Figure 12-1: Homicide deaths, Western Europe, US, and Mexico, 1300–2015
Sources: England, Netherlands & Belgium, Italy, 1300–1994: Eisner 2003, plotted in fig. 3–3 of Pinker 2011. England, 2000–2014: UK Office for National Statistics. Italy and Netherlands, 2010–2012: United Nations Office on Drugs and Crime 2014. New England (New England, whites only, 1636–1790, and Vermont and New Hampshire, 1780–1890): Roth 2009, plotted in fig. 3–13 of Pinker 2011; 2006 and 2014 from FBI Uniform Crime Reports. Southwest US (Arizona, Nevada, and New Mexico), 1850 and 1914: Roth 2009, plotted in fig. 3–16 of Pinker 2011; 2006 and 2014 from FBI Uniform Crime Reports. Mexico: Carlos Vilalta, personal communication, originally from Instituto Nacional de Estadística y Geografía 2016 and Botello 2016, averaged over decades until 2010.
In the case of the 1960s crime explosion, even the facts at hand refuted the root-cause theory. That was the decade of civil rights, with racism in steep decline (chapter 15), and of an economic boom, with levels of inequality and unemployment for which we are nostalgic.10 The 1930s, in contrast, was the decade of the Great Depression, Jim Crow laws, and monthly lynchings, yet the overall rate of violent crime plummeted. The root-cause theory was truly deracinated by a development that took everyone by surprise. Starting in 1992, the American homicide rate went into free fall during an era of steeply rising inequality, and then took another dive during the Great Recession beginning in 2007 (figure 12-2).11 England, Canada, and most other industrialized countries also saw their homicide rates fall in the past two decades. (Conversely, in Venezuela during the Chávez-Maduro regime, inequality fell while homicide soared.)12 Though numbers for the entire world exist only for this millennium and include heroic guesstimates for countries that are data deserts, the trend appears to be downward as well, from 8.8 homicides per 100,000 people in 2000 to 6.2 in 2012. That means there are 180,000 people walking around today who would have been murdered just in the last year if the global homicide rate had remained at its level of a dozen years before.13
Figure 12-2: Homicide deaths, 1967–2015
Sources: United States: FBI Uniform Crime Reports, https://ucr.fbi.gov/, and Federal Bureau of Investigation 2016. England (data include Wales): Office for National Statistics 2017. World, 2000: Krug et al. 2002. World, 2003–2011: United Nations Economic and Social Council 2014, fig. 1; the percentages were converted to homicide rates by setting the 2012 rate at 6.2, the estimate reported in United Nations Office on Drugs and Crime 2014, p. 12. The arrows point to the most recent years plotted in Pinker 2011 for the world (2004, fig. 3–9), US (2009, fig. 3–18), and England (2009, fig. 3–19).
Violent crime is a solvable problem. We may never get the homicide rate for the world down to the levels of Kuwait (0.4 per 100,000 per year), Iceland (0.3), or Singapore (0.2), let alone all the way to 0.14 But in 2014, Eisner, in consultation with the World Health Organization, proposed a goal of reducing the rate of global homicide by 50 percent within thirty years.15 The aspiration is not utopian but practical, based on two facts about the statistics of homicide.
The first is that the distribution of homicide is highly skewed at every level of granularity. The homicide rates of the most dangerous countries are several hundred times those of the safest, including Honduras (90.4 homicides per 100,000 per year), Venezuela (53.7), El Salvador (41.2), Jamaica (39.3), Lesotho (38), and South Africa (31).16 Half of the world’s homicides are committed in just twenty-three countries containing about a tenth of humanity, and a quarter are committed in just four: Brazil (25.2), Colombia (25.9), Mexico (12.9), and Venezuela. (The world’s two murder zones—northern Latin America and southern sub-Saharan Africa—are distinct from its war zones, which stretch from Nigeria through the Middle East into Pakistan.) The lopsidedness continues down the fractal scale. Within a country, most of the homicides cluster in a few cities, such as Caracas (120 per 100,000) and San Pedro Sula (in Honduras, 187). Within cities, the homicides cluster in a few neighborhoods; within neighborhoods, they cluster in a few blocks; and within blocks, many are carried out by a few individuals.17 In my hometown of Boston, 70 percent of the shootings take place in 5 percent of the city, and half the shootings were perpetrated by one percent of the youths.18
The other inspiration for the 50-30 goal is evident from figure 12-2: high rates of homicide can be brought down quickly. The most murderous affluent democracy, the United States, saw its homicide rate plunge by almost half in nine years; New York City’s decline during that time was even steeper, around 75 percent.19 Countries that are still more famous for their violence have also enjoyed steep declines, including Russia (from 19 per 100,000 in 2004 to 9.2 in 2012), South Africa (from 60.0 in 1995 to 31.2 in 2012), and Colombia (from 79.3 in 1991 to 25.9 in 2015).20 Among the eighty-eight countries with reliable data, sixty-seven have seen a decline in the last fifteen years.21 The unlucky ones (mostly in Latin America) have been ravaged by terrible increases, but even there, when leaders of cities and regions set their mind to reducing the bloodshed, they often succeed.22 Figure 12-1 shows that Mexico, after suffering a reversal from 2007 to 2011 (entirely attributable to organized crime), enjoyed a reversal of the reversal by 2014, including an almost 90 percent drop from 2010 to 2012 in notorious Juárez.23 Bogotá and Medellín saw declines by four-fifths in two decades, and São Paulo and the favelas of Rio de Janeiro saw declines by two-thirds.24 Even the world’s murder capital, San Pedro Sula, has seen homicide rates plunge by 62 percent in just two years.25
Now, combine the cockeyed distribution of violent crime with the proven possibility that high rates of violent crime can be brought down quickly, and the math is straightforward: a 50 percent reduction in thirty years is not just practicable but almost conservative.26 And it’s no statistical trick. The moral value of quantification is that it treats all lives as equally valuable, so actions that bring down the highest numbers of homicides prevent the greatest amount of human tragedy.
The lopsided skew of violent crime also points a flashing red arrow at the best way to reduce it.27 Forget root causes. Stay close to the symptoms—the neighborhoods and individuals responsible for the biggest wedges of violence—and chip away at the incentives and opportunities that drive them.
It begins with law enforcement. As Thomas Hobbes argued during the Age of Reason, zones of anarchy are always violent.28 It’s not because everyone wants to prey on everyone else, but because in the absence of a government the threat of violence can be self-inflating. If even a few potential predators lurk in the region or could show up on short notice, people must adopt an aggressive posture to deter them. This deterrent is credible if only they advertise their resolve by retaliating against any affront and avenging any depredation, regardless of the cost. This “Hobbesian trap,” as it is sometimes called, can easily set off cycles of feuding and vendetta: you have to be at least as violent as your adversaries lest you become their doormat. The largest category of homicide, and the one that varies the most across times and places, consists of confrontations between loosely acquainted young men over turf, reputation, or revenge. A disinterested third party with a monopoly on the legitimate use of force—that is, a state with a police force and judiciary—can nip this cycle in the bud. Not only does it disincentivize aggressors by the threat of punishment, but it reassures everyone else that the aggressors are disincentivized and thereby relieves them of the need for belligerent self-defense.
The most blatant evidence for the impact of law enforcement may be found in the sky-high rates of violence in the times and places where law enforcement is rudimentary, such as the upper left tips of the curves in figure 12-1. Equally persuasive is what happens when police go on strike: an eruption of looting and vigilantism.29 But crime rates can also soar when law enforcement is merely ineffective—when it is so inept, corrupt, or overwhelmed that people know they can break the law with impunity. That was a contributor to the 1960s crime boom, when the judicial system was no match for a wave of baby
boomers entering their crime-prone years, and it is a contributor to the high-crime regions of Latin America today.30 Conversely, an expansion of policing and criminal punishment (though with a big overshoot in incarceration) explains a good part of the Great American Crime Decline of the 1990s.31
Here is Eisner’s one-sentence summary of how to halve the homicide rate within three decades: “An effective rule of law, based on legitimate law enforcement, victim protection, swift and fair adjudication, moderate punishment, and humane prisons is critical to sustainable reductions in lethal violence.”32 The adjectives effective, legitimate, swift, fair, moderate, and humane differentiate his advice from the get-tough-on-crime rhetoric favored by right-wing politicians. The reasons were explained by Cesare Beccaria two hundred and fifty years ago. While the threat of ever-harsher punishments is both cheap and emotionally satisfying, it’s not particularly effective, because scofflaws just treat them like rare accidents—horrible, yes, but a risk that comes with the job. Punishments that are predictable, even if less draconian, are likelier to be factored into day-to-day choices.
Together with the presence of law enforcement, the legitimacy of the regime appears to matter, because people not only respect legitimate authority themselves but factor in the degree to which they expect their potential adversaries to respect it. Eisner, together with the historian Randolph Roth, notes that crime often shoots up in decades in which people question their society and government, including the American Civil War, the 1960s, and post-Soviet Russia.33
Recent reviews of what does and doesn’t work in crime prevention back up Eisner’s advisory, particularly a massive meta-analysis by the sociologists Thomas Abt and Christopher Winship of 2,300 studies evaluating just about every policy, plan, program, project, initiative, intervention, nostrum, and gimmick that has been tried in recent decades.34 They concluded that the single most effective tactic for reducing violent crime is focused deterrence. A “laser-like focus” must first be directed on the neighborhoods where crime is rampant or even just starting to creep up, with the “hot spots” identified by data gathered in real time. It must be further beamed at the individuals and gangs who are picking on victims or roaring for a fight. And it must deliver a simple and concrete message about the behavior that is expected of them, like “Stop shooting and we will help you, keep shooting and we will put you in prison.” Getting the message through, and then enforcing it, depends on the cooperation of other members of the community—the store owners, preachers, coaches, probation officers, and relatives.
Also provably effective is cognitive behavioral therapy. This has nothing to do with psychoanalyzing an offender’s childhood conflicts or propping his eyelids open while he retches to violent film clips like in A Clockwork Orange. It is a set of protocols designed to override the habits of thought and behavior that lead to criminal acts. Troublemakers are impulsive: they seize on sudden opportunities to steal or vandalize, and lash out at people who cross them, heedless of the long-term consequences.35 These temptations can be counteracted with therapies that teach strategies of self-control. Troublemakers also have narcissistic and sociopathic thought patterns, such as that they are always in the right, that they are entitled to universal deference, that disagreements are personal insults, and that other people have no feelings or interests. Though they cannot be “cured” of these delusions, they can be trained to recognize and counteract them.36 This swaggering mindset is amplified in a culture of honor, and it can be deconstructed in therapies of anger management and social-skills training as part of counseling for at-risk youth or programs to prevent recidivism.
Whether or not their impetuousness has been brought under control, potential miscreants can stay out of trouble simply because opportunities for instant gratification have been removed from their environments.37 When cars are harder to steal, houses are harder to burgle, goods are harder to pilfer and fence, pedestrians carry more credit cards than cash, and dark alleys are lit and video-monitored, would-be criminals don’t seek another outlet for their larcenous urges. The temptation passes, and a crime is not committed. Cheap consumer goods are another development that has turned weak-willed delinquents into law-abiding citizens despite themselves. Who nowadays would take the risk of breaking into an apartment just to steal a clock radio?
Together with anarchy, impulsiveness, and opportunity, a major trigger of criminal violence is contraband. Entrepreneurs in illegal goods and pastimes cannot file a lawsuit when they feel they have been swindled, or call the police when someone threatens them, so they have to protect their interests with a credible threat of violence. Violent crime exploded in the United States when alcohol was prohibited in the 1920s and when crack cocaine became popular in the late 1980s, and it is rampant in Latin American and Caribbean countries in which cocaine, heroin, and marijuana are trafficked today. Drug-fueled violence remains an unsolved international problem. Perhaps the ongoing decriminalization of marijuana, and in the future other drugs, will lift these industries out of their lawless underworld. In the meantime, Abt and Winship observe that “aggressive drug enforcement yields little anti-drug benefits and generally increases violence,” while “drug courts and treatment have a long history of effectiveness.”38
Any evidence-based reckoning is bound to pour cold water on programs that seemed promising in the theater of the imagination. Conspicuous by their absence from the list of what works are bold initiatives like slum clearance, gun buybacks, zero-tolerance policing, wilderness ordeals, three-strikes-and-you’re-out mandatory sentencing, police-led drug awareness classes, and “scared straight” programs in which at-risk youths are exposed to squalid prisons and badass convicts. And perhaps most disappointing to those who hold strong opinions without needing evidence are the equivocal effects of gun legislation. Neither right-to-carry laws favored by the right, nor bans and restrictions favored by the left, have been shown to make much difference—though there is much we don’t know, and political and practical impediments to finding out more.39
* * *
As I sought to explain various declines of violence in The Better Angels of Our Nature I put little stock in the idea that in the past “human life was cheap” and that over time it became more precious. It seemed woolly and untestable, almost circular, so I stuck to explanations that were closer to the phenomena, such as governance and trade. After sending in the manuscript, I had an experience that gave me second thoughts. To reward myself for completing that massive undertaking I decided to replace my rusty old car, and in the course of car shopping I bought the latest issue of Car and Driver magazine. The issue opened with an article called “Safety in Numbers: Traffic Deaths Fall to an All-Time Low,” and it was illustrated with a graph that was instantly familiar: time on the x-axis, rate of death on the y-axis, and a line that snaked from the top left to the bottom right.40 Between 1950 and 2009, the rate of death in traffic accidents fell sixfold. Staring up at me was yet another decline in violent death, but this time dominance and hatred had nothing to do with it. Some combination of forces had been working over the decades to reduce the risk of death from driving—as if, yes, life had become more precious. As society became richer, it spent more of its income, ingenuity, and moral passion on saving lives on the roads.
Later I learned that Car and Driver had been conservative. Had they plotted the dataset from its first year, 1921, it would have shown an almost twenty-four-fold reduction in the death rate. Figure 12-3 shows the full time line—though not even the full story, since for every person who died there were others who were crippled, disfigured, and racked with pain.
Figure 12-3: Motor vehicle accident deaths, US, 1921–2015
Sources: National Highway Traffic Safety Administration, accessed from http://www.informedforlife.org/demos/FCKeditor/UserFiles/File/TRAFFICFATALITIES(1899-2005).pdf, http://www-fars.nhtsa.dot.gov/Main/index.aspx, and https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/812384.
The ma
gazine graph was annotated with landmarks in auto safety which identified the technological, commercial, political, and moralistic forces at work. Over the short run they sometimes pushed against each other, but over the long run they collectively pulled the death rate down, down, down. At times there were moral crusades to reduce the carnage, with automobile manufacturers as the villains. In 1965 a young lawyer named Ralph Nader published Unsafe at Any Speed, a j’accuse of the industry for neglecting safety in automotive design. Soon after, the National Highway Traffic Safety Administration was established and legislation was passed requiring new cars to be equipped with a number of safety features. Yet the graph shows that steeper reductions came before the activism and the legislation, and the auto industry was sometimes ahead of its customers and regulators. A signpost in the graph pointing to 1956 notes, “Ford Motor Company offers the ‘Lifeguard’ package. . . . It includes seatbelts, a padded dash, padded visors, and a recessed steering-wheel hub designed to not turn drivers into a kebab during a collision. It is a sales failure.” It took a decade for those features to become mandatory.
Sprinkled along the slope were other episodes of push and pull among engineers, consumers, corporate suits, and government bureaucrats. At various times, crumple zones, four-wheel dual braking systems, collapsible steering columns, high-mounted center brake lights, buzzing and garroting seat belts, and air bags and stability control systems wended their way from the lab to the showroom. Another lifesaver was the paving of long ribbons of countryside into divided, reflectored, guard-railed, smooth-curved, and broad-shouldered interstate highways. In 1980 Mothers Against Drunk Driving was formed, and they lobbied for higher drinking ages, lowered legal blood alcohol levels, and the stigmatization of drunk driving, which popular culture had treated as a source of comedy (such as in the movies North by Northwest and Arthur). Crash testing, traffic law enforcement, and driver education (together with unintentional boons like congested roads and economic recessions) saved still more lives. A lot of lives: since 1980, about 650,000 Americans have lived who would have died if traffic death rates had remained the same.41 The numbers are all the more remarkable when we consider that with each passing decade, Americans drove more miles (55 billion in 1920, 458 billion in 1950, 1.5 trillion in 1980, and 3 trillion in 2013), so they were enjoying all the pleasures of leafy suburbs, soccer-playing children, seeing the USA in their Chevrolet, or just cruising down the streets, feeling out of sight, spending all their money on a Saturday night.42 The additional miles driven did not eat up the safety gains: automobile deaths per capita (as opposed to per vehicle mile) peaked in 1937 at close to 30 per 100,000 per year, and have been in steady decline since the late 1970s, hitting 10.2 in 2014, the lowest rate since 1917.43