overaggressive policing.

In the project of reconfiguring our criminal-justice system to do more good and less harm, the fiscal crunch that has led to cuts in police and prison budgets both helps and hurts. It seems unlikely that the country will ever again have as many police or as many prisoners as it did in 2010, which ought to force the pace of innovation. But fiscal pressure can also squeeze out new ideas and lead to the kind of short-term thinking that refuses to spend a dollar now to save $5 next year.

Crime Despite Punishment

The conservative approach to crime is simple: Punish it. The economic theory of crime developed by Gary Becker, Nobel laureate of the University of Chicago, holds that people will violate the law if and only if doing so is advantageous: if the expected value of the gains from criminal activity, less the expected value of the punishment imposed, exceeds the rewards of licit economic activity. The policy prescription is to make sure crime doesn’t pay by ramping up the punishments.

It’s an elegant theory, but, as Lt. Columbo would say, there’s just one thing: Even at the nadir of incarceration-to-offense ratios in the late 1970s, blue-collar crime has never paid. For every hour spent behind bars, the average burglar back then “earned” something less than $2; the wages of sin have always been well below the legal minimum. The subsequent increase in burglary-related incarceration and decrease in burglaries have driven wages for serving burglary time to the pennies-per-hour range. The surprise for those who believe that offenders are rational economic actors is how small the decrease in burglary has been even as the terms of trade worsened six-fold. (Crack-dealing wages have also fallen sharply, albeit from a higher level.)

Why do some people keep committing crimes, to their own evident disadvantage? Because they’re present-oriented and impulsive, with deficient capacities for shaping their current behavior in light of their future goals, and with poor judgment about their actual odds of getting caught: all characteristics, as noted above, likely to be produced by growing up in high-crime neighborhoods. (Neglectful and abusive parenting also contributes, of course. So does exposure to environmental lead; more on this below.)

If you’re looking for a single “root cause” of crime, look no further: The cause is bad decision-making by offenders. And the solution must lie in some combination of improving that decision-making process and devising deterrent threats that actually deter reckless, impulsive, short-term-oriented people, which the current regime of randomized draconian responses so dramatically fails to do.

Way back in the eighteenth century, Cesare Beccaria—the Italian criminologist from whom Jeremy Bentham borrowed not only the term “utility” but many of his ideas for criminal-justice reform—identified three characteristics that determine the deterrent efficacy of a threatened punishment: its swiftness, its certainty, and its severity. Of the three, severity is least important. If punishment is swift and certain, it need not be severe to be efficacious. If punishment is uncertain and delayed, it will not be efficacious even if it is severe. (It was only two and a half centuries after Beccaria that psychologists and behavioral economists discovered that some degree of excessive present-orientation, and excessive discounting of the risk of large losses, is normal.) The sort of bad gamble represented by most offenses tends to attract precisely those whose departures from rationality are most egregious.

Viewed from the perspective of deterrence, long prison terms are a bad bargain: The last 15 years of a 20-year prison sentence start five years from its beginning, a period distant enough to be beyond the planning horizon of the typical armed robber. And those long prison terms are no better viewed from the perspective of incapacitation—the purely mechanical effect of preventing crime by keeping the criminals locked up. Because crime is mostly a young man’s game, the amount of crime prevented per year incarcerated falls with the age of the prisoner; by the end of a long sentence even a prisoner who starts out young has grown old, as age is measured in the world of offending. Thanks to “three strikes” laws and absurdly long terms for drug dealing, the average prisoner is now in his (or, much more rarely, her) mid-30s while the average new crime is committed by someone in his early 20s. That’s a very costly mismatch.

Not only is severity an inadequate substitute for swiftness and certainty, it actually interferes with them. The more severe a punishment is, the more due process (leading to delay) is required to impose it, and (if severity is measured in sentence length) the less often it can be imposed before the prisons fill up. Unfortunately, increasing severity is easy for a legislature or sentencing commission to accomplish; increasing swiftness and certainty—for example, by adding police and improving police operations—is more complicated, especially given the need to balance intrusiveness against crime-control benefits, as in the case of New York’s sometimes overaggressive “stop-and-frisk” tactics.

Smarter Policing

Police work has been among the most innovative sectors of American public management over the past three decades. “Problem-oriented” and then “community” policing replaced the mindless use of “preventive patrol” and the drive to minimize response times. Instead, the new approach stressed attention to the neighborhood-level social forces and cooperation with the police. Then followed the CompStat revolution, in which police departments accepted, and met, the challenge of actually reducing crime rather than merely responding to it. Focusing on geographic and temporal “hot spots” and high-rate offenders pays big dividends, and the era of “predictive policing,” when data-driven modeling points officers to the place where the next crime is likely to occur, seems to be upon us.

All of this is still a work in progress. Policing, like education, is labor-intensive and hard to automate, and therefore subject to Baumol’s cost disease, under which the costs of activities with low productivity gains rise over time compared with the costs of more technologically progressive sectors. But falling information-processing costs can help cure the Baumol disease. We now know that a home that suffers a burglary is at heightened risk of being revictimized within weeks. Why not have the police officer who comes to investigate a burglary offer to install (now very cheap) police webcams over the doors and windows and leave them in operation for 60 days?

The revolution in information technology could also allow departments to massively crowdsource the task of monitoring conditions and observing crimes in progress. Virtually every pedestrian now carries what amounts to a TV camera; police will have to learn how to acquire, sort, and act on relevant information quickly and seamlessly. For $200, you can now buy a toy helicopter equipped with an HD camera and guided by a laptop. Learning how to use it takes minutes. Such devices could reinvigorate the concept of “neighborhood watch,” but only if the police can learn to make use of the results (and prevent the “watchers” from turning into stalkers or peeping Toms).

None of this is straightforward. But all of it is essential. New York City’s massive and aggressive police force has succeeded in reducing not only crime but incarceration. We need to learn how to do so at smaller costs in civil liberty and (given fiscal realities in cities without Wall Street) money. But that project will not be helped much by people whose response to “police” on a word-association test is “misconduct” or “brutality.”

The Alternative to Incarceration

At any one time, more than twice as many people are subject to community supervision—probation or parole—than are held in jails and prisons. Those populations are heterogeneous; they include a large number of low-risk, low-rate offenders, some high-rate minor offenders, and a relative handful of seriously dangerous people. Lots of people get caught once or twice; but it is those who offend again and again that account for most of the serious crimes and most of the prison headcount.

When not confined, these high-rate serious offenders are likely to be under community supervision; on their way up the career-criminal ladder, they will have spent time on juvenile probation. The more effective community supervision is, the less we need to rely on incarceration, both because judges and legislatures will be willing to imprison less often and because better supervision prevents re-offending.

Right now, though, those systems are dismal failures, and their reform offers the best prospect for reducing both victimization and days behind bars. Most contemporary community supervision reproduces the random severity of the larger criminal-justice system. Probationers and parolees are subject to myriad rules—not all of them clearly useful in reducing their risk of re-offending—but are so loosely monitored that most violations go undetected. (One scheduled visit a month to the probation office is about average, often with a drug test thrown in; since the test is also scheduled in advance, three days’ abstinence will produce a “clean” specimen for drugs other than cannabis.)

When a violation is detected—typically a positive drug test or a missed appointment—the usual response is…nothing. The probation officer says, “That’s against the rules. Don’t do it again.” In principle, any violation could be the basis for revocation of release and a trip to jail or prison for months or even years. But in practice, those sanctions are too drastic to be imposed often, and the paperwork surrounding them, combined with the sheer number of offenders on each officer’s caseload—150 to 200 is typical in high-crime jurisdictions—ensures that most violations will draw only warnings, and that the warnings will have little credibility. (Parole agents tend to have smaller caseloads of tougher offenders, so they tend to respond somewhat more aggressively; unlike probation officers, they don’t need to convince a judge to impose a sanction. But more than four out of five community-corrections clients are probationers, not parolees.)

As an individual accumulates more and more violations, the gravity of the warnings is likely to escalate, until at some unpredictable point—maybe the sixth violation, maybe the sixteenth—off he goes to see the judge and possibly spend a spell behind bars. Even though most violations go unpunished, revocation of community release is among the most common reasons for entering prison. It’s as if the system were deliberately creating a trap for the poorly self-controlled people subject to it.

The obvious (but hard-to-administer) common-sense alternative is to make the rules less numerous, the monitoring tighter, and the sanctions swift, certain, and reasonably mild, and to clearly tell each probationer and parolee exactly what the rules are and what exactly will happen, every time and right away, when a rule is broken. Mildness—or proportionality, if you like—is essential to making the threat credible, and severity turns out to be unnecessary. Experimental evidence from the HOPE program in Hawaii showed that two days in jail is as good a deterrent to drug use as six weeks, as long as the two days actually happen, and happen every time. We don’t know yet whether a day in jail, or a couple of hours in a holding cell, or a weekend of home confinement, or a week of a 9 p.m.-6 a.m. curfew, would do the trick, but we ought to learn.

But while severity is superfluous, mercy is toxic. The tendency to give a usually well-behaved client a break “just this once” destroys certainty. Worse, it converts what should be the inevitable consequence of the offender’s own actions—something he did, in effect, to himself—into something the officer could have chosen not to do, and thus the product of the officer’s whim.

Where swift-and-certain sanctions for enforcing community corrections rules have been tried—and where the system has been properly organized, first to deliver the clear warnings and then to deliver on the threats when necessary—the results have been astounding. Drug-using offenders supervised under those conditions achieve much bigger reductions in drug use than result from the mandatory drug treatment; 80 percent of long-term criminally active users of hard drugs turn out to be able to quit under steady pressure. That in turn leads to reductions of 50 percent or more in new crimes and, crucially, days spent behind bars. Those results make swift-and-certain sanctions programs—which, along with Hawaii, have been tried in Texas, Washington state, California, and (focusing on alcohol) South Dakota—easily the most promising approach to reducing crime rates and incarceration rates both relatively quickly (over the next few years) and dramatically.
Not only does swift-and-certain sanctioning succeed in changing the behavior of people with deficient self-command; there is some evidence—not nearly conclusive, but intriguing—that it may actually change the entire behavioral style of many clients by showing them that they do in fact have the capacity to manage their habits. That in turn increases the characteristic psychologists call “self-efficacy”—a person’s belief that he can exert control—which is a strong predictor of success in any attempt at behavior change, from quitting smoking to learning to restrain one’s anger.

The swift-and-certain-sanctions innovation is now spreading. It is doing so with support from the federal Office of Justice Programs, but against the fervent opposition of proponents of less effective and more expensive practices such as mandatory drug treatment. Research has been impeded by the refusal of the National Institute on Drug Abuse to fund studies of an approach that demonstrates the fatuity of the dogmas that drug abuse is merely a “brain disease” and that “punishment doesn’t work” (a conclusion based on studies where punishment was neither swift nor certain).

Whether the requisite public-management capacity can be mustered to bring it to full scale without loss of program integrity remains a question mark. Washington state is currently mounting the first statewide program, with encouraging preliminary results.

Meanwhile, GPS tracking makes pinpointing an offender’s location as easy—though not quite as cheap—as monitoring his drug use. That would allow the enforcement of curfews and stay-away orders, deter re-offending, and make the person wearing a monitor unwelcome in group criminal activity.

The combination of drug testing and position monitoring with swift-and-certain sanctioning could deliver most of the incapacitative effect of a prison cell at a fraction of the cost, and a still-smaller fraction of the suffering. Some proportion—perhaps a large one—of those currently behind bars could be safely released under those conditions. Now that we’re almost back to 1965 crime rates, perhaps we could start to move back to 1965 incarceration rates: that would mean an 80 percent reduction in the population behind bars.

Smart community corrections can save about $5 in incarceration spending for every dollar it costs. But the spending is now, and comes mostly from city and county budgets, while the savings are later, and mostly accrue to state budgets. The approach remains without a political champion, a funded lobby, or an enthusiastic foundation backer. Its supporters need to hope that good outcomes will speak for themselves.

Crime Control Without Punishment

Of course it would be better to control crime by helping people rather than harming them. But the list of proven social-service approaches to crime control is short. Even programs that have shown promise, including “coaching” of poorly educated first-time mothers, are slow acting. And intervening with a newborn won’t do much to crime rates in the next 15 years.

We could get quicker dividends by starting and ending the middle-school and high-school day a couple of hours later, shortening the burglary-friendly period between when school gets out and grown-ups get home from work. A later-starting school day would also benefit students’ health and academic performance by giving them more time to sleep. But changing school hours would also interfere with extracurricular activities and inconvenience various adults, so it seems to be a nonstarter.

Successful drug treatment is a big winner, but most drug-abusing offenders won’t go to treatment if offered it, even if nominally compelled to attend, or won’t stick with it long enough to matter. The probation machinery that is so ineffective in enforcing other conditions is no better at enforcing treatment mandates. Drug courts are too resource-intensive to be practicable at scale, and their good outcomes depend in large part on filtering out participants with histories of serious criminality. Spending the most attention on the least dangerous is hardly good policy.

Other than literacy programs, which are cheap and have some demonstrated benefits, most in-prison treatment has limited value. Changes in attitudes and behavior achieved “inside” rarely survive the transition back to the “free world.” Changing the capacity for self-command is hard.

No doubt a variety of social-service and health-care agencies, from Temporary Assistance for Needy Families offices to shock-trauma units, might be able to make substantial contributions to crime control. But that would mean designing and evaluating specific interventions with crime control in mind and holding officials responsible for crime-control outcomes, which would take us a long way from current practice. (How many high-school principals or housing-project managers could tell you what fraction of their students or tenants were arrested last year?) Crime control seems to be more a throwaway argument in support of programs favored on other grounds rather than something the social-service world and its political allies are ready to take seriously.

The wild card in the crime-control deck is lead abatement. The science is no longer in serious doubt (see Kevin Drum’s summing up in the January-February Mother Jones): Lead, even in very small doses, not only reduces IQ but also interferes directly with the systems in the brain that mediate self-command. Solid statistical evidence suggests that increasing lead exposure after World War II contributed to the great Boomer crime wave, and that EPA regulations starting in the 1970s that forced the conversion to unleaded gasoline were a major driver of the crime decline that started in the mid-1990s.

Today’s children are still being exposed to damaging levels of lead from the soil (a leftover from lead in gasoline) and from residential buildings, and especially old window casings. The cost of abatement would be moderately stiff: Drum estimates something like $20 billion per year over 20 years, and arguably the process should move faster than that. Still, at that pace, lead abatement, while it was going on, would cost about 10 percent of our current criminal-justice expenditure, and nothing after that, for a permanent crime reduction likely—based on past experience—to be at least 10 percent. All the other cognitive and health benefits would be gravy. It’s hard to imagine any other crime-control expenditure with anything like that much bang for the buck. And the benefits would be concentrated in poor, largely minority, urban neighborhoods. Unfortunately, an idea like lead abatement has no place in the current discourse on crime control: Foucauldians and sadists alike will find it not so much wrong as utterly beside the point.

Crime-control policy is not merely a practical problem; moral judgment is at its very core. And yet, with a little less heated rhetoric and a little more practical reasoning, we could have a lot less crime. Our failure to stop posturing—just for a moment—and do the actual work is deeply, unforgivably immoral.