The Better Angels of Our Nature: Why Violence Has Declined

by Steven Pinker

  The political scientists James Fearon and David Laitin have backed up such vignettes with data confirming that civil wars today are fought by small numbers of lightly armed men who use their knowledge of the local landscape to elude national forces and intimidate informants and government sympathizers. These insurgencies and rural guerrilla wars may have any number of pretexts, but at heart they are less ethnic, religious, or ideological contests than turf battles between street gangs or Mafiosi. In a regression analysis of 122 civil wars between 1945 and 1999, Fearon and Laitin found that, holding per capita income constant (which they interpret as a proxy for government resources), civil wars were not more likely to break out in countries that were ethnically or religiously diverse, that had policies which discriminated against minority religions or languages, or that had high levels of income inequality. Civil wars were more likely to break out in countries that had large populations, mountainous terrain, new or unstable governments, significant oil exports, and (perhaps) a large proportion of young males. Fearon and Laitin conclude, “Our theoretical interpretation is more Hobbesian than economic. Where states are relatively weak and capricious, both fears and opportunities encourage the rise of local would-be rulers who supply a rough justice while arrogating the power to ‘tax’ for themselves and, often, a larger cause.”48

  Just as the uptick in civil warfare arose from the decivilizing anarchy of decolonization, the recent decline may reflect a recivilizing process in which competent governments have begun to protect and serve their citizens rather than preying on them.49 Many African nations have traded in their Bokassa-style psychopaths for responsible democrats and, in the case of Nelson Mandela, one of history’s greatest statesmen.50

  The transition required an ideological change as well, not just in the affected countries but in the wider international community. The historian Gérard Prunier has noted that in 1960s Africa, independence from colonial rule became a messianic ideal. New nations made it a priority to adopt the trappings of sovereignty, such as airlines, palaces, and nationally branded institutions. Many were influenced by “dependency theorists” who advocated that third-world governments disengage from the global economy and cultivate self-sufficient industries and agrarian sectors, which most economists today consider a ticket to penury. Often economic nationalism was combined with a romantic militarism that glorified violent revolution, symbolized in two icons of the 1960s, the soft-color portrait of a glowing Mao and the hard-edged graphic of a dashing Che. When dictatorships by glorious revolutionaries lost their cachet, democratic elections became the new elixir. No one found much romance in the frumpy institutions of the Civilizing Process, namely a competent government and police force and a dependable infrastructure for trade and commerce. Yet history suggests that these institutions are necessary for the reduction of chronic violence, which is a prerequisite to every other social good.

  During the past two decades the great powers, donor nations, and intergovernmental organizations (such as the African Union) have begun to press the point. They have ostracized, penalized, shamed, and in some cases invaded states that have come under the control of incompetent tyrants.51 Measures to track and fight government corruption have become more common, as has the identification of barriers that penalize developing nations in global trade. Some combination of these unglamorous measures may have begun to reverse the governmental and social pathologies that had loosed civil wars on the developing world from the 1960s through the early 1990s.

  Decent governments tend to be reasonably democratic and market-oriented, and several regression studies have looked at datasets on civil conflict for signs of a Liberal Peace like the one that helps explain the avoidance of wars between developed nations. We have already seen that the first leg of the peace, democracy, does not reduce the number of civil conflicts, particularly when it comes in the rickety form of an anocracy. But it does seem to reduce their severity. The political scientist Bethany Lacina has found that civil wars in democracies have fewer than half the battle deaths of civil wars in nondemocracies, holding the usual variables constant. In his 2008 survey of the Liberal Peace, Gleditsch concluded that “democracies rarely experience large-scale civil wars.”52 The second leg of the Liberal Peace is even stronger. Openness to the global economy, including trade, foreign investment, aid with strings attached, and access to electronic media, appears to drive down both the likelihood and the severity of civil conflict.53

  The theory of the Kantian Peace places the weight of peace on three legs, the third of which is international organizations. One type of international organization in particular can claim much of the credit for driving down civil wars: international peacekeeping forces.54 In the postcolonial decades civil wars piled up not so much because they broke out at an increasing rate but because they broke out at a higher rate than they ended (2.2 outbreaks a year compared to 1.8 terminations), and thus began to accumulate.55 By 1999 an average civil war had been going on for fifteen years! That began to change in the late 1990s and 2000s, when civil wars started to fizzle out faster than new ones took their place. They also tended to end in negotiated settlements, without a clear victor, rather than being fought to the bitter end. Formerly these embers would smolder for a couple of years and then flare up again, but now they were more likely to die out for good.

  This burst of peace coincides with a burst of peacekeepers. Figure 6–6 shows that beginning in the late 1980s the international community stepped up its peacekeeping operations and, more importantly, staffed them with increasing numbers of peacekeepers so they could do their job properly. The end of the Cold War was a turning point, because at last the great powers were more interested in seeing a conflict end than in seeing their proxy win.56 The rise of peacekeeping is also a sign of the humanist times. War is increasingly seen as repugnant, and that includes wars that kill black and brown people.

  Peacekeeping is one of the things that the United Nations, for all its foibles, does well. (It doesn’t do so well at preventing wars in the first place.) In Does Peacekeeping Work? the political scientist Virginia Page Fortna answers the question in her title with “a clear and resounding yes.”57 Fortna assembled a dataset of 115 cease-fires in civil wars from 1944 to 1997 and examined whether the presence of a peacekeeping mission lowered the chances that the war would reignite. The dataset included missions by the UN, by permanent organizations such as NATO and the African Union, and by ad hoc coalitions of states. She found that the presence of peacekeepers reduced the risk of recidivism into another war by 80 percent. This doesn’t mean that peacekeeping missions are always successful—the genocides in Bosnia and Rwanda are two conspicuous failures—just that they prevent wars from restarting on average. Peacekeepers need not be substantial armies. Just as scrawny referees can pull apart brawling hockey players, lightly armed and even unarmed missions can get in between militias and induce them to lay down their weapons. And even when they don’t succeed at that, they can serve as a tripwire for bringing in the bigger guns. Nor do peacekeepers have to be blue-helmeted soldiers. Functionaries who scrutinize elections, reform the police, monitor human rights, and oversee the functioning of bad governments also make a difference.

  FIGURE 6–6. Growth of peacekeeping, 1948–2008

  Source: Graph from Gleditsch, 2008, based on research by Siri Rustad.

  Why does peacekeeping work? The first reason comes right out of Leviathan: the larger and better-armed missions can retaliate directly against violators of a peace agreement on either side, raising the costs of aggression. The imposed costs and benefits can be reputational as well as material. A member of a mission commented on what led Afonso Dhlakama and his RENAMO rebel force to sign a peace agreement with the government of Mozambique: “For Dhlakama, it meant a great deal to be taken seriously, to go to cocktail parties and be treated with respect. Through the UN he got the government to stop calling RENAMO ‘armed bandits.’ It felt good to be wooed.”58

  Even small missions can be effective at
keeping a peace because they can free the adversaries from a Hobbesian trap in which each side is tempted to attack out of fear of being attacked first. The very act of accepting intrusive peacekeepers is a costly (hence credible) signal that each side is serious about not attacking. Once the peacekeepers are in place, they can reinforce this security by monitoring compliance with the agreement, which allows them to credibly reassure each side that the other is not secretly rearming. They can also assume everyday policing activities, which deter the small acts of violence that can escalate into cycles of revenge. And they can identify the hotheads and spoilers who want to subvert the agreement. Even if a spoiler does launch a provocative attack, the peacekeepers can credibly reassure the target that it was a rogue act rather than the opening shot in a resumption of aggression.

  Peacekeeping initiatives have other levers of influence. They can try to stamp out the trade in contraband that finances rebels and warlords, who are often the same people. They can dangle pork-barrel funding as an incentive to leaders who abide by the peace, enhancing their power and electoral popularity. As one Sierra Leonean said of a presidential candidate, “If Kabbah go, white man go, UN go, money go.”59 Also, since third-world soldiers (like premodern soldiers) are often paid in opportunities to plunder, the money can be applied to “demobilization, disarmament, and reintegration” programs that aim to draw General Butt Naked and his comrades back into civil society. With guerrillas who have more of an ideological agenda, the fact that the bribes come from a neutral party rather than a despised enemy allows them to feel they have not sold out. Leverage can also be applied to force political leaders to open their governments to rival political or ethnic groups. As with the financial sweeteners, the fact that the concessions are made to a neutral party rather than to the hated foe provides the conceder with an opportunity to save face. Desmond Malloy, a UN worker in Sierra Leone, observed that “peacekeepers create an atmosphere for negotiations. [Concessions] become a point of pride—it’s a human trait. So you need a mechanism that allows negotiations without losing dignity and pride.”60

  For all these encouraging statistics, news readers who are familiar with the carnage in the Democratic Republic of the Congo, Iraq, Sudan, and other deathtraps may not be reassured. The PRIO/UCDP data we have been examining are limited in two ways. They include only state-based conflicts: wars in which at least one of the sides is a government. And they include only battle-related deaths: fatalities caused by battlefield weapons. What happens to the trends when we start looking for the keys that don’t fall under these lampposts?

  The first exclusion consists of the nonstate conflicts (also called intercommunal violence), in which warlords, militias, mafias, rebel groups, or paramilitaries, often affiliated with ethnic groups, go after each other. These conflicts usually occur in failed states, almost by definition. A war that doesn’t even bother to invite the government represents the ultimate failure of the state’s monopoly on violence.

  The problem with nonstate conflicts is that until recently war buffs just weren’t interested in them. No one kept track, so there’s nothing to count, and we cannot plot the trends. Even the United Nations, whose mission is to prevent “the scourge of war,” refuses to keep statistics on intercommunal violence (or on any other form of armed conflict), because its member states don’t want social scientists poking around inside their borders and exposing the violence that their murderous governments cause or their inept governments fail to prevent .61

  Nonetheless, a broad look at history suggests that nonstate conflicts today must be far fewer than they were in decades and centuries past, when less of the earth’s surface was controlled by states. Tribal battles, slave raids, pillagings by raiders and horse tribes, pirate attacks, and private wars by noblemen and warlords, all of them nonstate, were scourges of humanity for millennia. During China’s “warlord era” from 1916 to 1928, more than 900,000 people were killed by competing military chieftains in just a dozen years.62

  It was only in 2002 that nonstate conflicts began to be tabulated. Since then the UCDP has maintained a Non-State Conflict Dataset, and it contains three revelations. First, nonstate conflicts are in some years as numerous as state-based conflicts—which says more about the scarcity of war than about the prevalence of intercommunal combat. Most of them, not surprisingly, are in sub-Saharan Africa, though a growing number are in the Middle East (most prominently, Iraq). Second, nonstate conflicts kill far fewer people than conflicts that involve a government, perhaps a quarter as many. Again, this is not surprising, since governments almost by definition are in the violence business. Third, the trend in the death toll from 2002 to 2008 (the most recent year covered in the dataset) has been mostly downward, despite 2007’s being the deadliest year for intercommunal violence in Iraq.63 So as best as anyone can tell, it seems unlikely that nonstate conflicts kill enough people to stand as a counterexample to the decline in the worldwide toll of armed conflict that constitutes the New Peace.

  A more serious challenge is the number of indirect deaths of civilians from the hunger, disease, and lawlessness exacerbated by war. One often reads that a century ago only 10 percent of the deaths in war were suffered by civilians, but that today the figure is 90 percent. Consistent with this claim are new surveys by epidemiologists that reveal horrendous numbers of “excess deaths” (direct and indirect) among civilians. Rather than counting bodies from media reports and nongovernmental organizations, surveyors ask a sample of people whether they know someone who was killed, then extrapolate the proportion to the population as a whole. One of these surveys, published in the medical journal Lancet in 2006, estimated that 600,000 people died in the war in Iraq between 2003 and 2006—overwhelmingly more than the 80,000 to 90,000 battle deaths counted for that period by PRIO and by the Iraq Body Count, a respected nongovernmental organization.64 Another survey in the Democratic Republic of the Congo put the death toll from its civil war at 5.4 million—about thirty-five times the PRIO battle-death estimate, and more than half of the total of all the battle deaths it has recorded in all wars since 1946.65 Even granting that the PRIO figures are intended as lower bounds (because of the stringent requirements that deaths be attributed to a cause), this is quite a discrepancy, and raises doubts about whether, in the big picture, the decline in battle deaths can really be interpreted as an advance in peace.

  Casualty figures are always moralized, and it’s not surprising that these three numbers, which have been used to indict, respectively, the 20th century, Bush’s invasion of Iraq, and the world’s indifference to Africa, have been widely disseminated. But an objective look at the sources suggests that the revisionist estimates are not credible (which, needless to say, does not imply that anyone should be indifferent to civilian deaths in wartime).

  First off, the commonly cited 10-percent-to-90-percent reversal in civilian casualties turns out to be completely bogus. The political scientists Andrew Mack (of HSRP), Joshua Goldstein, and Adam Roberts have each tried to track down the source of this meme, since they all knew that the data needed to underpin it do not exist.66 They also knew that the claim fails basic sanity checks. For much of human history, peasants have subsisted on what they could grow, producing little in the way of a surplus. A horde of soldiers living off the land could easily tip a rural population into starvation. The Thirty Years’ War in particular saw not only numerous massacres of civilians but the deliberate destruction of homes, crops, livestock, and water supplies, adding up to truly horrendous civilian death tolls. The American Civil War, with its blockades, crop-burnings, and scorched-earth campaigns, caused an enormous number of civilian casualties (the historical reality behind Scarlett O’Hara’s vow in Gone With the Wind: “As God is my witness, I’ll never be hungry again”).67 During World War I the battlefront moved through populated areas, raining artillery shells on towns and villages, and each side tried to starve the other’s civilians with blockades. And as I have mentioned, if one includes the victims of the 1918 flu epidemic as in
direct deaths from the war, one could multiply the number of civilian casualties many times over. World War II, also in the first half of the 20th century, decimated civilians with a holocaust, a blitz, Slaughterhouse-Five–like firebombings of cities in Germany and Japan, and not one but two atomic explosions. It seems unlikely that today’s wars, however destructive to civilians, could be substantially worse.

  Goldstein, Roberts, and Mack traced the meme to a chain of garbled retellings in which different kinds of casualty estimates were mashed up: battle deaths in one era were compared with battle deaths, indirect deaths, injuries, and refugees in another. Mack and Goldstein estimate that civilians suffer around half of the battle deaths in war, and that the ratio varies from war to war but has not increased over time. Indeed, we shall see that it has recently decreased by a substantial margin.

  The most widely noted of the recent epidemiological estimates is the Lancet study of deaths in Iraq.68 A team of eight Iraqi health workers went door to door in eighteen regions and asked people about recent deaths in the family. The epidemiologists subtracted the death rate for the years before the 2003 invasion from the death rate for the years after, figuring that the difference could be attributed to the war, and multiplied that proportion by the size of the population of Iraq. This arithmetic suggested that 655,000 more Iraqis died than if the invasion had never taken place. And 92 percent of these excess deaths, the families indicated, were direct battle deaths from gunshots, airstrikes, and car bombs, not indirect deaths from disease or starvation. If so, the standard body counts would be underestimates by a factor of around seven.