In This Review

A Savage Order: How the World’s Deadliest Countries Can Forge a Path to Security
A Savage Order: How the World’s Deadliest Countries Can Forge a Path to Security
By Rachel Kleinfeld
496 pp, Pantheon, 2018

These are dangerous times: war in Syria and Yemen, bloody repression in Venezuela, ethnic cleansing in Myanmar. Yet by some measures, the world is safer than ever before. The rate of violent death has been falling, albeit unevenly, for decades, even centuries. Fewer people are killed on the battlefield, on the streets, and in homes. Led by the psychologist Steven Pinker, who has collected reams of evidence demonstrating that humanity has slowly but surely grown more peaceable, a new group of thinkers is urging policymakers and the public to consider not just what the world is doing wrong in terms of violence but also what it is doing right.

Rachel Kleinfeld is one such thinker, and her new book, A Savage Order, is to some extent an extension of Pinker’s work. Kleinfeld sets out to “study success” by examining regions where murderous violence has fallen. She also draws on her own research and experience promoting security and human rights to extract lessons for policymakers looking to pacify violent societies. Kleinfeld claims to have developed a “blueprint for action,” but that is an overstatement: her conclusions are too general to serve as a concrete guide for policymaking. Yet she does offer valuable insights into why some governments allow violence to fester and what societies can do to end it. As Kleinfeld shows, some countries have managed to cut crime and save lives. Others can and should learn from them.


Kleinfeld begins with an important observation: most violence takes place in everyday situations, not on the battlefield. Each year, individual attackers, organized criminal groups, and state security services kill four times as many people as do all the current wars put together. Leaders should continue working to prevent war, but most progress in promoting peace will depend on controlling the more mundane forms of brutality.

That progress has to start in the most violent countries, such as Brazil, which suffers from an annual national homicide rate (over 30 per 100,000 citizens) that is 15 times as high as that of Belgium or Canada. Scholars often attribute the violence in such states to the weakness of their governments, pointing to incompetent police forces, inadequate bureaucracies, and a lack of government control over large geographic areas. But Kleinfeld points out that although some violent states have governments that are weak or fragile in the conventional sense, others are stable enough to tackle violence—if they so chose. This distinction matters because it acknowledges that keeping the peace is fundamentally a matter of governance. Kleinfeld does away with the idea that violence is the product of culture, religion, or poverty and persuasively argues that domestic peace begins with a state monopoly on the use of force.

Kleinfeld’s work thus provides some much-needed accountability for governments that fail to protect their people. In most cases, the problem is not that governments are weak; it is that they are complicit, deliberately supporting criminal groups for the sake of both power and enrichment. In Bangladesh, Jamaica, and Nigeria, for example, political parties hire criminal gangs to herd their voters to the polls during elections and scare away the opposition, then protect the gangs between campaigns. Yet it’s worth noting that governments are rarely uniformly corrupt: criminal organizations may capture local or regional authorities but not federal ones, and the level of corruption often varies from place to place.

In most cases, the problem is not that governments are weak; it is that they are complicit.

Kleinfeld’s indictment of complicit governments leads her to another observation: governments and societies descend into violence and escape it together. When governments cripple their police and security services in order to exert control over them, they wind up normalizing everyday violence. Disputes among family members, friends, neighbors, and business associates turn bloody more easily. People resort to vigilantism to solve problems that the state will not. The resulting low-level violence often outpaces more organized criminality. For instance, even in Latin American countries where gang and drug violence run rampant, the majority of homicides still happen on weekend nights, often triggered by petty, alcohol-fueled disputes.

This kind of violence stems not just from the inadequacy of the security services but also from the perceived illegitimacy of the government and the criminal justice system. In poor communities, many residents quite reasonably believe that the police and other officials are unable or unwilling to help them. As a result, they refuse to report crimes, testify in court, or serve as jurors. What is more dangerous, some resort to violence to solve mundane disputes. (Sociologists often refer to this phenomenon as “legal cynicism.”) The more cynical people and communities are, the more violent they are likely to be.

A Colombian Special Forces soldier patrols a street in Buenaventura, March 2014
A Colombian Special Forces soldier patrols a street in Buenaventura, March 2014
John Vizcaino / REUTERS


Successful state transitions out of violence often depend on what Kleinfeld calls “dirty deals” between governments and violent groups. She deserves credit for stating publicly what many will only whisper behind closed doors: pragmatic leaders must sometimes make morally questionable tradeoffs to gain temporary truces. The “Medellín miracle” in Colombia—a massive decline in violence over the last two and a half decades in what had been the world’s most dangerous city—is often described as a triumph of so-called social urbanism. Commentators usually point to investments in infrastructure and services for neglected neighborhoods, such as rapid and reliable public transit and new libraries and parks. But Kleinfeld reveals that much of the city’s progress depended on hidden agreements, such as the one brokered in 2003 between the government and the paramilitary boss Diego Murillo Bejarano, known as Don Berna. In return for keeping the peace among Medellín’s many cartels, gangs, and other criminal groups, the government promised Don Berna and his fellow paramilitaries short prison sentences and protection from deportation to the United States, where they were wanted on drug and gang charges.

Although Medellín offers a success story, Kleinfeld is careful to note that such methods are fraught with risk. Deals can set dangerous precedents. In 2012, political leaders in El Salvador secretly reached a truce with the MS-13 and 18th Street gangs that resulted in a dramatic drop in homicides. Three years later, however, the truce had failed, and murder rates skyrocketed. The gangs’ leaders, confident that the authorities would bargain with them as they had in the past, ordered the killing of bus drivers in San Salvador to force the government back to the negotiating table.

Where leaders cannot bargain for peace, they may employ aggressive measures, including the use of military force, anti-racketeering and asset forfeiture laws, and electronic surveillance, to take back territory, purge corrupt officials, and break up criminal conspiracies. Tough military and law enforcement tactics, like dirty deals, come with risks, especially when they violate human and civil rights. State authority taken to extremes easily leads to abuse: many leaders begin as reformers but wind up turning into the kind of authoritarians they once opposed.

At their best, dirty deals and tough tactics buy the government time to implement democratic reforms, which leads to Kleinfeld’s final observation: lasting peace requires a new political settlement. Once the deals have been struck, governments must secure broad support for an agenda of reform that reasserts the rule of law and brings power back to the people. This finding, although undoubtedly true, is the least useful in the book. Deep reforms always require political support; the harder question is how to win it. Here, the book offers little concrete guidance. Kleinfeld also argues that “the middle class is the fulcrum of change” and that violence declines when “the people with enough voice and power to change the system mobilize.” But that description is equally true of elites, and many reform movements have depended in significant part on the support of business and government leaders. Mobilization by the middle class is necessary, but not sufficient, to spur and sustain reform.


Kleinfeld’s candor in describing the challenges of reforming corrupt and violent states is refreshing. The public in these places needs to be better informed about the potential risks and rewards of these tradeoffs if they are to empower their leaders to make difficult decisions and to hold them accountable for the results.

Kleinfeld presents her findings as authoritative, but in reality, they are merely persuasive. Most of them are based on observation and experience, not hard evidence. She is not to blame for this: there is little rigorous research addressing the intersection of crime, conflict, and governance. Kleinfeld does mention technocratic, evidence-informed policies that can improve social and economic welfare, but she doesn’t spend much time examining them; her book focuses on how to create the conditions under which such policies are possible. Frustratingly, that kind of change remains more an art than a science.

Governments need to ruthlessly zero in on homicide rates if they are to cut violence.

There are, however, specific actions that policymakers in poorer countries can take that would set the stage for more targeted interventions. First and foremost, to curb lethal violence, policymakers should focus their efforts specifically on such violence and abandon the false hope of addressing it indirectly via poverty, corruption, or other forms of crime. In fact, basic improvements in security make every other strategy to improve social and economic welfare more likely to succeed. In the United States, for instance, violence underpins concentrated urban poverty, locking into place poor schools, health care, and housing. This is true to an even greater extent in countries suffering from far higher levels of deadly crime.

The single most important and reliable measure of violence is homicide. Other crimes often go unreported, but the authorities know about most murders, because they usually produce an undeniable piece of physical evidence: a body. Homicide rates also serve as a useful indicator for other violent offenses, as they tend to rise and fall in unison over the long run. Governments, accordingly, need to ruthlessly zero in on homicide rates if they are to cut violence.

The best way to stop homicides is to work from the bottom up, improving everyday safety for the poor and working classes. Today, in many Latin American countries, such as Guatemala, Honduras, and Mexico, the authorities bring less than ten percent of killers to justice. Latin American leaders should focus on ending that impunity and abandon the failed “kingpin strategy,” which emphasizes targeting only the heads of criminal organizations. Fighting organized crime is important, but it should not justify neglecting the daily slaughter caused by disorganized crime.

The wrong target: El Chapo arrives at Long Island MacArthur airport in New York, January 2017
The wrong target: El Chapo arrives at Long Island MacArthur airport in New York, January 2017

Policymakers in developing countries can also learn from evidence-based programs that have reduced violence in rich countries such as the United States. “Focused deterrence,” which engages violent offenders with offers of assistance along with threats of punishment, has reduced shootings and killings in neighborhoods of Boston, New Orleans, Oakland, and elsewhere. “Hot spots policing,” which concentrates police resources on the specific addresses, city blocks, and streets with the highest crime rates, has led to significant improvements in Los Angeles, New York City, and Philadelphia. Local governments have cut crime and antisocial behavior by offering cognitive behavioral therapy to people who are at risk of reoffending. A group of approaches known as “multidimensional juvenile therapy,” which works to fix a wide range of problems in the lives of adolescent offenders, has been shown to consistently reduce criminal behavior and recidivism among young people. Drug treatment programs and drug courts, which deal only with drug offenders and focus on rehabilitation, have cut recidivism among people with substance abuse problems. And finally, alcohol restrictions, such as limits on the kinds of stores that can sell alcohol and the hours during which they can sell it, make fights and assaults on weekends and in the evenings less likely. Dozens of rigorous studies have shown that these approaches work in high-income countries. Many low- and middle-income countries are now experimenting with the same strategies (although some of these may not be as successful as they were in high-income countries without careful adaptation to local circumstances).

As important as specific evidence-based programs are, policymakers need to pair them with wider reforms. One of the most important institutional investments a state can make is to build systems that produce detailed and reliable crime data and research. Without the right information, governments will simply be thrashing about in the dark.

States suffering from high rates of homicidal violence compete with nonstate actors to gain control over society. Little by little, governments must increase their capacity and legitimacy, creating distance between them and their competitors. This race is a marathon, not a sprint. Outside actors—other countries, international organizations, philanthropic foundations—can do the most good by helping and pressuring these governments to make the long-term investments and reforms necessary to win the race. Yet as Kleinfeld demonstrates, countries marred by murder and mayhem must ultimately take responsibility for the problem themselves. In the end, only they can solve it.